Atlas of Gastrointestinal Pathology

A Pattern Based Approach to

Neoplastic Biopsies

DORA M. LAM-HIMLIN, MD
Associate Professor
Department of Laboratory Medicine and Pathology
Mayo Clinic
Scottsdale, Arizona

ELIZABETH A. MONTGOMERY, MD
Professor of Pathology, Oncology, and Orthopedic Surgery
Department of Pathology
Division of Gastrointestinal and Liver Pathology
Johns Hopkins Medical Institutions
Baltimore, Maryland

CHRISTINA A. ARNOLD, MD
Associate Professor
Department of Pathology
Division of Gastrointestinal and Liver Pathology
The Ohio State University Wexner Medical Center
Columbus, Ohio
Acquisitions Editor: Ryan Shaw
Editorial Coordinator: Lindsay Ries
Marketing Manager: Julie Sikora
Production Project Manager: Linda Van Pelt
Design Coordinator: Holly McLaughlin
Manufacturing Coordinator: Beth Welsh
Prepress Vendor: TNQ Technologies

Copyright © 2019 Wolters Kluwer

All rights reserved. This book is protected by copyright. No part of this book may be reproduced or transmitted in any
form or by any means, including as photocopies or scanned-in or other electronic copies, or utilized by any information
storage and retrieval system without written permission from the copyright owner, except for brief quotations embodied
in critical articles and reviews. Materials appearing in this book prepared by individuals as part of their official duties as
U.S. government employees are not covered by the above-mentioned copyright. To request permission, please contact
Wolters Kluwer at Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, via email at
[email protected], or via our website at shop.lww.com (products and services).

987654321

Printed in China

Library of Congress Cataloging-in-Publication Data

Names: Lam-Himlin, Dora M., author. | Montgomery, Elizabeth (Elizabeth A.), 1958- author. |
Arnold, Christina A., author.
Title: Atlas of gastrointestinal pathology : a pattern based approach to neoplastic biopsies / Dora M. Lam-Himlin,
Elizabeth A. Montgomery, Christina A. Arnold.
Description: Philadelphia : Wolters Kluwer Health, 2019. | Includes bibliographical references and index.
Identifiers: LCCN 2018046981 | eISBN 9781496367570
Subjects: | MESH: Gastrointestinal Tract–pathology | Gastrointestinal Diseases–pathology | Biopsy | Atlases
Classification: LCC RC802.9 | NLM WI 17 | DDC 616.3/307–dc23
LC record available at https://lccn.loc.gov/2018046981

This work is provided “as is,” and the publisher disclaims any and all warranties, express or implied, including any
warranties as to accuracy, comprehensiveness, or currency of the content of this work.

This work is no substitute for individual patient assessment based upon healthcare professionals’ examination of each
patient and consideration of, among other things, age, weight, gender, current or prior medical conditions, medication
history, laboratory data and other factors unique to the patient. The publisher does not provide medical advice or
guidance and this work is merely a reference tool. Healthcare professionals, and not the publisher, are solely
responsible for the use of this work including all medical judgments and for any resulting diagnosis and treatments.

Given continuous, rapid advances in medical science and health information, independent professional verification of
medical diagnoses, indications, appropriate pharmaceutical selections and dosages, and treatment options should be
made and healthcare professionals should consult a variety of sources. When prescribing medication, healthcare
professionals are advised to consult the product information sheet (the manufacturer’s package insert) accompanying
each drug to verify, among other things, conditions of use, warnings and side effects and identify any changes in
dosage schedule or contradictions, particularly if the medication to be administered is new, infrequently used or has a
narrow therapeutic range. To the maximum extent permitted under applicable law, no responsibility is assumed by the
publisher for any injury and/or damage to persons or property, as a matter of products liability, negligence law or
otherwise, or from any reference to or use by any person of this work.

shop.lww.com
 To my family:
Matt: For being my rock.
Madeline: Stay brave. Stay curious.
Matthew: Always have the heart of a gentleman.
Tommy: Yes, I’d rather be outside, too.
Dora M. Lam-Himlin, MD

To all the colleagues who love gastrointestinal pathology as much as we do.

Elizabeth A. Montgomery, MD

To my loving family:
Mary and Andy: Mike and I didn’t realize how much we needed parents, until
we became parents ourselves.
Thank you for always being there for us, and for reminding us about the
importance of family, love, and laughter.
Mom: Thank you for being my first best friend, believing in me always, and
teaching me the value of hard work.
Our Wednesday afternoon brownies are some of my favorite memories.
Jackson and Madelyn: Objects in the mirror are sometimes closer than they
appear. If it is important,
NEVER give up—hunker down and dig in!
Mike: The keys to a good road trip are fun, excitement, danger, and Johnny Cash.
OGN: This book and my career would not be possible without you. Although I
don’t deserve you,
I thank the heavens for you every day. Here’s to Baltimore, the Inca trail, and all
of tomorrow’s
mischievous adventures.
Christina A. Arnold, MD
 PREFACE

Diagnosis and reporting of neoplastic GI biopsies is a complex moving target, as

evidenced by evolving nomenclature, updates in society guidelines, recognition of new
therapeutic targets, and increasing requirements for prognostic elements and ancillary
testing interpretation. This follow-up textbook on GI neoplasia is the highly anticipated
companion volume to Atlas of Gastrointestinal Pathology: A Pattern Based Approach to
Non-Neoplastic Biopsies. This new book applies the now-familiar method of pattern
based learning to GI neoplasia and provides a systematic algorithm for tackling common
and uncommon interpretation challenges. Mirroring the previous text, this book
highlights tell-tale “red flags” found in the clinical chart, hidden clues in the slides, and
how to discern an exact diagnosis despite sometimes disabling artifacts.
New topics covered include a simple approach to the endoscopic mucosal resection
(EMR) specimen, the latest definition of Barrett esophagus and its reporting, an
algorithmic approach to serrated polyps, instructions for decoding the alphabet soup of
colorectal cancer molecular testing, the latest consensus guidelines on approaching anal
dysplasia and dysplasia arising in inflammatory bowel disease, and a handy guide to
syndromic polyps with emphasis on morphology, clinical considerations, genetics, and
practical reporting.
The illustrations extend beyond a handful of classic examples for each entity, and
more than 1600 images cover the full morphologic spectrum of the major patterns of GI
neoplasia. The sessile serrated adenoma/polyp, for example, is illustrated in more than
50 figures that include direct comparisons with differential diagnoses and borderline
cases, and each image is captioned with a careful description. The corresponding text
details information on how to classify the polyp, minimum diagnostic criteria, clinical
implications of surveillance intervals, and sample sign-out notes.
In this book, disease processes are grouped by their histologic patterns—an
approach that echoes the first volume and closely approximates the method by which
experienced pathologists mentally approach daily sign-out. Organized by these major
patterns, each chapter details neoplastic considerations for the esophagus, stomach,
small intestine, colon, anus, and soft tissue.
 The text is high yield and focused on checklists, key features, diagnostic pearls and
pitfalls, frequently asked questions, and sample notes—see the following descriptions.
We hope this collective experience leaves the reader with a familiarity of the major
patterns of GI neoplasia and confidence in navigating through the clinicopathologic
clues and pitfalls to arrive swiftly at the correct diagnosis. Select structural elements
are briefly introduced as follows.

• Each chapter opens with a “Chapter Outline” that outlines the enclosed structure and
allows the reader to quickly hone in on select patterns and pertinent differential
diagnostic considerations. Similar checklists are found throughout the chapter to neatly
organize complicated topics.
• “The Unremarkable X”: Normal histology is sometimes overlooked in textbooks
because it is assumed to be widely understood, much to the frustration of junior
trainees. A firm understanding of normal is essential to recognizing subtle injury
patterns. As such, each chapter begins with a brief discussion of normal histology to
contrast to the succeeding mucosal injury patterns and to highlight helpful diagnostic
clues.
• The “Pearls & Pitfalls” boxes include lessons from real life sign-out experience with
an emphasis on important diagnostic clues, mimics, and hazards.
• The “Frequently Asked Questions” sections stem from our busy consult service and
teaching sessions. In these sections, we discuss real-life diagnostic dilemmas and
offer diagnostic tips and tools to sort through commonly encountered sign-out
challenges.
• All major topics close with a “Key Features” section that summarizes the essential
elements of the subtopic for handy reference.
• A “Sample Note” section accompanies the more challenging topics. In these sections,
an example pathology report is included with the top-line diagnosis, pertinent
discussion, and salient references. These notes offer a template of how to synthesize
complicated topics and are based on real-life cases and interactions with clinicians.
The select references are included for those interested in further reading but also can
be included in pathology reports to help guide clinical management.
• Each chapter features a corresponding “Quiz” section in the appendix to emphasize
important teaching points. These sections offer the reader experience and confidence
with high-yield teaching topics. Questions are in the format of the board type
examinations and can also serve as useful board preparatory materials.
 ACKNOWLEDGMENTS

We thank our institutions, colleagues, and trainees for invaluable resources and support.
We are indebted to our inquisitive trainees and clinicians whose fresh perspectives and
lively discussions drove the direction of this book. We particularly thank our families
for understanding the numerous late night, early morning, and weekend marathon writing
sessions.
We thank our Acquisition Editor, Ryan Shaw, for taking a chance on this project, and
our Editorial Coordinator, Lindsay Ries, for working diligently with us to ensure timely
completion. We thank Frank M. Corl, MS, for the custom medical illustrations; Rick
Marshall for computer assistance in identifying pertinent teaching material; and Shawn
Scully for photography editing on select topics.
Lastly, we thank the production team led by Ramkumar Soundararajan for their
careful attention to detail.
 CONTENTS

1 ESOPHAGUS
The Unremarkable Esophagus
Some Esophageal Polyps
Barrett Esophagus
Squamous Neoplasia
Near Misses

2 STOMACH
The Unremarkable Stomach
Polyps
Adenocarcinoma
Well-Differentiated Neuroendocrine Tumors (Formerly “Carcinoid”)
MALT Lymphoma
Mesenchymal Lesions
Near Miss

3 SMALL BOWEL
The Unremarkable Small Bowel
Polyps
Adenocarcinoma
Neuroendocrine Tumors and Neuroendocrine Carcinomas
Near Miss

4 COLON
The Unremarkable Colon
Polyps
Dysplasia in Inflammatory Bowel Disease
 Adenocarcinoma
Near Miss

5 ANUS
The Unremarkable Anus
Nondysplastic Polyps
HPV-Associated Neoplasms
Adenocarcinoma
Extramammary Paget Disease
Malignant Melanoma
Near Miss

6 MESENCHYMAL LESIONS
The Layout of the Real Estate
Immunolabeling Comments
Mesenteric Lesions
Translocation Sarcomas Associated With the Gastrointestinal Tract
Other Spindle Cell Lesions
Summary of Gastrointestinal Tract Lesions Containing Many Inflammatory
Cells
Select Colorectal Lesions Likely to Be Encountered on Biopsies
Summary of Neural Polyps
Near Miss.
A Few Vascular Tumors.
Near Miss.

APPENDIX A
APPENDIX B
INDEX
 ESOPHAGUS 1

CHAPTER OUTLINE
The Unremarkable Esophagus
Some Esophageal Polyps
□ Granular Cell Tumor
□ Leiomyoma
□ Gastrointestinal Stromal Tumors of the Esophagus
Barrett Esophagus
□ Grading Barrett Dysplasia
• Barrett Esophagus, Negative for Dysplasia
• Barrett Esophagus Indefinite for Dysplasia
• Low-Grade Dysplasia
• High-Grade Dysplasia
• Intramucosal Adenocarcinoma and Adenocarcinoma
• Endoscopic Mucosal Resection Specimens
• Dysplasia Recapitulation
Squamous Neoplasia
□ Risk Factors
□ Squamous Carcinoma Precursor Lesions
• Squamous Dysplasia
• Epidermoid Metaplasia
• Squamous Cell Carcinoma
Near Misses
 □ What Went Wrong?
• Neuroendocrine Tumors of the Esophagus
• Adenosquamous Carcinoma
• Spread of Extraesophageal Carcinoma to the Esophagus
• Melanoma
• Esophageal Hematopoietic Disorders

THE UNREMARKABLE ESOPHAGUS

Normal esophageal mucosa is a common sample in the practice of gastrointestinal
pathology, and most of us are familiar with squamous mucosa on biopsies. Mucosa
consists of epithelium (stratified squamous), lamina propria, and muscularis mucosae.
Beneath those structures are the submucosa and muscularis propria. Assessing
resections and endoscopic mucosal resections (EMRs) helps us learn about these
layers.
The layers matter and there are some pitfalls! Cancers that invade only the lamina
propria are staged as T1a, whereas those that extend into the submucosa are T1b
neoplasms.1 There are some issues that can arise and result in confusion. In general,
mucosal biopsies grasp some epithelium and a little bit of lamina propria. Many
biopsies contain only squamous epithelium and lack even lamina propria; normal
squamous mucosa is slippery, and it is difficult for the endoscopist to easily grasp it to
obtain a large “bite,” so abundant lamina propria and/or muscularis mucosae tend to be
present in biopsies from damaged mucosa. Most biopsies do not grasp submucosa.
Note the indicated layers in the samples in Figs. 1.1 and 1.2. The layer just under the
epithelium is the lamina propria rather than the submucosa. This is easy to spot on well-
oriented samples such as the one seen in Fig. 1.1 but not so obvious at times on samples
such as those seen in Fig. 1.2. Furthermore, once the esophagus is damaged and the
epithelium is replaced by columnar epithelium, as in Fig. 1.3, the muscularis mucosae
becomes thick and disorganized, sometimes even forming two (duplicated) or three
(triplicated) layers with bits of lamina propria between them.2 The tissue between these
sloppy smooth layers is all lamina propria, not submucosa! The irregularities in the
muscularis mucosae following mucosal damage are further discussed in the section
concerning EMRs. In Fig. 1.3, there is a clue (in addition to the squamous epithelium)
that the sample is from the esophagus; an esophageal submucosal gland is present at the
lower right of the image, and a duct that is intended to lead from the submucosal gland
to the surface is indicated. The presence of esophageal submucosal glands and ducts in
a sample confirms that the sample is derived from the esophagus, but this is not a
common finding in mucosal biopsies.
The presence of so-called multilayered epithelium,3,4 discussed later, is also a clue
that a specimen is derived from the esophagus. Figure 1.4 shows a mucosal biopsy that
contains a submucosal gland, but this is unusual.

Figure 1-1. Esophageal mucosal biopsy. This sample is slightly tangentially embedded. Note
that the biopsy contains all three layers of the mucosa, namely, the squamous epithelium, the
loose lamina propria with a few delicate blood vessels, and the muscularis mucosae (this is
Latin for the muscle of the mucosa). The epithelium has only a few layers of the darker basal
cells, and the more superficial cells are pink (eosinophilic), with their long axes arranged
parallel to the basement membrane, which is normal polarity for squamous epithelium. For
columnar epithelium, the long axes of the nuclei are normally arranged perpendicular to the
basement membrane. This sample is essentially normal and fairly well oriented.
 Figure 1-2. Esophageal mucosal biopsy. This mucosal biopsy has been embedded in a
disorderly fashion such that it is a bit trickier to interpret than the sample shown in Fig. 1.1
There is a lymphoid aggregate at the left. The muscularis mucosae is tangentially embedded
and appears thick, but this is not muscularis propria. The loose connective tissue at the upper
right is lamina propria rather than submucosa.

Pancreatic acinar cell heterotopia is also common in esophageal biopsies and

resections and is an incidental finding. The resection images shown in Figs. 1.5 and 1.6
highlight this finding and compare it with the appearance of submucosal glands.
Pancreatic acinar heterotopia of the esophagus is generally encountered in the mucosa,
whereas submucosal glands, of course, are in the submucosa.
Figure 1-3. Endoscopic mucosal resection. Even though this specimen is not well oriented, it
shows the layers that can be seen. Submucosa appears at the lower right of the field, and a
submucosal gland is indicated. To the left of the submucosal gland, a thickened portion of
muscularis mucosae courses across the sample. A duct, which transports secretions from the
submucosal gland to lubricate the surface, is seen piercing through the muscularis mucosae.
The zone to the left beneath the mucosa is the lamina propria and not the submucosa.
Epithelium is seen at the left. Note that in the columnar cardiac mucosa portion, lamina propria
invests individual glands, whereas it is under the squamous epithelium in the zone with
squamous epithelium.
Figure 1-4. Biopsy of esophagus. This is an unusual case in that a submucosal gland is
present such that a small portion of submucosa is clearly present in the specimen. However,
the loose connective tissue in between the epithelium and the muscularis mucosae is lamina
propria rather than submucosa.
Figure 1-5. Gastroesophageal junctional tissue, resection specimen. In this resection
specimen, the muscularis propria curves across the bottom of the field, and an area of
squamous epithelium at the left coats the lamina propria, muscularis mucosae, and
submucosa, where two submucosal glands are marked with arrowheads. At the center and
right, the mucosa is of the cardiac type and cardiac glands are present as well as foci of
pancreatic heterotopia that are within the mucosa, and a delicate cord of muscularis mucosae
is beneath these foci of pancreatic heterotopia marked by the arrows.
 Figure 1-6. Gastroesophageal junctional tissue. This is a high-magnification image of one of the
foci of pancreatic heterotopia that are indicated in Fig. 1.5. A nodule of pancreatic heterotopia is
indicated by a single arrow and wisps of muscularis mucosae are beneath it. The double
arrows mark gastric cardiac glands, which produce mucin.

SOME ESOPHAGEAL POLYPS

GRANULAR CELL TUMOR
Granular cell tumors of the esophagus account for about 1% to 2% of all granular cell
tumors,5 and the esophagus is the most common gastrointestinal tract site.6 Most
esophageal granular cell tumors arise in the distal esophagus and about 5% to 10% are
multicentric. There is a female predominance, and these tumors are more common in
African-Americans than in whites. Occasional large examples require radical surgery,
and malignant examples are rare.7,8 Most esophageal granular cell tumors appear as
well-marginated masses on imaging studies such that they are interpreted as
gastrointestinal stromal tumors (GISTs). The important thing with granular cell tumors is
that they can be foolers. Do not be the next victim!
It is not clear why granular cell tumors are prone to elicit a pseudoepitheliomatous
reactive response in the overlying squamous epithelium (Fig. 1.7). “Epithelioma” is an
old term for carcinoma, so “pseudoepitheliomatous hyperplasia” simply means
“pseudocarcinomatous hyperplasia.” This benign response of squamous epithelium
occasionally leads to a misinterpretation of squamous cell carcinoma if the granular cell
tumor is not spotted. This phenomenon also applies to anal canal granular cell tumors.
Of course, pseudoepitheliomatous hyperplasia can be found on top of carcinomas as
well as other processes. Figs. 1.8 and 1.9 show pseudoepitheliomatous (squamous)
hyperplasia overlying an adenocarcinoma of the esophagus.
Like granular cell tumors in the skin and elsewhere, esophageal granular cell tumors
show strong S100 protein expression (Fig. 1.10). Remember that a high-quality S100
protein preparation should display both nuclear and cytoplasmic labeling.

LEIOMYOMA
Leiomyoma is by far the most common spindle cell tumor of the esophagus, but it is still
uncommon. Esophageal leiomyomas arise in young patients (well, at least compared
with one of the authors—median age, 35 years),9 with a male predominance. They
consist of cells with eosinophilic cytoplasm (Fig. 1.11) and express desmin and alpha-
smooth muscle actin, but not CD117 and CD34. The important pitfall to be aware of in
diagnosing gastrointestinal tract leiomyomas is that, if one performs immunolabeling for
CD117 and DOG1, these stains label Cajal cells that are either entrapped in or
proliferating along with the lesion (Fig. 1.12). For this reason, confident morphologists
avoid these stains. Esophageal leiomyomas are easy to diagnose on staining with
hematoxylin and eosin (H&E)—they are hypocellular, pink, and benign.
Figure 1-7. Granular cell tumor. The tumor consists of plump eosinophilic cells with granular
cytoplasm and small nuclei. This type of tumor is notorious for stimulating hyperplasia of the
overlying squamous epithelium (so-called pseudoepitheliomatous hyperplasia), which can be
mistaken for carcinoma.
Figure 1-8. Pseudoepitheliomatous hyperplasia associated with an adenocarcinoma. The
adenocarcinoma at the bottom of the field has undermined the squamous epithelium at the top
of the field. The squamous process is benign and reactive.
Figure 1-9. Pseudoepitheliomatous hyperplasia associated with an adenocarcinoma. This is a
high-magnification image of the same lesion as that seen in Fig. 1.8. The surface squamous
component is benign and simply reacting to the adenocarcinoma beneath it.
Figure 1-10. Granular cell tumor. This stunning S100 protein immunohistochemical stain shows
striking nuclear and cytoplasmic expression.
Figure 1-11. Esophageal leiomyoma. The lesional cells are brightly eosinophilic, and the tumor
has low cellularity. The cytoplasm is fibrillary, and paranuclear vacuoles are present.
 Figure 1-12. Esophageal leiomyoma. This is a CD117/KIT stain. Do not be fooled by the
scattered labeled cells. Whether these are entrapped Cajal cells or an integral part of the
leiomyoma is not clear, but they should not result in an interpretation of gastrointestinal stromal
tumor.

GASTROINTESTINAL STROMAL TUMORS OF THE

ESOPHAGUS
GISTs predominate in the stomach and intestines, but they are vanishingly rare in the
esophagus. Even the combined files of the former Armed Forces Institute of Pathology
and the Haartman Institute of the University of Helsinki yielded only 17 examples of
esophageal GISTs!9 They arose in the lower third of the esophagi of adults (12 men and
5 women), with a median age of 63 years). Patients most commonly presented with
dysphagia. Compared with leiomyomas, GISTs have an overall basophilic appearance
and combinations of solid, myxoid, and perivascular collarlike patterns (Figs.
1.13–1.15). GISTs are discussed in “Stomach” and “Mesenchymal Lesions” chapters in
more detail, but those in the esophagus are truly rare.
Figure 1-13. Esophageal gastrointestinal stromal tumor. These lesions are rare in contrast to
gastric gastrointestinal stromal tumor. Note the prominent cellularity, a striking difference from
the cellularity of the leiomyoma in Figs. 1.11 and 1.12.
Figure 1-14. Esophageal gastrointestinal stromal tumor. This is a CD117/KIT stain.
 Figure 1-15. Esophageal gastrointestinal stromal tumor. This is a DOG1 stain.

Most patients in Miettinen’s early series died of their tumors, but the study was from
the era before targeted therapy. Regardless, in contrast to GISTs of the stomach,
esophageal examples should be considered aggressive.

BARRETT ESOPHAGUS
Evaluating for Barrett esophagus and the neoplasia associated with it is a large part of
the practice of Western pathologists. Columnar metaplasia in the esophagus is the source
of anxiety for patients and pathologists alike. This section is approached with a short
list of questions before dysplasia is discussed.

FAQ: How is Barrett mucosa defined?

Answer: It depends on whom you ask! In the United Kingdom and Japan the
definition differs from that in the United States, and at this writing, there is
 some flux at play in the United States.
British (and Japanese) definition of Barrett mucosa 2014:
• Columnar epithelium with or without goblet cells extending ≥1 cm above the
gastric folds10
American Gastroenterological Association definition of Barrett mucosa 2011:
• Columnar epithelium in the esophagus that contains goblet cells—no length
requirement11
American College of Gastroenterologists’ definition of Barrett mucosa 2016:
• Columnar epithelium with goblet cells extending ≥1 cm above the top of the
gastric folds12

The last definition (the 2016 one from the American College of Gastroenterologists)
makes the life of the pathologist challenging. In some instances, we have a good idea
about the length of a segment of columnar epithelium in question, whereas in others, the
only information we have is “esophagus.” Obviously, if we have a sample labeled
“esophagus, 40 cm” and we see intestinal metaplasia and we have a second sample that
is labeled “esophagus, 34 cm” and we see intestinal metaplasia, it is clear that the
affected segment of lesion measures at least 1 cm. Interestingly, the gastroenterology
colleagues who compiled the recommendations even went so far as to caution our
endoscopy colleagues to refrain from biopsies of the gastroesophageal junction unless
there was a visible alteration. At least in our hospitals, there seems to be little
compliance with the latter suggestion. The American College of Gastroenterology
suggested the term “specialized intestinal metaplasia of the esophagogastric junction”
for lesions that contain goblet cells but do not satisfy the requirement for the mucosal
irregularity to extend at least 1 cm above the top of the gastric folds.12
To get around the length issue, we have developed two notes that we find useful for
situations for which (1) we see intestinal metaplasia and we do not know the segment
length or (2) the sample is labeled “gastroesophageal junction” and we see intestinal
metaplasia.

SAMPLE NOTES: Barrett Mucosa

Situation A: Barrett Mucosa, Negative for Dysplasia. See Note

Note: The diagnosis of Barrett esophagus is made owing to the presence of goblet cells
(intestinal metaplasia), with the assumption that the biopsies were obtained from
columnar mucosa in the distal esophagus and the mucosal irregularity extends at least 1
cm above the top of the gastric folds as per 2016 American College of Gastroenterology
(ACG) guidelines.

Reference:
Shaheen NJ, Falk GW, Iyer PG, Gerson LB; American College of Gastroenterology.
ACG clinical guideline: diagnosis and management of Barrett’s esophagus. Am J
Gastroenterol. 2016;111(1):30-50.

Situation B: Cardiac Mucosa With Intestinal Metaplasia. See Note

Note: This biopsy shows gastric-type mucosa with scattered goblet cells. The diagnosis
in this case depends on the location of this biopsy. If this biopsy was taken from the
tubular esophagus and the mucosal irregularity extends at least 1 cm above the top of the
gastric folds, the diagnosis is Barrett mucosa of the distinctive type. If this biopsy was
taken from the gastric cardia, the diagnosis is intestinal metaplasia of the gastric cardia.

Reference:
Shaheen NJ, Falk GW, Iyer PG, Gerson LB; American College of Gastroenterology.
ACG clinical guideline: diagnosis and management of Barrett’s esophagus. Am J
Gastroenterol. 2016;111(1):30-50.

FAQ: The ACG 2016 criteria state: BE should be diagnosed when there is
extension of salmon-colored mucosa into the tubular esophagus
extending ≥1 cm above the top of the gastric folds with biopsy
confirmation of IM (strong recommendation, low level of evidence). So, if
there is a 0.5-cm island that is >1 cm above the top of the gastric folds
that contains intestinal metaplasia, is it Barrett esophagus?

Answer: This is a contentious area in gastroenterology circles. The ≥1 cm

length of mucosal irregularity rule was created to differentiate irregular Z lines
with intestinal metaplasia in cardiac mucosa from true Barrett mucosa, which is
a lesion of the tubular esophagus. By the rules of the ACG, the described 0.5-
cm island >1 cm above the top of the gastric folds is “intestinal metaplasia in
the esophagus” and not Barrett mucosa. This is in fact a shortcoming of the
Prague classification as well. By common sense, such an island is biologically
Barrett mucosa, but it is not formally recognized as such according to the
newest guidelines. However, recent data suggest that such a lesion should be
managed as per Barrett esophagus.13 As such, report what you see on the
slide and work with your endoscopist to plan follow-up.

FAQ: Why do some observers want to eliminate the requirement for goblet
cells for a diagnosis of Barrett mucosa in the United States?

Answer: There are some studies that suggest that most esophageal
adenocarcinomas that are detected arise in the absence of intestinal
metaplasia. In one of them, the authors used EMR samples and found adjoining
intestinal metaplasia in less than half of the samples with early cancers.14
However, these colleagues did not attempt to learn if the patients had separate
samples that contained intestinal metaplasia. In two West Coast US studies,
intestinal metaplasia essentially always accompanied high-grade columnar
epithelial dysplasia and carcinomas.15,16 We found similar results in an East
Coast study17 and would endorse retaining the requirement for goblet cells, but
others have suggested eliminating the requirement.18 Regardless, there are a
few cases of esophageal adenocarcinomas that are unassociated with
intestinal metaplasia, but they are not numerous in our Western population.
Elimination of the requirement for goblet cells to diagnose Barrett esophagus in
the United States would even further reduce the cost-effectiveness of
surveillance of patients with Barrett esophagus because it would open the
floodgates to patients with reflux. A key issue is that, when cardiac mucosa is
found (without intestinal metaplasia), it could be a normal finding, so the
presence of intestinal metaplasia offers some assurance that at least abnormal
tissue has been biopsied.

FAQ: Which histologic patterns can mimic Barrett mucosa?

Answer: This topic has been addressed in the non-neoplastic volume preceding
this volume (Atlas of Gastrointestinal Pathology. A pattern based approach to
non-neoplastic biopsies), although we have changed our practice patterns. We
performed up front periodic acid–Schiff/alcian blue staining on upper
gastrointestinal tract samples (including all biopsies of the esophagus) in the
past; we no longer follow this practice, but there are occasional cases for
which it is instructive. Both pancreatic acinar heterotopia (Figs. 1.5, 1.6, 1.16
and 1.17) and some gastric foveolar cells (Figs. 1.18 and 1.19) can suggest
intestinal metaplasia. Pancreatic heterotopia contains cells with granules, so
these cells can be easily distinguished from goblet cells, but some gastric
foveolar cells can take on a bluish hue or show weak alcianophilia on alcian
blue staining. These latter cells lack the true intestinal differentiation that we will
see later.

Figure 1-16. Pancreatic acinar cell heterotopia. This image allows us to further consider the
layer in which pancreatic acinar cell heterotopia is found. In this case, there is squamous rather
than cardiac-type mucosa associated with the pancreatic acinar cell heterotopia. Note that it is
found in the lamina propria. There is squamous epithelium over it, cardiac glands to the right,
and muscularis mucosae below it.
Figure 1-17. Pancreatic acinar cell heterotopia. This high-magnification image shows zymogen
granules in the cells to full advantage.
Figure 1-18. “Fake” goblet cells. A few of the cells in this cardiac gland have a bluish tint but
lack the demarcation of goblet cells.
 Figure 1-19. “Fake” goblet cells. This is a periodic acid–Schiff/alcian blue stain. It shows a bit of
nonspecific alcianophilia, but no goblet cells are present. This does not qualify as Barrett
mucosa.

FAQ: What is the difference between complete and incomplete intestinal

metaplasia and do they matter?

Answer: Complete intestinal metaplasia indicates that the metaplasia perfectly

recapitulates intestine in that there are goblet cells separated by absorptive
cells that have a brush border but not any cytoplasmic mucin. Incomplete
intestinal metaplasia has only incompletely converted from gastric cardiac
mucosa to intestinal-type mucosa. As such, there are goblet cells separated by
cells that resemble gastric foveolar cells. An example of a sample with both
types juxtaposed is shown in Figs. 1.20 and 1.21. The complete type is far
more likely to be found in gastric mucosa than in intestinal metaplasia of the
esophagus, and its presence can suggest that the sample is from the stomach
rather than from the esophagus. However, complete intestinal metaplasia is not
specific for gastric intestinal metaplasia. The incomplete type,
epidemiologically, is more likely to progress to adenocarcinoma than the
complete type, but reporting it is not particularly relevant in any given patient,
and it is not necessary to report which type is seen unless the purpose is a
study protocol.

Figure 1-20. Barrett esophagus with complete and incomplete intestinal metaplasia. In
incomplete intestinal metaplasia, the mucosa has incompletely transformed from gastric
cardiac type to an intestinal phenotype such that there are goblet cells (see the arrow beneath
the “incomplete” annotation) interspersed with gastric foveolar cells with apical mucin. In
contrast, the complete type of intestinal metaplasia fully recapitulates intestinal epithelium and
cells with brush borders are seen between goblet cells (a goblet cell is marked with an arrow
underneath the zone marked as “complete”).
 Figure 1-21. Barrett esophagus with complete and incomplete intestinal metaplasia. This is a
high-magnification image of the area shown in Fig. 1.20. It is not necessary to specify the type
of metaplasia in reports unless there is a specific protocol, but, on epidemiologic grounds, the
incomplete type is more likely to progress to carcinoma. In the individual case, it is not an
important factor.

FAQ: What is multilayered epithelium?

Answer: Multilayered epithelium is a type of epithelium that is associated with

gastroesophageal reflux and that has some properties of intestinal-type
epithelium and some properties of squamous epithelium.3,4,19,20 Because of
these features, many observers have suggested that it is a precursor to Barrett
mucosa (a precursor to a precursor).21 Essentially, it is only important to avoid
diagnosing it as dysplasia and not worry too much about it! Because it
correlates with clinical reflux disease and can be found in association with
Barrett mucosa, it is a topic of interest. In addition, the presence of
multilayered epithelium is a clue that a sample has come from the esophagus
rather than from the stomach. For those who use special stains to detect
Barrett mucosa, some examples of multilayered epithelium react with these
stains, including CDX2 and MUC2, but not diffusely. There is no need to either
use these stains or be concerned about the labeling if you do.
 The multilayered epithelium has an appearance similar to that of immature
squamous metaplasia in the uterine cervix. The surface has a few cells that
have a blush of mucin, and the base has a squamous appearance. The mucin
can have a basophilic appearance akin to that of goblet cells but is not crisply
delineated like that in goblet cells. An example is seen in Figs. 1.22 and 1.23.

Figure 1-22. Multilayered epithelium. This is a high-magnification image showing epithelium and
a tiny sliver of lamina propria at the bottom of the image. At first glance, the basal cells appear
hyperchromatic, but a quick glance at the lymphocytes at the lower right is reassuring. The
epithelial cell nuclei are only slightly larger than the lymphocyte nuclei, so this is not squamous
dysplasia. Note that the cells at the surface contain bubbly mucin-filled cytoplasm that is bluish,
a feature similar to the mucin in goblet cells, but the mucin does not form a single discreet
droplet in the manner of goblet cells. The appearance of esophageal multilayered epithelium is
similar to that of immature squamous metaplasia of the uterine cervix. This type of epithelium is
found in patients with reflux and can also accompany conventional Barrett mucosa. There is
also some experimental evidence that this type of epithelium gives rise to conventional Barrett
mucosa.21 We do not currently report this finding.
 Figure 1-23. Multilayered epithelium. This is a periodic acid–Schiff/alcian blue stain. Some of
the mucin-laden cells display alcianophilia (have a blue appearance), but their morphology is
different from that of actual goblet cells like those seen in Fig. 1.21.

FAQ: How does Barrett mucosa appear endoscopically?

Answer: Barrett mucosa appears as velvety “salmon-colored” epithelium that

extends as “tongues” above the gastric folds (Figs. 1.24 and 1.25). When an
area of Barrett mucosa is encircled by squamous mucosa, the columnar zone is
referred to as an “island.”
Figure 1-24. Barrett esophagus, endoscopic image. There is a tongue at the lower right, which
means an extension of columnar epithelium above the gastric folds into the tubular portion of
the esophagus. In this image, we are looking down from the incisors and the dark area at the
left center of the image is the stomach. The gray mucosa at the lower left of the image is
squamous mucosa. The word “tongue” is written on the squamous epithelium to indicate the
tongue of columnar epithelium just to the right of the word.
 Figure 1-25. Barrett esophagus, endoscopic image. There are tongues of salmon-colored
columnar mucosa at the top, lower center, and lower left of the image. The gastric folds can be
seen in the center of the image. One area of columnar epithelium is present at the lower right. It
is salmon colored but surrounded on all sides by gray pearly squamous mucosa. This is
referred to as an “island” of columnar epithelium. A biopsy is needed to learn whether this area
has intestinal metaplasia.

Our endoscopy colleagues use a wonderful system to describe the extent of Barrett
esophagus called the Prague system (of course they got to have a fancy meeting in
Prague, probably with nice concerts and fancy food and a posh hotel, whereas we have
to meet in Baltimore or Columbus, but that is fine because we get to sit down when we
do our daily work) in which the distance of the circumferential length of Barrett mucosa
(C) and the maximum length (M) are recorded.22 This method has allowed
standardization of endoscopy reports and is important for pathologists to also
understand. When these data are provided to us, they can help us understand whether the
samples are from long-segment of short-segment Barrett mucosa—long segment means
more than 3 cm. This scheme is seen in Fig. 1.26.
 Figure 1-26. Schematic for the Prague classification for Barrett mucosa. In this scheme, which
is used worldwide by endoscopy colleagues, the distance of the circumferential length of
Barrett mucosa (C) and the maximum length are recorded (M). If the information is provided to
the pathologist, it can be used to decide whether the findings in any given case meet criteria for
Barrett mucosa because a minimal length criterion has been introduced.

FAQ: Are special stains needed to diagnose Barrett esophagus?

Answer: The short answer is “nope,” but, of course, we all occasionally need a
bit of help. However, we would note that, when Panarelli and Yantiss reviewed
the topic, they concluded that neither histochemical nor immunohistochemical
stains add value over H&E stains because they tend to produce false-positive
results.23
In the past, the idea that CK7/CK20 stains could be used to differentiate
esophageal intestinal metaplasia from gastric cardiac intestinal metaplasia was
entertained but did not really catch on.24,25 Long-segment Barrett esophagus
cases (>3 cm) were characterized by superficial and deep CK7
immunoreactivity in the intestinalized mucosa, with only superficial CK20
staining in the intestinalized zones. In contrast, distal gastric intestinal
metaplasia was characterized by patchy, superficial, and deep CK20 staining in
areas of incomplete intestinal metaplasia; strong, superficial, and deep CK20
staining in areas of complete intestinal metaplasia; and patchy or absent CK7
staining in either type of gastric intestinal metaplasia. Many other immunostains
have been studied, including mucin core (MUC) polypeptides, to better
characterize gastric cardiac versus esophageal intestinal metaplasia. The MUC
polypeptides seem to be of little practical value in any given patient and are
presently of academic interest only. They include MUC5 (gastric foveolar
mucin), MUC6 (cardiac glands, antral glands, Brunner glands), and MUC2
(goblet cells). Some have used CDX2 staining to label areas of intestinal
metaplasia.26 Others have used it to note that some cases lacking goblet cells
express these markers (illustrated later) anyway to make the point that the
United States should drop the requirement for goblet cells to diagnose Barrett
esophagus.27 Others have suggested that hepatocyte antigen (Hepar-1,
carbamoyl phosphate synthetase 1) is helpful in detecting intestinal metaplasia
or processes that are intestinalized in the absence of goblet cells.28 It seems
more practical to simply search for goblet cells because that can be done
consistently in any laboratory without the added costs of immunolabeling. In
fact, none of the immunostains intended to detect goblet cells offers added
value over H&E stains.23 Fig. 1.27 was prepared from a consultation case. It is
a CDX2 stain of Barrett mucosa in which every single nucleus is labeled,
including both those of goblet cells and those of gastric foveolar cells.
 Figure 1-27. Esophageal columnar mucosa with scattered goblet cells, CDX2 stain. This is
interesting in that cells that are clearly goblet cells, such as the one near the tip of the villuslike
structure at the upper left, and all other columnar cells are reactive. This finding suggests that
the cardiac-type epithelium already has some traits that overlap with those of the intestinal-type
epithelium. However, at least in a US population, goblet cells are probably a better marker for
patients who require surveillance than columnar epithelium alone.17

GRADING BARRETT DYSPLASIA

This should be a piece of cake and easy, right? In fact, most of the time it is easy
because most samples from Barrett lack dysplasia! However, it is not always easy and
knowing a few tricks and tips can help. It is well known that observer variation is an
issue,29 but we can take steps to reduce this. In the past, at least in Johns Hopkins, we
have reported low-grade dysplasia and “indefinite for dysplasia” too liberally but have
been able to tighten our criteria in the last few years and this has not resulted in patient
harm by missing important lesions (Data unpublished at presstime; search Waters et al.),
but as a general rule, an institution should not be reporting more than 5% low-grade
dysplasia in patients presenting to clinic. However, clinics devoted to dysplasia would
be expected to have a higher percentage of dysplasia cases. Here are the categories that
are used on biopsies30:

Negative for dysplasia

Indefinite for dysplasia
Low-grade dysplasia
High-grade dysplasia
Adenocarcinoma (intramucosal carcinoma for the purpose of this discussion)

In evaluating Barrett mucosa, it is a good idea to have a systematic approach.

Finding goblet cells is important and then the epithelial changes can be studied. Check
for surface maturation and glandular crowding, examine the cytologic features, and
decide if inflammation is an obscuring factor before making a diagnosis.

Barrett Esophagus, Negative for Dysplasia

Nondysplastic Barrett mucosa should feature surface maturation, which can be difficult
to assess in badly embedded samples. It is normal to encounter some degree of nuclear
alterations in the bases of the metaplastic pits and this can be ignored. Remember not to
focus too much on these pits when you see surface epithelium on all sides of a biopsy
fragment and the center portion shows areas with larger nuclei. This is normal for
nondysplastic Barrett mucosa. Ignore it!
The best way to add confidence to diagnosing dysplasia in Barrett mucosa is to
always think about the polarity of the epithelial cells and how they are arranged with
respect to their neighbors. An extremely useful indication of maintained overall cellular
polarity can be referred to as “the four lines.” This appearance indicates preserved
polarity of epithelial cells in both gastric cardiac mucosa and in Barrett mucosa (with
goblet cells). Figs. 1.28–1.30 show cardiac mucosa (Fig. 1.28) and Barrett mucosa
(Figs. 1.29 and 1.30) annotated to demonstrate the four lines at high magnification, but
they can be spotted easily at 4X. The top line consists of the apical mucin cap of neutral
mucin typical of Barrett mucosa (with incomplete intestinal metaplasia) and cardiac
mucosa. The second line is formed by the bases of the mucin caps. The third line is
formed by the aligned cytoplasm and the fourth line by the row of nuclei. These lines are
not the same in the rare cases of complete intestinal metaplasia in the esophagus, but
these lesions are generally readily classified for the purposes of grading dysplasia.
Complete intestinal metaplasia has two lines, one for the cytoplasm and one for the row
of nuclei. Awareness of these patterns makes it easy to dismiss many cases as reactive.
Figs. 1.31–1.33 show cardiac mucosa with nuclear stratification that can be regarded as
reactive. This same principle applies to Figs. 1.34–1.36, a Barrett case (negative for
dysplasia). There is surface maturation, and the lines are preserved despite nuclear
stratification. Using the magic lines as a criterion can help reduce the use of indefinite
for dysplasia and overdiagnosis of low-grade dysplasia.
Figure 1-28. Gastric cardiac epithelium, high magnification. This image shows the four lines.
Their presence indicates that the cells are arranged in an orderly fashion with respect to one
another. This concept is also shown in “Colon” chapter. The top line or row consists of the
apical cap of neutral mucin that characterizes gastric foveolar-type cells. The second line is the
base of the cap, the third line is the row of cytoplasm, and the fourth is the row of well-aligned
nuclei.
Figure 1-29. Barrett mucosa. There is a goblet cell at the upper left of the left villiform structure
and another one at the left side of the villiform structure on the right. Nondysplastic Barrett
mucosa retains the same overall cell-to-cell polarity as that seen in gastric cardiac mucosa.
Again, the top line is the apical mucin cap of the foveolar cells, the second line is the base of
this mucin vacuole, the third line is the cytoplasm, and the fourth line is the row of nuclei. The
goblet cell cytoplasm punctuates but does not disorder this basic polarity.
Figure 1-30. Barrett mucosa. Despite some tangential embedding, the lines are all present,
even though they are broken up into dashed lines by goblet cells. Between the goblet cells, line
one consists of the droplet of gastric foveolar mucin, line two consists of the base of the mucin
cap, line three consists of the cytoplasm of the foveolar-type cells, and line four consists of the
row of nuclei. This is nondysplastic Barrett mucosa. The nuclei appear a bit reactive, as
evidenced by the presence of perfectly round nucleoli, but these nuclei are not hyperchromatic
and their nuclear membranes are smooth.
Figure 1-31. Reactive gastric cardiac mucosa adjoining multilayered epithelium. There is a lot
of multilayered epithelium at the left of the image. Some of the basal zone cells in this area
have an appearance like that of squamous epithelium, but there are surface columnar
characteristics. The cardiac epithelium contains cells with mild nuclear enlargement, but it is
easy to spot the four lines, especially at the right side of the image. This epithelium can
confidently be regarded as reactive.
Figure 1-32. Reactive columnar mucosa. There appear to be a few goblet cells in the gland at
the lower left, so this could be Barrett mucosa or intestinal metaplasia of the cardia depending
on the endoscopic appearance. Importantly, although the nuclei are slightly stratified, the lines
are intact. This is reactive mucosa.
Figure 1-33. Reactive gastric cardiac mucosa. Although some of the nuclei are elongated and
have nucleoli, the cells have preserved relationships to one another and the four lines are intact.
This focus can be regarded as reactive. Note that there is some mitotic activity in the deep part
of the gland at the right. This is acceptable, especially because the nuclei gradually mature
(become smaller) toward the surface.
Figure 1-34. Barrett mucosa, negative for dysplasia. There is plenty of lamina propria between
the glands. The glands at the deep part of the sample have larger nuclei than those seen at the
surface. This is perfectly acceptable. Notice that the four lines are perfectly seen at the surface.
However, note that where there is tangential sectioning, they are more difficult to make out.
Figure 1-35. Barrett mucosa, negative for dysplasia. The lines are essentially intact. The nuclei
become smaller and a bit less open at the surface, where slight tangential embedding slightly
obscures the lines. However, the lines are easy to see on the sides of the glands in this image.
Figure 1-36. Barrett mucosa, negative for dysplasia. Note that the nuclei are smaller and more
condensed at the surface, as they prepare to slough into the lumen.
 Figure 1-37. Barrett mucosa with reactive changes and inflammation, negative for dysplasia.
This image shows a few tricks! The findings are all reactive. Note at the surface that the
columnar epithelium directly adjoining the squamous component is typically a bit
hyperchromatic and slightly disorderly, but as the surface is scanned in the image, the lines are
present everywhere that the epithelium is well oriented. The arrow indicates a deeper gland that
displays slight nuclear enlargement, but it does not “jump out” from the other deep glands.
Spending a lot of time looking at such glands at very high magnification can quickly lead to an
overinterpretation as basal pattern dysplasia.

The case shown in Figs. 1.37–1.40 has reactive features and illustrates another
pitfall. The deep glands have slightly enlarged hyperchromatic nuclei in a context of
inflammation and gradual surface maturation. A TP53 (P53) stain is also shown for this
case and shows a wild-type pattern. Many laboratories endorse the use of TP53
immunolabeling in all Barrett cases, and this is a common practice in Europe. We have
not adopted this practice, but if P53 labeling is always used in your practice, it is
important to think before reacting to light staining. This protein is a tumor suppressor
protein, and thus it is active during cell proliferation and has a quick half-life. Because
it is busy preventing cancer during cell proliferation, a bit of labeling is to be expected
in the proliferative compartment of the mucosa. As such, there is always a bit of nuclear
labeling in the basal layer of the squamous epithelium and in the pits of the stomach.
When the TP53 gene is mutated, this results in a TP53 protein with an extra-long half-
life, so it accumulates in the nuclei of the cells and can be detected by immunolabeling
and is a wonderful marker for dysplasia, but intense staining should be used to confirm
dysplasia. Fig. 1.41 shows an area of intense labeling in a zone that could be interpreted
as basal pattern dysplasia (basal crypt dysplasia31), which is discussed later. The so-
called null pattern of P53 labeling is discussed in the “High-Grade Dysplasia” section.

Figure 1-38. Barrett mucosa with reactive changes and inflammation, negative for dysplasia.
This is a very-high-magnification image of the gland indicated in Fig. 1.37. This was not a good
idea! Mitoses are present (acceptable in deep glands), and the nuclei appear somewhat
jumbled and hyperchromatic out of context.
Figure 1-39. Barrett mucosa with reactive changes and inflammation, negative for dysplasia.
This is another high-magnification image of some of the deep glands in the case seen in Fig.
1.37. The nucleoli in the gland with cells containing abundant eosinophilic cytoplasm and a
somewhat syncytial arrangement have all of the features of a reparative process. This sort of
appearance can be seen after any type of injury (including mucosal ablation). This field shows
reactive nuclei rather than dysplastic ones.
 Figure 1-40. Barrett mucosa with reactive changes and inflammation, negative for dysplasia.
This is a TP53 (P53) immunohistochemical stain. It shows a wild-type (not mutated) pattern
that would be expected in mucosa with reparative/reactive changes. The P53 protein is a tumor
suppressor protein encoded by a tumor suppressor gene. Its job is to suppress neoplastic
transformation while cells are going through the proliferation process every time our mucosa is
made and remade. As such, there should be a little of this protein in cells that are proliferating
or regenerating an area (our mucosa turns over constantly). As such, there is staining in the
basal layer of the squamous epithelium and in the deeper glands. The normal gene produces a
protein that has a short half-life, so we just see just a little bit of it accumulate in proliferating
nuclei. Notice that there is none at the columnar surface, and only the basal squamous nuclei
are labeled. A few of the lymphocyte nuclei are also labeled lightly. When the TP53 gene is
mutated, it produces a protein that is not properly degraded and has a longer half-life, so the
cells with mutant TP53 show prominent nuclear labeling. Do not equate light labeling such as
that in this image as evidence of dysplasia.

Barrett Esophagus Indefinite for Dysplasia

This is everyone’s least favorite category. The term was introduced in a 1983 study of
epithelial changes in inflammatory bowel disease that is further discussed in “Colon”
chapter (which concerns the colon) and was a radical change32 because it was baffling
to our clinical colleagues that sometimes we simply have no idea whether epithelial
changes are dysplastic/neoplastic or not. The term was later used in evaluating
columnar epithelial changes in Barrett mucosa,29,30 and until recently, gastroenterology
societies did not even mention how to address it in guidelines. This category can be
regarded as a holding diagnosis until the findings in the patient are sorted out by a
course of treatment to reduce any inflammation followed by repeat sampling.12
No matter which images we show to illustrate our conception of cases that we
would regard as indefinite for dysplasia, some colleagues will dismiss the illustrated
changes as reactive and others will be concerned that dysplasia has been overlooked.
That is the entire point of this category. It merits follow-up, but the patient should not
receive definitive treatment (mucosal ablation therapy) until the findings are clarified.
Because, by definition, the indefinite category indicates that the pathologist is
uncertain whether the area in question manifests reparative changes or
dysplastic/neoplastic ones, it is difficult to write criteria. However, one can view cases
that show glands with nuclear alterations in the pits that are concerning for dysplasia but
that mature toward the surface as indefinite for dysplasia or lesions that lose the four
lines at the surface but lack nuclear hyperchromasia as indefinite for dysplasia. When
the issue concerns epithelial changes restricted to the bases of the pits, the question
always concerns whether to regard the changes as those of basal pattern dysplasia
(discussed later) or to diagnose them as indefinite for dysplasia. However, the main
issues revolve around inflamed samples with nuclear alterations or poorly embedded
samples showing diathermy (cautery) artifact.
 Figure 1-41. Basal pattern dysplasia, P53 immunostain. There is strong nuclear labeling,
support for dysplasia, in five glands in the lower half of the image. Those in the upper half show
a wild-type pattern. This is from a case of basal pattern dysplasia and included at this point to
compare and contrast with Fig. 1.40.

Figs. 1.42–1.45 are images taken from biopsies from a single patient. Are the
features reactive? In some areas the four lines are present, but in other areas these lines
are jumbled and some of the surface nuclei are arranged in a disorderly fashion. Are the
nuclei hyperchromatic? Maybe. In one area a deeper gland appears hyperchromatic but
perhaps it is crushed. Based on the uncertainty, this process was interpreted as
indefinite for dysplasia. Figs. 1.46–1.51 show similar alterations in the setting of
inflammation that were interpreted as indefinite for dysplasia. There is no shame in
diagnosing lesions as indefinite for dysplasia, but the number can be reduced by
showing colleagues the case, and sometimes the presence of a wild-type P53
immunolabeling pattern can be reassuring. Sometimes fresh eyes can clarify the
findings! This diagnosis should be used in only a small percentage of cases (up to 3% to
5%). Some colleagues (personal communications) essentially never use this category.
Perhaps these colleagues are very good and always know. We wish we were that good!
Figure 1-42. Barrett esophagus, indefinite for dysplasia. In this case, the findings are difficult to
interpret. There is a large complex gland at the right and adjoining squamous epithelium at the
left. Many of the nuclei on the surface at the right are hyperchromatic, but it is not clear if they
are reactive. The lines are obscured at the upper right, but the nuclei are not particularly
enlarged. Perhaps this is all reactive, but it is difficult to be entirely certain.

Figure 1-43. Barrett esophagus, indefinite for dysplasia. This is from the same case as that
seen in Fig. 1.42. One could argue that the lines are intact at the upper right and assume that
the findings at the upper left reflect tangential embedding. However, the image seen in Fig. 1.44
is from the same case as well.
Figure 1-44. Barrett esophagus, indefinite for dysplasia. There is an atypical gland at the left of
center, but the section appears thick in that focus. The overlying nuclei are not particularly
enlarged, but the lines are obscured. Is this reactive? Not sure. Is this dysplastic (adenoma
like)? Not sure.
Figure 1-45. Barrett esophagus, indefinite for dysplasia. This is a higher magnification of the
field seen in Fig. 1.44. The nucleoli in the glands at the left suggest that the findings are
reactive, but the jumbled nuclei on the surface at the left are concerning, but they gradually
merge with an area to the right that has the four lines.
Figure 1-46. Barrett esophagus, indefinite for dysplasia. In this sample, the surface shows
prominent nuclear stratification and the lines are obscured. However, the nuclei are smaller
than the ones in the deep glands. The features are adenoma-like (low-grade dysplasia), but the
key issue is obscuring inflammation. It is no problem to consider high-grade dysplasia in the
setting of lots of inflammation, but obscuring inflammation is problematic for diagnosing low-
grade dysplasia.
Figure 1-47. Barrett esophagus, indefinite for dysplasia. This is a high-magnification image of
the lesion seen in Fig. 1.46. The nucleoli at the surface do suggest that the findings are reactive
in the setting of obscuring inflammation, but the features at low magnification are striking such
that a “treat and repeat” scenario is not unreasonable.
Figure 1-48. Barrett esophagus, indefinite for dysplasia. This image is from a different case
from the one seen in Figs. 1.44–1.47. In this example, inflammation is prominent as is “chatter
artifact,” but there seem to be some surface alterations that are concerning for low-grade
dysplasia.
Figure 1-49. Barrett esophagus, indefinite for dysplasia. This image is from the same case as
the one in Fig. 1.48. In this area, the surface lines are obscured but there are many neutrophils
embedded in the epithelium. The nuclei are dark but not particularly enlarged. Compare their
sizes with those of the inflammatory cell nuclei.
Figure 1-50. Barrett esophagus, indefinite for dysplasia. This is from the same case as that
seen in Figs. 1.48 and 1.49. The specimen is tangentially embedded and there are apparent
surface hyperchromatic nuclei, but it is not clear if in fact the surface is well represented.
 Figure 1-51. Barrett esophagus, indefinite for dysplasia. Although the lines are obscured in this
zone, the adjoining areas seem to have preservation of cell polarity. The nuclei are enlarged but
not crisply distinct from nuclei that appear unremarkable.

Low-Grade Dysplasia
Low-grade dysplasia should be clearly neoplastic (adenoma-like) and should involve
the surface epithelium. It is important not to overdiagnose it because current guidelines
endorse mucosal ablation for low-grade dysplasia and mucosal ablation confers a risk
for stricture formation even in skilled hands. In prospectively evaluated patients the
incidence of low-grade dysplasia should be on the order of 2% to 3% but not more than
5%.33
The nuclei in the cells of low-grade dysplasia are larger than those of normal
Barrett mucosa, and generally, there is little inflammation in samples confidently
diagnosed with low-grade dysplasia. A helpful clue that low-grade dysplasia is present
is that one observes an abrupt transition between the dysplastic zone and adjoining
zones that are clearly not dysplastic. The surface lines that help to confirm
nondysplastic Barrett mucosa are effaced. Classic examples of low-grade dysplasia
with intestinal differentiation can resemble colorectal tubular adenomas. When this
happens, it is important to consider the possibility of a sample switch with an actual
colorectal adenoma before reporting the case. Examples of low-grade dysplasia appear
in Figs. 1.52–1.61. In general, low-grade dysplasia should demonstrate loss of the four
lines (an overall indication of altered cell architecture and cell polarity) but not loss of
nuclear polarity. The long axes of the nuclei should remain more or less perpendicular
to the basement membrane. This general concept is further discussed and illustrated in
“Colon” chapter with the construct of well aligned rows of nuclei. There is some
subjectivity in differentiating low-grade dysplasia from high-grade dysplasia, but this is
less important than it was in the past because all dysplasia are currently managed by
endoscopic ablation, although gastroenterology societies do suggest peer review of
dysplasia cases before mucosal ablation is performed.10,12
Staining for TP53 is sometimes useful in confirming an interpretation of low-grade
dysplasia—a few darkly stained nuclei on the surface can be identified and the basal
glands are labeled. In contrast, alpha-methylacyl-CoA racemase and other markers have
not proven useful, at least in our hands.
Most dysplasia cases show intestinal differentiation in that the epithelium is
stratified and punctuated by goblet cells in a fashion similar to the appearance of
colorectal adenomas, but not all do.34 On the order of 10% of low-grade dysplasia
cases can display gastric-type differentiation, which can manifest either as an
appearance similar to that of gastric pyloric gland adenoma (see “Stomach” chapter) or
a lesion that resembles a gastric foveolar-type adenoma. In such lesions in Barrett
mucosa the background may or may not contain goblet cells but the surface shows
dysplastic-appearing nuclei coated by an apical mucin cap akin to that seen in gastric
foveolar epithelium, except that the lines are absent. This type of dysplasia is
characterized by smaller nuclei than those in intestinal-type dysplasia, but they are
typically slightly hyperchromatic. Examples of this pattern of dysplasia are seen in Figs.
1.62–1.65. Unfortunately, immunolabeling for TP53 is not particularly helpful.
Figure 1-52. Low-grade dysplasia. This case is easy! This lesion is adenoma-like. The surface
is involved, and the nuclei at the surface appear similar to those in the glands. The lines are
completely obscured throughout the process and most of the nuclei are lined up with their long
axes oriented perpendicular to the basement membrane. With a case such as this, some
observers might prefer a diagnosis of high-grade dysplasia, but, importantly, presumably all
observers would regard this process as dysplastic and endoscopic treatment would be offered.
Figure 1-53. Low-grade dysplasia. This is a high-magnification image of the lesion depicted in
Fig. 1.52. Overall, the nuclei are oriented perpendicular to the basement membranes.
Figure 1-54. Low-grade dysplasia. This is a P53 stain from the lesion depicted in Figs. 1.52 and
1.53. It shows a wild-type pattern and was not particularly helpful in confirming the
interpretation.
Figure 1-55. Low-grade dysplasia. Note the abrupt demarcation between the low-grade
dysplasia and the nondysplastic mucosa. Note also that the lines are obscured and most of the
nuclei are oriented in an alignment that is perpendicular to the basement membrane. This latter
alignment is lost in high-grade dysplasia.
Figure 1-56. Low-grade dysplasia. A sharp demarcation is present between cardiac-type and
Barrett mucosa at the left of the image versus the dysplasia at the right.
Figure 1-57. Low-grade dysplasia. This field shows that, despite the stratification and loss of
the overall cellular polarity (loss of the lines), the nuclear polarity is maintained with the
elongated nuclei aligned perpendicularly to the basement membrane.
Figure 1-58. Low-grade dysplasia. This high magnification intends to show the alignment of the
nuclei with respect to the basement membrane.
Figure 1-59. Low-grade dysplasia. This image shows areas with sharp demarcations between
foci of low-grade dysplasia versus nondysplastic epithelium.
Figure 1-60. Low-grade dysplasia. The arrow indicates a zone of demarcation between
dysplastic and nondysplastic epithelium. Note that the dysplastic nuclei lack nucleoli in most
cells.

Figure 1-61. Low-grade dysplasia. This very-high-magnification image shows the nuclear
features of low-grade dysplasia.
Figure 1-62. Low-grade dysplasia. Rare examples of low-grade dysplasia show gastric foveolar
differentiation. The lines are obscured and the nuclei are hyperchromatic, but there are no
goblet cells in the area of dysplasia in contrast to the situation in Figs. 1.52–1.61. This pattern of
dysplasia can be encountered in a background of Barrett mucosa with goblet cells, but the
dysplasia itself shows gastric foveolar differentiation.
Figure 1-63. Low-grade dysplasia. This is a high-magnification image of the foveolar-type
dysplasia seen in Fig. 1.62. Note that the nuclei themselves appear similar to those in the
intestinal-type dysplasia seen in Figs. 1.60 and 1.61.
Figure 1-64. Foveolar pattern dysplasia. This case is a bit controversial. It could be interpreted
as low-grade dysplasia based on the disorderly alignment of the cells, but some cells have lost
their relationship to the membrane (a feature of high-grade dysplasia). This was interpreted as
low-grade dysplasia because the nuclei are not particularly enlarged. There are no goblet cells
in this area.
 Figure 1-65. Foveolar pattern dysplasia. This is a high-magnification image of the lesion seen in
Fig. 1.64.

High-Grade Dysplasia
Generally, high-grade dysplasia is not difficult to recognize and is difficult to overlook.
At low magnification the area appears hyperchromatic and stands out from any
nondysplastic mucosa in the sample. It is generally not particularly inflamed, but even
examples showing inflammation appear extremely hyperchromatic at low magnification.
The alterations usually can be detected in the surface epithelium. Prominent nucleoli are
not a usual feature of high-grade dysplasia, but there are exceptions. In most cases, there
is still plenty of lamina propria between the glands. Figs. 1.66 and 1.67 show a
characteristic example of high-grade dysplasia with all the key features. Although there
are a few neutrophils (indicated), the nuclear hyperchromasia is in excess of that which
can be explained by a reparative process. Furthermore, there are often prominent
nucleoli in a reactive process and the nuclei shown are quite dense appearing. Many
nuclei both in the pits and at the surface have completely lost their relationship to the
basement membrane and have rounded up and are arranged in a jumbled configuration.
A similar lesion is shown in Figs. 1.68 and 1.69. Nuclear hyperchromasia is the key
finding, although the glands are crowded in this example as well. Some examples of
high-grade dysplasia feature markedly enlarged nuclei, sometimes in the absence of
glandular crowding, such as the case shown in Figs. 1.70 and 1.71.

Figure 1-66. High-grade dysplasia. In a case like this, finding acute inflammation (arrow) need
not detract from the diagnosis. In this case, the degree of nuclear hyperchromasia is in excess
of that which can be attributed to inflammation. Note that the nuclei at the surface are
hyperchromatic and many have lost their relationship. There are also many nuclei in the glands
that are hyperchromatic and have rounded up and lost their alignment with the basement
membrane.
Figure 1-67. High-grade dysplasia. This is a high-magnification image of the lesion seen in Fig.
1.66. The rounded hyperchromatic nuclei are the key feature. Note that nucleoli are not a
prominent feature.
Figure 1-68. High-grade dysplasia. The key finding is striking nuclear hyperchromasia.
 Figure 1-69. High-grade dysplasia. This is a P53 immunostain. This strong diffuse nuclear
labeling pattern is characteristic of high-grade dysplasia. Some examples of low-grade
dysplasia show strong labeling, but usually just a few surface cells are labeled. In this example,
the surface nuclei are strongly reactive.

Most cases of high-grade dysplasia show some degree of nuclear elongation and
stratification akin to the features in colorectal adenomas. A subset of cases shows an
unusual pattern of small tubules, each lined by a monolayer of hyperchromatic nuclei
(Figs. 1.72 and 1.73). Because the monolayer appearance is unusual and lacks nuclear
stratification, it has been referred to as a “nonadenomatous” form of high-grade
dysplasia by some,35 but others have considered this pattern as evidence for gastric
differentiation in high-grade dysplasia.36,37 None of this has any effect on management,
but such cases are still part of high-grade dysplasia.
Figure 1-70. High-grade dysplasia. Even though there is no glandular crowding, the giant
hyperchromatic nuclei in some of the glands and on the surface at the left part of the image are
sufficient for a high-grade dysplasia diagnosis.
Figure 1-71. High-grade dysplasia. This is a very-high-magnification image of the lesion
depicted in Fig. 1.70. The nuclei are extremely hyperchromatic. Compare their sizes (including
the smaller ones) to the sizes of the lamina propria inflammatory cells.
Figure 1-72. High-grade dysplasia. Although the surface shows loss of nuclear polarity and
somewhat stratified nuclei, the arrow indicates some deep glands that also show high-grade
dysplasia in a pattern consisting of small tubules each lined by a monolayer of tiny
hyperchromatic nuclei that are not elongated like those of classic dysplasia. Because of the
lack of nuclear stratification, some colleagues have referred to this pattern as
“nonadenomatous” and others have regarded it as a form of gastric-type differentiation. The
main thing is to be aware that some examples of high-grade dysplasia lack nuclear
stratification.
 Figure 1-73. High-grade dysplasia. This is a high-magnification image of the deep dysplastic
glands seen in Fig. 1.72.

Immunolabeling for TP53 can be helpful in confirming an impression of high-grade

dysplasia when there is doubt. If the histologic features are already classic, there is no
reason to perform immunolabeling, but about 85% to 90% of the time, if
immunolabeling is added, strong labeling like that seen in Fig. 1.69 will confirm the
impression. Thankfully, such strong labeling is typically present in the “small cell”
pattern depicted in Figs. 1.72 and 1.73. As noted earlier, TP53 is a tumor suppressor
protein that is active during cell division as a normal component and has a short half-
life. Because TP53 is a molecule put in place to prevent cancer development during the
course of normal cell proliferation, some labeling is to be expected in the proliferative
compartment of the mucosa. As such, there is always nuclear labeling in the basal layer
of the squamous epithelium and in the pits of the stomach. When the TP53 gene is
mutated, this usually results in a TP53 protein with an extra-long half-life, so it
accumulates in the nuclei of the cells and can be detected by immunolabeling. However,
there is a subset of high-grade dysplasia cases in which there is biallelic loss of the
TP53 gene. In this form of dysplasia, the cells have absolutely no TP53 to depend on
and there is complete absence of staining in the dysplastic nuclei. This pattern has been
termed the “null pattern” and can be exploited for diagnosis just as well as the strongly
reactive pattern. The dysplasia/neoplasia cases shown in Figs. 1.74–1.78 display the
null pattern of TP53 labeling that confirms the diagnosis. In addition, Figs. 1.77 and
1.78 show “buried dysplasia.” This means that there is a coating of squamous
epithelium on top of the dysplasia. There have been articles that express concern that
this is a huge issue in patients who have had mucosal ablation procedures but it is not.
We would not have this image if our endoscopy colleagues did not find and biopsy this
area. Furthermore, there is nearly always separate surface dysplasia in patients with
buried dysplasia and exceptions are rare.38 The “Endoscopic Mucosal Resections”
section also features examples of buried dysplasia.
The line between high-grade dysplasia and early carcinoma can be difficult to draw,
and some features that are helpful in confirming an interpretation of intramucosal
carcinoma are discussed later. We diagnosed the lesion seen in Figs. 1.79–1.81 as high-
grade dysplasia rather than early carcinoma, but the irregular growth pattern of the
dilated deeper glands could be interpreted as evidence of early invasion.39
Lastly, there are cases of high-grade dysplasia that display either gastric foveolar
differentiation or differentiation similar to that seen in pyloric gland adenomas of the
stomach, as discussed in “Stomach” chapter. This type of dysplasia can arise in a
background of intestinal metaplasia, but it still has gastric-type differentiation.34 The
important point is that the nuclei in this pattern of dysplasia are paler than those seen in
conventional-type dysplasia and this form of dysplasia is diagnosed by attention to
nuclear multilayering and loss of nuclear polarity (Figs. 1.82–1.85). For this pattern,
TP53 immunolabeling often fails to add information.
Figure 1-74. Intramucosal carcinoma. Such cases are controversial, and some colleagues
might regard this lesion as high-grade dysplasia. The glands in the center seem to be growing
parallel to the surface, and the luminal necrosis in the center is a feature of the earliest invasion
into the lamina propria. However, this case is shown because it has a special pattern of P53
labeling, as seen in Fig. 1.75.
Figure 1-75. Intramucosal carcinoma. This is a P53 stain from the lesion seen in Fig. 1.75.
There are cardiac glands in the lower left of both images, but a few additional
nondysplastic/nonneoplastic glands have popped into the center of the field. These cardiac
glands all show light nuclear wild-type P53 labeling. The dysplasia/early carcinoma, however,
shows complete absence of P53 labeling, which is just as useful for confirming a diagnosis as
finding strong labeling. This finding indicates that the TP53 gene and any P53 protein it might
have produced to keep neoplasia at bay has been wiped out on both copies of the gene in the
neoplastic cells. This can be from deletions/allelic loss. Biallelic inactivation of the gene is at
work in the dysplasia.
Figure 1-76. Intramucosal carcinoma. This is a very-high-magnification image of the lesion
seen in Figs. 1.74 and 1.75. The two glands at the bottom of the image have luminal necrosis.
Figure 1-77. Intramucosal carcinoma. This example is seen beneath reactive squamous
epithelium with reactive (pseudoepitheliomatous) squamous epithelial changes. The glands
contain luminal debris and several of the nuclei contain large nucleoli, a feature of early
invasion. This area was detected endoscopically, even though it is coated with squamous
epithelium. Case reports of “buried” neoplasia after mucosal ablation raised initial concerns, but
this concern is overstated based on accumulated data.
Figure 1-78. Intramucosal carcinoma. This P53 stain shows labeling in reactive squamous
epithelium and a few reactive cardiac-type glands but is entirely nonreactive in the large
dysplastic/neoplastic nuclei such that it is a helpful stain in confirming the impression of
neoplasia.
Figure 1-79. High-grade dysplasia versus intramucosal carcinoma. The glands at the right and
left and surface of the lesion in this image are composed of cells with hyperchromatic enlarged
nuclei. In contrast, there are some dilated glands in the center that grow in an abnormal
arrangement and are lined by cells with small hyperchromatic nuclei, a pattern of high-grade
dysplasia. Some would regard this pattern as that of intramucosal carcinoma because the
glands are convoluted and growing parallel (instead of perpendicular) to the surface. In the
modern era of endoscopic treatment of both dysplasia and early carcinoma, this distinction is
not critical.
Figure 1-80. High-grade dysplasia versus intramucosal carcinoma. This is a high-magnification
image of the area in Fig. 1.79. We report this pattern as intramucosal carcinoma, but it is
acceptable to report it as high-grade dysplasia to forestall overtreatment in centers with
overzealous surgeons.
Figure 1-81. High-grade dysplasia versus intramucosal carcinoma. This is a higher-
magnification image of the key area seen in Figs. 1.70 and 1.80. The nuclei are abnormal and
hyperchromatic in a gland with an abnormal infiltrative growth pattern.
Figure 1-82. High-grade dysplasia, foveolar differentiation. Many of the cells in this image have
apical mucin caps in the fashion of gastric foveolar cells. This example is easy to diagnose as
high-grade dysplasia because the nuclei are enlarged and hyperchromatic.
Figure 1-83. High-grade dysplasia, foveolar differentiation. There are no goblet cells, and there
are cells with gastric-type mucin at the left.
Figure 1-84. High-grade dysplasia, foveolar differentiation. Cases like this are tricky to diagnose
because the nuclei are not hyperchromatic. Many cells have gastric-type mucin, but the key to
diagnosis is noting the loss of nuclear polarity.
 Figure 1-85. High-grade dysplasia, foveolar differentiation. This high-magnification image shows
the gastric foveolar-type mucin to advantage. The nuclei are jumbled and some have nucleoli.

PEARLS & PITFALLS: Basal “Crypt” Dysplasia

This is a controversial area. This pattern was described in 2006 using the
above mentioned terminology (the glands in the stomach are pits and glands
rather than crypts, but because the mucosa is intestinalized the term “crypts”
makes sense).31,40 This pattern of course makes sense. Biology is a continuum
such that we would expect that occasional cases of dysplasia would be
sampled before the findings have appeared on the surface. The trouble is that it
is subjective to differentiate epithelial changes that simply reflect reactive
glands (Figs. 1.37–1.40) from cases in which there is dysplasia without a
surface component. Often these issues can be resolved by adding a few
recuts. We will admit to being inconsistent in how we handle such cases. Using
the indefinite category for such lesions will result in follow-up with resampling
and using the basal dysplasia category may result in mucosal ablation, so
caution is advised. In some cases, the nuclear alterations are striking and akin
to those of classic high-grade dysplasia, and these cases can be diagnosed as
“high-grade dysplasia, basal pattern” with no qualms concerning whether the
patient receives mucosal ablation. However, it is unclear which lesions are low-
grade dysplasia, basal pattern. In the initial study, the authors suggested
“lumping” these cases in the low-grade category, but at the time of the
publication, universal ablation for low-grade dysplasia had yet to be endorsed,
so the stakes were low. For this reason, some observers prefer not to use this
category.
These are the cases for which P53 immunolabeling can help for a diagnosis
of dysplasia. Without strong labeling, we are more likely to classify the findings
as indefinite for dysplasia. Some examples of cases are shown in Figs. 1.41
1.86–1.91. Fortunately this issue is not common. Our advice is to err on the
side of indefinite for dysplasia unless the nuclear alterations are really striking,
especially in the absence of P53 immunolabeling. The patient will be followed
but not ablated. Here are some approaches:
1. Atypical basal glands with surface maturation, slight nuclear alterations—
Barrett mucosa, negative for dysplasia
2. Atypical basal glands with surface maturation, prominent atypical nuclei,
gradual maturation to surface and gradual transition to glands that are
clearly reactive, wild-type P53 labeling—Barrett mucosa, negative for
dysplasia
3. Atypical basal glands with surface maturation, prominent atypical nuclei,
abrupt transition to more mature nuclei with maturation at surface, wild-
type P53 labeling—Safer to regard as indefinite for dysplasia to avoid
overtreatment
4. Atypical basal glands with surface maturation, prominent atypical nuclei,
abrupt transition to more mature nuclei with maturation at surface,
abnormal P53 labeling—basal pattern dysplasia. If there is no loss of
nuclear polarity in the glands in question—low grade. If there is loss of
nuclear polarity in the glands in question—then high-grade basal pattern
dysplasia
Figure 1-86. Basal pattern dysplasia. Note that this sample is badly embedded and that there is
surface epithelium on all sides. Most of the surface epithelium has the characteristic lines of
nondysplastic Barrett mucosa, but at the upper right, this is not clear. There is a zone of glands
with hyperchromatic nuclei in the lower left, but the surface is not fully evaluable in this zone.
Figure 1-87. Basal pattern dysplasia. This is a P53 stain from the case shown in Fig. 1.86. A
zone of strong nuclear labeling conforms to the area in Fig. 1.86. Note that several glands at the
right of the field have wild-type staining. This pattern altogether can be diagnosed as basal
pattern dysplasia, but it acceptable to diagnose it as indefinite for dysplasia, which will prompt
close follow-up and additional sampling.
Figure 1-88. Basal pattern dysplasia. This is a high-magnification image of the area seen in
Figs. 1.86 and 1.87. The nuclei are enlarged and hyperchromatic compared with the nuclei in
the gland to the lower right. This lesion could be interpreted as low-grade dysplasia, basal
pattern. Were there striking loss of nuclear polarity and striking nuclear enlargement, this would
be high-grade dysplasia, basal pattern. However, if there is any doubt, the indefinite category
should be used because a dysplasia diagnosis (low or high grade) often results in mucosal
ablation.
Figure 1-89. Basal pattern dysplasia. This is a high-magnification image of the P53 preparation
from the case seen in Figs. 1.86–1.88.
Figure 1-90. Intramucosal carcinoma with apparent surface maturation. There is a carcinoma
invading the lamina propria at the left side of the tissue fragment. Seeing a mature surface in
this zone may simply mean that the lesion has spread laterally and an involved surface a small
distance away was simply not sampled. Recuts of apparent basal pattern lesions sometimes
show zones of surface involvement.
 Figure 1-91. Intramucosal carcinoma with apparent surface maturation. This is a high-
magnification image of the lesion seen in Fig. 1.90. Atypical glands with angulated contours and
luminal necrotic debris are present in the lamina propria beneath a surface of nondysplastic
Barrett mucosa.

Intramucosal Adenocarcinoma and Adenocarcinoma

In theory it is not possible to diagnose intramucosal carcinoma on mucosal biopsy
samples because the tissue beneath the mucosa has not been evaluated, but in practical
terms, one can have a good idea based on the findings in the mucosal biopsy alone!
Because of this, we do render diagnoses of intramucosal carcinoma (invasion into the
lamina propria) on mucosal samples, in part because we know that this interpretation
will prompt an EMR and if deeper invasion is found at this time, an esophagectomy will
follow. It is always nice to give a patient a chance at endoscopic treatment! Some
institutions diagnose adenocarcinoma with a note that the patient should be evaluated
clinically to determine the depth of invasion and whether the lesion is amenable to
endoscopic treatment. Either approach is reasonable depending on who might read the
report, but we prefer the first, because it could prevent overtreatment by some surgeons
who continue to practice using standards from an era when endoscopic treatments were
not widely available and accepted in the United States.
Intramucosal carcinoma indicates invasion into the mucosa and is staged as T1a1,
which differs from the approach in the colon wherein mucosal invasion is staged as Tis
because there is negligible lymphatic access in the lamina propria of the colon, as
discussed further in “Colon” chapter. Esophageal carcinomas that invade the submucosa
are staged as T1b1. The problem with intramucosal carcinoma is that desmoplasia is not
well developed. Features associated with early invasion include luminal necrosis of the
glands, an architecture with back-to-back glands, the presence of nucleoli in atypical
glands, and glands that grow parallel to the surface.41,42 Examples of intramucosal
carcinoma appear in Figs. 1.92–1.99, illustrating the criteria noted earlier. After a
diagnosis of intramucosal carcinoma is given on mucosal biopsies, the current standard
of care is mucosal resection and ablation, which is discussed later. Intramucosal
adenocarcinoma is adenocarcinoma, it is simply T1a and thereby usually associated
with a good outcome following endoscopic treatment.43 Even some lesions that invade
the superficial submucosa can be managed with endoscopic treatments.44,45
There are two hints that deep invasion is present when a mucosal biopsy is
reviewed.
The first: If there is prominent desmoplasia (Fig. 1.100), invasion into at least the
submucosa is likely. Desmoplasia essentially means scarring and is discussed in more
detail in the “Colon” chapter.
Figure 1-92. Intramucosal carcinoma. The glands are crowded together and have effaced the
lamina propria. Many contain luminal debris. Some contain more open nuclei than those
encountered in high-grade dysplasia. This open appearance is imparted by nucleoli.
Figure 1-93. Intramucosal carcinoma. Many of the nuclei have prominent nucleoli. This is a
high-magnification image of the lesion seen in Fig. 1.92.
Figure 1-94. Intramucosal carcinoma. This example is coated by squamous epithelium (but the
endoscopist still spotted this area and sampled it). Although desmoplasia is not well developed,
the glands grow in an angulated manner and several run parallel to the surface. The carcinoma
is seen in the lamina propria. The wisps of muscle below it are wisps of muscularis mucosae,
probably duplicated muscularis mucosae.
 Figure 1-95. Intramucosal carcinoma. It is difficult to determine where one glands starts and
another stops in this lesion that effaces the lamina propria.

The second: If there is pagetoid extension of single adenocarcinoma cells into the
squamous epithelium (Figs. 1.101 and 1.102), there is invariably an associated deeply
invasive underlying carcinoma, but this pattern is uncommon in biopsies.42,46
Most adenocarcinomas that have invaded beyond the lamina propria are
straightforward to diagnose and not subtle. They can have mucinous and signet cell
patterns akin to those in gastric carcinoma, and some are poorly differentiated and can
require immunolabeling to be classified (Figs. 1.103–1.105). At present, HER2 testing
is added for esophageal adenocarcinomas and the scheme for scoring is the same as that
for gastric carcinomas and addressed in detail in “Stomach” chapter.
Figure 1-96. Intramucosal carcinoma. This lesion has features of high-grade dysplasia with
hyperchromatic nuclei, but some of the glands bud off from larger ones, some have luminal
necrosis, and some grow parallel to the surface. Cases such as this are often interpreted as
high-grade dysplasia, which is acceptable for the purposes of modern treatments, as lesions
such as this are unlikely to metastasize.
Figure 1-97. Intramucosal carcinoma. This is a high-magnification image of the lesion seen in
Fig. 1.96. It is the complex architecture of the process that merits an interpretation of
intramucosal carcinoma rather than high-grade dysplasia. The glands do not form individual
tubules but instead interanastomose.
Figure 1-98. Intramucosal carcinoma. This is another high-magnification image of the lesion
shown in Figs. 1.96 and 1.97 intended to show the luminal necrosis at the lower right and the
abnormal complex architecture of the glands.
Figure 1-99. Intramucosal carcinoma. This is a CDX2 stain from the case seen in Figs.
1.96–1.98. There was no reason to perform this labeling—the laboratory that handled the
sample was in the habit of adding CDX2 labeling to all esophageal biopsies. The strong nuclear
labeling certainly supports that the lesion displays intestinal differentiation.
Figure 1-100. Adenocarcinoma. This adenocarcinoma is accompanied by desmoplasia
(scarring in response to the lesion). This finding on a biopsy sample suggests that there may
be deeper invasion than into the lamina propria.
Figure 1-101. Adenocarcinoma. In this case, there is pagetoid extension of single
adenocarcinoma cells into accompanying squamous mucosa in the sample. This finding
suggests that a deeply invasive carcinoma is present and is an ominous sign.
Figure 1-102. Adenocarcinoma. This is PAS/AB stain from the case seen in Fig. 1.101. The
PAS highlights glycogen in the superficial squamous epithelium (the basal cells lack glycogen),
but the adenocarcinoma cells contain bluish alcian blue–reactive mucin.
Figure 1-103. Poorly differentiated carcinoma associated with columnar epithelium in the
esophagus. This lesion has a suggestion of gland formation (adenocarcinoma), but it is difficult
to assure that it is not a squamous cell carcinoma.
Figure 1-104. Poorly differentiated carcinoma. This is a high-magnification image of the lesion
seen in Fig. 1.103. The cells are arranged in sheets and some have prominent nucleoli.
 Figure 1-105. Poorly differentiated carcinoma. This strongly reactive BEREP4 stain supports an
interpretation of poorly differentiated adenocarcinoma. This staining was performed to
determine if HER2 testing should be done. HER2 testing is discussed in the “Stomach”
chapter, but the same criteria are used for gastric and esophageal adenocarcinomas.

FAQ: How are dysplasia and early carcinoma managed?

Answer: Before we discuss EMR samples, it is worth considering how

dysplasia and early carcinomas are managed. The 2016 American College of
Gastroenterologists suggestions for management and follow-up of Barrett
mucosa and estimates for likelihood of progression to carcinoma are given in
Table 1.1. Essentially, all dysplasia is currently managed by mucosal ablation,
including low-grade dysplasia, so endoscopy societies have suggested
expert pathology review before treatment. Unfortunately, criteria for whom to
regard as an expert are not well established,47 but interobserver variability can
be an issue in assessing dysplasia no matter who is reviewing such that it is
always a good idea to review cases with colleagues29; it is seen as prudent to
have peer review as a matter of course before performing mucosal ablation.
Generally, flat/invisible dysplasia is ablated using radiofrequency ablation,
whereas dysplasia that forms a visible lesion is resected endoscopically either
 with EMR or endoscopic submucosal dissection. The latter technique provides
a better en bloc sample that typically has negative lateral margins, but the
former is quicker and several adjoining areas can be removed piecemeal.48
Most mucosal ablation is performed using radiofrequency ablation,49 but there
are a few other techniques that can be used, including cryotherapy.50

TABLE 1.1: American College of Gastroenterology (ACG)

Estimated
Likelihood of
Dysplasia Grade Further Documentation Follow-up, ACG/AGA Progression

None None 3–5 y 0.2%–0.5%/year

initially, later up to
9%

Indefinite Repeat after optimization of acid 12 months Data unclear

suppression in 3–6 months. If
another indefinite, follow up

Low grade Expert confirmation, ablation Every 6 months during 0.7%/year

the first year, then
annually

High-grade or Expert confirmation. EMR Every 3 months for the 7%/year (HGD)
intramucosal recommended if biopsies are first year, every 6
carcinoma taken from an area of mucosal months for the second
irregularity coupled with RFA year, then annually

EMR, endoscopic mucosal resection; HGD, high-grade dysplasia; RFA, radiofrequency ablation.
Adapted from Shaheen NJ, Falk GW, Iyer PG, Gerson LB; American College of Gastroenterology. ACG
clinical guideline: diagnosis and management of Barrett’s esophagus. Am J Gastroenterol.
2016;111(1):30-50.
Endoscopic Mucosal Resection Specimens
EMR specimens are the more common samples received compared with submucosal
dissections and are more prone to interpretation issues because they are often simply
tossed into a jar of formalin, which causes them to curl. Ideally, they should be pinned
to a corkboard or wax board, but they are often too tiny to pin without ruining them. The
main thing is to have a way to orient them before they are sectioned. Ideally, they are
breadloafed and embedded so that each slice shows the surface mucosa and some deep
submucosa.44 It can be helpful to ink the deep (submucosal) margin, but if this is not
done, diathermy (cautery) is usually apparent to offer a clue about the location of the
margins.
There are two forms of artifact that are often encountered in EMRs, and once they
are understood, most EMRs can be readily assessed. The first is that the surface is often
iatrogenically damaged. This is because, to perform an EMR, often the surface mucosa
is sucked into a plastic cap after injecting the submucosa to create a polyplike lesion
that is then removed by polypectomy. As such, the surface is rubbed and sometimes the
most superficial epithelium is denuded. The second artifact is that, once the sample is
removed and tossed into fixative, the sample curls as noted earlier. This gives an
appearance that the lateral (mucosal) margin is a deep margin.
It is especially important when evaluating EMRs to be aware that, in patients with
Barrett mucosa, the muscularis mucosae becomes disorganized and duplicated or even
triplicated in response to cycles of damage and repair.2,51-53 This can easily result in
misinterpretation of lamina propria as submucosa and overstaging of the lesion as T1b
when the stage is in fact T1a. Remember that T1a indicates invasion into lamina propria
and T1b indicates invasion into submucosa.1 The space between the duplicated and
original muscularis mucosae is still lamina propria!!! A number of EMR samples,
some annotated to highlight the issue with the muscularis mucosae, are shown in Figs.
1.106–1.120.
For clinical treatment purposes, it does not matter which layer of the mucosa is
invaded, but clinical colleagues like to know a depth of invasion because they like a
number and a quantitation. They also like to know the status of the sample margins.
There are some formalized schemes for reporting this, but in our experience, it is safest
to simply describe the depth of invasion in words. For example, “the tumor invades into
the space between the initial and duplicated muscularis mucosae.” Table 1.2 shows the
various schemes that have been developed, and several authors have used “m1-m3” or
something similar to account for the various depths that a tumor can invade the mucosa.
We usually avoid the classifications with “m1-m3” or “m1-m4” because an oncologist
invariably misinterprets the “m” for “mucosal invasion” as the “M” for “metastasis” in
the staging manual. Argh.

Figure 1-106. Endoscopic mucosal resection (EMR) specimen. This image illustrates some
key issues in interpreting these samples. There is a lesion present in the upper right portion of
the specimen. Note that the right lesion is covered by squamous epithelium (buried or
pseudoregression pattern). The sample was not pinned to a corkboard such that it has curled,
but it is well embedded. A thick black line separates the muscularis mucosae from the
submucosa. The original muscularis mucosae is annotated, as is the duplicated muscularis
mucosae. The loose tissue between these two muscle layers in not submucosa, it is part of the
lamina propria! Note also that the surface epithelium appears damaged—this is probably from
the procedure to perform the EMR, in which a cap is applied to the lesion’s surface and suction
is applied. In this example, the muscularis mucosae layers are so thick that the endoscopist
barely obtained any submucosa.
Figure 1-107. Endoscopic mucosal resection (EMR) specimen. This is a higher-magnification
image of the right side of the lesion seen in Fig. 1.106. There is an intramucosal carcinoma that
invades the lamina propria. The squamous epithelium and the accompanying lamina propria
has curled such that the lateral margin masquerades as a deep margin. The actual deep
margin does not begin until the middle of the image, where a submucosal vessel is seen; the
muscularis mucosae is draped over it!
Figure 1-108. Endoscopic mucosal resection (EMR) specimen. This is a higher-magnification
image of the lesion seen in Figs. 1.106 and 1.107. The intramucosal carcinoma consists of
glands with luminal necrosis and angulated growth. Note that the surface at the left part of the
image has been altered during the EMR operation itself.

Figure 1-109. Endoscopic mucosal resection (EMR) specimen. In this example, the junction
between the mucosa and submucosa is delineated with a thick black line. It drapes over some
submucosal glands; their presence is proof that the cordoned off area is indeed submucosa!
The thick green line thus is the true deep (submucosal) margin, and the rest of the apparent
“deep” area is in fact lateral margin! The squamous epithelium itself has curled around to the
apparent deep margin. This is an artifact of tissue retraction associated with diathermy
(cautery) used in the EMR. In the top center, a few glands are arranged parallel to the surface in
the lamina propria (not the submucosa).
Figure 1-110. Endoscopic mucosal resection (EMR) specimen. This is a higher-magnification
image of the left side of the tissue seen in Fig. 1.109. The thick black line separates the
mucosa and submucosa. The thick muscularis mucosae has draped over the submucosal
gland to the left. The deep (submucosal) margin is indicated by the thick green line, but the
apparent deep tissue inked with black India ink is in fact a lateral margin consisting of curled
squamous epithelium and muscularis mucosae. There is an intramucosal carcinoma invading
into duplicated disorganized muscularis mucosae at the upper right of the field.
Figure 1-111. Endoscopic mucosal resection (EMR) specimen. This is a high-magnification
image of the lateral margin seen at the lower left of Fig. 1.110. Note the muscularis mucosae
curling over the submucosal gland at the upper right and painted with India ink at the bottom of
the image.
Figure 1-112. Endoscopic mucosal resection (EMR) specimen. This is a high-magnification
image of the lesion seen in Figs. 1.109–1.111. The intramucosal carcinoma has invaded into
the disorganized muscularis mucosae at the upper right of the image and is in the space
between the original muscularis mucosae and the superficial duplicated muscularis mucosae.
Note that there is no desmoplastic response to the invasion in this intramucosal carcinoma. In
this case, it is easy to see the extent of the lesion because this is an EMR sample but this field
should serve as a caution—were this a small biopsy, it would be easy to incorrectly assume
that there was submucosal invasion and that the smooth muscle at the bottom was muscularis
propria. However, a moment’s thought should rectify this idea because the muscularis propria
is practically never seen on mucosal biopsies of columnar esophagus.
Figure 1-113. Endoscopic mucosal resection (EMR) specimen. This example was pinned well
and has not curled much. As such, the bottom of the sample is submucosa and the thick band
of smooth muscle near the bottom is the muscularis mucosae. There is additional disorganized
smooth muscle above the thick muscularis mucosae.
Figure 1-114. Endoscopic mucosal resection (EMR) specimen. This is a higher-magnification
image of the EMR seen in Fig. 1.113. This particular EMR lacks dysplasia but shows only
reactive Barrett mucosa. However, it gives a good look at the thickened original (lower part of
field) muscularis mucosae and disorganized duplicated muscularis mucosae in the center of
the image above the original muscularis mucosae.
Figure 1-115. Endoscopic mucosal resection (EMR) specimen. This EMR is curled and
contains a carcinoma that has invaded the space between the original and duplicated
muscularis mucosae (T1a lesion). A thick black line separates the mucosa and submucosa. A
thick green line shows the deep (submucosal) margin, and thick blue lines show the lateral
margins. There is intramucosal carcinoma at the lateral (mucosal) margin on the left. This is
not at the deep margin! This is a common occurrence because endoscopists often resect
lesion by means of several side-by-side EMRs. In this case, much of the lesion is buried under
squamous epithelium but the endoscopist found it anyway.
Figure 1-116. Endoscopic mucosal resection (EMR) specimen. This EMR contains an
intramucosal carcinoma that has invaded nearly to the submucosa. A black line separates the
mucosa and submucosa. A thick green line shows the deep (submucosal) margin, and a blue
line shows the lateral margin. This T1a neoplasm is present at the lateral margin.
Figure 1-117. Endoscopic mucosal resection (EMR) specimen. This is a high-magnification
image of the involved lateral margin from the lesion seen in Fig. 1.115. The thick black line
shows the demarcation between the mucosa and submucosa. A submucosal gland is present
at the right side of the image beneath the black line. The green line shows the deep
(submucosal) margin, and the blue line shows the lateral (mucosal) margin. The muscularis
mucosae can be seen draping over a large submucosal vessel in the lower central part of the
field.
Figure 1-118. Endoscopic mucosal resection (EMR) specimen. This EMR shows a carcinoma
that invades the submucosa in the center of the image. In this case, it is worthwhile to measure
the depth of invasion starting from the bottom of the original muscularis mucosae.
Figure 1-119. Endoscopic mucosal resection (EMR) specimen. This is a higher-magnification
image of the lesion seen in Fig. 1.118. A bit of muscularis mucosae is seen in the center of the
image, and the carcinoma at the left is invading into the submucosa. Large submucosal
vessels are seen at the lower right.
 Figure 1-120. Endoscopic mucosal resection (EMR) specimen. A carcinoma has entered the
submucosa. Note the large thick vessels in this space!

When tumors invade the submucosa, it is worthwhile to provide a depth of invasion

as measured from the base of the muscularis mucosae because these items are in the
balance of whether to perform a follow-up esophagectomy. A depth of more than half a
millimeter (500 µm) prompts discussion of esophagectomy unless the patient has
significant comorbidities.45 Finding a poorly differentiated component (lots of single
cells rather than cells forming glands) or vascular space invasion also prompts this
discussion. Such decisions are best made with a clinical team and in light of known
comorbidities in the patient.

TABLE 1.2: Reported Subclassification Schemes for Intramucosal

Carcinoma (T1a)
 Designation,
Descriptive Designation Westerterp Designation, Designation,
of Depth of Invasion Scheme 54 Vieth Scheme 55 Kaneshiro Scheme 52

None (Tis, high-grade m1 HGD HGD

dysplasia, HGD)

Tumor cells invading beyond m2 m1 LP

basement membrane into
lamina propria

Tumor cells invading (inner) m2 m2 IMM

duplicated muscularis
mucosae

Tumor cells in the space m2 m3 BMM

between the duplicated
muscularis mucosae and
original muscularis
mucosae

Tumor cells into (outer) m3 m4 OMM

original muscularis
mucosae

SAMPLE NOTE: Carcinoma Confined to the Lamina

Propria
Esophagus, 35 cm (EMR): Intramucosal carcinoma in a background of Barrett mucosa
(T1a). The carcinoma invades into the lamina propria. No vascular space invasion seen.
No poorly differentiated component seen. The lesion is present at the lateral (mucosal)
margin. The deep (submucosal) margin is uninvolved.

SAMPLE NOTE: Carcinoma Invading Into the

Submucosa
Esophagus, 37 cm (EMR): Moderately differentiated adenocarcinoma invading into the
superficial submucosa (T1b) in a background of Barrett mucosa with extensive high-
grade dysplasia. No poorly differentiated component seen. No vascular space invasion.
Carcinoma invades to a depth of 0.4 mm measured from the base of the muscularis
mucosae and is 1 mm from the deep margin. The deep (submucosal) and lateral
(mucosal) margins are uninvolved.

Dysplasia Recapitulation
Composite Fig. 1.121 summarizes the key issues of Barrett dysplasia grading:

Negative for dysplasia—There is surface maturation, preserved overall cell polarity

(the four lines), and maintained nuclear polarity (long axes of nuclei are
perpendicular to the basement membrane). Lamina propria is plentiful.
Indefinite for dysplasia—There is equivocal surface maturation, equivocally preserved
overall cell polarity (the four lines), and maintained nuclear polarity (long axes of
nuclei are perpendicular to the basement membrane). Lamina propria is plentiful.
Inflammation is often a factor in using the indefinite category.
Low-grade dysplasia—The surface shows areas that lack surface maturation, and there
is loss of overall cell polarity (the lines are obscured). Nuclei are hyperchromatic
but have overall maintained polarity. Lamina propria is plentiful. Cases with gastric-
type differentiation are different and discussed in the earlier text.
Basal pattern dysplasia—There are nuclear features of dysplasia restricted to the pits.
Often recut sections will disclose zones of surface alterations.
 Figure 1-121. Summary of Barrett esophagus and neoplasia. A: Negative for dysplasia—there
is surface maturation, preserved overall cell polarity (the four lines), and maintained nuclear
polarity (long axes of nuclei are perpendicular to the basement membrane). Lamina propria is
abundant. B: Barrett mucosa with reactive changes. The polarity of the cells is maintained (the
four lines), and deeper glands may display mild nuclear alterations. C: Indefinite for dysplasia—
there is equivocal surface maturation, equivocally preserved overall cell polarity (the four lines),
and maintained nuclear polarity (long axes of nuclei are perpendicular to the basement
membrane). Lamina propria is plentiful. Inflammation is often a factor in using the indefinite
category. D: Low-grade dysplasia—the surface shows areas that lack surface maturation, and
there is loss of overall cell polarity (the lines are obscured). Nuclei are hyperchromatic but have
overall maintained polarity. Lamina propria is plentiful. E: Basal pattern dysplasia—there are
nuclear features of dysplasia restricted to the glands. Often recut sections will disclose zones
of surface alterations. F: High-grade dysplasia—the surface shows lack of maturation, and
there is loss of overall cell polarity (the lines are obscured). Nuclei are hyperchromatic with loss
of polarity. Nucleoli are inconspicuous. Lamina propria is often plentiful (but not always). G:
Intramucosal carcinoma (lamina propria invasion T1a)—this is the earliest invasive carcinoma
and invasion only into the lamina propria can be suggested even on mucosal biopsies because
desmoplasia is not yet well developed. The surface shows lack of maturation, and there is loss
of overall cell polarity (the lines are obscured). Many nuclei are hyperchromatic with loss of
polarity. Large nucleoli are present. Lamina propria is overrun by glands. H: Adenocarcinoma
invasive into at least the submucosa (at least T1b as seen on mucosal biopsies)—all of the
features of intramucosal carcinoma are present as well as angulated glands with desmoplasia.
Occasional cases show pagetoid extension of single cells into squamous epithelium.

High-grade dysplasia—The surface shows lack of maturation, and there is loss of

overall cell polarity (the lines are obscured). Nuclei are hyperchromatic with loss of
polarity. Nucleoli are usually inconspicuous. Lamina propria is often plentiful (but
not always). Cases with gastric-type differentiation are different and discussed in the
earlier text.
Intramucosal carcinoma (lamina propria invasion T1a)—This is the earliest invasive
carcinoma, and invasion only into the lamina propria can be suggested even on
mucosal biopsies because desmoplasia is not yet well developed. The surface shows
lack of maturation, and there is loss of overall cell polarity (the lines are obscured).
 Many nuclei are hyperchromatic with loss of polarity. Large nucleoli are present.
Lamina propria is overrun by glands.
Adenocarcinoma invasive into at least submucosa (at least T1b as seen on mucosal
biopsies)—all of the features of intramucosal carcinoma are present as well as
angulated glands with desmoplasia. Occasional cases show pagetoid extension of
single cells into squamous epithelium.

FAQ: What is pseudoregression (buried Barrett mucosa)?

Answer: This is a pattern seen after either injury from reflux or injury from
mucosal ablation in which squamous mucosa grows on top of columnar mucosa
(Barrett mucosa). There were early case reports of carcinomas appearing
beneath the squamous mucosa following ablation for dysplasia, but it is
currently believed that this concern was overblown and not an issue. In fact,
there is virtually always a surface lesion accompanying buried lesions
(pseudoregression lesions).38 In addition, using modern high-resolution
endoscopes, many endoscopists are able to see areas that harbor buried
lesions. We have already seen examples of these in Figs. 1.94, 1.106, 1.107
1.109, 1.110, 1.115, 1.116, 1.118, and 1.122. In some of these examples, the
endoscopist detected the lesions even though they were buried beneath
squamous epithelium and even labeled them “lesion” on the specimen jars. Of
course, she is a good endoscopist, but the point is that such lesions can be
found by careful gross (endoscopic) examination.
 Figure 1-122. “Buried” Barrett-associated neoplasia/pseudoregression. This was once believed
to be an issue, but additional data do not support such a concern. In this example, an
intramucosal carcinoma, is seen buried under the squamous epithelium.

PEARLS & PITFALLS: Do not Be Fooled by Duplicated Muscularis Mucosae!

Disorganized double and even triple layers of muscularis mucosae are
commonly encountered in the injured esophagus with Barrett metaplasia.2,53
Correctly identifying the layers is easy on resections, but on superficial
biopsies, remember that it is unusual to see submucosa, and tissue seen deep
to muscularis mucosae on superficial samples is probably not submucosa but
instead lamina propria between the duplicated muscularis mucosae and the
deeper original muscularis mucosae. Fig. 1.106 shows an EMR sample with a
very thick duplicated muscularis mucosae. Imagine a biopsy that is superficial
and encompasses the space between the original muscularis mucosae and the
duplicated muscularis mucosae. The savvy pathologist knows that that space is
not submucosa and reports accordingly.

FAQ: What changes can we expect in samples obtained in patients who have
 had mucosal ablation?

Answer: The objective of mucosal ablation is to eliminate the dysplasia and,

better still, all columnar epithelium. In some patients, follow-up samples show
findings indistinguishable from those in undamaged squamous mucosa, but, in
many patients, mild fibrosis of the lamina propria, prominent eosinophils,
lymphocytosis,56 and lichenoid changes (prominent intraepithelial lymphocytes
and apoptotic squamous epithelial cells) may be encountered. Of course, do
not forget to search for candidiasis and viral cytopathic changes.

SQUAMOUS NEOPLASIA
RISK FACTORS
Esophageal squamous neoplasia is uncommon in the United States compared with other
regions of the world. Most patients with squamous carcinomas are men, and most are
adults at least in their fifties. In contrast to the demographics for adenocarcinomas,
which typically affect white men, squamous carcinoma predominates among African-
American men.57 Recent advances in improvement in detecting and treating
adenocarcinomas are not mirrored for squamous carcinoma.58
The incidence of squamous cell carcinoma is diminishing in comparison with
esophageal adenocarcinomas in the United States. In contrast, squamous carcinomas
have a high incidence in developing countries, e.g., in southern Africa and China. In
Southeast Asians, polymorphisms in ALDH, the gene that encodes aldehyde
dehydrogenase, are associated with esophageal squamous cell cancer.59,60 The effects of
these polymorphisms are synergistic with alcohol and smoking. ALDH polymorphisms
also result in accumulation in acetaldehyde, which causes flushing upon ingestion of
alcohol in about a third of East Asians (Chinese, Japanese, and Koreans).
Any factor that causes chronic irritation and inflammation of the esophageal mucosa
predisposes one to esophageal squamous cell carcinoma. Even skin disease affecting
the esophagus can initiate the development of dysplasia/carcinoma of the esophagus. An
example of this is lichen planus. However, substantial alcohol intake, especially in
combination with smoking, exponentially increases the risk of squamous cell carcinoma
(but not adenocarcinoma) and may account for the vast majority of squamous cell
carcinoma of the esophagus in the developed world. The combination of smoking and
alcohol abuse also results in an increased risk of head and neck cancer. Squamous cell
carcinoma of the esophagus is in fact discovered incidentally in 1% to 2% of patients
with head and neck cancers.61
Chronic esophageal irritation can also result from achalasia and esophageal
diverticula such that food is retained and decomposes, releasing various chemical
irritants. Frequent consumption of extremely hot beverages seems to increase the
incidence of squamous cell carcinoma. Lastly, persons who have ingested lye or other
caustic fluids require lifelong surveillance for the development of this cancer.62
Nonepidermolytic palmoplantar keratoderma (tylosis) is a rare autosomal dominant
disorder defined by RHBDF2 mutations on chromosome 17q2563 associated with
squamous cell carcinoma of the esophagus.64 Patients have hyperkeratosis of the palms
and soles and thickening of the oral mucosa. Although it confers up to a 95% risk of
squamous cell carcinoma of the esophagus by the age of 70 years, it is a rare syndrome.
Squamous cell carcinoma (but not adenocarcinoma) is associated with low
socioeconomic status, presumably a reflection of poor nutrition and other lifestyle
factors. However, deficiency syndromes associated with this cancer, such as the
Plummer-Vinson syndrome (dysphagia, iron-deficiency anemia, and esophageal webs),
are becoming uncommon in the developed world as overall nutrition improves. A role
for human papillomavirus (HPV) in the development of esophageal squamous cell
carcinoma is debatable, even though it is well-established in the anal canal, as
discussed in “Anus” chapter. Although HPV DNA detection rates are minimal (0% to
2%) in some studies from low-incidence areas,65,66 higher rates are reported in high-
incidence areas, such as China and Iran.67,68 However, one Mexican study, an area of
low tumor incidence, reported the presence of high-risk HPV DNA in 25% of
esophageal squamous cell carcinomas.69 The rate in the United States is about 10%.70

SQUAMOUS CARCINOMA PRECURSOR LESIONS

Squamous Dysplasia
Background squamous epithelial dysplasia (intraepithelial neoplasia), low or high
grade (including “carcinoma in situ”), can often be encountered at the periphery of
invasive squamous cell carcinomas (Figs. 1.123–1.132). As in other sites, the
convention is to regard epithelial changes in the bottom half of the epithelium as low-
grade dysplasia (the intraepithelial neoplasia terminology is preferred in Europe) and
into the top half as high-grade dysplasia. When there is absolutely no maturation at the
surface, the term carcinoma in situ can be used, but it really does not matter for
treatment purposes. Likewise, it can be difficult to determine when there is very early
invasion into the lamina propria because this does not invoke a desmoplastic reaction.
However, severe epithelial changes in the bottom half of the epithelium often suggest an
adjoining unsampled carcinoma, as noted later. Sometimes reactive conditions can
mimic dysplasia, and it is especially important to be cautious in the setting of
inflammatory conditions (which themselves predispose to squamous neoplasia).
Examples of lichenoid esophagitis71 and pill-associated esophagitis with prominent
epithelial changes are seen in Figs. 1.133–1.136. The reader is also referred to volume
1 of this series.

Figure 1-123. Low-grade squamous dysplasia. These lesions are subtle and often difficult to
differentiate from reparative changes. The nuclei are hyperchromatic, and there may be
mitoses and apoptotic nuclei as seen in the left part of the image. The epithelial changes are
restricted to the lower half of the epithelium. In the setting of prominent inflammation, it can be
impossible to determine if epithelial changes are reactive or dysplastic. Note that nucleoli are
not prominent.
Figure 1-124. Low-grade squamous dysplasia. This case is difficult, but there are no
inflammatory features to explain the findings and there are a few apoptotic bodies in the
proliferation, which occupies just under half of the epithelial thickness.
Figure 1-125. Low-grade squamous dysplasia. This is a high-magnification image of the case
seen in Fig. 1.123.
Figure 1-126. Low-grade squamous dysplasia. The nuclei in the basal portion of the epithelium
are markedly hyperchromatic.

Figure 1-127. Low-grade squamous dysplasia. The basal zone is thickened, and the nuclei are
enlarged and hyperchromatic.
Figure 1-128. Low-grade squamous dysplasia. High magnification shows scattered mitoses
and apoptotic bodies.
Figure 1-129. Low-grade squamous dysplasia. Note the marked nuclear hyperchromasia in the
center of the image. There is no inflammatory process to explain the findings.
Figure 1-130. Low-grade squamous dysplasia. This is a high-magnification image of the
process seen in Fig. 1.129.
Figure 1-131. High-grade squamous dysplasia. Despite the inflammation, the epithelial changes
are in excess of those attributable to inflammation. In Japan, lesions such as this would be
regarded as intramucosal carcinomas.
Figure 1-132. High-grade squamous dysplasia. The keratinization is concerning for invasion.
Figure 1-133. Lichenoid esophagitis with marked reactive epithelial changes. The findings are
similar to those of low-grade dysplasia, but the striking inflammatory process can explain them.
Patients with lichen planus of the esophagus and related lesions are at risk for squamous cell
carcinomas and are monitored accordingly, so if these changes are in fact dysplastic and
interpreted as reactive, this will not end patient surveillance.
Figure 1-134. Lichenoid esophagitis with marked reactive epithelial changes. The prominent
nucleoli suggest a reactive process.
Figure 1-135. Pill esophagitis with striking reactive changes. In the context of the exudate and
embedded pill material, these changes are best regarded as reparative.
 Figure 1-136. Pill esophagitis with striking reactive changes. This is a high-magnification image
of the lesion seen in Fig. 1.135. Note the prominent intracellular edema at the lower right. The
edema causes the intercellular bridges to appear prominent.

Epidermoid Metaplasia
A peculiar pattern of hyperkeratosis and hypergranulosis, which can be termed
“esophageal leukoplakia” (because the endoscopist notes white plaques), but more
accurately “epidermoid metaplasia,” can be encountered as an isolated finding, but it
can also be seen associated with samples showing squamous dysplasia and
carcinoma.72,73 We suspect that this is a precursor lesion. There is also some more
scientific rather than histologic evidence that epidermoid metaplasia is a precursor
lesion; we have noted that molecular alterations detected by next-generation sequencing
mirror those in the associated dysplasias and carcinomas.74
Epidermoid metaplasia is easy to overlook unless one is in the know because it
appears normal at first glance except that it is metaplastic. Unfortunately, the magnitude
of the risk for associated squamous cell dysplasia and carcinoma is not known. The
subtle finding to be sought is simply the presence of a granular layer in esophageal
squamous epithelium, which is not normal for the esophagus but perfectly normal in
skin! Examples of epidermoid metaplasia, one of which is associated with neoplasia,
are shown in Figs. 1.137–1.142.
Figure 1-137. Epidermoid metaplasia. This sample appears rather unremarkable at first glance
but note the granular layer at the left. Note also that the lesion is sharply demarcated from the
uninvolved squamous epithelium.

Figure 1-138. Epidermoid metaplasia. This is a high-magnification image of the lesion seen in
Fig. 1.137. It shows the sharp demarcation between the normal area and the zone with
epidermoid metaplasia.
Figure 1-139. Epidermoid metaplasia. Note the granular layer and the surface hyperkeratosis.
Figure 1-140. Epidermoid metaplasia. This example is from a resection specimen—the
indication for resection was squamous cell carcinoma and there was extensive epidermoid
metaplasia in the adjoining tissue.
Figure 1-141. Squamous cell carcinoma associated with epidermoid metaplasia. This
squamous cell carcinoma was associated with epidermoid metaplasia.
Figure 1-142. Squamous cell carcinoma associated with epidermoid metaplasia. This is a high-
magnification image of the carcinoma seen in Fig. 1.141.
Figure 1-143. Epidermoid metaplasia. This is a tricky case. There is epidermoid metaplasia and
apparently dysplasia in the lower half of the mucosa. But, wait, there is more! The patient was
taking a taxane medication for a separate carcinoma (breast), and in fact there is taxane effect
(note the ring mitoses) in association with epidermoid metaplasia.
Figure 1-144. Taxane effect. In this tangentially embedded focus, it is a real mimicker of
dysplasia but note the mitotic arrest.
Figure 1-145. Taxane effect. This field is from the same biopsy as the images seen in Figs.
1.143 and 1.144. The mitotic arrest (ring mitoses) is apparent.
 Figure 1-146. Esophageal squamous cell carcinoma. Squamous cell carcinoma of the
esophagus has the same appearance as it does elsewhere in the body with abnormally
keratinized overtly malignant cells.

However, in the spirit of trickery and to introduce the near misses section of this
chapter, note that Figs. 1.143–1.145 show a lesion in which a medication effect (taxane
effect75) mimics dysplasia in a patient with esophageal epidermoid metaplasia! The
clue is the ring mitoses.

Squamous Cell Carcinoma

Esophageal squamous cell carcinomas display essentially the same appearance as
squamous cell carcinoma in the rest of the body and are usually not a diagnostic
problem (Figs. 1.146–1.149). Most are well differentiated with prominent
keratinization, but they can be basaloid, spindled, papillary, or verrucous; the latter can
be impossible to diagnose on superficial samples but is instead diagnosed in the context
of the endoscopic finding—a mass is seen and the pathologist is faced with a bland-
appearing thickened mucosa.
Figure 1-147. Esophageal squamous cell carcinoma. This example shows whorls of malignant
squamous cells.
Figure 1-148. Esophageal squamous cell carcinoma. This biopsy is scant but shows abnormal
keratinization in cells arranged in squamous pearls with cytologically malignant nuclei.
 Figure 1-149. Esophageal squamous cell carcinoma. Even though this sample is superficial, it
was from a mass lesion; in this context it can be diagnosed as squamous cell carcinoma.

Squamous carcinomas can assume a prominent spindle cell appearance as well—

please remember that a spindle cell malignant neoplasm of the esophagus is almost
never a sarcoma.
Like squamous carcinomas elsewhere, esophageal examples express p63, p40,
CK5/6, and a host of epithelial markers. Most examples do not pose diagnostic
problems, and it is not difficult to obtain repeat samples if an initial biopsy is not fully
diagnostic. This is usually a result of superficial sampling. However, when these
carcinomas are spindled, melanoma and sarcomas must be excluded. In resection
specimens, it is easy to do this by sampling as much of the overlying squamous mucosa
as possible to detect an in situ component or a zone of conventional-appearing
squamous cell carcinoma. Otherwise, S100 protein (which should be negative) and
various keratins are most useful. Do not even think of wasting tissue on a vimentin!
Repeat, never waste tissue on vimentin staining in mucosal biopsies of the
gastrointestinal tract. Rare examples of esophageal squamous carcinoma are
associated with Epstein-Barr virus, but such cases are more commonly encountered in
the stomach.
Lastly, before moving on to case studies that we regard as “near misses,” we would
like to point out a pattern that can be seen that superficially mimics low-grade dysplasia
that is instead a marker of an unsampled squamous cell carcinoma. Figs. 1.150 and
1.151 are from a case of squamous cell carcinoma of the esophagus. Note that, in one of
the fragments, the lower half of the epithelium is replaced by strikingly atypical
squamous cells. Were only this fragment sampled, it could be interpreted as low-grade
dysplasia because the epithelial changes encompass less than half the thickness of the
epithelium. However, the cytologic features are in excess of those expected for low-
grade squamous dysplasia. This pattern reflects intraepithelial invasion of the nearby
squamous cell carcinoma. It can be termed a lateral spread pattern, but it is important to
call attention to it to assure that the endoscopist brings the patient back for additional
sampling to search for a subtle unsampled invasive carcinoma.
Table 1.3 outlines some of the demographic and other differences between
esophageal adenocarcinoma and squamous cell carcinoma.

Figure 1-150. Esophageal squamous cell carcinoma. This is a portion of a biopsy of a

squamous cell carcinoma that illustrates an important phenomenon. Notice that the carcinoma
(not well demonstrated in this field) has spread into the space between the basement
membrane and the normal squamous epithelium in the fragment seen at the left and in the
center. This results in severely atypical cell being seen in the lower half of the epithelium, a
feature of low-grade dysplasia. However, the cytologic changes are far in excess of the subtle
ones that form low-grade squamous dysplasia. This pattern has been called a lateral spread
pattern.
 Figure 1-151. Lateral spread pattern. This is a high-magnification image of the fragment seen at
the center and at the left in Fig. 1.150. Strikingly atypical cells are seen in the lower half of the
epithelium. This finding suggests intraepithelial invasion of an adjoining carcinoma. If only this
pattern is seen on a biopsy, the endoscopist should be encouraged to perform extensive
resampling with the goal of detecting an occult invasive carcinoma.

TABLE 1.3: Esophageal Adenocarcinoma Versus Squamous Cell Carcinoma

(US Population)

Feature Adenocarcinoma Squamous Cell Carcinoma

Typical demographics Older white men Older black men

Risk factors Reflux, obesity, smoking Smoking, alcohol

Precursors Barrett esophagus; Barrett Epidermoid metaplasia

esophagus with dysplasia (probable), squamous dysplasia

NEAR MISSES
Case 1
Figs. 1.152 and 1.153 were taken from the esophagus and diagnosed as adenocarcinoma
arising in association with Barrett mucosa with high-grade dysplasia. As a result, the
patient was referred for endoscopic treatment and the biopsies were rereviewed in
preparation for this.

WHAT WENT WRONG?

The nuclei are far too hyperchromatic for the usual adenocarcinoma and are arranged in
sheets rather than remotely forming glands and have no component of mucin. Some
adenocarcinomas and squamous carcinomas are poorly differentiated, but this case is
unusual. In Fig. 1.153, where the tumor undermines the squamous epithelium, it shows
nuclear molding. Because the carcinoma arises in association with high-grade columnar
epithelium (Fig. 1.152), the colleague assumed that it was an adenocarcinoma. It is
instead a high-grade neuroendocrine carcinoma, small cell type. Figs. 1.154 and 1.155
show a synaptophysin stain, and Fig. 1.155 highlights the characteristic dotlike staining
seen with synaptophysin in small cell carcinomas from all sites. In addition, the
extremely high proliferation index using Ki-67 immunolabeling seen in Fig. 1.156 is
characteristic of small cell carcinoma. Esophageal high-grade neuroendocrine
carcinomas are not common but tend to arise in association with columnar epithelial
dysplasia in the US population and in association with high-grade squamous dysplasia
in populations in which esophageal squamous lesions are common. This is similar in
some ways to high-grade neuroendocrine carcinomas encountered in the colon, which
are often associated with an adenomatous (columnar) precursor. In fact, sometimes
neuroendocrine differentiation can be seen in Barrett esophagus with high-grade
dysplasia (Figs. 1.157 and 1.158)—we would assume that these cells are the precursors
to high-grade neuroendocrine carcinomas of the esophagus. An example of an
esophageal high-grade neuroendocrine carcinoma of the large cell type is seen in Fig.
1.159.

Neuroendocrine Tumors of the Esophagus

Esophageal neuroendocrine tumors (NETs) are currently classified as differentiated
NETs, grade 1 and 2 (carcinoid tumor, atypical carcinoid tumor), and neuroendocrine
carcinomas (large and small cell types), grade 3.76 They are rare (on the order of 100
cases have been reported, and most are high-grade carcinomas/small cell carcinomas).
In a large series of carcinoid tumors, <0.1% involved the esophagus.77 They are only
rarely encountered on biopsies. Well-differentiated NETs (carcinoids) of the esophagus
resemble those seen in other sites, although in the esophagus most are polypoid and thus
appear in the lamina propria on biopsies. They consist of uniform bland tumor cells
with an insular pattern and solid to cribriform growth (Figs. 1.160–1.164). On
immunolabeling, they express keratin, synaptophysin, and chromogranin. Interestingly,
they can be associated with heterotopic oxyntic mucosa.

Figure 1-152. High-grade neuroendocrine carcinoma associated with high-grade columnar

epithelial dysplasia. A malignant lesion composed of extremely hyperchromatic cells has filled
the lamina propria between glands that show high-grade dysplasia features.
Figure 1-153. High-grade neuroendocrine carcinoma, small cell type. The carcinoma fills the
lamina propria beneath the squamous epithelium. The malignant cells show prominent nuclear
molding. This is another field from the same lesion as that depicted in Fig. 1.152.
Figure 1-154. High-grade neuroendocrine carcinoma, small cell type. This is a synaptophysin
stain, which labels the carcinoma but not many cells in the high-grade columnar epithelial
precursor.
Figure 1-155. High-grade neuroendocrine carcinoma, small cell type. This is a high-
magnification image of the lesion seen in Figs. 1.152–1.154. The arrowhead indicates dotlike
reactivity with the synaptophysin stain, a characteristic feature of small cell carcinoma.
Figure 1-156. High-grade neuroendocrine carcinoma, small cell type. This is a Ki-67 stain of the
lesion seen in Figs. 1.152–1.155.
Figure 1-157. High-grade columnar epithelial dysplasia. Note the prominent neuroendocrine
cells (Kulchitsky cells) in this example! They are characterized by red granules that point away
from the gland lumina.
Figure 1-158. High-grade columnar epithelial dysplasia. This example shows cells with
neuroendocrine features. It is a different lesion from the one seen in Fig. 1.157.

Figure 1-159. High-grade neuroendocrine carcinoma, large cell type. The cells are nested and
some are large with large nucleoli.
 Like well-differentiated NETs elsewhere, they can be graded by assessing the
mitotic activity (to the extent possible in small biopsies, which means that they cannot
really be graded) or by performing Ki-67 immunolabeling.76 As of the 2010 WHO
classification of gastrointestinal tumors, well-differentiated NETs were graded as G1 or
G2 and the rare lesions with well-differentiated morphology but a high Ki-67 labeling
index were included with high-grade neuroendocrine carcinoma, but this has sometimes
proved unsatisfactory (because there is different treatment of small cell carcinoma and
large cell neuroendocrine carcinoma). As such, the 2017 WHO Classification of
Tumours of Endocrine Organs, at least with good data for pancreatic lesions, has added
a new category, termed NET G3.78 This latter category encompasses neoplasms that
have an appearance like that of a well-differentiated neuroendocrine (carcinoid) tumor
but that have a high proliferation index when Ki-67 labeling is performed. Because this
latter category is rare and NET of the esophagus is rare as well, we have not
encountered such a lesion. To summarize for neuroendocrine tumor76,78:

Figure 1-160. Well-differentiated neuroendocrine (carcinoid) tumor of the esophagus. These

are rare (rarer than high-grade neuroendocrine carcinomas) in the esophagus. This lesion was
found in the tubular esophagus in an area with cardiac-type mucosa.
Figure 1-161. Well-differentiated neuroendocrine (carcinoid) tumor of the esophagus. The
lesion is to the left. A bit of multilayered epithelium is also present, confirming that the process
indeed involved the esophagus rather than the stomach!
Figure 1-162. Well-differentiated neuroendocrine (carcinoid) tumor of the esophagus. The
lesion is bland appearing at high magnification.
Figure 1-163. Well-differentiated neuroendocrine (carcinoid) tumor of the esophagus. This is a
chromogranin stain.
Figure 1-164. Well-differentiated neuroendocrine (carcinoid) tumor of the esophagus. This is a
synaptophysin stain.
 Figure 1-165. Adenosquamous carcinoma of the esophagus. The lesion seems to spill from the
base on nondysplastic squamous epithelium.

NET G1—<2 mitoses per 10 high-power fields or Ki-67 labeling <3%

NET G2—2 to 20 mitoses per 10 high-power fields or Ki-67 labeling 3% to 20%
NET G3—>20 mitoses per 10 high-power fields or Ki-67 labeling >20.

On the other hand, classic neuroendocrine carcinomas (always grade 3) are high-
grade lesions and can be classified as “small cell” or “large cell” types.76,78 Both types
are aggressive lesions. Most arise in men (although they are rare), sometimes in
association with Barrett mucosa. These are rare tumors that have not been well studied,
but in one Chinese series of patients with resectable disease, the mean age was about 60
years and most patients (about 70%) were men. About 60% were of the small cell
rather than of the large cell type. Although nearly 70% were alive after a year, only a
third were alive at 3 years, even though the patients had resectable disease.79 Patients
with these carcinomas are often treated with platinum-based chemotherapy. Large cell
neuroendocrine carcinomas manifest an organoid pattern with solid nests or acinar
structures, whereas small cell carcinomas form solid sheets and nests and are composed
of cells with small dark nuclei and minimal cytoplasm or larger cells with more
cytoplasm. Adjoining zones of adenocarcinoma or squamous cell carcinoma may be
present (the latter may require p63/p40 or CK5/6 staining to detect). They express
keratins, synaptophysin, and chromogranin. Some observers also use CD56
immunolabeling, which is hampered by its lack of specificity. Both small cell and large
cell neuroendocrine carcinomas are G3 neoplasms:

G3 neuroendocrine carcinoma—mitotic count >20 per 10 high-power field or >20%

Ki-67 index.

Case 2
Figs. 1.165 and 1.166 are from an esophageal biopsy. A mass was seen at
endoscopy in the middle third of the esophagus. The sample was interpreted as
squamous cell carcinoma. Indeed, the lesion does appear to “drip,” but there is
something amiss for squamous cell carcinoma. Fig. 1.166 gives the clue in that there is
mucin in some of the cells. This is an adenosquamous carcinoma. However, it is not
horrible to miss this feature—both types (adenocarcinoma and squamous cell
carcinoma) are treated the same way, although HER2 testing might be added for
adenosquamous carcinoma. This is different from the situation in the prior case—
detecting a high-grade neuroendocrine carcinoma results in a different chemotherapy
regimen than that for adenocarcinoma.
Figure 1-166. Adenosquamous carcinoma of the esophagus. Note the mixture of squamous
cells and cells that contain mucin.
Figure 1-167. Thyroid papillary carcinoma extending directly into the esophagus. The
carcinoma has an odd appearance for an esophageal adenocarcinoma.
 Figure 1-168. Thyroid papillary carcinoma extending directly into the esophagus. A few grooves
and intranuclear inclusions are present, but they are easier to see if you know that the patient
has a history of a large thyroid carcinoma!

Adenosquamous Carcinoma
Adenosquamous carcinoma arises only rarely in the esophagus. It consists of a mixture
of infiltrating squamous cell carcinoma and adenocarcinoma elements. Like squamous
cell carcinoma, this variant preferentially affects the middle third of the esophagus.
Adenosquamous carcinoma also seems to present at an earlier stage than pure
adenocarcinoma or squamous cell carcinoma.80 The squamous component tends to be
more conspicuous than the glandular areas with gradual transitions between the two.
Areas of accompanying Barrett mucosa may rarely be identified. The outcome is more
favorable than that associated with pure squamous or adenocarcinoma, presumably a
result of the smaller size and lower stage at presentation.80

Case 3
Figs. 1.167 and 1.168 are images taken from an esophageal biopsy in a 58-year-old
woman who complained of dysphagia. The sample was diagnosed as well-
differentiated adenocarcinoma, and the patient was referred to one of our institutions for
treatment. On review, it was noted that the nuclei appeared rather small compared with
those that usually form esophageal adenocarcinomas and also that nucleoli were not a
feature. This prompted extensive queries into the clinical history, and we learned that
the patient had a history of papillary thyroid carcinoma. Based on this history, we added
a thyroglobulin stain, which is seen in Fig. 1.169. In this case, the presence of unusual
features in the lesion prompted some digging in the patient’s records. Once we
considered thyroid carcinoma, the nuclear inclusions and grooves became apparent to
us, even though the sample was suboptimal.

Spread of Extraesophageal Carcinoma to the Esophagus

The extraesophageal neoplasms that spread to the esophagus are primarily lung
carcinoma (Figs. 1.170 and 1.171) and breast carcinoma, but thyroid carcinoma is
known, and rarely, cervix carcinoma can spread to the lung (this requires HPV studies
to differentiate from primary esophageal squamous carcinoma), as well as ovary,
prostate, and renal carcinoma. Recognition can be impossible without knowledge of
pertinent history, but a clue to at least consider an extraesophageal primary is the
absence of a precursor component. Immunolabeling can be useful but has limitations.
For example, both esophageal and lung adenocarcinoma can express TTF1 and napsin if
a polyclonal antibody is used.81 However, esophageal adenocarcinomas generally lack
expression of hormone receptors (estrogen and progesterone receptors) and GATA3,
whereas expression of these can suggest metastatic breast carcinoma.
Figure 1-169. Thyroid papillary carcinoma extending directly into the esophagus. The lesion
shows expression of thyroglobulin.
Figure 1-170. Lung carcinoma extending into the lamina propria of the esophagus. This is
diagnosed in the context of the history!
 Figure 1-171. Lung carcinoma extending into the lamina propria of the esophagus. Note the
reactive changes in the overlying inflamed squamous epithelium.

Case 4
Fig. 1.172 shows a spindle cell lesion that was biopsied from the esophagus. On
immunolabeling, it was KIT reactive but CD34 nonreactive. A diagnosis of esophageal
GIST was made. The mitotic count was high in keeping with a high-risk lesion, and a
resection was performed.

What Went Wrong?

This is a melanoma! Remember that many melanomas show KIT expression and about
20% of mucosal melanomas in fact have KIT mutations, the presence of which can be
exploited using targeted therapy.82-85 This is an important pitfall, especially in spindle
cell melanomas. Melanoma can masquerade in many forms. It is usually CD34 negative
as well. In this particular case, simply looking carefully at the slides gave the correct
diagnosis because there was an obvious in situ pigmented component that had been
overlooked (Fig. 1.173).

Melanoma
Primary esophageal melanoma is rare, with only about 300 cases reported. It is
encountered in adults with a mean age of about 60 years.86-91 There is a male
predominance but no racial predominance. Most esophageal melanomas arise in the
distal esophagus. At endoscopy, most are polypoid and pigmented (about 85%). Imaging
studies show bulky polypoid masses that bulge intraluminally without associated
obstruction. Many examples are pigmented, an obvious diagnostic clue. The rest are
whitish and poorly marginated. Primary melanomas may display an in situ component.
When present, this finding is extremely useful in establishing the esophagus as the
primary site. The malignant cells are spindled to epithelioid with variable pigment and
prominent nucleoli, and sometimes prominent intranuclear pseudoinclusions can be
found.
The key differential diagnosis is with poorly differentiated carcinoma, which is far
more common, and also with high-grade lymphomas. Immunohistochemistry can be
critical in establishing the diagnosis of esophageal melanoma, as it is with melanomas
elsewhere. Useful antibodies include S100 protein, Sox 10, MART, HMB45, and Melan
A. Remember that melanomas are also reactive with CD117/KIT antibodies about 40%
of the time, so a panel approach is best; overall, melanomas are usually more
pleomorphic than GISTs. The host of masqueraders is usually unmasked using a panel
including lymphoid markers (among which is CD30 for anaplastic lymphoma) and
pankeratins. Sarcomas are rarely in the differential diagnosis, but, of course, spindled
melanomas may lack other melanoma markers aside from S100 protein and Sox 10.
Another spindle cell tumor that is strongly S100 protein reactive is cellular (benign)
schwannoma. The distinction is on cytologic grounds, with attention to nuclear
pleomorphism and large nucleoli, features of melanoma but not of cellular schwannoma.
These unusual primary esophageal melanomas have had a dismal prognosis. Based on
the literature, only rare patients, whose tumors present early, can be cured. However,
newer treatments exploiting PDL1/PD1 blockade may improve the situation, but there
are no outcome data available at this point for esophageal melanoma specifically. In one
case report a good tumor response was achieved with nivolumab.92
Figure 1-172. Esophageal melanoma with spindle cell features. This lesion displays the
monotonous cytologic features that are often seen in gastrointestinal stromal tumors.
 Figure 1-173. Melanoma extending into the lamina propria of the esophagus. In this case, a
pigmented in situ component gives away the diagnosis.

Case 5
Fig. 1.174 shows a biopsy from a mass lesion of the distal esophagus. It was
submitted on a rush protocol on a Friday night by a surgeon eager to plan an operation
for Monday morning and was reported as poorly differentiated carcinoma. The surgeon
scheduled the patient for 7 AM the following Monday. The resection resulted in a sinking
feeling.

What Went Wrong?

This is a diffuse large B cell lymphoma, which of course is not treated surgically.
On small biopsies of overtly malignant neoplasms for which limited tissue in
present in the sample, it can be prudent to begin with a limited immunolabeling panel
and an order of unstained sections up front. A panel that can work well consists of S100
protein (to address melanoma), a pankeratin (to address carcinoma), and a CD20 (to
address diffuse large B cell lymphoma, the most likely type of lymphoma encountered).
If all three of these stains are negative on the first round of immunolabeling, the
morphology can be revisited and a second attempt made, which might be directed at
GIST (KIT and DOG1). This sequential approach can allow a diagnosis before the
tissue is exhausted.

Esophageal Hematopoietic Disorders

Primary esophageal hematopoietic lesions are truly rare, but the majority are extranodal
lymphomas. Esophageal lymphoma is defined as an extranodal lymphoma arising in the
esophagus itself rather than extending into the esophagus from the mediastinum, stomach,
or a contiguous lymph node. The lesion shown in Fig. 1.174 might be regarded as a
gastric lymphoma because it is associated with cardiac-type mucosa. However, when
lymphomas are found in the esophagus, most are large B cell lymphomas or mucosa-
associated lymphoid tissue lymphomas. Any type of lymphoma can be detected in the
esophagus, and a discussion of every type is beyond the scope of this section.93

Case 6
Fig. 1.175 is from a biopsy taken from a 75-year-old man with dysphagia, and a
subtle early lesion was seen by the endoscopist. A diagnosis of intramucosal
adenocarcinoma was made. The next day, the endoscopist called and said that she was
confused because the patient’s symptoms were out of proportion to the histologic
diagnosis and to the results of endoscopic ultrasound performed at the time of the
endoscopic evaluation.
Figure 1-174. Diffuse large B cell lymphoma involving the esophagus. It is easy to consider a
poorly differentiated carcinoma in this location.
 Figure 1-175. Esophageal intramucosal carcinoma with extensive associated inflammation. Be
sure to check the areas of exudate!

What Went Wrong?

On rereview, herpes simplex viral cytopathic effect was noted in an eroded area that
had been overlooked! The classic viral cytopathic effect is seen in Fig. 1.176 and, just
for fun, immunolabeling for herpes simplex virus was performed and is shown in Figs.
1.177 and 1.178.
The message in this case is that we must always remember to look for additional
diagnoses once we have made our key diagnosis. It is easy forget to check erosions and
ulcers for infectious agents when we diagnose neoplasms, but sometimes treating the
infection can make the patient more comfortable while the main issue is dealt with.
Figure 1-176. Esophageal intramucosal carcinoma with extensive associated inflammation. A
careful search in the inflamed zones yielded foci of herpes simplex virus cytopathic effect.
Figure 1-177. Esophageal intramucosal carcinoma with extensive associated inflammation and
herpes simplex virus infection. This is a herpes simplex virus immunostain.
 Figure 1-178. Esophageal intramucosal carcinoma with extensive associated inflammation and
herpes simplex virus infection. This is a high-magnification image of the preparation shown in
Fig. 1.177.

References
1. Amin M, Edge S, Greene F, et al. AJCC Cancer Staging Manual. 8th ed. Springer
International Publishing AG Switzerland; 2017.
2. Abraham SC, Krasinskas AM, Correa AM, et al. Duplication of the muscularis
mucosae in Barrett esophagus: an underrecognized feature and its implication for
staging of adenocarcinoma. Am J Surg Pathol. 2007;31(11):1719-1725.
3. Glickman JN, Chen YY, Wang HH, Antonioli DA, Odze RD. Phenotypic
characteristics of a distinctive multilayered epithelium suggests that it is a
precursor in the development of Barrett’s esophagus. Am J Surg Pathol.
2001;25(5):569-578.
4. Glickman JN, Spechler SJ, Souza RF, Lunsford T, Lee E, Odze RD. Multilayered
epithelium in mucosal biopsy specimens from the gastroesophageal junction region
is a histologic marker of gastroesophageal reflux disease. Am J Surg Pathol.
2009;33(6):818-825.
5. Lack EE, Worsham GF, Callihan MD, et al. Granular cell tumor: a
 clinicopathologic study of 110 patients. J Surg Oncol. 1980;13(4):301-316.
6. Johnston J, Helwig EB. Granular cell tumors of the gastrointestinal tract and
perianal region: a study of 74 cases. Dig Dis Sci. 1981;26(9):807-816.
7. Ohmori T, Arita N, Uraga N, Tabei R, Tani M, Okamura H. Malignant granular cell
tumor of the esophagus. A case report with light and electron microscopic,
histochemical, and immunohistochemical study. Acta Pathol Jpn. 1987;37(5):775-
783.
8. Yoshizawa A, Ota H, Sakaguchi N, et al. Malignant granular cell tumor of the
esophagus. Virchows Arch. 2004;444(3):304-306.
9. Miettinen M, Sarlomo-Rikala M, Sobin LH, Lasota J. Esophageal stromal tumors:
a clinicopathologic, immunohistochemical, and molecular genetic study of 17 cases
and comparison with esophageal leiomyomas and leiomyosarcomas. Am J Surg
Pathol. 2000;24(2):211-222.
10. Fitzgerald RC, di Pietro M, Ragunath K, et al. British Society of Gastroenterology
guidelines on the diagnosis and management of Barrett’s oesophagus. Gut.
2014;63(1):7-42.
11. American Gastroenterological Association, Spechler SJ, Sharma P, Souza RF,
Inadomi JM, Shaheen NJ. American Gastroenterological Association medical
position statement on the management of Barrett’s esophagus. Gastroenterology.
2011;140(3):1084-1091.
12. Shaheen NJ, Falk GW, Iyer PG, Gerson LB; American College of
Gastroenterology. ACG clinical guideline: diagnosis and management of Barrett’s
esophagus. Am J Gastroenterol. 2016;111(1):30-50; quiz 51.
13. Epstein JA, Cosby H, Falk GW, et al. Columnar islands in Barrett’s esophagus: do
they impact Prague C&M criteria and dysplasia grade? J Gastroenterol Hepatol.
2017;32(9):1598-1603.
14. Aida J, Vieth M, Shepherd NA, et al. Is carcinoma in columnar-lined esophagus
always located adjacent to intestinal metaplasia?: a histopathologic assessment.
Am J Surg Pathol. 2015;39(2):188-196.
15. Chandrasoma P, Wijetunge S, DeMeester S, et al. Columnar-lined esophagus
without intestinal metaplasia has no proven risk of adenocarcinoma. Am J Surg
Pathol. 2012;36(1):1-7.
16. Smith J, Garcia A, Zhang R, DeMeester S, Vallone J, Chandrasoma P. Intestinal
metaplasia is present in most if not all patients who have undergone endoscopic
mucosal resection for esophageal adenocarcinoma. Am J Surg Pathol.
2016;40(4):537-543.
17. Salimian KJ, Waters KM, Eze O, et al. Definition of Barrett esophagus in the
United States: support for retention of a requirement for goblet cells. Am J Surg
Pathol. 2018;42(2):264-268.
18. Bennett C, Moayyedi P, Corley DA, et al. BOB CAT: a large-scale review and
Delphi consensus for management of Barrett’s esophagus with no dysplasia,
indefinite for, or low-grade dysplasia. Am J Gastroenterol. 2015;110(5):662-682;
quiz 683.
19. Shields HM, Rosenberg SJ, Zwas FR, Ransil BJ, Lembo AJ, Odze R. Prospective
evaluation of multilayered epithelium in Barrett’s esophagus. Am J Gastroenterol.
2001;96(12):3268-3273.
20. Takubo K, Honma N, Arai T. Multilayered epithelium in Barrett’s esophagus. Am J
Surg Pathol. 2001;25(11):1460-1461.
21. Jiang M, Li H, Zhang Y, et al. Transitional basal cells at the squamous-columnar
junction generate Barrett’s oesophagus. Nature. 2017;550(7677):529-533.
22. Sharma P, Dent J, Armstrong D, et al. The development and validation of an
endoscopic grading system for Barrett’s esophagus: the Prague C & M criteria.
Gastroenterology. 2006;131(5):1392-1399.
23. Panarelli NC, Yantiss RK. Do ancillary studies aid detection and classification of
Barrett esophagus? Am J Surg Pathol. 2016;40(8):e83-e93.
24. Ormsby AH, Goldblum JR, Rice TW, et al. Cytokeratin subsets can reliably
distinguish Barrett’s esophagus from intestinal metaplasia of the stomach. Hum
Pathol. 1999;30(3):288-294.
25. Ormsby AH, Vaezi MF, Richter JE, et al. Cytokeratin immunoreactivity patterns in
the diagnosis of short-segment Barrett’s esophagus. Gastroenterology.
2000;119(3):683-690.
26. Phillips RW, Frierson HF Jr, Moskaluk CA. Cdx2 as a marker of epithelial
intestinal differentiation in the esophagus. Am J Surg Pathol. 2003;27(11):1442-
1447.
27. Hahn HP, Blount PL, Ayub K, et al. Intestinal differentiation in metaplastic,
nongoblet columnar epithelium in the esophagus. Am J Surg Pathol.
2009;33(7):1006-1015.
28. Chu PG, Jiang Z, Weiss LM. Hepatocyte antigen as a marker of intestinal
metaplasia. Am J Surg Pathol. 2003;27(7):952-959.
29. Montgomery E, Bronner MP, Goldblum JR, et al. Reproducibility of the diagnosis
of dysplasia in Barrett esophagus: a reaffirmation. Hum Pathol. 2001;32(4):368-
378.
30. Reid BJ, Haggitt RC, Rubin CE, et al. Observer variation in the diagnosis of
dysplasia in Barrett’s esophagus. Hum Pathol. 1988;19(2):166-178.
31. Lomo LC, Blount PL, Sanchez CA, et al. Crypt dysplasia with surface maturation:
a clinical, pathologic, and molecular study of a Barrett’s esophagus cohort. Am J
Surg Pathol. 2006;30(4):423-435.
32. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel
disease: standardized classification with provisional clinical applications. Hum
Pathol. 1983;14(11):931-968.
33. Curvers WL, ten Kate FJ, Krishnadath KK, et al. Low-grade dysplasia in Barrett’s
esophagus: overdiagnosed and underestimated. Am J Gastroenterol.
2010;105(7):1523-1530.
34. Vieth M, Montgomery EA, Riddell RH. Observations of different patterns of
dysplasia in Barrett’s esophagus - a first step to harmonize grading. Cesk Patol.
2016;52(3):154-163.
35. Rucker-Schmidt RL, Sanchez CA, Blount PL, et al. Nonadenomatous dysplasia in
Barrett esophagus: a clinical, pathologic, and DNA content flow cytometric study.
Am J Surg Pathol. 2009;33(6):886-893.
36. Mahajan D, Bennett AE, Liu X, Bena J, Bronner MP. Grading of gastric foveolar-
type dysplasia in Barrett’s esophagus. Mod Pathol. 2010;23(1):1-11.
37. Patil DT, Bennett AE, Mahajan D, Bronner MP. Distinguishing Barrett gastric
foveolar dysplasia from reactive cardiac mucosa in gastroesophageal reflux
disease. Hum Pathol. 2013;44(6):1146-1153.
38. Bronner MP, Overholt BF, Taylor SL, et al. Squamous overgrowth is not a safety
concern for photodynamic therapy for Barrett’s esophagus with high-grade
dysplasia. Gastroenterology. 2009;136(1):56-64; quiz 351-352.
39. Vieth M, Riddell RH, Montgomery EA. High-grade dysplasia versus carcinoma:
east is east and west is west, but does it need to be that way? Am J Surg Pathol.
2014;38(11):1453-1456.
40. Coco DP, Goldblum JR, Hornick JL, et al. Interobserver variability in the
diagnosis of crypt dysplasia in Barrett esophagus. Am J Surg Pathol.
2011;35(1):45-54.
41. Voltaggio L, Montgomery EA. Diagnosis and management of Barrett-related
neoplasia in the modern era. Surgical Pathol Clin. 2017;10(4):781-800.
42. Zhu W, Appelman HD, Greenson JK, et al. A histologically defined subset of high-
grade dysplasia in Barrett mucosa is predictive of associated carcinoma. Am J
Clin Pathol. 2009;132(1):94-100.
43. Pech O, May A, Manner H, et al. Long-term efficacy and safety of endoscopic
resection for patients with mucosal adenocarcinoma of the esophagus.
Gastroenterology. 2014;146(3):652.e651-660.e651.
44. Greene CL, Worrell SG, Attwood SE, et al. Emerging concepts for the endoscopic
management of superficial esophageal adenocarcinoma. J Gastrointest Surg.
2016;20(4):851-860.
45. Manner H, Pech O, Heldmann Y, et al. The frequency of lymph node metastasis in
early-stage adenocarcinoma of the esophagus with incipient submucosal invasion
(pT1b sm1) depending on histological risk patterns. Surg Endosc.
2015;29(7):1888-1896.
46. Abraham SC, Wang H, Wang KK, Wu TT. Paget cells in the esophagus: assessment
of their histopathologic features and near-universal association with underlying
esophageal adenocarcinoma. Am J Surg Pathol. 2008;32(7):1068-1074.
47. van der Wel MJ, Jansen M, Vieth M, Meijer SL. What makes an expert Barrett’s
histopathologist? Adv Exp Med Biol. 2016;908:137-159.
48. Terheggen G, Horn EM, Vieth M, et al. A randomised trial of endoscopic
submucosal dissection versus endoscopic mucosal resection for early Barrett’s
neoplasia. Gut. 2017;66(5):783-793.
49. Shaheen NJ, Sharma P, Overholt BF, et al. Radiofrequency ablation in Barrett’s
esophagus with dysplasia. N Engl J Med. 2009;360(22):2277-2288.
50. Canto MI, Shin EJ, Khashab MA, et al. Safety and efficacy of carbon dioxide
cryotherapy for treatment of neoplastic Barrett’s esophagus. Endoscopy.
2015;47(7):591.
51. Estrella JS, Hofstetter WL, Correa AM, et al. Duplicated muscularis mucosae
invasion has similar risk of lymph node metastasis and recurrence-free survival as
intramucosal esophageal adenocarcinoma. Am J Surg Pathol. 2011;35(7):1045-
1053.
52. Kaneshiro DK, Post JC, Rybicki L, Rice TW, Goldblum JR. Clinical significance
of the duplicated muscularis mucosae in Barrett esophagus-related superficial
adenocarcinoma. Am J Surg Pathol. 2011;35(5):697-700.
53. Lewis JT, Wang KK, Abraham SC. Muscularis mucosae duplication and the
musculo-fibrous anomaly in endoscopic mucosal resections for Barrett esophagus:
implications for staging of adenocarcinoma. Am J Surg Pathol. 2008;32(4):566-
571.
54. Westerterp M, Koppert LB, Buskens CJ, et al. Outcome of surgical treatment for
early adenocarcinoma of the esophagus or gastro-esophageal junction. Virchows
 Arch. 2005;446(5):497-504.
55. Vieth M, Stolte M. Pathology of early upper GI cancers. Best Pract Res Clin
Gastroenterol. 2005;19(6):857-869.
56. Kissiedu J, Thota PN, Gohel T, Lopez R, Gordon IO. Post-ablation lymphocytic
esophagitis in Barrett esophagus with high grade dysplasia or intramucosal
carcinoma. Mod Pathol. 2016;29(6):599-606.
57. Brown LM, Hoover R, Silverman D, et al. Excess incidence of squamous cell
esophageal cancer among US black men: role of social class and other risk factors.
Am J Epidemiol. 2001;153(2):114-122.
58. Njei B, McCarty TR, Birk JW. Trends in esophageal cancer survival in United
States adults from 1973 to 2009: a SEER database analysis. J Gastroenterol
Hepatol. 2016;31(6):1141-1146.
59. Brooks PJ, Enoch MA, Goldman D, Li TK, Yokoyama A. The alcohol flushing
response: an unrecognized risk factor for esophageal cancer from alcohol
consumption. PLoS Med. 2009;6(3):e50.
60. Cui R, Kamatani Y, Takahashi A, et al. Functional variants in ADH1B and ALDH2
coupled with alcohol and smoking synergistically enhance esophageal cancer risk.
Gastroenterology. 2009;137(5):1768-1775.
61. Erkal HS, Mendenhall WM, Amdur RJ, Villaret DB, Stringer SP. Synchronous and
metachronous squamous cell carcinomas of the head and neck mucosal sites. J Clin
Oncol. 2001;19(5):1358-1362.
62. Csikos M, Horvath O, Petri A, Petri I, Imre J. Late malignant transformation of
chronic corrosive oesophageal strictures. Langenbecks Arch Chir.
1985;365(4):231-238.
63. Blaydon DC, Etheridge SL, Risk JM, et al. RHBDF2 mutations are associated with
tylosis, a familial esophageal cancer syndrome. Am J Hum Genet. 2012;90(2):340-
346.
64. Ellis A, Risk JM, Maruthappu T, Kelsell DP. Tylosis with oesophageal cancer:
diagnosis, management and molecular mechanisms. Orphanet J Rare Dis.
2015;10:126.
65. Turner JR, Shen LH, Crum CP, Dean PJ, Odze RD. Low prevalence of human
papillomavirus infection in esophageal squamous cell carcinomas from North
America: analysis by a highly sensitive and specific polymerase chain reaction-
based approach. Hum Pathol. 1997;28(2):174-178.
66. Poljak M, Cerar A, Seme K. Human papillomavirus infection in esophageal
carcinomas: a study of 121 lesions using multiple broad-spectrum polymerase
 chain reactions and literature review. Hum Pathol. 1998;29(3):266-271.
67. Chang F, Syrjanen S, Shen Q, et al. Human papillomavirus involvement in
esophageal carcinogenesis in the high-incidence area of China. A study of 700
cases by screening and type-specific in situ hybridization. Scand J Gastroenterol.
2000;35(2):123-130.
68. Farhadi M, Tahmasebi Z, Merat S, Kamangar F, Nasrollahzadeh D, Malekzadeh R.
Human papillomavirus in squamous cell carcinoma of esophagus in a high-risk
population. World J Gastroenterol. 2005;11(8):1200-1203.
69. Herrera-Goepfert R, Lizano M, Akiba S, Carrillo-Garcia A, Becker-D’Acosta M.
Human papilloma virus and esophageal carcinoma in a Latin-American region.
World J Gastroenterol. 2009;15(25):3142-3147.
70. Syrjanen K. Geographic origin is a significant determinant of human
papillomavirus prevalence in oesophageal squamous cell carcinoma: systematic
review and meta-analysis. Scand J Infect Dis. 2013;45(1):1-18.
71. Salaria SN, Abu Alfa AK, Cruise MW, Wood LD, Montgomery EA. Lichenoid
esophagitis: clinicopathologic overlap with established esophageal lichen planus.
Am J Surg Pathol. 2013;37(12):1889-1894.
72. Singhi AD, Arnold CA, Crowder CD, Lam-Himlin DM, Voltaggio L, Montgomery
EA. Esophageal leukoplakia or epidermoid metaplasia: a clinicopathological study
of 18 patients. Mod Pathol. 2014;27(1):38-43.
73. Taggart MW, Rashid A, Ross WA, Abraham SC. Oesophageal hyperkeratosis:
clinicopathological associations. Histopathology. 2013;63(4):463-473.
74. Singhi AD, Arnold CA, Lam-Himlin DM, et al. Targeted next-generation
sequencing supports epidermoid metaplasia of the esophagus as a precursor to
esophageal squamous neoplasia. Mod Pathol. 2017;30(11):1613-1621.
75. Daniels JA, Gibson MK, Xu L, et al. Gastrointestinal tract epithelial changes
associated with taxanes: marker of drug toxicity versus effect. Am J Surg Pathol.
2008;32(3):473-477.
76. Bosman F, Carneiro F, Hruban R, Theise N, eds. WHO Classification of Tumours
of the Digestive System. Lyon: IARC; 2010. Bosman F, Jaffee E, Lakhani S, Ohgaki
H, eds. World Health Organization Classification of Tumours.
77. Modlin IM, Sandor A. An analysis of 8305 cases of carcinoid tumors. Cancer.
1997;79(4):813-829.
78. Lloyd R, Osamura R, Kloppel G, Rosai J. WHO Classification of Tumours of
Endocrine Organs. Lyon: IARC Press; 2017.
79. Deng HY, Ni PZ, Wang YC, Wang WP, Chen LQ. Neuroendocrine carcinoma of the
 esophagus: clinical characteristics and prognostic evaluation of 49 cases with
surgical resection. J Thorac Dis. 2016;8(6):1250-1256.
80. Yachida S, Nakanishi Y, Shimoda T, et al. Adenosquamous carcinoma of the
esophagus. Clinicopathologic study of 18 cases. Oncology. 2004;66(3):218-225.
81. Aulakh K, Chisholm C, Speights V. TTF-1 and napsin-A frequently positive in
esophageal and pulmonary adenocarcinomas: dispelling myths and offering new
insights. Mod Pathol. 2011;24(suppl 1):142A; (Abstract 596).
82. Antonescu CR, Busam KJ, Francone TD, et al. L576P KIT mutation in anal
melanomas correlates with KIT protein expression and is sensitive to specific
kinase inhibition. Int J Canc. 2007;121(2):257-264.
83. Carvajal RD, Antonescu CR, Wolchok JD, et al. KIT as a therapeutic target in
metastatic melanoma. JAMA. 2011;305(22):2327-2334.
84. Carvajal RD, Hamid O, Antonescu CR. Selecting patients for KIT inhibition in
melanoma. Meth Mol Biol. 2014;1102:137-162.
85. Carvajal RD, Lawrence DP, Weber JS, et al. Phase II study of nilotinib in
melanoma harboring KIT alterations following progression to prior KIT inhibition.
Clin Canc Res. 2015;21(10):2289-2296.
86. Gollub MJ, Prowda JC. Primary melanoma of the esophagus: radiologic and
clinical findings in six patients. Radiology. 1999;213(1):97-100.
87. Syrigos KN, Konstadoulakis MM, Ricaniades N, Leandros M, Karakousis CP.
Primary malignant melanoma of the esophagus: report of two cases and review of
the literature. In Vivo. 1999;13(5):421-422.
88. Takubo K, Kanda Y, Ishii M, et al. Primary malignant melanoma of the esophagus.
Hum Pathol. 1983;14(8):727-730.
89. Volpin E, Sauvanet A, Couvelard A, Belghiti J. Primary malignant melanoma of the
esophagus: a case report and review of the literature. Dis Esophagus.
2002;15(3):244-249.
90. Wayman J, Irving M, Russell N, Nicoll J, Raimes SA. Intraluminal radiotherapy
and Nd:YAG laser photoablation for primary malignant melanoma of the
esophagus. Gastrointest Endosc. 2004;59(7):927-929.
91. Yoo CC, Levine MS, McLarney JK, Lowry MA. Primary malignant melanoma of
the esophagus: radiographic findings in seven patients. Radiology.
1998;209(2):455-459.
92. Inadomi K, Kumagai H, Arita S, et al. Bi-cytopenia possibly induced by anti-PD-1
antibody for primary malignant melanoma of the esophagus: a case report.
Medicine. 2016;95(29):e4283.
93. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health
Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-
2390.
 STOMACH 2

CHAPTER OUTLINE
The Unremarkable Stomach
Polyps
□ Pancreatic Heterotopia
□ Hamartomatous Polyps and Syndromic Considerations
• Peutz-Jeghers Polyp
• Juvenile Polyp
• Cronkhite-Canada Syndrome Polyp
• PTEN Hamartoma Tumor Syndrome and Cowden Syndrome
Polyp
□ Epithelial Polyps
• Hyperplastic Polyp
• Fundic Gland Polyp
□ Adenomatous Polyps
• Gastric Adenoma, Intestinal Type
• Gastric Adenoma, Foveolar Type
• Pyloric Gland Adenoma
• Oxyntic Gland Polyp/Adenoma
Adenocarcinoma
□ Approach to the Biopsy
• Tumor Location and Staging: Esophageal Versus Gastric
□ Tumor Classification: Lauren, Ming, and WHO Morphologic
 Variants
□ Risk Factors and Genetic Considerations
• Background Mucosa
• Environmental Risk Factors
• Familial Predisposition
□ Biomarker Testing 101
• Programed Death Receptor-1/Programed Death Ligand-1 (PD-
1/PD-L1)
• Human Epidermal Growth Factor Receptor 2 (HER2)
Well-Differentiated Neuroendocrine Tumors (Formerly “Carcinoid”)
□ Gastric Well-Differentiated Neuroendocrine Tumor, Type I
□ Gastric Well-Differentiated Neuroendocrine Tumor, Type II
□ Gastric Well-Differentiated Neuroendocrine Tumor, Type III
MALT Lymphoma
Mesenchymal Lesions
□ Inflammatory Fibroid Polyp
□ Gastrointestinal Stromal Tumor
□ Leiomyoma
□ Granular Cell Tumor
Near Miss
□ Metastatic Lobular Breast Carcinoma
□ Gastric Xanthoma
□ Gastritis Cystica Profunda
□ Inactive Chronic Gastritis Hides Tumors

THE UNREMARKABLE STOMACH

The stomach consists of the cardiac opening, fundus, body, antrum, and pylorus (Fig.
2.1). These sites perform different physiologic functions and accordingly have variable
histologic characteristics. Familiarity with these features will aid in both gastric polyp
classification and the prognostication of neoplastic processes affected by the presence
of background gastritis. For example, the sporadic gastric neuroendocrine tumor (NET)
(type III) has a relatively worse prognosis than the NET found in autoimmune gastritis
(type I); this important distinction relies upon adequate regional sampling and the
information gleaned therein. See also “Well-Differentiated Neuroendocrine Tumors”
section.
The gastric wall includes the mucosa, submucosa, muscularis propria, and serosa
(Fig. 2.2). The mucosa consists of the epithelium, lamina propria, and muscularis
mucosae. Unlike the deep glands, the surface epithelial component is consistent
throughout the stomach and is composed of mucus-secreting foveolar cells that line
superficial pits (foveolae). Deep to these pits are mucous neck cells and the gastric
glands, whose function and cell composition are region dependent (Fig. 2.3). For
example, deep glands found in the gastric cardia are composed of mucus-secreting cells
with abundant clear foamy cytoplasm (Fig. 2.4). By comparison, deep glands of the
body and fundus contain mixed bluish-purple chief cells, pink parietal cells, and
scattered amphophilic enterochromaffinlike (ECL) endocrine cells (Fig. 2.5). In the
antrum and pylorus, the deep glands are again mucus secreting and are similar in
histology to those in the gastric cardia (Figs. 2.6 and 2.7). Transition zones between
these regions may show a variable mixture of gland histology. For a more detailed
discussion of the unremarkable stomach, see also “Stomach” chapter, Atlas of
Gastrointestinal Pathology: A Pattern Based Approach to Non-Neoplastic Biopsies.
Figure 2-1. The stomach consists of the cardiac opening, fundus, corpus or body, antrum, and
pylorus. The fundus and body are the bulk of the stomach, whereas the gastric antrum
comprises merely 10%.
Figure 2-2. A full-thickness section of the stomach wall shows mucosa composed of epithelium
and lamina propria (E&L) and muscularis mucosae (MM). Beyond the submucosa is the
muscularis propria, which contains three layers of muscle, all of which is covered by serosa
(visceral peritoneum).
Figure 2-3. The mucosa throughout the stomach is lined by surface foveolar mucin cells, which
extend down the gastric pits. Mucous neck cells transition to deep glands, the composition of
which varies by gastric subsite. The muscularis mucosae is the deepest layer of the mucosa
before reaching submucosa. This section is through the gastric corpus (body and fundus) in
which the oxyntic glands comprise approximately 80% of the mucosal thickness, and the
foveolar pits are relatively shallow.
Figure 2-4. The deep glands of the gastric cardia vary between oxyntic type (chief cells and
parietal cells, not pictured) and cardiac type, which are lined by cuboidal cells with abundant
clear cytoplasm and small flattened nuclei pushed toward the basement membrane
(arrowhead ). These cells are indistinguishable from the pyloric glands found in the gastric
antrum. The luminal surface and foveolar pits are lined by foveolar mucin cells (arrow), as it is
throughout the stomach.
Figure 2-5. The deep glands of the gastric body and fundus are lined by oxyntic glands
composed of pink parietal cells and blue chief cells (arrows). Mucous neck cells with clear
foamy cytoplasm and eccentric nuclei (arrowhead ) are normal and should not be mistaken for
signet ring cell carcinoma.
Figure 2-6. The pits of the gastric antrum are lined by foveolar mucin cells and are deeper as
compared with the pits in the corpus, extending to 50% of the mucosal thickness. The deep
pyloric glands of the antrum are cuboidal with clear cytoplasm.
 Figure 2-7. The deep glands of the antrum and pylorus contain pyloric-type glands, which are
similar in appearance as cardiac-type glands, with cuboidal cells containing abundant clear
foamy cytoplasm and small flattened nuclei displaced toward the basement membrane.

The lamina propria between the pits and glands is normally devoid of inflammatory
cells and contains inconspicuous lymphovascular channels accessible for metastatic
spread of tumor cells. The muscularis mucosae is composed of a thin delicate layer of
smooth muscle cells, separating the mucosa from the underlying submucosa, which
contains abundant larger lymphatic and vascular structures, also readily able to
facilitate metastasis. The muscularis propria of the stomach is composed of three sets of
smooth muscle fibers: longitudinal, circular, and oblique. The entire organ is encased by
the mesothelial-derived serosa, which forms the boundary to the peritoneal space.
Staging of invasive tumors requires accurate identification of each of these layers.

POLYPS
Several features contribute to challenges in gastric polyp classification, even for the
skilled pathologist. For example, gastric polyps are far less common than colonic
polyps, resulting in a more recent and sparse body of literature; they show significant
histologic overlap with one another; and classification is influenced by features of the
background flat mucosa. Unlike polyps elsewhere in the tubular gastrointestinal (GI)
tract, which are frequently isolated findings, polyps of the stomach often arise in
association with an inflammatory backdrop or a polyposis syndrome. Thus, although
they may be troublesome to classify, careful attention to the background mucosa with a
deliberate effort toward an integrated interpretation will provide important information
about prognosis or risk for familial syndromes. Gastric polyps are found in 6% of upper
endoscopies and are seen as projections above the adjacent flat mucosa (Fig. 2.8).1
These proliferative lesions may arise from the epithelium (most common) or from other
compartments in the mucosa and submucosa. Table 2.1 lists the most common gastric
polyps by chief proliferative compartment and serves as an outline for this segment.
This section provides a diagnostic approach for epithelial and hamartomatous polyps;
mesenchymal lesions are discussed in a dedicated chapter (see “Mesenchymal Lesions”
chapter). An understanding of the normal regions and histologic compartments,
reviewed at the beginning of this chapter, will facilitate application of this approach.

Figure 2-8. Endoscopically, gastric polyps appear exophytic and are mucosal or submucosal
based. All gastric polyps require histologic diagnosis.
 TABLE 2.1: Gastric Polyps: Categorized by Proliferative Compartment

Heterotopic Polyps
• Pancreatic acinar heterotopia

Hamartomatous Polyps
• Peutz-Jeghers polypa
• Juvenile polypa
• Cronkhite-Canada syndrome–associated polypa
• PTEN hamartoma tumor syndrome and Cowden syndrome–associated polypa

Epithelial Polyps and Hyperplasias

• Hyperplastic polyp, polypoid foveolar hyperplasia, foveolar polyp
• Fundic gland polypa
• Adenomatous polyps
• Intestinal type
• Gastric type, foveolar adenoma
• Gastric type, pyloric gland adenoma
• Gastric type, oxyntic gland polyp/adenoma
• Carcinomatous polyp, primary or metastatic
• Neuroendocrine tumors

Mesenchymal
• Inflammatory fibroid polyp
• Gastrointestinal stromal tumor
• Leiomyoma
• Vascular lesions
• Granular cell tumor

Other findings that may appear polypoid

• Lymphoid hyperplasia
• Lymphoma
• Xanthoma
• Granuloma
• Amyloidosis
• Hemosiderosis
• Calcium deposit
• Gastritis cystica profunda

aPolyps associated with clinical syndromes.

PTEN, phosphatase and tensin homolog.

PANCREATIC HETEROTOPIA
Heterotopic pancreatic tissue is an incidental finding most commonly seen in the wall of
the gastric antrum. Endoscopically, it can appear as a submucosal nodule with central
umbilication and surface erosion (Fig. 2.9). These are commonly biopsied with clinical
concern for a gastrointestinal stromal tumor (GIST), which is endoscopically similar in
appearance. However, because the bulk of these lesions are submucosal, superficial
biopsies may lack diagnostic material and rebiopsy may be necessary. When well
sampled or mucosally resected, these lesions consist of a variable admixture of benign
pancreatic acinar cells, islets, and ducts, the latter of which may even connect to the
mucosal surface for drainage in larger lesions (Figs. 2.10–2.14). These structures
attempt to recapitulate normal pancreatic tissue and maintain a lobular architecture but
may appear slightly disorganized. Although the pancreatic acinar cells may be sparse,
they are identified by their typical triangular shape, eccentric round nuclei, and abundant
eosinophilic to amphophilic granular cytoplasm. Importantly, there is no cytologic
atypia, infiltrative border, or desmoplastic stromal reaction. Adjacent or overlying
mucosa may show reactive changes or intestinal metaplasia (IM) owing to the digestive
secretions of the pancreatic acinar cells.

FAQ: Is there significant difference between pancreatic heterotopia and

pancreatic acinar cell metaplasia?

Yes! Pancreatic heterotopia is incidental, and, aside from possible gastric

outlet obstruction or ulceration in larger examples, the finding is generally
inconsequential and not associated with inflammatory conditions. By
comparison, the finding of pancreatic acinar cell metaplasia (PAM) should alert
the pathologist to examine the background mucosa more closely for possible
autoimmune metaplastic atrophic gastritis (AMAG) as nearly half of AMAG
cases contain PAM.2 Histologic distinction between heterotopia and metaplasia
is not always possible, but some clues can favor one over the other. For
example, because pancreatic heterotopia is ectopic pancreatic tissue, it often
involves the submucosa and contains heterogeneous pancreatic cell types (i.e.,
some combination of acini, ducts, and islets). In contrast, PAM is the result of
single cell transformation (gastric cells to acinar cells) and thus is limited to the
mucosa and is homogeneous (composed of acinar cells only) (Fig. 2.15).

PEARLS & PITFALLS

When stumped by a gastric nodule that resembles a well-differentiated NET but
fails to stain for neuroendocrine markers such as chromogranin and
synaptophysin, consider PAM or pancreatic heterotopia. Often, simply
remembering this entity results in an “ah-ha” moment of diagnosis. Pancreatic
heterotopia may contain small ducts, a helpful clue. In difficult cases,
immunohistochemistry for trypsin or chymotrypsin is positive in pancreatic
acinar cells.

Figure 2-9. Pancreatic heterotopia. A submucosal nodule with a central umbilication is seen in
the gastric antrum.
Figure 2-10. Pancreatic heterotopia. Pancreatic heterotopia is found in the wall of the stomach.
This EMR specimen shows deep submucosal location as well as mucosal involvement. Note
the lobulated appearance of the glands.

Figure 2-11. Pancreatic heterotopia. This example contains a few lobules of pancreatic acini in
the submucosa. The surface epithelium shows mild erosive and reactive changes.
Figure 2-12. Pancreatic heterotopia. Most examples are straightforward with well-developed
pancreatic acinar lobules, as seen here. In challenging cases, the presence of a duct (arrow)
can be a helpful clue.
Figure 2-13. Pancreatic heterotopia. Some cases can resemble neuroendocrine tumors, but
careful examination for a population of pancreatic acinar cells with eosinophilic granular
cytoplasm (arrows) can steer one away from this pitfall. Trypsin or chymotrypsin immunostains
would also be reactive in pancreatic acinar cells.
Figure 2-14. Pancreatic heterotopia. A small pancreatic duct (arrow) is a helpful clue in
differentiating pancreatic heterotopia from pancreatic metaplasia and neuroendocrine tumor.
 Figure 2-15. Pancreatic acinar metaplasia (PAM). These pancreatic acinar cells (arrow) are
admixed with normal antral glands and lack the lobulated organization of pancreatic heterotopia.
The acinar cells of either PAM or pancreatic heterotopia may secret enzymes causing adjacent
intestinal metaplasia (arrowheads).

HAMARTOMATOUS POLYPS AND SYNDROMIC

CONSIDERATIONS
Hamartomatous polyps result from disordered growth of tissues native to the site and
can arise from any of the three embryonic layers. These lesions are frequently, but not
universally, associated with a clinical syndrome. Compared with their small bowel and
colonic counterparts, hamartomatous polyps in the gastric mucosa have nonspecific
histology and may be indistinguishable from hyperplastic polyps even in a patient with a
known polyposis syndrome.3 Therefore, this section covers common diagnostic
challenges of gastric hamartomatous polyps and some of the syndromic considerations,
and additional discussion of polyposis syndromes can be found in the “Small Bowel”
chapter.

Peutz-Jeghers Polyp
Peutz-Jeghers syndrome is characterized by hamartomatous polyps of the GI tract and
melanocytic mucocutaneous hyperpigmentation.4 Up to 25% of documented cases are
sporadic, but this condition is best known as an autosomal dominant inherited syndrome
with 80% of affected families harboring a germline mutation in the STK11/LKB1
gene.5,6 Patients with Peutz-Jeghers syndrome have a 93% cumulative lifetime risk for
cancer, including carcinomas of the GI tract, breast, ovary, and testis.7-9 In this context,
early recognition of the syndrome allows for appropriate screening and surveillance for
patients and family members. World Health Organization (WHO) criteria for the clinical
diagnosis of Peutz-Jeghers syndrome are:

1. Detection of three or more histologically confirmed Peutz-Jeghers polyps.

or
2. The presence of any number of Peutz-Jeghers polyps in a patient with a family
history of the syndrome.
or
3. Detection of characteristic, prominent mucocutaneous pigmentation in the patient
with a family history of the syndrome.
or
4. Detection of any number of Peutz-Jeghers polyps in a patient with prominent
mucocutaneous pigmentation.10

1Three of the above four potential methods of diagnosis include histologic

identification of Peutz-Jeghers polyps, making pathologic recognition a necessary skill.
These hamartomatous polyps have a characteristic appearance, showing compactly
spaced glands supported by an arborizing framework of well-developed smooth muscle
that is contiguous with the muscularis mucosae (Figs. 2.16–2.23). Gastric polyps, which
occur in about 15% to 30% of syndromic patients, are less frequent than small bowel
(64%) or colonic polyps (53%).11 These lesions of the small bowel and colon are not
only more common but also highly distinctive; their intact lamina propria with
arborizing smooth muscle fibers helps differentiate them from juvenile polyps. Gastric
Peutz-Jeghers polyps, on the other hand, are routinely indistinguishable from
nonspecific gastric hyperplastic polyps or other syndromic gastric polyps. Clues, such
as unexplained wisps of smooth muscle in the lamina propria and lobular architecture of
the glands (Figs. 2.20–2.23), can improve the sensitivity for hamartomatous polyps of
Peutz-Jeghers type.3,12 Dysplasia is rarely found in these polyps, but patients with the
syndrome have significant risk for malignancy elsewhere, including gastric
adenocarcinoma outside of the polyp.13 Given these diagnostic challenges and the
clinical implications for both patients and families, a low threshold should be
maintained for diagnosis of gastric hamartomatous polyps. Suggesting further
bidirectional endoscopy may provide more definitive histologic evidence for Peutz-
Jeghers polyps at other sites.
Figure 2-16. Well-developed gastric Peutz-Jeghers polyp. At scanning magnification, an
arborizing smooth muscle base is evident along with disorganized, branching, and cystic
glands.
Figure 2-17. Higher magnification of the previous figure. Arborizing bundles of smooth muscle
splay out at various angles.
Figure 2-18. Gastric Peutz-Jeghers polyp. This example shows broad areas of foveolar
hyperplasia, lamina propria edema, and disorganized gland architecture with dilated glands and
abundant admixed smooth muscle.
Figure 2-19. Gastric Peutz-Jeghers polyp. This polyp in the gastric body shows a prominent
core of disorganized smooth muscle with branches extending at various angles.

Figure 2-20. Higher magnification of the previous figure. Disorganized small bundles and wisps
of smooth muscle are present in the lamina propria and stream at intersecting angles.
Figure 2-21. Gastric Peutz-Jeghers polyp. This polyp is not as well developed as previous
examples, but it was retrieved from a patient known to have Peutz-Jeghers syndrome. The
glands appear organized and lack dilation or architectural changes, but note the presence of
smooth muscle within the lamina propria (arrowheads).
Figure 2-22. Higher magnification of the previous figure. Small disorganized bundles and wisps
of smooth muscle expand the lamina propria and branch in different directions.

Figure 2-23. Gastric Peutz-Jeghers polyp. Wisps of intersecting smooth muscle surround
glands creating small lobules, a helpful clue in some examples.
 PEARLS & PITFALLS
Although a low threshold for diagnosis of hamartomatous polyps is advocated,
caution is advised against using gastric polyps to fulfill any of the WHO criteria
requiring “histologically confirmed” Peutz-Jeghers polyps. In the stomach, the
histologic features are not reliable enough to differentiate syndromic polyps
from reactive lesions such as inflammatory/hyperplastic polyps, and the
implications are considerable. Instead, compose a descriptive sign out with an
explanatory note (see the following note). If the patient has prior GI polyps, it
may be worthwhile to review the histology.

SAMPLE NOTE: Gastric Polyp With Features

Suggesting Hamartomatous Versus
Inflammatory/Hyperplastic Polyp
Stomach, Antrum, Polyp, Biopsy
• Gastric polyp with features suggestive of hamartomatous polyp, see Comment.

Comment
There are no reliable histologic features to distinguish gastric hamartomatous polyps
from reactive lesions (i.e., gastric inflammatory/hyperplastic polyps) or to reliably
differentiate subtypes of hamartomatous polyps. Nevertheless, the current specimen
contains some features that suggest hamartomatous development, such as cystically
dilated irregular glands and admixed smooth muscle. Syndromes involving
hamartomatous GI polyps (e.g., Peutz-Jeghers syndrome, juvenile polyposis syndrome,
Cowden syndrome) should be ruled out clinically. In addition, bidirectional endoscopy
is advised, with biopsy of any polypoid lesions in either the upper or lower GI tract, as
polyps found outside the stomach are more likely to retain pathognomonic features.

FAQ: I have a patient with single classic Peutz-Jeghers polyp, but this patient
has no other polyps or family history of Peutz-Jeghers syndrome. What does
this mean?

Rare instances of patients with isolated sporadic Peutz-Jeghers polyps

have been documented. In nearly all instances, these patients had clinical
 histories suggesting Peutz-Jeghers syndrome (e.g., concurrent pancreatic
cancer, family GI cancer history, metachronous tumors) but failed to meet the
WHO criteria. Thus, isolated or sporadic Peutz-Jeghers polyps may occur, but
clinicians should be advised that these patients appear to have a cumulative
lifetime risk of malignancy similar to those with the syndrome14 (Figs. 2.24 and
2.25).

Juvenile Polyp
Juvenile polyposis syndrome, the most common of the hamartomatous polyposis
syndromes, affects one in 100,000. The syndrome is largely sporadic but can be
inherited as autosomal dominant familial syndrome (30%).15 Both inherited and
sporadic forms share similar genetics, with germline mutations in SMAD4 (also known
as DPC4) (15%) and the related gene BMPR1A (25%), whereas ENG is associated
with early childhood presentation.16,17 These genetic changes cause a disruption in the
transforming growth factor beta signal transduction pathway and result in an increased
risk of malignancy. The overall risk of GI malignancies in these patients is 55%, with
colorectal cancer presenting at an average age of 37 years.18,19
Figure 2-24. Gastric hamartomatous polyp. This gastric polyp was retrieved from a patient
without a known syndrome but shows brightly eosinophilic disorganized smooth muscle
bundles within the lamina propria suggestive of a hamartomatous polyp.
Figure 2-25. Gastric hamartomatous polyp, higher magnification of the previous figure. The
foveolar epithelium is separated by disorganized smooth muscle bundles that stream at
intersecting angles. Isolated sporadic Peutz-Jeghers polyps are rare but can occur. These
patients share the same cumulative lifetime risk of cancer as patients with Peutz-Jeghers
syndrome.
Figure 2-26. Juvenile polyp. This gastric polyp was retrieved from a patient known to have
juvenile polyposis syndrome. There are broad fingerlike projections with an excess of lamina
propria and inflammation. The glands are abnormally shaped and dilated.

Figure 2-27. Juvenile polyp. These polyps can resemble inflammatory polyps owing to their
edematous lamina propria, chronic inflammation, dilated glands, and surface erosion.
 WHO criteria for the clinical diagnosis of juvenile polyposis syndrome are:

1. More than three to five juvenile polyps of the colorectum

or
2. Juvenile polyps throughout the GI tract
or
3. Any number of juvenile polyps with a family history of juvenile polyposis20

In the stomach, juvenile polyps are usually 3- to 20-mm solitary pedunculated

lesions with a predilection for the gastric antrum. Like Peutz-Jeghers polyps, juvenile
polyps are classified as hamartomatous lesions, but their histologic profiles differ.
Juvenile polyps consist primarily of an excess of lamina propria and show abundant
distorted and dilated glands (Figs. 2.26–2.30). The combination of lamina propria
edema and abundance of distended, mucus-filled glands in combination with
inflammatory cells is occasionally mistaken for “inflammatory” or “retention” polyp.
This is understandable because solitary or sporadic juvenile polyps may be
indistinguishable from inflammatory/retention polyps in either the upper or lower GI
tract. Fortunately, and in contrast to sporadic Peutz-Jeghers polyps, there is no
documented increased lifetime risk for malignancy reported for sporadic juvenile
polyps. Nevertheless, foci of dysplasia are regularly seen in juvenile polyps,
underscoring their neoplastic potential and the value of recognition. In the absence of
sufficient clinical history, and when the histology precludes definitive classification,
patients benefit from a more inclusive diagnosis, such as “juvenile/inflammatory polyp”
and a careful note (see the earlier sample note for Peutz-Jeghers syndrome). When
multiple similar polyps are encountered, the possibility of a polyposis syndrome should
be stated. On the other hand, take care not to label these patients prematurely, as
classification of single gastric polyps is often irresolvable. Other syndromes involving
hamartomatous GI polyps should be ruled out clinically or by pathologic examination
(Figs. 2.31–2.34).
Figure 2-28. Juvenile polyp. Broad and blunt projections characterize this polyp from a patient
with known juvenile polyposis syndrome. The lamina propria is edematous and contains
inflammatory cells. Surface erosion is present with a fibroinflammatory layer.
Figure 2-29. Juvenile polyp. Excess lamina propria and disorganized dilated glands are present.
Figure 2-30. Juvenile polyp. Sparse abnormally shaped glands are present with abundant
edematous lamina propria, surface erosion, and inflammation. The histologic features are not
specific to juvenile polyp, but this example is from a patient known to have the syndrome.
Figure 2-31. Juvenile polyp. Surface erosion, granulation tissue formation, and a
fibroinflammatory coating are common nonspecific findings. These changes are also found in
inflammatory polyps and large hyperplastic polyps.
Figure 2-32. Juvenile polyp. This gastric polyp was retrieved from a patient known to have
juvenile polyposis syndrome, and it shows features characteristic of juvenile polyps: dilated
glands, surface foveolar hyperplasia, and lamina propria edema with inflammation. However, it
also contains bundles of smooth muscle scattered in the lamina propria (arrowheads), a
feature that overlaps with Peutz-Jeghers polyps.
Figure 2-33. Juvenile polyp, higher magnification of the previous figure. The lamina propria
contains smooth muscle bundles (arrowheads) streaming at various angles. One might
consider a diagnosis of Peutz-Jeghers polyps based on this finding; the patient, however, was
known to have juvenile polyposis syndrome.
 Figure 2-34. Juvenile polyp, separate high-magnification area of the previous figure. Small
disorganized bundles of smooth muscle within gastric polyps, such as seen here, are a red flag
for hamartomatous polyps, particularly the Peutz-Jeghers type. This example, however, is from
a patient with juvenile polyposis syndrome, emphasizing the histologic overlap among gastric
syndromic polyps.

PEARLS & PITFALLS: An Approach to Unusual or Potentially Syndromic

Gastric Polyps
Akin to gastric Peutz-Jeghers polyps, the gastric juvenile polyp can defy
classification and may be indistinguishable from the polyps of other syndromes
(e.g., Peutz-Jeghers and Cronkhite-Canada syndrome) and common
hyperplastic polyps. When confronted with an unusual gastric polyp the
following approach is recommended: (1) use a low threshold to report juvenile
polyps (or any hamartomatous polyp); (2) investigate into the patient’s medical
record for family history, clinical presentation, and endoscopic findings; (3)
review any prior pathology specimens, such as previous GI polyps; and (4)
communicate with the clinician for a multidisciplinary approach to follow-up or
surveillance.

Cronkhite-Canada Syndrome Polyp

Cronkhite-Canada syndrome is a rare, noninherited clinical condition characterized by
GI hamartomatous polyposis and the dermatologic triad of alopecia, onychodystrophy,
and hyperpigmentation.21,22 Consider this syndrome when numerous biopsies show
juvenile polyp–like features: cystically dilated and tortuous glands containing
proteinaceous fluid or inspissated mucus with a background lamina propria showing
marked edema and chronic inflammation (Figs. 2.35–2.41).23 At first glance, the
changes resemble inflammatory-type polyps and are nonspecific in the absence of
clinical information, but the tip-off will be the diffuse nature of the changes and the lack
of intervening normal mucosa. Given the 55% 5-year mortality rate, consideration for
and recognition of this condition is critical such that patients receive appropriate
treatment.

Figure 2-35. Cronkhite-Canada syndrome. Diffuse polyposis is seen in the stomach (pictured)
and throughout the upper and lower gastrointestinal tract.
Figure 2-36. Cronkhite-Canada syndrome, gastric antrum. Multiple biopsies of the stomach
show the diffuse nature of this disease. The mucosa is atrophic with abundant edematous
lamina propria, similar to that seen in juvenile polyps. The similarities between the antrum,
body, and fundus (see the next two figures) emphasize the diffuse process.
Figure 2-37. Cronkhite-Canada syndrome, gastric body. These biopsies from the gastric body
show cystically dilated glands and marked lamina propria edema. At first glance, they resemble
juvenile polyps. However, these changes are found in random samples of nonpolypoid mucosa.
One of the key diagnostic clues to Cronkhite-Canada syndrome is the involvement of
nonpolypoid mucosa.
Figure 2-38. Cronkhite-Canada syndrome, gastric fundus. As in the previous two figures,
multiple biopsies of the gastric fundus show diffuse changes involving all tissue fragments.
There is atrophy of the oxyntic glands, abundant lamina propria edema, and cystically dilated
glands with inspissated proteinaceous material.
Figure 2-39. Cronkhite-Canada syndrome. Polypoid and nonpolypoid mucosa is histologically
similar to juvenile polyps. The edematous lamina propria contains variable inflammation, and
focal areas may show erosion and reactive epithelial changes.
Figure 2-40. Cronkhite-Canada syndrome. There is an excess of edematous lamina propria
and dilated irregular glands. In isolation, the findings suggest a juvenile polyp or inflammatory
polyp. When these changes are found diffusely in both polypoid and nonpolypoid mucosa,
consider Cronkhite-Canada syndrome and look for the clinical triad of alopecia,
onychodystrophy (changes in nail color or quality), and hyperpigmentation.
Figure 2-41. Cronkhite-Canada syndrome. The mucosa contains cystically dilated glands and
edematous lamina propria with sparse inflammatory cells, similar in appearance to juvenile
polyps.
 Figure 2-42. Cowden syndrome gastric polyp. Gastrointestinal lesions in PTEN hamartoma
tumor syndrome or Cowden syndrome include hamartomatous polyps, lipomas,
ganglioneuromas, and inflammatory polyps. This gastric hamartomatous polyp from a patient
with Cowden syndrome contains mucosal adipocytes (arrowhead ) and dilated, distorted
glands.

PTEN Hamartoma Tumor Syndrome and Cowden Syndrome Polyp

Cowden syndrome is inherited in an autosomal dominant fashion and is the best
described phosphatase and tensin homolog (PTEN) hamartoma syndrome with 66% to
100% of affected individuals reported to have gastric or duodenal polyps.24-26 The
range of clinical manifestations of Cowden syndrome includes mucocutaneous and
extracutaneous hamartomatous tumors in multiple organ systems and characteristic
dermatologic manifestations, such as trichilemmomas, oral fibromas, and punctate
palmoplantar keratoses. This syndrome is highly associated with an increased risk of
carcinomas of the breast (cumulative risk as high as 85%), endometrium (13% to 28%),
thyroid (3% to 35%), kidney (13% to 34%), and colorectum (66% to 93%).25-27 This
increased risk of malignancy underscores the importance of reporting any
hamartomatous polyp within the gastric mucosa and thus alerting clinicians to screen for
a polyposis syndrome. Gastrointestinal lesions include hamartomatous polyps, lipomas,
ganglioneuromas, and inflammatory polyps (Figs. 2.42–2.44). These lesions may also
be found in Bannayan-Riley Ruvalcaba syndrome and adult Lhermitte-Duclos disease,
both rare disorders characterized by PTEN mutations. Proteus and proteus-like
syndromes, although also belonging to the PTEN mutation family, do not feature GI
hamartomas as a prominent finding.

Figure 2-43. Cowden syndrome gastric polyp. This polyp shows lobules of glands, similar to
that of Peutz-Jeghers polyps, but the presence of adipocytes (arrowhead ) and neural tissue
(arrow) argue against this diagnosis. This mixed hamartomatous polyp is from a patient with
Cowden syndrome.
 Figure 2-44. Cowden syndrome gastric polyp. The lamina propria between these distorted
gastric glands is myxoid and contains hamartomatous elements, such as adipose tissue.

PEARLS & PITFALLS: An Algorithm for Separating Gastric Hyperplastic and

Hamartomatous Polyps
When encountering an unusual hyperplasticlike gastric polyp (e.g., containing
foveolar hyperplasia, dilated and distorted glands, lamina propria edema, and
chronic inflammation), always consider the possibility of a hamartomatous polyp
or syndrome, even in the absence of any obvious hamartomatous element. An
algorithmic approach is provided in Fig. 2.45.

EPITHELIAL POLYPS
Epithelial polyps are the most commonly encountered gastric polyps. These include
gastric hyperplastic polyps, fundic gland polyps (FGPs), and adenomatous polyps, all of
which are associated with distinctly different clinical contexts, as discussed later. Less
common epithelial lesions manifesting as polyps include NETs, discussed separately in
this chapter.

Hyperplastic Polyp
These benign epithelial proliferations are the second most common type of gastric
polyp, and, in contrast to the incidental colonic hyperplastic polyp, gastric hyperplastic
polyps are highly associated with background mucosal injury (85%), including
Helicobacter infection (25%), reactive/chemical gastropathy (21%), AMAG (12%),
and environmental metaplastic atrophic gastritis (8%).28 Other associated conditions
include mucosal ulcerations and erosions, ostomy sites, and gastroesophageal reflux
disease, which, notably, are also forms of mucosal injury. Although they may be found at
any age, gastric hyperplastic polyps occur more frequently with increasing age (mean
age 65 to 75 years) and are female predominant. The lesions are solitary in 75% of
cases and, owing to their foveolar origin, are found in all regions of the stomach with
fairly even distribution. When large, these lesions may cause gastric outlet obstruction
(Fig. 2.46); recurrence following polypectomy or endoscopic resection is common (up
to 50%).29,30

Figure 2-45. A flow diagram for interpreting unusual gastric polyps. The histologic overlap of
hyperplastic and syndromic gastric polyps can be a challenge during routine signout, but most
can be handled by applying the simple algorithmic approach illustrated in this figure.
 Figure 2-46. Gastric hyperplastic polyp. This antral-based gastric polyp protrudes into the
lumen with a lobulated appearance. The entire lesion should be resected for histologic
evaluation, and a separate jar containing tandem biopsies of the surrounding flat mucosa
should also be submitted by the endoscopist to evaluate for background etiologic factors (e.g.,
H. pylori, AMAG).

Histologically, the polypoid mucosa often has a broad pedicle and shows elongated
and distorted pits lined by a single layer of foveolar epithelial cells. There is wide
histologic variability: the gastric pits and glands may show cystic dilation separated by
edematous lamina propria with mixed inflammatory cells (Fig. 2.47), or there may be
glandular crowding with gastric foveolar hyperplasia (Fig. 2.48). Surface erosion is
common and may be accompanied by reactive epithelial changes (Figs. 2.49 and 2.50).
Always ensure that a quick search for IM and dysplasia is performed (Figs. 2.51 and
2.52). Given the high association with surrounding mucosal inflammation and damage,
and the potential for neoplastic progression, pathologists should carefully review the
background flat mucosa when such material is available. Because not all endoscopists
are in the habit of doing so, this necessitates advising them to submit separate samples
of flat mucosa every time a gastric polyp is encountered. Recommended biopsy
protocols exist, the most widely accepted of which is the Sydney protocol, which
includes five mucosal samples: two each from greater and lesser curvatures (to include
both antrum and body) and one from incisura (transition zone).
Figure 2-47. Gastric hyperplastic polyp. Gastric hyperplastic polyps have a wide range of
appearances. This example shows a broad, rounded fingerlike projection at low magnification,
with tortuous distorted and dilated glands, abundant lamina propria, and surface erosion. The
features are similar to the juvenile polyp in Fig. 2.25, but this patient does not have a history of
juvenile polyposis syndrome.
Figure 2-48. Gastric hyperplastic polyp. The histologic spectrum ranges from stroma-rich, like
the previous example, to gland-rich, as seen here. This polyp contains tightly packed glands
and marked foveolar hyperplasia with elongated tortuous foveolar pits, similar to
chemical/reactive gastropathy. In this case, the gastric pits comprise about 80% of the mucosal
thickness (normal is up to 25% in the body/fundus and up to 50% in the antrum).
Figure 2-49. Reactive atypia in a gastric hyperplastic polyp. Epithelial cells may show reactive
changes, such as mild nuclear hyperchromasia and enlargement due to inflammation or
erosion.
Figure 2-50. Reactive atypia in a gastric hyperplastic polyp, higher magnification of the previous
figure. Reassuring features for reactive atypia are the presence of preserved nuclear to
cytoplasmic ratio, indistinct nucleoli, lack of nuclear crowding, and absence of an abrupt
transition from normal.
Figure 2-51. Low-grade dysplasia arising in a hyperplastic polyp. Most dysplasia can be
identified at low magnification by the presence of stark hyperchromasia and nuclear crowding.
An abrupt transition from normal to dysplastic (arrow) is a helpful clue. The nuclei in this
example remain elongated, and the glandular architecture is simple, supporting classification
as low-grade dysplasia.
 Figure 2-52. High-grade dysplasia. Both glands and cells are crowded in this example. The
glands are back to back and architecturally complex. The cells are no longer elongated and
show loss of nuclear polarity. Prominent nucleoli and irregular nuclear contours are present.
Frequent mitoses and apoptoses are evident.

PEARLS & PITFALLS: The Background Flat Mucosa May Reveal the Etiology
of the Hyperplastic Polyp
Always search for an etiologic factor when encountering these reactive lesions.
Treatment of an underlying condition, such as eradication of Helicobacter
infection, can prevent further polyp formation and recurrence.

SAMPLE NOTE: Gastric Hyperplastic Polyp

Submitted Without Background Flat Mucosa
Stomach, Antrum, Polyp, Biopsy
• Gastric hyperplastic polyp, see Comment.

Comment
Gastric hyperplastic polyps are reactive lesions highly associated with mucosal injury
and inflammation. Should repeat endoscopy be performed, separately submitted
biopsies of the background flat mucosa to evaluate the gastric environment would be of
interest.

Despite their classification as a regenerative growth, large (>2 cm) polyps may
harbor IM or dysplasia and subsequent risk of malignant transformation (2% to
20%).28,31-33 Similar to other areas of the GI tract, gastric dysplasia is categorized as
negative for dysplasia, low-grade dysplasia (LGD), high-grade dysplasia (HGD), or
indefinite for dysplasia. The microscopic features also parallel those seen in other areas
of the GI tract with LGD showing epithelial changes of hyperchromasia, nuclear
elongation, and pseudostratification extending to the mucosal surface (Fig. 2.51). Other
features include nuclear atypia, prominent nucleoli, and increased mitoses and
apoptoses. High-grade epithelial dysplasia is characterized by greater nuclear
pleomorphism, anisonucleosis, loss of cell polarity, and architectural complexity with
back-to-back or cribriform gland formation (Fig. 2.52). These descriptive terms sound
quite similar to the words used to describe the conventional tubular adenoma, yet
gastric dysplasia remains slightly more challenging to interpret. Contributors to this
struggle include the array of gastric dysplastic lesions (e.g., gastric adenomas include
foveolar, intestinal, pyloric gland, and oxyntic gland types; see later discussion) and the
frequency of background gastric inflammatory changes, from which one must
differentiate reactive and dysplastic lesions. The abrupt histologic transition of
dysplasia remains the best practical tool in dysplasia assessment and is evident even at
low magnification.
Figure 2-53. Gastric antral prolapse. Gastric antral prolapse can produce polypoid areas, not to
be mistaken for hyperplastic polyps. The presence of tortuous gastric pits and exuberant
smooth muscle (arrows) streaming perpendicular to the luminal surface are features of antral
prolapse.
 Figure 2-54. Gastric hyperplastic polyp. This example is small, rounded and contains an
excess of pale edematous lamina propria, variably distorted glands, and inflammation.
Compare this example with the juvenile polyp from Fig. 2.26. There are no reliable histologic
features to differentiate the two entities aside from the presence of background mucosal
damage (e.g., H. pylori, AMAG, gastritis), which favors a gastric hyperplastic/inflammatory
polyp.

The gastric hyperplastic polyp is plagued with challenges in differential diagnoses.

Gastric mucosal prolapse may form polypoid areas, but this finding is usually limited to
the gastric antrum. Histologically, mucosal prolapse lacks the highly edematous stroma
and cystic dilated glands/pits of hyperplastic polyps. Instead, the lamina propria
contains streaming strands of smooth muscle arranged perpendicular to the gastric
surface, and pits show corkscrew tortuosity such as that seen in chemical gastropathy
(Fig. 2.53). Several other conditions enter into the differential diagnosis, most of which
show so much histologic overlap as to be indistinguishable from one another. The most
common of these is the inflammatory polyp, which contains variable degrees of
inflammation in combination with features also seen in hyperplastic polyps, such as
foveolar hyperplasia, cystic and distorted glands, lamina propria edema, and surface
ulceration/erosion. In some instances, these features show so much histologic overlap
with the hyperplastic polyp that the inflammatory polyp probably represents the same
process at different temporal intervals. That is to say, inflamed hyperplastic polyps may
be indistinguishable from inflammatory polyps, and when the inflammation of an
inflammatory polyp subsides, it may resemble a hyperplastic polyp or juvenile polyp
(Figs. 2.47 and 2.54). Interpretation and sign out of these overlapping gastric lesions
can be handled in a fashion similar to that for the gastric hamartomatous polyps (see
later discussion). Some observers prefer to lump these lesions together, signing them as
“inflammatory/hyperplastic polyps.”
Other differential diagnoses include hamartomatous polyps such as juvenile polyps
and Peutz-Jeghers polyps (Fig. 2.55), which also can be histologically indistinguishable
in the stomach despite having highly characteristic morphology in the colon and small
bowel.3 Thus, if there is suspicion for a hamartomatous polyposis syndrome after the
review of gastric biopsies, include a diagnosis comment suggesting that bidirectional
endoscopy could identify other sites of involvement and more characteristic histology.
Other hamartomatous tumor syndromes, such as the PTEN hamartomatous tumor
syndromes (i.e., Cowden syndrome, Bannayan-Riley-Ruvalcaba syndrome, and adult
Lhermitte-Duclos disease) and Cronkhite-Canada syndrome also display polyps that
overlap histologically with gastric hyperplastic polyps. However, attention to clues
such as nonepithelial components, such as adipocytes or neural proliferations within the
polyps, should alert one to the possibility of a PTEN syndrome (Figs. 2.42–2.44).
Cronkhite-Canada syndrome, although often referred to as a hamartomatous polyposis
syndrome, actually represents diffuse mucosal changes that endoscopically appear
nodular and/or polypoid owing to intervening areas of mucosal atrophy. These
submitted “polyps” are similar to hyperplastic polyps as they contain dilated and cystic
glands, lamina propria edema, and variable inflammation. Clues to diagnosis include
diffuse changes throughout the upper and lower GI tract in both the polypoid and
nonpolypoid mucosa and the finding of other ectodermal features (e.g., onychodystrophy
and alopecia) (Figs. 2.35–2.41). Changes found in Menetrier disease are limited to the
gastric body and fundus but are diffuse with striking foveolar hyperplasia and an
absence of intervening normal mucosa (Fig. 2.56). The provided algorithmic approach
(Fig. 2.45) can help differentiate these challenging cases (Fig. 2.57).
Figure 2-55. Gastric hyperplastic polyp. This hyperplastic polyp features crowded glands with
architectural distortion and foveolar hyperplasia, in contrast to the previous example, which was
more edematous and less cellular. The small disorganized bundles of smooth muscle
(arrowheads) in the lamina propria raise the possibility of a hamartomatous polyp, but the
background mucosa in this case showed erosive iron pill gastritis (not pictured), supporting this
as a reactive lesion over a hamartomatous polyp.
 Figure 2-56. Menetrier disease. Compared with the foveolar hyperplasia seen in the previous
figure, the marked foveolar hyperplasia in Menetrier disease is diffuse, giving rise to giant
gastric folds throughout the stomach.

CHECKLIST: Diagnostic Considerations for Gastric Hyperplastic Polyps

□ Always check background flat mucosa for inflammatory causes, such as Helicobacter
pylori or autoimmune gastritis
□ Gastric hyperplastic polyp
□ Inflammatory polyp
□ Gastric antral prolapse
□ Gastric Peutz-Jeghers polyp
□ Gastric juvenile polyp
□ PTEN syndromes (Cowden syndrome, Bannayan-Riley-Ruvalcaba syndrome, adult
Lhermitte-Duclos disease)
□ Cronkhite-Canada syndrome
□ Menetrier disease

Fundic Gland Polyp

Fundic gland polyps (FGPs) comprise 77% of all gastric polyps,1 are found exclusively
in areas where fundic glands reside (i.e., body and fundus only), and are typically
sessile and small (<2 mm, and rarely >1 cm). They may arise either sporadically or
associated with a syndrome, the best known of which is familial adenomatous polyposis
(FAP) syndrome, and this results in unique clinicopathologic characteristics for each
group (see Table 2.2). Sporadic FGPs are usually single or few in number (Fig. 2.58)
and are asymptomatic incidental findings that, in contrast to hyperplastic polyps, are not
associated with an inflammatory or atrophic mucosal backdrop. In fact, their presence is
inversely correlated with Helicobacter infection and active gastritis.34 In the sporadic
setting, it is widely accepted that proton pump inhibitor (PPI) use is associated with
FGPs,35 and although these lesions reportedly regress with medication cessation, no
proven causal pathogenetic relationship between PPIs and FGPs exists.36 Other clinical
associations include gastroesophageal reflux disease, gastric heterotopia, and colonic
polyps (hyperplastic in men and adenomas in women).

TABLE 2.2: Fundic Gland Polyps: Sporadic Versus FAP Associated

Features Sporadic FAP-Associated

Number Single; few when multiple (40%) Multiple (90%); carpet of

hundreds to thousands

Male to female ratio F>M F=M

Age 5th–6th decade; no children 3rd decade; can occur in children

Mutations β-catenin APC gene mutation; CTNNB1/β-

catenin

Frequency of LGD Low (<1%) High (nearly half)

Progression to carcinoma None reported Case reports only

Surveillance None, even when dysplasia Upper endoscopy every 1–3

detected years for detection of duodenal
and periampullary lesions; no
specific surveillance for FGPs,
even when dysplasia detected

F, female; FAP, familial adenomatous polyposis; FGP, fundic gland polyp; LGD, low-grade dysplasia; M,
male.
 Figure 2-57. Gastric hyperplastic polyp. Lamina propria edema, mild foveolar hyperplasia, and
scattered chronic inflammatory cells can be seen in both hyperplastic and inflammatory polyps.
This small polyp arose in a backdrop of portal hypertensive gastropathy (not pictured) and
contains ectatic vessels in the lamina propria (arrowheads). Findings in the background flat
mucosa aid in classifying gastric polyps.

Histologically, these polyps consist of dilated cystic oxyntic glands with distorted
glandular architecture admixed with normal-appearing glands (Figs. 2.59 and 2.60).
Parietal cells balloon into the lumen with snoutlike protuberances, sometimes resulting
in exfoliated anucleate blebs with eosinophilic granules that clog the gland outlets (Figs.
2.61 and 2.62). Some cystic spaces may be lined by columnar epithelium as a result of
adjacent gastric pit dilation (Fig. 2.63), but most are lined by chief cells and parietal
cells. The findings, for practical purposes, are identical to that of PPI effect, and
distinction requires correlation with an endoscopic lesion (Fig. 2.64). Likewise, there
are no differentiating histologic features between syndromic and sporadic FGPs, but the
clinicopathologic features are distinct. For example, nearly all patients with FAP have
FGPs (12.5% to 88% of patients with FAP, depending on age at time of endoscopy) and
these polyps are more numerous than sporadic FGPs, often with hundreds or thousands
in a carpetlike distribution.37 FAP-associated FGPs also occur earlier than in the
sporadic setting (third decade vs. fifth to sixth decade) and can be found among the
pediatric population, which is exceptionally rare for sporadic FGPs. A key distinction
of FAP-associated FGPs is their association with low-grade epithelial dysplasia, which
is reported in up to half of cases (Figs. 2.65–2.69). Despite the high prevalence of
dysplasia in these lesions, the risk of malignant transformation is exceptionally low,
with only case reports in the literature.38,39 By comparison, dysplasia develops in less
than 1% of sporadic FGPs and has never been associated with progression to
carcinoma.34,40 Given the rarity of both HGD and carcinoma in either sporadic or FAP
settings, surveillance is unnecessary for LGD in an FGP and surgical resection is never
advised. Patients with FAP do require continued surveillance for other risk factors,
however, such as duodenal polyposis and periampullary adenomas and
adenocarcinomas.41 Upper endoscopy and biopsy of selected polyps is recommended
every 1 to 3 years, with closer screening warranted when a gastric adenoma is detected.

Figure 2-58. Fundic gland polyps (FGP). Multiple small sessile lesions (arrowheads) are seen
in the gastric body. Sporadic FGPs may be multiple but usually few in number. When
innumerable FGPs carpet the gastric corpus, consider a syndromic cause, such as familial
adenomatous polyposis syndrome, MutYH-associated polyposis, or gastric adenocarcinoma
and proximal polyposis syndrome.
Figure 2-59. Fundic gland polyps, scanning magnification. These polypoid tissue fragments
show dilated cystic glands with distorted architecture. Multiple polyps are frequently found,
although endoscopists typically only sample one or two.
Figure 2-60. Fundic gland polyp. These small round polyps are expanded by cystically dilated
oxyntic glands, and there is no significant increase in stroma. The surface epithelium is foveolar
and unremarkable.

Figure 2-61. Fundic gland polyp. The cystic spaces are dilated oxyntic glands and are lined by
pink parietal cells and blue chief cells. The parietal cells balloon into the lumen with snoutlike
protuberances, and the lumen contains exfoliated cells.
Figure 2-62. Fundic gland polyp. At high magnification, the parietal cells have brightly
eosinophilic granular cytoplasm that protrudes toward the lumen, similar to apocrine metaplasia
in the breast. The apical snouts produce anucleate granular blebs within the cystic space.
Figure 2-63. Fundic gland polyp. Some cystic spaces are partially lined by foveolar mucous
cells (arrowhead ), but this variation does not indicate a mixed polyp. The background dilated
oxyntic glands and apical snouting of parietal cells are characteristic for fundic gland polyp.
Figure 2-64. Proton pump inhibitor (PPI) effect. PPIs are associated with diffuse oxyntic gland
dilation and apical snouting of parietal cells, similar to that seen in fundic gland polyps.
Distinction between the two requires endoscopic correlation with a polypoid lesion or with
medication history.
Figure 2-65. Low-grade dysplasia in a fundic gland polyp. Low-grade dysplasia is best identified
as an area of eye-catching hyperchromasia at low magnification. An abrupt transition
(arrowhead ) between normal foveolar epithelium and an area of hyperchromasia is a helpful
clue.
Figure 2-66. Low-grade dysplasia in a fundic gland polyp. Dysplasia in FGPs occurs in half of
patients with FAP syndrome. Dysplasia can also be found in sporadic FGPs but is seen in
<1%. An abrupt transition (arrow) from nondysplastic to dysplastic epithelium can usually be
identified. The dysplastic foveolar epithelial cells show pseudostratification and nuclear
crowding, imparting an area of distinct hyperchromasia.
Figure 2-67. Low-grade dysplasia in a fundic gland polyp. Dysplasia affects the surface foveolar
cells overlying the dilated oxyntic glands and extends down the foveolar-lined gastric pits. An
abrupt surface transition (arrowhead ) from nondysplastic cells is present.
Figure 2-68. Low-grade dysplasia in a fundic gland polyp. Dysplasia affects the foveolar
epithelium overlying the fundic glands and shows cytologic features of pseudostratification,
nuclear crowding, increased mitotic activity, and apoptotic activity (arrowhead ).
 Figure 2-69. Low- and high-grade dysplasia in a fundic gland polyp. A distinct change in
architecture and cytology is seen in the right side of the figure and can be arguably interpreted
as high-grade dysplasia, in contrast to the characteristic low-grade dysplasia seen on the left.
Dysplasia in fundic gland polyps, whether low or high grade, has negligible risk for malignant
transformation.

An exception in FAP that is discussed later in this chapter is the syndrome termed
gastric adenocarcinoma and proximal polyposis syndrome (GAPPS). GAPPS is a rare
subvariant of FAP characterized by many FGPs and gastric adenocarcinoma.

FAQ: What is the surveillance protocol for LGD in an FGP? Are there
differences in follow-up for LGD in sporadic versus FAP-associated FGPs?

Clinicians may be advised that no surveillance is necessary for LGD arising

in FGPs, whether FAP associated or sporadic. This is a frequent finding,
occurring in nearly half of patients with FAP, but progression to HGD or
carcinoma is so rare as to be case report material. Carcinoma arising in a
sporadic FGP has never been reported. Thus, specific follow-up for LGD
arising in FGPs is unnecessary. However, other than the special situation of
GAPPS noted earlier and later in the chapter, patients with FAP should continue
their scheduled interval surveillance for duodenal polyposis and periampullary
lesions, as these lesions carry risk for malignant progression.

PEARLS & PITFALLS

Diffuse fundic gland hyperplasia is commonly ascribed to PPI usage, but do not
overlook the possibility of Zollinger-Ellison (ZE) syndrome when massive oxyntic
gland hyperplasia is seen. This tumor-mediated syndrome is the result of a
gastrinoma of the small bowel or pancreas causing elevated serum gastrin,
which stimulates oxyntic gland hyperplasia and excessive gastric acid secretion,
resulting in subsequent ulcers of the stomach or small bowel. This same gastrin
stimulation can also give rise to gastric NETs (type 2), as a result of ECL-cell
hyperplasia, discussed later in this chapter. Although most ZE syndrome cases
are sporadic, 30% are the result of a defect in the MEN1 tumor suppressor
gene associated with the autosomal dominant syndrome multiple endocrine
neoplasia type 1 (MEN1). As the name implies, it is important to identify these
patients because they are at risk for neoplasia in multiple endocrine organs.
Neoplasms of the pituitary, parathyroid, and pancreas (and small bowel) have
conferred the moniker the “3P syndrome.” Clinical testing for serum gastrin
levels ≥10 times normal can suggest ZE syndrome, but more than half of
patients with ZE have nondiagnostic serum gastrin results. In these cases,
consider other studies such as gastric pH, imaging studies of the small bowel
and pancreas, somatostatin receptor scintigraphy, and MEN1 genotyping.

FAQ: What causes innumerable carpetlike FGPs?

When FGPs are innumerable, consider syndromic etiologies such as FAP,

attenuated FAP, GAPPS, MutYH-associated polyposis, and sporadic
FGPosis.42,43 Differentiating these requires correlation with family history,
physical examination, endoscopic findings, and genetic testing. Mutations of the
APC gene are at the root of FAP and attenuated FAP. Point mutations of this
same APC gene have also been identified in the more recently described
GAPPS, which is now regarded as an FAP variant.43 Mutations in MutYH
characterize MutYH-associated polyposis, and mutations of CTNNB1 (the gene
that encodes for β-catenin) are seen in sporadic fundic gland polyposis (as well
as nonsyndromic sporadic FGPs).42,44
KEY FEATURES: FGPs
• FGPs are the most common type of gastric polyp.
• Endoscopic finding of a discreet lesion separates FGPs from oxyntic gland
hyperplasia such as seen in PPI use and ZE syndrome.
• Histologically, they are composed of dilated cystic oxyntic glands with hypertrophic
parietal cells, some of which have apocrine-like snouts.
• Sporadic FGPs are typically single, whereas innumerable polyps may be seen in the
syndromic setting.
• Syndromic associations include FAP syndrome, attenuated FAP, GAPPS, MutYH-
associated polyposis, and sporadic fundic gland polyposis syndrome.
• LGD is common in FAP-associated FGPs (∼50%) and uncommon in sporadic FGPs
(<1%), but all have negligible risk for malignant transformation and do not require
surveillance.

ADENOMATOUS POLYPS
By convention, polypoid dysplasia in the stomach is designated “gastric adenoma,” and
the degree of dysplasia is graded similar to other parts of the GI tract: low grade, high
grade, or indefinite for dysplasia (Figs. 2.70–2.77). Similar to colon tubular adenomas,
a diagnosis of gastric adenoma implies at least LGD. Gastric adenomas come in several
varieties, including the intestinal type, foveolar type, pyloric gland adenoma (PGA), and
oxyntic gland polyp/adenoma (see Table 2.3). Our understanding of this area continues
to evolve, and much of the previous literature is difficult to interpret owing to several
factors: (1) Divergent histologic criteria between Japanese and Western practices;
Japanese pathologists, who contributed a significant bulk of the 20th century literature
on this subject, did not require invasion to diagnose “carcinoma,” whereas invasion is
an explicit criterion of Western practice.45 (2) Revision of nomenclature as a result of
this discrepancy; definitions of gastric epithelial neoplasia were harmonized in the
Vienna classification published in 2000, resulting in a system congruent with Western
understanding.46 (3) Subsequent ongoing revisions of polyp classification and
nomenclature; oxyntic gland polyp/adenoma was previously classified as “gastric
adenocarcinoma with chief cell differentiation.”47 (4) Lack of recognition or
underreporting of certain gastric adenomas; for example, in one study, PGA was the
third most common gastric polyp, yet only rare case reports were published on this
entity previously.48 (5) Histologic blends of “hybrid” polyps; these defy specific
classification and continue to muddle the literature. Ongoing study in this area will most
certainly yield new updates before a revised edition of this text. In the meantime, the
following section depicts our current understanding for each of the gastric adenomas.
Some bear risk for malignant progression, but each is remarkably different in prognosis
and clinicopathologic associations, discussed later. As with gastric hyperplastic polyps,
sampling the background flat mucosa is important to identify other risk factors (e.g., H.
pylori, IM) and associated conditions such as AMAG.

Figure 2-70. Reactive changes, negative for dysplasia. Areas of intestinal metaplasia
(arrowhead ) may be hyperchromatic at low power. However, this polyp shows a gradual
gradient toward the lighter mature surface. Surface maturation is a reassuring feature for
nondysplastic reactive changes.
Figure 2-71. Gastric adenoma, intestinal type (low-grade dysplasia implied). The designation as
“gastric adenoma” implies at least low-grade dysplasia. The hyperchromasia is striking at low
magnification and extends to the mucosal surface. Note the abrupt transition from
nondysplastic adjacent mucosa.
Figure 2-72. Gastric adenoma, intestinal type. LGD is dramatically hyperchromatic at low
magnification. The crowded cells extend to the mucosal surface, and there is an abrupt
transition from nondysplastic epithelium. Reactive changes can mimic LGD, and this
juxtaposition is a nice contrast, showing slight hyperchromasia but no nuclear crowding. Also
note the gradual color gradient in the reactive area versus the abrupt change in LGD.
Figure 2-73. Gastric adenoma, intestinal type. Involvement of the surface epithelium is a helpful
clue to dysplasia, but beware of tangential embedding. In this example, the area of LGD is
hyperchromatic and eye-catching but does not involve the overlying surface epithelium. Search
out areas of background intestinal metaplasia (arrowhead ) for comparison. The marked
contrast in hyperchromasia between the two areas supports LGD.
Figure 2-74. Low-grade dysplasia. Gastric adenoma, intestinal type (goblet cells not seen in this
field). Dysplasia in this example is conventional, similar to that seen in a colonic tubular
adenoma. The architecture is simple with well-formed glands, and the nuclei are elongated,
pseudostratified, and crowded. Mitotic activity (arrow) and apoptoses (arrowhead ) are
increased.
Figure 2-75. Gastric adenoma, intestinal type (goblet cells not seen in this field). The cells at the
top of this photo are starting to show more rounded nuclei (loss of nuclear polarity) instead of
elongated nuclei. The combination of architectural complexity, loss of nuclear polarity, and
increased cytologic atypia are criteria for high-grade dysplasia.
Figure 2-76. High-grade dysplasia and low-grade dysplasia in gastric adenoma, intestinal type.
A diagnosis of gastric adenoma implies low-grade dysplasia. When high-grade dysplasia is
seen, it is listed first. The cells in high-grade dysplasia show disorganization with loss of
nuclear polarity and marked nuclear atypia characterized by variability in both size and shape
as compared with neighboring cells. In contrast, the cells of low-grade dysplasia retain nuclear
polarity and appear organized and relatively uniform, despite the hyperchromasia and nuclear
crowding.
 Figure 2-77. High-grade dysplasia in gastric adenoma, intestinal type (goblet cells not seen in
this field). These cells have an increased nuclear to cytoplasmic ratio, appear disorganized
owing to loss of nuclear polarity, and show marked nuclear variability in size and shape.

PEARLS & PITFALLS: A Diagnosis of “Gastric Adenoma” Implies Low-Grade

Dysplasia, Similar to Colon Tubular Adenomas
By convention, a designation of “gastric adenoma” implies the presence of at
least LGD, similar to colon tubular adenomas. Unlike HGD in colon tubular
adenomas, which requires both cytologic and architectural complexity, these
features are not required for HGD in gastric adenomas. In this way, dysplasia
grading of gastric adenomas is more similar to dysplasia grading in Barrett
mucosa, whereby loss of nuclear polarity, glandular crowding, and
pleomorphism are sufficient for HGD; architectural complexity is not required
for HGD in gastric adenomas. Furthermore, because “adenoma” and “low-
grade dysplasia” are synonymous, it is unnecessary to use the redundant
nomenclature of “low-grade dysplasia arising in a gastric adenoma.” Rather, a
simplified diagnosis of “gastric adenoma” is sufficient and conveys the presence
of LGD. When HGD is identified, however, it is worthwhile to list the HGD first:
“high-grade dysplasia in a gastric adenoma, foveolar type.”
See also “Esophagus” chapter and “Colon” chapter.
Gastric Adenoma, Intestinal Type
Intestinal-type gastric adenomas comprise the bulk of all gastric adenomas, the others
being relatively rare. They are single, well circumscribed, sessile, or pedunculated and
are typically <2 cm. Although they can arise in any area of the stomach, they have a
predilection for the gastric antrum/pylorus (61%) and are found more commonly in men
than women (3:1), occurring in the sixth and seventh decades.49 Histologically, the
epithelial lining of this adenoma contains goblet cells and/or Paneth cells and by
definition is dysplastic. The dysplasia is conventional (similar to that seen in a tubular
adenoma) and is characterized by the presence of hyperchromatic mucin-depleted
elongated cells with crowding and nuclear pseudostratification extending to involve the
surface (Figs. 2.70–2.75). LGD has simple architecture with small glands, whereas
HGD shows architectural complexity (cribriforming, budding, branching, or crowding)
in addition to other cytological changes such as ovoid nuclei with loss of nuclear
polarity, increased nuclear to cytoplasmic ratio (N:C), prominent nucleoli, clumped
chromatin, and irregular nuclear contours (Figs. 2.76 and 2.77). Approximately 40% of
intestinal-type gastric adenomas harbor HGD and nearly one-fourth progress to
adenocarcinoma.49 This is in contrast to the gastric foveolar-type adenoma, discussed
next, which rarely harbors HGD or carcinoma. Nearly all intestinal-type adenomas
arise in association with background IM, gastric atrophy, and gastritis, again
emphasizing the importance of separately sampling the background gastric mucosa
(Figs. 2.78–2.88).

FAQ: We do not see gastric adenomas often. What should I advise my

clinician?

Clinicians should be aware that gastric adenomas with intestinal subtype

confer a high rate of synchronous HGD (40%) and invasive adenocarcinoma
(24%). Accordingly, complete endoscopic excision should be assured, and the
lesion should be processed entirely to examine for areas of invasive carcinoma.
Furthermore, additional biopsies of the background flat mucosa should be
submitted separately, as these lesions are associated with separate foci of
adenocarcinoma (16%), flat dysplasia (6%), and IM (97%). These background
gastric biopsies also serve to identify other highly associated conditions, such
as H. pylori (42%) and background gastritis (71%; environmental metaplastic
atrophic gastritis [EMAG] 52% and AMAG 19%).49
TABLE 2.3: Gastric Adenomatous Polyps

Pyloric
Intestinal Foveolar Gland Oxyntic Gland
Features Type Type Adenoma Polyp/Adenoma

Morphology low power Fig. 2.77 Fig. 2.93 Fig. 2.99 Fig. 2.119

Morphology high power Fig. 2.83 Fig. 2.96 Fig. 2.110 Fig. 2.122

Size <2 cm <1 cm 1–2 cm 0.2–0.8 cm

Location Antrum > Body/fundus Body Fundus and

body/fundus ≫ antrum cardia

Mean age 60s 40s 70s 60s

Male to female ratio M ≫ F (3:1) F=M F ≫ M (3:1) F=M

HGD ++ (44%) - ++ (51%) -

Adenocarcinoma + (24%) - + (12%– -

30%)

Background IM Yes (virtually No Yes (60%) No

100%)

Associated conditions H. pylori FAP (70%) AMAG (40%) GERD (100%)

(42%)
EMAG
(52%)
AMAG
(19%)
Flat
dysplasia
(6%)

IHC staining MUC5AC - + + -

characteristics (Foveolar)
MUC6
(Pyloric)
MUC 2 &
 CDX2
(Intestinal)

- - + +

+ - - -
(Background
IM will stain)

AMAG, autoimmune metaplastic atrophic gastritis; EMAG, environmental metaplastic atrophic gastritis; F,
female; FAP, familial adenomatous polyposis; GERD, gastroesophageal reflux disease; HGD, high-grade
dysplasia; IHC, immunohistochemistry; IM, intestinal metaplasia; LGD, low-grade dysplasia; M, male.

Figure 2-78. Gastric adenoma, intestinal type. The most common gastric adenoma subtype is
the intestinal type. It is fairly easy to identify owing to the marked hyperchromasia at low
magnification and presence of goblet cells.
Figure 2-79. Gastric adenoma, intestinal type. There is an abrupt transition (arrow) from
nondysplastic to dysplastic epithelium.
Figure 2-80. Gastric adenoma, intestinal type. The marked hyperchromasia and presence of
goblet cells make intestinal-type adenomas easy to classify as compared with other subtypes.
Figure 2-81. Gastric adenoma, intestinal type. Goblet cells may be sparse in some intestinal
type adenomas, but the adjacent mucosa may provide clues, such as foci of intestinal
metaplasia (arrow).
Figure 2-82. Gastric adenoma, intestinal type. The cytologic features of intestinal-type gastric
adenomas are similar to the conventional colonic tubular adenoma, with elongated and
crowded pseudostratified nuclei. Goblet cells (arrow) are usually present.
Figure 2-83. Gastric adenoma, intestinal type. These lesions are readily diagnosed at low
magnification, with strong resemblance to the colonic tubular adenoma. The presence of goblet
cells (arrow) are also a helpful clue to classification.
Figure 2-84. Gastric adenoma, intestinal type. By definition, gastric adenomas are at least low-
grade dysplasia. Although crowded and pseudostratified, the nuclei remain polarized and
organized. There may be increased mitoses and apoptoses (arrow). The presence of goblet
cells (arrowheads) classifies this gastric adenoma as intestinal type.
Figure 2-85. Gastric adenoma, intestinal type. Cytoplasmic features are key to differentiating
the gastric adenoma subtypes. For intestinal-type gastric adenomas, the presence of goblet
cells (arrowheads) is the easiest clue to spot. However, also note the smooth pale eosinophilic
cytoplasm (arrows) in the intervening cells. The luminal surface has a stiff eosinophilic border,
and the cytoplasm lacks clearing, foaminess, or a mucin cap.
Figure 2-86. Gastric adenoma, intestinal type. These cells share a stiff eosinophilic border at
the luminal surface, and a microvillous brush border can sometimes be seen (arrowheads).
The cytoplasm is smooth, pale, and eosinophilic. The absence of clearing, foaminess, and
mucin distinguish this from other gastric adenoma subtypes, especially when goblet cells are
absent.
Figure 2-87. Gastric adenoma, intestinal type. When goblet cells are abundant, classification of
the intestinal-type gastric adenoma is straightforward. Take a moment to note the cytoplasmic
qualities of the intervening cells, which appear pale and smooth. The stiff luminal border
(arrowheads) is distinctly eosinophilic.
 Figure 2-88. Gastric adenoma, intestinal type. A microvillous brush border (arrow) is present in
this intestinal-type gastric adenoma. In the absence of goblet cells, the smooth pink cytoplasm
and lack of a mucin cap or cytoplasmic clearing are helpful features.

Gastric Adenoma, Foveolar Type

Adenomas composed entirely of dysplastic foveolar-type epithelium are rare. They are
typically solitary, arise more commonly in the body/fundus, and show equal gender
distribution with a mean age of 44 years.49,50 These lesions usually contain no more than
LGD and are lined by gastric epithelial mucin cells with a neutral mucin cap (Figs.
2.88–2.99). Immunohistochemistry for MUC5AC (gastric mucin marker) can confirm
gastric differentiation, whereas negative markers include MUC6 (pyloric mucin
marker), MUC2, and CDX2 (intestinal markers). In contrast to intestinal type, foveolar-
type adenomas tend to occur in otherwise normal, nonatrophic gastric mucosa and rarely
harbor HGD or carcinoma.49,50 Furthermore, background IM is the exception (found in
<3% of cases) and no flat epithelial dysplasia is found in the nonpolypoid areas.49 This
stark dissimilarity between the intestinal and foveolar adenomas have led investigators
to postulate divergent genetic pathways for each, but although most harbor some
detectable genetic alteration, no statistically significant differences in any particular
genetic alteration (APC, CTNNB1, KRAS mutations, and MSI) were found between the
intestinal and foveolar types.51
FAQ: Are there any known associations for foveolar-type gastric adenomas?

A genetic background of FAP syndrome can be found in 68% of foveolar-

type gastric adenomas. These syndrome-associated gastric foveolar
adenomas are found more commonly in the gastric antrum and herald severe
duodenal adenomatosis.52

Figure 2-89. Gastric adenoma, foveolar type (low-grade dysplasia implied). By definition, gastric
adenomas are at least low-grade dysplastic. At low magnification, this gastric polyp is distinctly
hyperchromatic and easy to spot as a dysplastic lesion.
Figure 2-90. Gastric adenoma, foveolar type. This gastric polyp is hyperchromatic a low
magnification due to the presence of low grade dysplasia.
Figure 2-91. Gastric adenoma, foveolar type. The cells of the foveolar type gastric adenoma
recapitulate normal foveolar cells, which have a clear mucin cap (arrowheads). This
cytoplasmic feature distinguishes foveolar adenomas from other gastric adenomas.
Figure 2-92. Gastric adenoma, foveolar type. The nuclei of this foveolar-type adenoma are
slightly crowded and disorganized, but they remain basally located, without the degree of
pseudostratification seen in intestinal-type gastric adenomas. The key to classification,
however, is the presence of a pale pink to clear apical mucin cap along the luminal surface
(arrowheads). Mitotic activity is present (arrow).
Figure 2-93. Gastric adenoma, foveolar type. The nuclei of this foveolar-type adenoma are
slightly elongated and crowded but basally located and lack the degree of elongation and
pseudostratification seen in intestinal-type gastric adenomas. The presence of pale pink to
clear cytoplasmic mucin at the apical border (arrowheads) is key to classification. Mitotic
activity is present (arrow).
Figure 2-94. Gastric adenoma, foveolar type. At low magnification, this polyp is distinctly
hyperchromatic, indicating that it contains at least low-grade dysplasia and should be
categorized as a gastric adenoma. The clear mucinous cap along the surface epithelial cells,
evident even at low magnification, indicates foveolar subtype.
Figure 2-95. Gastric adenoma, foveolar type, higher magnification of the previous figure. The
cells are columnar with crowded and basally located nuclei. The presence of an apical mucin
cap indicates foveolar differentiation.
Figure 2-96. Gastric adenoma, foveolar type. These columnar cells have crowded nuclei and
some pseudostratification but not to the degree seen in intestinal-type gastric adenomas. The
absence of goblet cells and the presence of a pale pink to clear apical mucin cap identify this
gastric adenoma as foveolar subtype.
Figure 2-97. Gastric adenoma, foveolar type. Compare these epithelial cells to those found in
intestinal-type gastric adenomas. These nuclei are crowded and pseudostratified, but the key
differentiating feature is the cytoplasmic quality. Foveolar-type gastric adenomas have a pale
pink to clear apical mucin cap (arrows and arrowheads) not found in intestinal-type gastric
adenomas.
Figure 2-98. Gastric adenoma, foveolar type. This low-grade dysplastic gland is lined by cells
with crowded pseudostratified nuclei and apical mucin caps (arrows). Adjacent nondysplastic
foveolar epithelium (arrowhead ) is in the field for comparison.
 Figure 2-99. Gastric adenoma, foveolar type. The apical mucin cap on these dysplastic cells
identifies this polyp as a foveolar-type gastric adenoma. Unlike intestinal-type and pyloric-type
gastric adenomas, these lesions rarely harbor high-grade dysplasia or carcinoma, and the
background mucosa is usually normal and nonatrophic.

Pyloric Gland Adenoma

PGAs are neoplastic lesions with malignant potential and can be found in a variety of
GI sites, with the stomach being the most common, followed by the duodenal bulb, bile
duct, gallbladder, duodenum, and main pancreatic duct.48,53 In the stomach, they are 1 to
2 cm, have a marked female predominance (75%), and arise in older individuals (mean
age, 73 years).48,50 Histologically, they are composed of closely packed pyloric gland–
like tubules with a single layer of cuboidal to low columnar epithelial cells containing
round nuclei and pale to eosinophilic cytoplasm with ground glass appearance. These
lesions can be histologically heterogenous (Figs. 2.100–2.117), and this variability, in
combination with the overall infrequency of these lesions (2.7% of all gastric polyps),
causes challenges in recognition.48,50 Application of immunohistochemical stains is
helpful in problematic cases: pyloric gland adenomas (PGAs) demonstrate
coexpression of MUC5AC (foveolar mucin marker) and MUC6 (pyloric mucin marker)
(Figs. 2.118 and 2.119), and are nonreactive for MUC2 or CDX2 (both intestinal mucin
markers).50 By comparison, gastric foveolar-type adenomas are reactive for MUC5AC
only, and gastric intestinal-type adenomas show reactivity for MUC2 and/or CDX2.
Most centers, however, do not have these helpful MUC stains readily available, and so
a good number of pathologists will rely upon routine hematoxylin and eosin (H&E)
stains for this diagnosis. Helpful histologic clues include the lack of an apical mucin
cap, distinguishing it from gastric foveolar-type adenoma, and the absence of
intralesional goblet cells, distinguishing it from gastric intestinal-type adenoma. Based
on morphologic features, some observers suggest that PGAs and oxyntic gland
polyp/adenomas are closely related lesions.54,55
Although PGAs lack intralesional goblet cells, 60% of PGAs show associated
background IM and 40% are associated with AMAG,50 a key feature of which is IM.
Thus, adequate mucosal sampling is important to identify the gastric environment in
which these lesions arise. These lesions are composed of a bland monolayer of pyloric-
type cells (Fig. 2.110), well described in early papers, and conventional LGD (similar
to that seen in tubular adenomas) is found in slightly more than half of PGAs (63%)
(Fig. 2.111).50 Note that all PGAs, whether they contain conventional dysplasia or not,
are considered at least low-grade dysplastic, a view that has developed over time since
the original descriptions.56 This notion is further supported by the frequent association
with HGD (51%) (Figs. 2.113–2.115) and adjacent adenocarcinoma (12% to 30%)
(Figs. 2.116 and 2.117), a feature that also underscores the importance of adequate
sampling50; complete excision is the treatment of choice. PGAs have been reported to
occur in FAP syndrome (6% of patients), but these lesions show similar genetic
background as sporadic PGAs (i.e., KRAS and GNAS mutations).54,55,57

PEARLS & PITFALLS

IM is seen in association with 60% of PGAs. This is important to note for
several reasons.
First, the presence of background IM can mislead one to a diagnosis of
gastric intestinal-type adenoma; avoid this misstep by carefully scrutinizing the
lesional tissue, which, in a PGA, will contain predominantly cuboidal and low
columnar cells with pale to eosinophilic ground glass cytoplasm and lack
intralesional goblet cells.
Second, 40% of PGAs arise in the setting of AMAG,48,50 a key feature of
which is IM. Thus, the finding of IM in association with a PGA should trigger a
workup for AMAG (i.e., gastrin and chromogranin immunohistochemical stains)
when sufficient background gastric mucosa is available. A diagnosis of AMAG
may be rendered when a constellation of features is found limited to the gastric
body and fundus (the gastrin immunohistochemistry confirms this location by
the absence of staining for G-cells):
 • Intestinal metaplasia
• “Antralization” (atrophy of oxyntic glands and pyloric metaplasia)
• Linear or nodular ECL-cell hyperplasia, confirmed by chromogranin
• Inactive chronic gastritis, characterized by a low-lying lymphocytic infiltrate
By comparison, the gastric antrum (confirmed by gastrin
immunohistochemistry highlighting abundant G-cells), will be near normal (at
most, reactive/chemical gastropathy) (see “Gastric Well-Differentiated
Neuroendocrine Tumors, Type 1” section for more details).
Finally, for those fortunate enough to have access to MUC2, MUC5AC, and
MUC6 immunohistochemistry, areas of transition from PGA to background IM
can demonstrate MUC2 reactivity. Take care to interpret the stain results in
areas of lesional tissue. PGA should be reactive for MUC5AC and MUC6 and
nonreactive for MUC2 (Figs. 2.118 and 2.119).

FAQ: What is the recommended surveillance for PGA?

No specific surveillance recommendations exist for gastric PGA, but

clinicians should be advised that PGAs are neoplastic lesions with malignant
potential (12% to 30% are associated with adjacent adenocarcinoma). As
such, complete excision is recommended, along with sampling of the
background flat mucosa to evaluate for AMAG, which is found in 40% of cases.
Figure 2-100. Pyloric gland adenoma (low-grade dysplasia is implied). The morphologic
heterogeneity of PGAs contributes to lack of recognition, but several key unifying features can
aid in diagnosis. For example, at low magnification, this polyp is predominantly composed of
closely packed pyloric gland–like tubules. As with colon tubular adenomas, the diagnosis of
PGA implies at least low-grade dysplasia.
Figure 2-101. Pyloric gland adenoma. Some areas may show back-to-back well-formed
tubules, whereas other areas may have dilated and distorted glands, but the cells lining these
areas are the same. Despite the bland appearance, these lesions are all considered at least
low-grade dysplasia.
Figure 2-102. Pyloric gland adenoma. PGAs are a composed of a monolayer of low columnar
or cuboidal cells with abundant clear to foamy cytoplasm and basally located nuclei. These
polyps are considered low grade dysplasia despite the lack of conventional dysplasia (such as
that seen in colon tubular adenomas).
Figure 2-103. Pyloric gland adenoma. At low magnification, this PGA shows closely packed
pyloric gland–like tubules, some of which are cystically dilated and distorted. The lesion lacks
stroma-rich or edematous areas, and the presence of back-to-back glands indicates an
epithelial proliferative process. PGA, by definition, has at least low-grade dysplasia.
Figure 2-104. Pyloric gland adenoma. These back-to-back glands are composed of uniform low
columnar to cuboidal cells with abundant ground glass or foamy cytoplasm. The nuclei are
small and basally located. Original descriptions did not consider areas containing a monolayer
of bland cells dysplastic, but this view has evolved over time, and all PGAs are now considered
at least low-grade dysplasia regardless of whether conventional dysplasia is present.
Figure 2-105. Pyloric gland adenoma. The proliferative tubules are pyloric gland–like, with
uniform basally located small nuclei and abundant clear foamy cytoplasm.
Figure 2-106. Pyloric gland adenoma. These cells are low columnar with round basally located
nuclei. The cytoplasmic quality helps subclassify gastric adenomas, and PGAs have
eosinophilic to clear cytoplasm with ground glass or foamy appearance.
Figure 2-107. Pyloric gland adenoma. At low magnification, this polyp contains more cystically
dilated glands than previous examples. There is absence of stroma-rich areas or lamina
propria edema, and focal areas show closely packed smaller tubules.
Figure 2-108. Pyloric gland adenoma. PGAs can be morphologically heterogeneous, and this
example shows a PGA lined by both columnar cells with eosinophilic cytoplasm (arrow) and
cuboidal cells with clear foamy cytoplasm (arrowhead).
Figure 2-109. Pyloric gland adenoma. Another area in the previous polyp shows more of the
columnar cells with eosinophilic ground glass cytoplasm. This morphologic heterogeneity
contributes to difficulty in diagnostic recognition. These polyps are considered low-grade
dysplasia despite the lack of conventional dysplasia (such as seen in tubular adenomas).
Figure 2-110. Pyloric gland adenoma. Yet a different area of the same polyp shows the more
characteristic cuboidal cells with round basally located nuclei and abundant foamy clear
cytoplasm. One can appreciate the similarity between these areas and the normal pyloric
glands found in the gastric antrum and pylorus. Original descriptions did not consider areas
containing a monolayer of bland cells dysplastic, but this view has evolved over time, and all
PGAs are now considered at least low-grade dysplasia regardless of whether conventional
dysplasia is present.
Figure 2-111. Pyloric gland adenoma with conventional low-grade dysplasia. This example
shows more cytologic atypia, especially along the surface epithelium where the cells are
overlapping and show pseudostratification. There is some disorganization in glandular
architecture, but the nuclei maintain polarity and there is abundant cytoplasm. Although all
PGAs are considered low-grade dysplasia, this example shows conventional dysplasia like that
seen in a typical colonic tubular adenoma.
Figure 2-112. Pyloric gland adenoma. This is another example of the variability in cytoplasmic
quality found in PGAs. Rather than clear and foamy, this example shows eosinophilic ground
glass cytoplasm.
Figure 2-113. High-grade dysplasia in a pyloric gland adenoma. The nuclei in this example are
highly atypical with variation in size and shape. The nuclei are no longer situated in an orderly
fashion along the basement membrane but instead are haphazardly arranged.
Figure 2-114. High-grade dysplasia in a pyloric gland adenoma. High-grade dysplasia is found in
half of PGAs. These glands are cribriforming, and the cells show marked nuclear atypia. There
is loss of nuclear polarity, variation in size and shape, and prominent nucleoli.
Figure 2-115. High-grade dysplasia in a pyloric gland adenoma. The glandular architecture is
complex with crowding and cribriforming. The nuclei show variation in size and shape, have
irregular nuclear borders, and contain prominent nucleoli.
Figure 2-116. Adenocarcinoma in a pyloric gland adenoma. Associated or adjacent
adenocarcinoma is reported in up to one-third of PGAs. Look carefully in busy areas with
complex architecture. Single infiltrating cells (arrows) are diagnostic for adenocarcinoma
invasive into the lamina propria (pT1).
Figure 2-117. Adenocarcinoma in a pyloric gland adenoma. This adenocarcinoma is poorly
differentiated and arising from a PGA.
Figure 2-118. Pyloric gland adenoma, MUC5AC expression. MUC5AC is a marker for foveolar
differentiation and is expressed in PGAs along with MUC6, which indicates pyloric
differentiation. These immunostains may be helpful in challenging cases, but most PGAs can
be distinguished by H&E.
 Figure 2-119. Pyloric gland adenoma, MUC6 expression. PGAs coexpress foveolar (MUC5AC)
and pyloric (MUC6) immunohistochemistry. They are negative for intestinal markers MUC2 and
CDX2. These immunostains may be applied in challenging cases, but they are not readily
available in most laboratories, making H&E recognition essential.

Oxyntic Gland Polyp/Adenoma

The oxyntic gland polyp/adenoma is one example of the challenges in the nomenclature
of gastric neoplasia. This uncommon gastric lesion was previously classified as both an
adenocarcinoma (the so-called gastric adenocarcinoma of fundic gland type/chief cell
predominant type or gastric adenocarcinoma with chief cell differentiation) and a
benign “chief cell hyperplasia” or FGP variant by several authors.58-60 However, the
bland cytologic features of the lesion and low Ki-67 proliferation rate (<2%) in
combination with the benign clinical course prompted a proposal for reclassification as
“oxyntic gland polyp/adenoma.”47 These rare lesions are encountered as small (0.2 to
0.8 cm) single polypoid growths in the gastric fundus (70%) or cardia (30%), with
equal distribution among men and women, and are found uniformly in the setting of
gastroesophageal reflux disease.47 Histologically, they are characterized by a deep
proliferation of oxyntic glands arranged in cords and clusters (Figs. 2.120–2.125). Thin
wisps of smooth muscle separate the glands (Figs. 2.123 and 2.124), but there is no
desmoplastic stromal response to suggest invasion. The predominant cells are
monotonous pale gray-blue (basophilic or amphophilic) columnar cells resembling
chief cells; interspersed are smaller numbers of eosinophilic parietal cells and clear
mucous cells. The cells lack overt cytologic atypia, mitotic activity, or necrosis.
Assessment for dysplasia in these lesions and whether they are best classified as polyps
or adenomas remains unclear until further studies clarify the pathogenesis. However, in
their series, Singhi and colleagues point out that no reported cases have shown true
recurrence or progression of disease. Similar to PGAs, these lesions are associated
with GNAS mutations61 and are reactive for MUC6 (a pyloric marker) by
immunohistochemistry (Fig. 2.126). They can be differentiated from PGAs by their
negative MUC5AC (a foveolar marker) immunoreactivity (Fig. 2.127).

FAQ: What is the appropriate follow-up for this lesion?

The nomenclature ranges from benign (oxyntic gland polyp, chief cell
hyperplasia, chief cell hamartoma) to benign neoplasia (oxyntic gland
polyp/adenoma) to malignant (gastric adenocarcinoma of fundic gland/chief cell
predominant type and gastric adenocarcinoma with chief cell differentiation),
but clinicians may be assured that despite the evolving nomenclature, the data
are quite clear that these lesions are biologically benign. These lesions have
been reported to persist when incompletely resected, but no reported cases
have shown true recurrence or progression of disease. Thus, complete
endoscopic excision of these lesions is sensible, whereas surgical intervention
is overkill.
Figure 2-120. Oxyntic gland polyp/adenoma. These are typically small subcentimeter lesions
found in the gastric fundus with a proliferation of cords and clusters of oxyntic glands. The
surface foveolar epithelium is unchanged.
Figure 2-121. Oxyntic gland polyp/adenoma. This example shows an abrupt change from
normal oxyntic gland architecture at the left to distorted cords of oxyntic cells at the right. The
surface foveolar epithelium is normal.
Figure 2-122. Oxyntic gland polyp/adenoma. There is a deep proliferation of oxyntic glands
arranged in cords and clusters, whereas the overlying foveolar epithelium is unchanged. The
far right of this biopsy shows normal oxyntic mucosa for comparison.
Figure 2-123. Oxyntic gland polyp/adenoma. The cells in this example are bland and exquisitely
uniform. They are arranged in long tubules or cords separated by smooth muscle strands in the
lamina propria (arrowhead ).
Figure 2-124. Oxyntic gland polyp/adenoma. Another example demonstrates some
heterogeneity in the cells of this lesion. Although the cells are oxyntic in origin, this may not be
evident at high-power views. The arrangement of the cells in cords separated by smooth
muscle strands (arrowheads) and the deep location within oxyntic mucosa are clues to the
diagnosis.
Figure 2-125. Oxyntic gland polyp/adenoma. Note the deep proliferation of glands (bracket) that
has embedded smooth muscle between the cords of cells. This example arises in a backdrop
of autoimmune metaplastic atrophic gastritis.
Figure 2-126. Oxyntic gland polyp/adenoma, MUC6 immunohistochemistry. These lesions
demonstrate MUC6 (pyloric differentiation) immunoreactivity and are negative for MUC5AC
(foveolar).

Figure 2-127. Oxyntic gland polyp/adenoma, MUC5AC immunohistochemistry. There is no

expression of MUC5AC foveolar differentiation in the tumor cells. Note the reactivity of the
normal overlying foveolar epithelial cells, which serve as internal control.
KEY FEATURES: Gastric Adenomas (See Table 2.3)
• Polypoid dysplasia is designated as gastric adenoma and is classified as low-grade or
high-grade dysplasia.
• Varieties include intestinal type, foveolar type, PGA, and oxyntic gland
polyp/adenoma.
• Intestinal-type gastric adenoma:
• Contains goblet cells with or without Paneth cells
• Frequently associated with HGD (44%) and adenocarcinoma (24%)
• High frequency of background gastritis (e.g. H. pylori, AMAG, EMAG)
• Immunohistochemistry (IHC) profile: MUC2+, CDX2+, MUC5AC−, MUC6−
• Foveolar-type gastric adenoma:
• Lined by foveolar epithelial cells, identified by its mucin cap
• Associated with FAP (70%)
• Low risk of adenocarcinoma (<1%)
• IHC profile: MUC5AC+, MUC6−, MUC2−
• Pyloric gland adenoma:
• Closely packed tubules lined by cuboidal pyloric gland cells with ground glass
cytoplasm
• Frequently associated with HGD (51%) and adenocarcinoma (30%)
• Strong association with AMAG (40%) and background IM (60%)
• IHC profile: MUC5AC+, MUC6+, MUC2−, CDX2−
• Oxyntic gland polyp/adenoma:
• Deep cords and clusters of oxyntic glands without cytologic atypia or desmoplasia
• Biologically benign
• Associated with gastroesophageal reflux disease (100%)
• IHC profile: MUC6+, MUC5AC− MUC2−, CDX2−
• Regardless of subtype, polyps should be managed by complete excision; additional
biopsies should be taken of mucosa away from the polyp to evaluate for gastritis,
IM, flat dysplasia, and other synchronous lesions.

ADENOCARCINOMA
Globally, gastric cancer is among the top five leading causes of cancer and a leading
cause of cancer death. The highest incidence worldwide is in Asia, central Europe, and
South America.62,63 In the United States, the overall incidence rates are modest but
remain greater than those of esophageal cancer, and when compared with Caucasians,
the incidence is increased in all non-Caucasian ethnic and racial groups including
Hispanics, Asian, and African-Americans.64 Although gastric cancer remains a deadly
disease, the incidence and mortality rates have fallen dramatically over the past 80
years.62 Advances in endoscopic techniques have improved detection of early gastric
cancer (defined as pT1), and adjuvant chemotherapy contributes to improved overall
mortality and 5-year survival rates.65,66 Stage remains the most important prognostic
indicator, but a number of considerations affect biopsy interpretation, such as tumor
location and stage, Lauren classification, morphologic variants, background mucosa
characteristics, genetic considerations, and biomarker profile, as discussed later.

APPROACH TO THE BIOPSY

Tumor Location and Staging: Esophageal Versus Gastric
Organ subsite remains an important prognostic indicator, and treatment indications
differ if the carcinoma involves the gastric cardia or gastroesophageal junction (GEJ).
For example, for proximal tumors involving the GEJ or gastric cardia, radiation therapy
may be added to adjuvant chemotherapy before resection. These proximal tumors are on
the rise in developed countries and are associated with risk factors for
gastroesophageal reflux.66 By comparison, distal gastric cancers, which are associated
with diet and H. pylori infection, have been decreasing worldwide as a result of
improved diet and health care. For staging purposes, the esophagus and stomach are
differentiated by the American Joint Committee on Cancer staging manual (eighth
edition) based on involvement of and proximity to the GEJ.67 Carcinomas entirely on
either side of the GEJ are easily classified as either esophageal or gastric cancers,
respectively. For tumors involving the GEJ, these are considered gastric when the tumor
midpoint is greater than 2 cm into the stomach, whereas those with tumor midpoints 2
cm or less into the stomach are staged as esophageal cancers (Fig. 2.128).
Staging of gastric cancers is based on depth of invasion and is independent of tumor
size. Any invasion into the lamina propria, muscularis mucosae, or submucosa is
considered pT1. Recall, these intramucosal carcinomas have access to the rich
lymphatics within the gastric lamina propria, allowing for lymphatic spread and lymph
node metastases. The T1 category is further divided into pT1a (invasion into the lamina
propria or muscularis mucosae) and pT1b (invasion into the submucosa). pT2 tumors
invade the muscularis propria, pT3 penetrate the subserosal connective tissue, and
tumors that involve the visceral peritoneum are considered pT4. This last category is
further subdivided as pT4a (involvement of visceral peritoneum only) and pT4b
(involvement of adjacent organs or structures) (Fig. 2.129). Early gastric cancers are
tumors with invasion limited to the lamina propria, muscularis mucosae, or submucosa
(pT1) and may be amenable to conservative endoscopic mucosal resection (EMR) if
they meet specific criteria. Early studies were limited to small lesions with specific
endoscopic features (e.g., exophytic vs depressed), whereas recent studies cite more
liberal criteria that include early submucosal invasion. Examples of tumors amenable to
EMR include (1) intramucosal nonulcerated differentiated tumors >2 cm, (2)
intramucosal ulcerated differentiated tumors ≤3 cm, (3) intramucosal nonulcerated
undifferentiated tumors ≤2 cm, and (4) submucosal invasion <500 μm (sm1)
differentiated tumor ≤3 cm.68 Thus, tumor differentiation, as well as depth of invasion,
is an important feature to include in the pathology report.

FAQ: What anatomic landmark distinguishes esophageal from gastric tumors?

Tumors entirely proximal or entirely distal to the GEJ are simply classified
as esophageal or gastric, respectively. For tumors that involve the GEJ, the
distance of the tumor’s midpoint is taken into account. A somewhat arbitrary
distance of 2 cm is the cutoff: the tumor is considered esophageal if its
midpoint is ≤2 cm into the stomach, and gastric if >2 cm (see Fig. 2.128).

PEARLS & PITFALLS: GEJ Is Not Interchangeable With Squamocolumnar

Junction
Determining whether a tumor is gastric versus esophageal relies upon its
positional relationship with the GEJ and not the squamocolumnar junction
(SCJ); these two are not synonymous (see Fig. 2.128). The GEJ is defined
anatomically, whereas the SCJ is defined histologically, and these two do not
always coincide. Take, for example, in Barrett esophagus, where the SCJ has
moved caudally into the tubular esophagus as the result of columnar
metaplasia. This distinction is important because the American Joint Committee
on Cancer definition of gastric cancer utilizes the GEJ and not the SCJ. These
landmarks can be identified endoscopically or at the grossing bench by the
following distinctive features:
GEJ: Proximal end of gastric folds, or the notch where the pouched stomach
meets the tubular esophagus.
SCJ: The line where squamous mucosa (which has a pale glossy appearance)
meets the columnar mucosa (which appears reddish and has a coarse texture).
PEARLS & PITFALLS: Staging Challenges in pT3 and pT4 Tumors
It is possible for a tumor to invade through the muscularis propria (pT3) of the
stomach and directly extend into the gastrocolic or gastrohepatic ligament, or
the greater or lesser omentum, without perforating the visceral peritoneum. In
these cases, the lesion should be classified as pT3. Note these structures and
organs are covered in visceral peritoneum, and pT4 designation is reserved for
tumors with invasion into this visceral peritoneum.
Figure 2-128. Criteria for gastric versus esophageal classification. The gastroesophageal
junction (GEJ), not the squamocolumnar junction, serves as a reference point for tumor
classification as gastric or esophageal. Any tumor entirely distal to the GEJ (not involving the
GEJ) is classified as a gastric tumor (A) regardless of midpoint location. For tumors involving
the GEJ, do not classify the tumor based on the location of the bulk of the tumor. Instead, the
tumor midpoint (stars) is the important reference point. Those with a tumor midpoint ≤2 cm into
the stomach (B) are esophageal, and those with tumor midpoint >2 cm into the stomach (C)
are gastric.
 Figure 2-129. Pathologic T staging of gastric cancer, AJCC eighth edition. T1 is divided into
lamina propria invasion (pT1a) and invasion into submucosa (pT1b). Involvement of muscularis
propria is pT2, whereas invasion beyond the muscularis propria into the subserosal tissue is
pT3. Penetration of the serosa (visceral peritoneum) is pT4a, and when adjacent structures are
involved the stage is designated pT4b. (Used with permission of the American Joint Committee
on Cancer [AJCC], Chicago, IL. The original source for this material is the AJCC Cancer
Staging Atlas. [2006]. Greene FL, Compton CC, Fritz AG, et al. [eds.]. Springer Science and
Business Media, LLC. www.springerlink.com.)

TUMOR CLASSIFICATION: LAUREN, MING, AND WHO

MORPHOLOGIC VARIANTS
In addition to providing basic histologic typing, the Lauren tumor type offers significant
prognostic information and is simple to apply.69 The Lauren tumor type is divided into
two chief categories: intestinal-type and diffuse-type gastric adenocarcinomas; tumors
that do not fit into these two categories are considered indeterminate unclassified types.
Mass-forming and expansile gastric adenocarcinomas that arise in a backdrop of IM are
classified as intestinal type (Figs. 2.130 and 2.131). By comparison, adenocarcinomas
that have an infiltrative growth pattern and arise without the backdrop of IM are
classified as diffuse type. The classic example for diffuse type is the signet ring cell
carcinoma with endoscopic characteristics of poor insufflation (linitis plastica) and the
absence of an endoscopically visible mass lesion (Figs. 2.132–2.134), features that
contribute to the worse prognosis as compared with the intestinal type. The relative
frequencies are 50% to 67% for intestinal type, 29% to 35% for diffuse type, and 3% to
21% for indeterminate type.70 A similar two-tiered Ming grading system is used by
some pathologists, in which the categories “expanding type” and “infiltrative type”
correlate with Lauren’s intestinal and diffuse types, respectively.
The Lauren classification is used most widely by pathologists but does not specify
the morphologic variants. The WHO 2010 classification of gastric carcinomas includes
seven major tumor varieties, including adenocarcinoma, adenosquamous carcinoma,
carcinoma with lymphoid stroma (medullary carcinoma), hepatoid adenocarcinoma,
squamous cell carcinoma, undifferentiated carcinoma, and neuroendocrine carcinoma
(NEC).71 The WHO adenocarcinoma group is further divided into additional
morphologic subtypes, including papillary, tubular, mucinous, poorly cohesive
carcinomas, and mixed carcinoma; the differentiating histologic features are summarized
in Table 2.4. Although most multivariate analyses show no effect of these tumor types
(aside from the poorly performing small cell carcinoma) on prognosis independent of
grade or stage,72 this morphologic classification is useful to include in a biopsy report
to assist pathologist colleagues who will review surgical margin frozen sections or
work up metastatic deposits. Of the adenocarcinoma variants, two WHO morphologic
subtypes correlate with the Lauren diffuse type: (1) mucinous adenocarcinoma, defined
as having at least 50% extracellular mucin pools (with or without signet ring cells), and
(2) poorly cohesive adenocarcinoma, which includes the signet ring cell variant as well
as other variants. In general, these WHO, Lauren, and Ming groups correlate closely.73
Figure 2-130. Gastric adenocarcinoma, intestinal type, gross specimen. The gross pathology of
gastric cancer correlates closely with histologic classification in the Lauren and Ming systems.
This large, mass-forming exophytic growth represents an intestinal type (Lauren) or expanding
type (Ming) even without histologic review.
Figure 2-131. Gastric adenocarcinoma, representative histology of the previous figure.
Histologic features show the mass lesion is composed of malignant glands with an expansile
growth pattern.
Figure 2-132. Stomach linitis plastica, endoscopic view. The stomach shows loss of rugal folds
and has a stiff “leather bottle” quality that results in poor insufflation and distensibility. No mass
lesion is visible, but this endoscopic feature is synonymous with gastric adenocarcinoma,
diffuse type (Lauren) or infiltrating type (Ming).
Figure 2-133. Stomach linitis plastica, gross specimen cross section. The gastric wall is
thickened and stiff without an obvious mass lesion. Histologic sections reveal diffuse infiltration
of the gastric wall by sheets of discohesive malignant cells.
 Figure 2-134. Gastric adenocarcinoma, representative histology of the previous figure.
Histologic features show architecturally intact superficial mucosa without an expansile mass
lesion. Instead, single discohesive malignant cells (brackets) infiltrate the mucosa without
much disruption of the glands. This pattern is classified as diffuse type by Lauren or infiltrating
type by Ming, both simple bimodal classification systems. By WHO classification, this tumor is
a poorly cohesive carcinoma variant.

Signet ring cell carcinoma is a specific subset of poorly cohesive carcinomas that is
composed predominantly of discohesive single cells with distended cytoplasmic mucin
that displaces and eccentrically compresses the nucleus. This designation should be
restricted to tumors composed predominantly of the signet ring cell type, as other
variants of poorly cohesive carcinomas may also infiltrate in a diffuse or single cell
pattern, but are mucin poor (Figs. 2.135–2.163). In either case, these are the carcinomas
that provide the most challenge on endoscopic biopsies because they lack the helpful
red flag of an endoscopic mass, and stromal desmoplasia is frequently absent. Examine
gastric biopsies with careful attention to areas of pallor, crush artifact, or slightly
increased cellularity. Make sure to account for every cell in the tissue and call out any
possible mimickers before moving on (e.g., crushed oxyntic glands, mucous neck cells,
xanthoma cells). If there is any uncertainty, a pancytokeratin immunostain can aid in
highlighting the architecture of epithelial cells (Figs. 2.142 and 2.152), and consultation
with a fellow surgical pathologist may clarify matters. Squamous cell carcinomas may
be either keratinizing or nonkeratinizing, and adenosquamous carcinomas should have at
least 25% squamous component by volume mixed with a glandular component. Biopsy
material may not be wholly representative, and final morphologic classification can be
either reserved for or revised following review of the resection specimen. The
carcinoma with lymphoid stroma (also referred to as medullary carcinoma or
lymphoepithelioma-like carcinoma) may resemble well differentiated tubular
adenocarcinomas or undifferentiated carcinomas with poorly formed sheets of tumor
cells, but in all instances it contains a prominent stromal lymphoid infiltrate, sometimes
nearly obscuring the carcinoma itself (Figs. 2.164–2.171). This tumor has reactivity for
Epstein-Barr virus–encoded RNA by in situ hybridization (EBER ISH) and is more
commonly seen in the proximal stomach, among men and in a younger age group. Some
observers have reported significantly better prognosis via longer disease-free survival
and overall cancer survival attributed to the younger age at presentation and less lymph
node metastasis.74,75 Another interesting variant, hepatoid adenocarcinoma,
morphologically resembles hepatocellular carcinoma with large polygonal cells
containing abundant eosinophilic granular cytoplasm (Figs. 2.172–2.175). These
hepatoid gastric adenocarcinomas can raise concern for metastatic hepatocellular
carcinoma and may even express immunoreactivity for alpha-fetoprotein (AFP) and
glypican-3, which are also reactive in hepatocellular carcinomas.76 SALL4
immunohistochemistry may help differentiate these tumors, as it is expressed in 89% of
AFP-positive hepatoid gastric adenocarcinomas and negative in most hepatocellular
carcinoma.76 However, we would caution that a subset of primary liver hepatocellular
carcinomas are SALL4 reactive; these latter tumors tend to be aggressive but, thankfully,
targetable.77 Tumors demonstrating features that are not classifiable into one of the other
WHO categories are considered undifferentiated carcinomas.

TABLE 2.4: WHO 2010 Classification of Gastric Carcinomas

Tumor Type Histologic Features

Adenocarcinoma

Papillary Exophytic tumor with finger-like or frond-like projections containing

fibrovascular cores
Usually low grade

Tubular Branching and anastomosing tubules with dilated or slit-like lumen

Usually low grade

Mucinous >50% extracellular mucin pools

May contain signet ring cells
Lauren classification = diffuse type
Poorly cohesive carcinomas Infiltrating single cells or small aggregates
(signet ring cell and other Signet ring cell type is a specific subset composed predominantly of
variants) signet ring cells containing a clear droplet of cytoplasmic mucin
displacing the nucleus
Lauren classification = diffuse type

Mixed Mixture of morphologically identifiable components (e.g., tubular,

papillary, poorly cohesive)

Adenosquamous ≥25% squamous component mixed with glandular

carcinoma

Carcinoma with lymphoid Poorly differentiated carcinoma with prominent lymphoid infiltrate
stroma Associated with Epstein-Barr virus
Better prognosis

Hepatoid adenocarcinoma Resembles hepatocellular carcinoma with large polygonal cells

containing abundant eosinophilic cytoplasm
May express alpha-fetoprotein

Squamous cell carcinoma Either keratinizing or nonkeratinizing

Undifferentiated High-grade carcinoma not classifiable among other categories

carcinoma

Neuroendocrine carcinoma (NEC)

Large cell NEC Poorly differentiated Cells are large and pleomorphic
High grade Moderate amount of cytoplasm
Marked nuclear atypia Prominent nucleoli
Synaptophysin+ or
Small cell NEC chromogranin+ Cells are small
May have: Finely granular chromatin (“salt
Focal necrosis and pepper”)
Ki-67 >20% Indistinct nucleoli
>20 mitoses per 10 HPF
Mixed adenoneuroendocrine >30% each of gland-forming and
carcinoma (MANEC) neuroendocrine areas
(Adenocarcinomas showing
immunoreactivity for
neuroendocrine markers is not
sufficient for diagnosis.)
Figure 2-135. Benign gastric mucous neck cells. Do not overcall carcinoma when benign
gastric mucous neck cells are seen. These mucous filled cells found in the neck of gastric
pits/glands are inconspicuous in normal tissues but can appear concerning when tissue is
crushed, fragmented, or reactive. Do not make a diagnosis of signet ring carcinoma unless
there is 100% certainty; suggest rebiopsy if the diagnosis is unclear.
Figure 2-136. Gastric adenocarcinoma, diffuse type. These discohesive cells are infiltrating
beneath an intact and benign surface epithelium.
Figure 2-137. Gastric adenocarcinoma, diffuse type. Single discohesive and non-gland-forming
cells infiltrate the wall of the stomach. The cytoplasm is filled with mucin and pushes the
nucleus off to one side, similar in profile to a signet ring.
Figure 2-138. Gastric adenocarcinoma, diffuse type, with signet ring cells. True signet ring cells
(arrow) have cytoplasm distended with a clear vacuole of mucin that displaces and
compresses the nucleus.
Figure 2-139. Metastatic lobular breast carcinoma. These tumor cells appear similar to signet
ring cells, but lobular breast carcinoma cells do not contain mucin. Instead, there is a single
round cytoplasmic vacuole with a sharply demarcated border, which sometimes contains a
dense hyaline eosinophilic body (arrow) imparting a targetoid appearance.
Figure 2-140. Gastric adenocarcinoma, diffuse type, with signet ring cells. Signet ring cells
have distended clear cytoplasm filled with mucin that displaces the nucleus to the periphery.
This example shows marked variation in size and shape among the signet ring cells.
Figure 2-141. Gastric adenocarcinoma, diffuse type. At low magnification, one can appreciate
how the malignant cells diffusely infiltrate the gastric mucosa without forming an obvious mass
lesion. The surface epithelium is intact, and the gastric pits remain relatively evenly spaced.
The absence of significant architectural disturbance makes the diffuse type of gastric
adenocarcinoma more difficult to spot than the intestinal mass-forming type.
Figure 2-142. Gastric adenocarcinoma, diffuse type, cytokeratin 7 immunohistochemistry of the
previous figure. A cytokeratin immunostain can be helpful to assess the architecture in difficult
cases. In this example, CK7 highlights a lack of normal gland formation and single infiltrating
cells at the base of the biopsy.
Figure 2-143. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure.
The cells in this example are markedly atypical with large, irregularly shaped nuclei that vary in
size.
Figure 2-144. Gastric adenocarcinoma, diffuse type. Take care to look closely at “busy” gastric
biopsies. At low magnification, the inflammatory cells and reactive gastropathy changes
obscure a focus of gastric adenocarcinoma (arrow). One good rule of thumb is to always
review gastric biopsies at high magnification and mentally account for all cells present.
Figure 2-145. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure.
Amid a background of acute and chronic inflammation, this cluster of signet ring cells (arrow)
could be easily overlooked. Do not forget to always look beyond the first (obvious) diagnosis.
Figure 2-146. Gastric adenocarcinoma, diffuse type. This low-power view shows a “busy”
gastric biopsy with surface erosion and abundant inflammation. Always look closer at areas
that appear abnormal at low magnification.
Figure 2-147. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure.
Even at midpower magnification, the adenocarcinoma hiding in this busy tissue fragment is not
clear. The malignant cells blend into the background of inflammation which is distracting one’s
eye.
Figure 2-148. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure.
These sneaky signet ring cell clusters (arrows) are only visible at high magnification,
underscoring the importance of reviewing each gastric biopsy at high power and mentally
registering each and every cell.

Figure 2-149. Gastric adenocarcinoma, diffuse type. This low-power view shows gastric
mucosa with an area of edema and lamina propria hemorrhage. It looks harmless at this power,
but any area of abnormality should always trigger a higher-power perusal.
Figure 2-150. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure.
Impossible to see at low magnification, these small malignant cells infiltrate between the benign
glands, leaving the architecture relatively intact.
Figure 2-151. Gastric adenocarcinoma, diffuse type. Areas of gastric erosion and chronic
inflammation should always trigger a closer examination. Inflammation can be particularly
devious in hiding gastric cancer.
Figure 2-152. Gastric adenocarcinoma, diffuse type, pancytokeratin immunostain of the
previous figure. A pancytokeratin stain highlights a broad area containing single infiltrating cells
(bracket). The extent of invasive carcinoma is frequently surprising when highlighted in this
manner and demonstrates how challenging it is to discern these cells on H&E. Contrast the
malignant area against the normal glandular architecture to the left.
Figure 2-153. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure.
The malignant cells are obscured by the inflammatory backdrop. Some remnant benign glands
add to the difficulty in diagnosis. Even at this magnification, it can be challenging to point out the
malignant cells. Always take a moment to go to high magnification and mentally account for
each cell.
Figure 2-154. Gastric adenocarcinoma, diffuse type, higher magnification of the previous figure
. High magnification reveals many malignant cells admixed with acute and chronic inflammatory
cells. At this magnification, it is possible to name each cell present as benign or malignant.
Figure 2-155. Gastric adenocarcinoma, diffuse type. These discohesive tumor cells are
characteristic of diffuse-type adenocarcinoma by Lauren classification, infiltrative type by Ming
classification, and poorly cohesive carcinoma by WHO classification. The cells lack clear
cytoplasmic mucin vacuoles and thus are not true signet ring cells.
Figure 2-156. Gastric adenocarcinoma, diffuse type. Another example of the infiltrative
discohesive tumor cells found in diffuse-type adenocarcinoma. These cells permeate the
gastric wall without forming a mass lesion and frequently have no background gastritis or other
gastric pathology.
Figure 2-157. Gastric adenocarcinoma, diffuse type, eosinophilic variant. The cells in this
unusual example of diffuse-type adenocarcinoma are deeply eosinophilic and lack
intracytoplasmic mucin. This is classified as a poorly cohesive carcinoma by the WHO
classification, which includes signet ring cell and other variants. The tumor contains <50%
glands and is therefore poorly differentiated.
Figure 2-158. Gastric adenocarcinoma, diffuse type, eosinophilic variant. The cells are
markedly atypical with eosinophilic cytoplasm and a myxoid backdrop. These are not signet ring
cells because they lack the characteristic clear vacuole of intracytoplasmic mucin.
Figure 2-159. Gastric adenocarcinoma, diffuse type, eosinophilic variant. This unusual variant
contains cells that have abundant eosinophilic cytoplasm without mucin. Some areas show
cohesive malignant cells arranged in cords.
Figure 2-160. Gastric adenocarcinoma, diffuse type, eosinophilic variant. The tumor cells
appear plasmacytoid with irregular nuclei containing conspicuous nucleoli. These should not be
mistaken for signet ring cells, as they lack the characteristic compressed nucleus and
cytoplasmic mucin vacuole.
Figure 2-161. Gastric adenocarcinoma, WHO mucinous type. These tumors contain >50%
extracellular mucin pools and are considered a diffuse type by Lauren classification and are
poorly differentiated by definition.

Figure 2-162. Gastric adenocarcinoma, WHO mucinous type, Lauren diffuse type. There is
abundant intracellular and extracellular mucin, the latter of which comprises >50% of the tumor.
Figure 2-163. Gastric adenocarcinoma, WHO mucinous type, Lauren diffuse type. Mucinous
adenocarcinoma may or may not contain signet ring cells (arrows).
Figure 2-164. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma). At low magnification, this mass-forming lesion shows a prominent lymphoid
infiltrate that obscures the glands. At first glance, one might consider a lymphoid neoplasm
such as MALT lymphoma.
Figure 2-165. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma). Higher magnification of the previous figure shows abnormal distorted and
abortive glands with prominent intraepithelial lymphocytes. These features are reminiscent of
lymphoepithelial lesions found in MALT lymphoma, but do not be fooled, and take a closer look
at higher power.
Figure 2-166. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma). The intact glands show abundant intraepithelial lymphocytes, but the
abnormal glands are composed of malignant cells, indicating this is in fact a carcinoma and not
a lymphoma.
Figure 2-167. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma), pancytokeratin immunostain. Immunohistochemistry highlights the poorly
formed glands, which are more abundant than visualized on H&E and infiltrative in architecture.
Figure 2-168. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma), in situ hybridization for Epstein-Barr virus encoded RNA (EBER ISH). A
subset of adenocarcinomas with lymphoid stroma will show positivity for EBER ISH localized to
the tumor nuclei, as in this example.
Figure 2-169. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma). Another example of a poorly differentiated gastric carcinoma with abundant
lymphoid cells. The lymphocytes in close association with these sheets of malignant cells are a
clue to the diagnosis of EBV-associated gastric cancer and a relatively better prognosis.
Figure 2-170. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma). Higher magnification of the previous figure shows syncytial sheets of tumor
cells with indistinct cell borders and numerous intratumoral lymphocytes. Previous names for
this tumor include medullary carcinoma or lymphoepitheliallike carcinoma. The tumor is
considered poorly differentiated because it has <50% gland formation.
Figure 2-171. Gastric adenocarcinoma with lymphoid stroma (EBV-associated gastric
adenocarcinoma). In situ hybridization for Epstein-Barr encoded RNA shows the tumor nuclei
contain EBV RNA. This finding is associated with a better prognosis.
Figure 2-172. Gastric adenocarcinoma, WHO hepatoid variant. This tumor resembles
hepatocellular carcinoma with large polygonal cells containing abundant granular cytoplasm.
Pseudoacini may be present, and this example even shows bile formation.
Figure 2-173. Gastric adenocarcinoma, WHO hepatoid variant. This gastric cancer shows
sheets of malignant cells with abundant eosinophilic granular cytoplasm and prominent nucleoli
arranged in sheets or wide trabeculae with fine endothelial spaces simulating hepatic
sinusoids. The cytology and architecture are hepatoid. Exclusion of metastatic hepatocellular
carcinoma to the stomach may require correlation with radiologic findings.
Figure 2-174. Gastric adenocarcinoma, WHO hepatoid variant. The tumor is composed of large
cells with abundant eosinophilic cytoplasm and round nuclei with prominent nucleoli arranged in
vague trabeculae. The delicate vasculature reinforces the similarity to widened hepatic plates
and sinusoids.
 Figure 2-175. Gastric adenocarcinoma, WHO hepatoid variant, hepatocyte-specific antigen
(HSA). Hepatoid variant of primary gastric carcinoma may be reactive for HSA (pictured), AFP,
and glypican-3, all of which are also reactive in hepatocellular carcinoma. SALL-4 can help
differentiate these tumors, as it is positive in most gastric hepatoid carcinomas and negative in
hepatocellular carcinoma.

NECs are classified as such by their poorly differentiated high-grade cytology and
marked nuclear atypia, although this was not always the case. The now outdated WHO
2010 classification of neuroendocrine neoplasms defined three grades based purely on
mitotic rate and Ki-67 index, with the high grade (G3) category defined as >20 mitoses
per 10 HPF or Ki-67 proliferation index >20% without regard to histologic atypia.71
This older WHO 2010 classification regarded G3 neoplasms as synonymous with
poorly differentiated NECs. However, when investigators divided the G3 group into
morphologically well- and poorly differentiated tumors, the poorly differentiated tumors
had different etiologies, genetic alterations, and response to treatment with worse
survival outcomes.78-82 For these reasons, the updated classification of endocrine
tumors (WHO 2017) now relies first on morphologic features; a tumor should be
identified as well differentiated or poorly differentiated before further stratification by
mitotic count and Ki-67 proliferative index. In the case of poorly differentiated
morphology, these should always be classified as NECs and staged as carcinomas.83
This group can be further divided into (1) small cell type, in which the cells are small
with finely granular “salt and pepper” chromatin and inconspicuous nucleoli (Figs.
2.176–2.179); (2) large cell type, in which the cells are large with a moderate amount
of cytoplasm and prominent nucleoli (Figs. 2.180–2.181); and (3) mixed
neuroendocrine-nonneuroendocrine neoplasm (MiNEN; previously “mixed
adenoneuroendocrine carcinoma” or MANEC), in which the tumor is composed of at
least 30% both NET and adenocarcinoma or other high-grade carcinoma.
Immunoreactivity for neuroendocrine markers within an adenocarcinoma does not
indicate MiNEN.

FAQ: What is the preferred classification system: Lauren, Ming, or WHO

subtypes?

The Lauren system is the most commonly used classification system,

dividing tumors into either “intestinal type” (mass-forming lesions with a
backdrop of IM, such as seen in Figs. 2.130 and 2.131) or “diffuse type” (non–
mass forming infiltrative tumors without an identifiable precursor lesion, such as
seen in Figs. 2.132–2.134). The Ming classification is similar with “expanding”
and “infiltrating” types that correlate with the Lauren’s intestinal and diffuse
types, respectively. All three (Lauren, Ming, and WHO) correlate closely, as
described in the earlier text, and morphologic classification is helpful for frozen
section interpretation and metastatic workups. Aside from the poorly
performing small cell carcinoma, multivariate analyses show no effect of tumor
type on prognosis, independent of grade or stage.
Figure 2-176. Gastric neuroendocrine carcinoma, small cell type. At low magnification, this
gastric biopsy is markedly abnormal with a deeply basophilic infiltrate and considerations
include lymphoma and poorly differentiated carcinoma.
Figure 2-177. Gastric neuroendocrine carcinoma, small cell type. Higher magnification of the
previous figure shows infiltrative highly atypical cells with prominent crush artifact.
Figure 2-178. Gastric neuroendocrine carcinoma, small cell type, CAM 5.2
immunohistochemistry. Immunohistochemical stains show the tumor cells are positive for
epithelial differentiation, thus ruling out lymphoma. Chromogranin (not pictured) was also
positive, supporting neuroendocrine differentiation. Based on the high-grade morphology, this
tumor qualifies as a poorly differentiated neuroendocrine carcinoma, G3, and is classified as
small cell type because of the typical small cell features of high nuclear to cytoplasmic ratio,
scant cytoplasm, absence of nucleoli, prominent nuclear molding, and crush artifact.
Figure 2-179. Gastric neuroendocrine carcinoma, small cell type, Ki-67 immunohistochemistry.
The proliferation index of the previous figures shows near 100% Ki-67 labeling in the tumor
cells. This further supports a diagnosis of NEC.
Figure 2-180. Gastric neuroendocrine carcinoma, large cell type. The high-grade cytologic
features of this neuroendocrine neoplasm (confirmed by chromogranin and synaptophysin, not
pictured) classify it as a poorly differentiated neuroendocrine carcinoma (G3) and not a well-
differentiated neuroendocrine tumor. These cells are large with variation in size and shape,
angulated nuclei, prominent nucleoli, and visible cytoplasm, differentiating it from small cell
type. Frequent mitoses (arrowhead ) and areas of necrosis (arrow) are present.
 Figure 2-181. Gastric neuroendocrine carcinoma, large cell type. Another example shows large
pleomorphic cells with variable amounts of cytoplasm and prominent nucleoli. A chromogranin
immunostain (not pictured) confirms neuroendocrine differentiation, and the cytologic features
classify this as a poorly differentiated neuroendocrine carcinoma, G3, large cell type. Note the
frequent mitotic activity and areas of punctate necrosis.

FAQ: What does it mean when a signet ring cell carcinoma is not staining for
mucin?

Not all tumors with single infiltrating cells are signet ring cell carcinomas.
True signet ring cells have cytoplasm distended with a clear vacuole of mucin
that displaces and compresses the nucleus (Fig. 2.138). Other poorly cohesive
or diffuse type variants without intracytoplasmic mucin may have eosinophilic
cytoplasm instead (Figs. 2.157–2.160) and should be classified as diffuse type
gastric adenocarcinoma without the “signet ring” designation. Do not forget to
consider metastatic lobular breast carcinoma, which may also mimic signet ring
cells. Lobular breast carcinoma cells contain intracytoplasmic lumina, which can
mimic the mucin vacuole of signet ring cells. However, the single round vacuole
found in lobular breast carcinoma cells have thick sharply demarcated edges,
sometimes contain a dense hyaline eosinophilic body imparting a “targetoid”
appearance (Fig. 2.139), and are not mucicarminophilic. They can be
differentiated from gastric cancer with immunohistochemistry for breast
markers such as GCDFP or GATA3.

FAQ: How does one differentiate NET from NEC?

Based on WHO 2017 classification, NECs are identified by their poorly

differentiated high-grade morphology (Figs. 2.180 and 2.181), independent of
mitotic count or Ki-67 proliferative index. This is a departure from previous
WHO 2010 classification for neuroendocrine neoplasms, which stratified tumors
based purely on mitoses and Ki-67. The updated system not only indicates
prognosis and response to therapy more reliably but is also more intuitive to
pathologists. Tumors that are morphologically well differentiated are classified
as well-differentiated NETs and can be further graded based on mitotic and Ki-
67 cutoffs. See later in this chapter for a discussion on well-differentiated NETs.

PEARLS & PITFALLS: Histologic Grading of Gastric Adenocarcinomas

The histologic grading of adenocarcinomas is a three-tiered system and is
based on the extent of glandular differentiation. Tumors with >95% gland
formation are G1 (well differentiated), 50% to 95% are G2 (moderately
differentiated), and <50% G3 (poorly differentiated or undifferentiated) (Figs.
2.182–2.185). The following tumor subtypes are automatically G3: Lauren
diffuse type, WHO mucinous (Figs. 2.161–2.163), WHO poorly cohesive
(including signet ring and other variants), and WHO undifferentiated.
Figure 2-182. Gastric adenocarcinoma, intestinal type, G1 well differentiated. Gastric
adenocarcinomas are graded histologically by the degree of gland formation. Tumors with
>95% gland formation are G1 (well differentiated).
Figure 2-183. Gastric adenocarcinoma, intestinal type, G2 moderately differentiated. Tumors
with 50%–95% gland formation are graded as G2 (moderately differentiated). Some areas of
this carcinoma show poor gland formation and sheets of tumor cells (arrow).
Figure 2-184. Gastric adenocarcinoma, intestinal type, G2 moderately differentiated. Glands
make up most of this tumor (in the 50%–95% range), whereas some areas show no gland
formation (bracket). This tumor is best graded as G2, moderately differentiated.
 Figure 2-185. Gastric adenocarcinoma, intestinal type, G3 poorly differentiated. Tumors with
<50% gland formation are graded as G3, poorly differentiated. This example shows only rare
glandular lumina.

RISK FACTORS AND GENETIC CONSIDERATIONS

Background Mucosa
One cannot overstate the importance of examining the background gastric mucosa
whenever a gastric lesion is encountered. This background mucosa can reveal
additional information about a patient’s risk factors for gastric adenocarcinoma and it
should be sampled routinely, either by the endoscopist at the time of biopsy or, in the
case of a gastrectomy, by the prosector at the grossing bench. Features to look for
include background IM, Helicobacter infection, AMAG, and precursor polyps or areas
of flat dysplasia.
Analogous to the adenoma-carcinoma sequence in colorectal cancer, gastric
carcinogenesis is a multistep and multifactorial process, with a sequence of precursor
histologic findings that progress from IM to LGD, HGD, and carcinoma. Although IM is
a known risk factor for gastric cancer, the overall risk of gastric cancer in a patient with
IM is extremely low compared with the risk of adenocarcinoma in a patient with Barrett
esophagus.84 Known risk factors for IM can be divided into two categories: (1) EMAG,
which encompasses both H. pylori gastritis (typically an active chronic superficial
plasmacytic inflammation) and chemical/reactive gastropathy (characterized by tortuous
“corkscrew” hyperplasia of the antral foveolae and pits, lamina propria smooth muscle
streaming toward the surface, and attenuation of foveolar mucin) related to high salt
intake, smoking, alcohol consumption, and chronic bile reflux; and (2) AMAG. Recall,
the IM (and other histologic abnormalities) found in EMAG is antral predominant,
which is in direct contrast to the body-predominant IM (and other histologic
abnormalities) found in AMAG. For more detailed discussion on EMAG and AMAG,
see Atlas of Gastrointestinal Pathology: A Pattern-Based Approach to Non-Neoplastic
Biopsies.
In 1994, based on epidemiologic evidence, the International Agency for Research on
Cancer recognized H. pylori as a class 1 carcinogen and primary cause of gastric
adenocarcinoma. Only a small minority of infected individuals develop gastric cancer
(three cases per year for every 10,000 infected persons), which is predominantly
intestinal type, and progression is multifactorial owing to influences such as host
susceptibility, environmental forces, and bacterial strain (e.g., CagA strains are
associated with a higher frequency of precancerous lesions and gastric cancer). Left
untreated, H. pylori infection results in a long latency period (four or more decades) and
progresses from chronic active nonatrophic gastritis to multifocal atrophic gastritis, and
then to IM, dysplasia, and intestinal type invasive carcinoma.85,86 Patients with AMAG
are also at risk for gastric neoplasia, including both NETs and adenocarcinoma. The
incidence of adenocarcinoma in these patients is seven times more frequent than in the
general population, with an overall prevalence of 2%.87 Do not forget to check the
background mucosa for other associated precursor lesions, such as gastric adenomas,
hyperplastic polyps, or syndromic hamartomatous polyps. Flat dysplasia is an
uncommon finding in the stomach usually associated with chronic atrophy (H. pylori or
AMAG) and can be graded in a two-tiered system similar to gastric adenomas (LGD
and HGD).

CHECKLIST: Before Signing Out a Gastric Adenocarcinoma, Examine the

Background Mucosa Away From the Lesion for the Following
□ Intestinal metaplasia
○ Environmental metaplastic atrophic gastritis (antral predominant)
■ H. pylori (active and chronic superficial plasmacytic gastritis)
■ Chemical/reactive gastropathy (e.g., high-salt diet, smoking, alcohol, bile reflux)
○ AMAG (body/fundus predominant)
□ Precursor polyps
 ○ Gastric adenomas
○ Hyperplastic polyps
○ Syndromic hamartomatous polyps
□ Flat dysplasia

FAQ: What is H. pylori CagA and how does it relate to gastric cancer?

Bacteria have developed several mechanisms to secrete proteins or to

inject toxins into target cells. H. pylori injects the oncoprotein cytotoxin-
associated antigen A (CagA) into host cells. Once inside the cell, CagA is
phosphorylated and acts as a scaffold or hub protein that disrupts multiple host
signaling pathways and targets the apical junctional complex of the epithelial
cell. Different domains of the CagA protein interfere with signaling cascades,
which results in cytoskeleton rearrangements and degradation of signal
transduction pathways that maintain normal epithelial differentiation, including
cell adhesion, cell polarity, and the inhibition of cell migration. CagA is also a
highly antigenic protein that elicits interleukin-8 production resulting in a
pronounced inflammatory response. These factors all play into the development
of gastric adenocarcinoma and mucosa-associated lymphoid tissue (MALT)
extranodal marginal zone lymphoma.

Environmental Risk Factors

There are distinctive differences in the geographic and ethnic incidence of gastric
cancer. Emigrants acquire risk similar to that of their destination population, and such
findings strongly suggest that exposure to environmental factors plays an important role
in gastric cancer development. A number of dietary factors also play a significant role,
including salt and salt-preserved foods, such as salted fish, cured meat, and salted-
pickled vegetables. Modern refrigeration has reduced the need for this kind of food
preservation and is cited as a reason for the decreasing incidence of gastric cancers.88,89
However, regional and ethnic consumption patterns persist and continue to correlate
with gastric cancer incidence. Another contributor is N-nitroso compounds, which are
formed endogenously and following ingestion of dietary nitrates (found in some cheeses
and cured meats but largely in natural foods, such as vegetables and potatoes). Diets
high in fried food, processed meat and fish, and alcohol are associated with an
increased risk of gastric carcinoma, whereas fruits, vegetables, and fiber appear
protective. Obesity, smoking, occupational exposures, previous history of stomach
surgery, and blood group A all contribute, but the most important environmental risk
factor is H. pylori. Epstein-Barr virus also contributes to 5% to 10% of gastric cancers,
as discussed earlier.

FAQ: What is the link between nitrates, nitrites, and nitrosamines and gastric
cancer? Which of these ingredients do I need to avoid, exactly?

Do not quit your hot dog habit just yet. High consumption of processed
meats, such as ham, bacon, sausages, and hot dogs, is linked to an increased
gastric cancer risk, and many attribute this to the food additives nitrates and
nitrites that retard microbial spoilage, preserve meat’s recognizable
appearance, and enhance flavor. As a group, these compounds containing an –
NO group are referred to as N-nitroso compounds, which all become the same
end product when consumed. Nitrates (NO3) are inert until they are reduced to
nitrites (NO2) by oral bacteria, are swallowed, and, upon hitting the acidic
gastric juices, are converted to nitrous acid (HNO2), which then binds to
amines, amides, and amino acids to form nitrosamines. Major sources of
human exposure to N-nitroso compounds include diet, occupational exposure,
and smoking, but in vivo formation accounts for up to 75% (that is right, your
body is making most of it!).90 Several studies have investigated the potential
association between dietary consumption of nitrates, nitrites, and nitrosamines
with gastric cancer. Increased consumption of nitrites and N
nitrosodimethylamine are linked to an increased risk for gastric cancer,91
whereas several studies have demonstrated a protective effect of nitrates,91-93
and this was found related to the high levels of nitrates found in green leafy
vegetables. Researchers conclude that higher intake of antioxidants relative to
nitrates offers protective effects. So, although the hype surrounding nitrate-free
products is not entirely a load of bologna, a few bites of broccoli may be kinder
to your GI tract.

Familial Predisposition
Most gastric cancers are sporadic, but about 10% of cases occur with aggregation
within families. Truly hereditary familial gastric cancer accounts for 1% to 3% of the
global burden of gastric cancer and comprises at least two major syndromes: hereditary
diffuse gastric cancer (HDGC) and GAPPS of the stomach (essentially a variant of
FAP).
Hereditary Diffuse Gastric Cancer
HDGC is inherited in an autosomal dominant pattern with high penetrance. Nearly 50%
of HDGC is associated with germline truncating mutations of the CDH1 gene, located
on chromosome 16q22.1, which was first identified in three Maori families from New
Zealand that were predisposed to diffuse gastric cancer. More than 75 families with
nearly 4,000 probands have since been identified, and we now know these mutations
are not concentrated in a single hotspot but are evenly distributed along the CDH1 gene
in several different exons and, to date, more than 155 different germline CDH1
mutations have been identified.94,95
Quick Fact: CDH1 is a tumor suppressor gene that requires a second hit for tumor
formation. The CDH1 gene provides instructions for making the E-cadherin protein,
which is a membrane-bound cell-adhesion molecule that also acts as a tumor suppressor
protein, preventing cells from growing or dividing too rapidly. Simply put, CDH1 and
E-cadherin are important in controlling cellular cohesion (cells sticking together) and
division. The mechanism by which the second allele of E-cadherin is inactivated is
diverse and includes promoter hypermethylation, mutation, and loss of heterozygosity,
any of which results in loss of E-cadherin expression. Functionally, the resulting loss of
cellular cohesion leads to an increased ability for tumor cells to invade and migrate, a
feature seen in diffuse gastric cancers and invasive lobular breast carcinomas.
The lifetime risk of gastric cancer in individuals from these families is 70% for men
and 56% for women, and the average age of onset is 38 years (range 14 to 82
years).94,95 Asymptomatic carriers of the mutation are recommended prophylactic total
gastrectomy generally between ages 20 and 30 years, during which the risk of gastric
cancer rises from <1% to 4%.95 Women in these affected families also have a 42%
cumulative risk of lobular breast carcinoma, with increased risk before age 30.94 Cases
of signet ring cell appendiceal and colorectal cancers have also been reported, but these
do not appear increased as compared with those in non-CDH1-mutated populations.94
Histologically, the tumors in these patients are identical to sporadic diffuse gastric
cancers and are similarly challenging to identify (Figs. 2.186–2.195). In situ lesions or
pagetoid spread of signet ring cells are described commonly in the literature with
CDH1-mutated diffuse gastric cancer, but identifying these nearly invisible lesions
requires skillful experience and, in most cases, prior knowledge of the patient’s history.

FAQ: When should I raise the possibility of HDGC in my report?

Consensus guidelines for CDH1 mutation testing include the following criteria:
1. Two gastric cancer cases in a family regardless of age (at least on
 confirmed diffuse type)
2. Diffuse gastric cancer in an individual <40 years
3. Personal or family history (first- or second-degree relative) of diffuse gastric
cancer and lobular breast cancer, one diagnosed <50 years.95
Other patients in whom testing should be considered include:
1. Bilateral lobular breast cancer or family history (first- or second-degree
relative) of ≥2 cases of lobular breast cancer <50 years
2. Personal or family history (first- or second-degree relative) of cleft lip/palate
in a patient with diffuse gastric cancer
3. Any individual with in situ signet ring cells and/or pagetoid spread of signet
ring cell on a gastric biopsy

Figure 2-186. Hereditary diffuse gastric cancer, CDH1 mutated. A nearly invisible focus of early
gastric cancer (arrow) is found in this prophylactic gastrectomy specimen from a patient known
to carry a CDH1 germline mutation. Without prior knowledge of the patient’s history, such a
small focus could easily be missed.
Figure 2-187. Hereditary diffuse gastric cancer, CDH1 mutated, higher power of the previous
figure. The small cluster of signet ring cells (bracket) invade into the lamina propria without
disturbing the overall architecture or surface epithelium, making it extremely difficult to detect.
Figure 2-188. Hereditary diffuse gastric cancer, CDH1 mutated. Another example of early
gastric cancer in a patient with CDH1 gene mutation shows subepithelial signet ring cells
(brackets) infiltrating the lamina propria without disturbing the overall architecture or surface
mucosa, features that would normally alert the pathologist to take a closer look.
Figure 2-189. Hereditary diffuse gastric cancer, CDH1 mutated, higher magnification of the
previous figure. These pale signet ring cells (brackets) infiltrate the lamina propria and cause no
epithelial or stromal reaction. Note the absence of background gastritis or intestinal metaplasia.
Figure 2-190. Hereditary diffuse gastric cancer, CDH1 mutated. Single cells (arrowheads) are
so subtle as to be nearly invisible. Even with knowledge of the patient’s clinical history, finding
these foci on a prophylactic gastrectomy can be extremely challenging.
Figure 2-191. Hereditary diffuse gastric cancer, CDH1 mutated. Signet ring cells are easier to
identify when sufficient numbers cluster together. The absence of an expansile lesion, surface
change, or stromal changes is highly characteristic.
Figure 2-192. Hereditary diffuse gastric cancer, CDH1 mutated. This example has a backdrop
of inactive chronic gastritis that obscures an already difficult diagnosis. Single malignant signet
ring cells (arrows) are surrounded by chronic inflammatory cells.
Figure 2-193. Hereditary diffuse gastric cancer, CDH1 mutated, PAS stain. Some experts
suggest performing PAS stain in lieu of H&E to screen prophylactic gastrectomy specimens.
The PAS stain highlights tumor cells (arrow), providing a better contrast as compared with
H&E. This technique can also be helpful in biopsy material.
Figure 2-194. Hereditary diffuse gastric cancer, CDH1 mutated, PAS stain, higher magnification
of the previous figure. The mucin within the signet ring cells (arrow) are PAS positive.
 Figure 2-195. Hereditary diffuse gastric cancer, CDH1 mutated, PAS stain, higher magnification
of the previous figure. Individual signet ring cells are highlighted by PAS. Although many cells
contain intracytoplasmic mucin, benign cells maintain normal architecture and gland formation.
By comparison, the signet ring cells are single, discohesive, and disorganized.

FAQ: How do I process the prophylactic gastrectomy specimen?

Prophylactic gastrectomy is performed for cancer risk reduction, typically

between 20 and 30 years of age in patients with known CDH1 mutations.
Examination of the entire mucosa is essential, and one study showed at least
five occult tumor foci in all cases.96 Systematic submission of the entire mucosa
can be performed by cutting transverse strips and submitting the stomach
sequentially from proximal to distal. When multiple sections are embedded in
each block, the average case may yield 200 to 300 blocks. Coupled with a
gross photograph, these blocks can be mapped to assist in pinpointing the
exact tumor location at the time of signout. Mapping also allows for
concentrated screening in high-risk areas, such as the proximal stomach, where
nearly two-thirds of tumors are found (37% anterior proximal fundus and 27%
cardia/proximal fundus).96 Histologically, early and in situ lesions are difficult to
detect because the surface epithelium remains undisturbed. Some observers
recommend periodic acid–Schiff (PAS) stain in lieu of H&E (Figs. 2.193–2.195),
to improve detection of lesions and reduce the time required to screen cases.97

PEARLS & PITFALLS: Complete Submission of a Prophylactic Total

Gastrectomy May Yield Several Hundred Blocks!
Examination of the entire mucosa is important to identify areas of occult diffuse
gastric cancer. The typical prophylactic gastrectomy results in about 500
sections, and even when multiple sections are embedded within a single block,
the average case may yield 200 to 300 blocks. Undoubtedly, for most centers,
this will be a significant increase in block/slide production, and strategic
planning with the histology laboratory is critical to prevent bottlenecking of other
urgent clinical work. For these gastrectomies, completing the slide production
over several days while also prioritizing other patient material can ease the
burden on the laboratory and provide a more humane signout for the
pathologist of record. Although this increases the turnaround time for final
reporting, the delay is inconsequential, as prophylactic cases have little clinical
urgency.

FAQ: Can E-cadherin immunohistochemistry be used to identify CDH1-related

HDGC?

Germline mutations in the CDH1 gene encoding E-cadherin are detected in

nearly half of patients with HDGC. Among these patients, E-cadherin
immunohistochemistry shows reduced or absent expression in the tumor foci
and retained expression in the intervening nonneoplastic mucosa.98 However,
sporadic tumors may also show decreased or absent E-cadherin
immunoreactivity, making this stain ineffective as a diagnostic marker for CDH1
related HDGC.

FAQ: Is there a role for endoscopic surveillance for patients with HDGC?

There are no reliable screening tests that allow for early detection of diffuse
gastric cancer. Prophylactic gastrectomy is the treatment of choice for CDH1
mutation carriers after age 20 years, but annual endoscopic surveillance may
be considered for patients under the age of 20 years, patients who refuse or
 postpone gastrectomy (e.g., fertility concerns), and individuals who have
genetic variants of undetermined significance. Note, however, that because
diffuse gastric carcinomas do not form endoscopically visible lesions,
endoscopic surveillance is likely to have extremely low detection rates for
cancers in these patients. Random mapping biopsies may be sent to pathology,
but the estimated number of biopsies necessary to capture a single focus of
cancer (90% detection rate) is theoretically projected at 1768!96

KEY FEATURES: Hereditary Diffuse Gastric Cancer

• Autosomal dominant with high penetrance
• Nearly 50% attributed to mutations in the CDH1 gene encoding E-cadherin
• Lifetime risk of gastric cancer is 70% for men and 56% for women, with mean age 38
years (range 14 to 82 years)
• Endoscopic surveillance is not endorsed owing to ineffective detection of cancer
• Prophylactic total gastrectomy is recommended between ages 20 and 30 years
• The entire specimen should be mapped and submitted for histologic evaluation
• Nearly 70% of tumors are found in the proximal stomach
• PAS may be more helpful than H&E in the detection of early and small lesions (e.g., in
situ or pagetoid spread)

Gastric Adenocarcinoma and Proximal Polyposis of the Stomach

GAPPS was initially identified in 2012 and is characterized by the autosomal dominant
transmission of fundic gland polyposis with dysplasia and intestinal-type gastric
adenocarcinoma that are restricted to the proximal stomach, with no evidence of
duodenal or colorectal polyposis or other hereditary GI cancer syndrome.99,100 The
defining point mutations in exon 1B of the APC gene may also be found in some patients
with FAP,43 suggesting that GAPPS may represent a variant of FAP.101 Fewer than a
dozen families have been identified in the literature, and GAPPS is characterized by
incomplete penetrance, with polyposis presenting in patients as young as 10 years and
the median age for adenocarcinoma arising at 50 years (range 33 to 75 years), making it
important for pathologists to consider this entity.99,100,102 Proposed diagnostic criteria
include (1) gastric polyps limited to the body/fundus without colorectal or duodenal
polyposis; (2) >100 gastric polyps (or >30 polyps if known first-degree relative with
GAPPS); (3) polyps are primarily FGPs, some of which have dysplasia (or any family
member with dysplastic FGP or gastric carcinoma); and (4) autosomal dominant pattern
of inheritance.99

KEY FEATURES: GAPPS

• Gastric adenocarcinoma and proximal polyposis of the stomach is a variant of FAP
• Point mutation in exon 1B of APC gene
• Autosomal dominant inheritance with incomplete penetrance
• Carpet of >100 FGPs, some dysplastic, in body/fundus, as early as age 10 years
• Gastric adenocarcinoma, usually intestinal type, presents at median age 50 years
(33 to 75 years)
• Absence of colorectal or duodenal polyps is required
• Absence of other inherited polyposis syndromes is also required (Figs.
2.196–2.202)

Figure 2-196. Gastric adenocarcinoma and proximal polyposis syndrome, endoscopic view.
Diagnostic criteria include >100 gastric polyps limited to the gastric body/fundus (endoscopic
view pictured) without colorectal or duodenal polyposis and autosomal dominant inheritance.
Figure 2-197. Fundic gland polyps in a patient with GAPPS. The polyps in GAPPS are primarily
fundic gland polyps (pictured).

Figure 2-198. Fundic gland polyp in a patient with GAPPS. Most of the polyps found in GAPPS
are fundic gland polyps. This example shows cystically dilated glands amid oxyntic glands and
overlying normal foveolar epithelium.
Figure 2-199. Pyloric gland adenoma in a patient with GAPPS. This patient had a large gastric
pyloric gland adenoma (pictured) giving rise to adenocarcinoma, in addition to hundreds of
fundic gland polyps.
Figure 2-200. Pyloric gland adenoma (PGA) in a patient with GAPPS, higher magnification of
the previous figure. Architecturally, PGAs are composed of back-to-back tubules, which when
large can form papillary or frondlike extensions.
Figure 2-201. Pyloric gland adenoma (PGA) in a patient with GAPPS, higher magnification of
the previous figure. The cells of PGA can be columnar to cuboidal and have abundant clear to
eosinophilic cytoplasm with a ground glass or foamy appearance.
 Figure 2-202. Pyloric gland adenoma in a patient with GAPPS. Other areas closely resemble
pyloric glands, with closely packed small glands composed of cuboidal cells with clear foamy
cytoplasm and nuclei pushed toward the basement membrane.

Other Hereditary Cancer Syndromes

Gastric cancer has also been described in association with certain other inherited
cancer syndromes, including Lynch syndrome (hereditary nonpolyposis colorectal
cancer), FAP, Li-Fraumeni syndrome, Peutz-Jeghers syndrome, juvenile polyposis,
hereditary breast and ovarian cancer syndrome, and possibly PTEN hamartoma tumor
(Cowden) syndrome, but these frequently present with benign gastric polyps, when
present, and are all fairly rare causes of gastric cancer.

BIOMARKER TESTING 101

Programed Death Receptor-1/Programed Death Ligand-1 (PD-1/PD-L1)
The PD-1/PD-L1 pathway is involved in immune checkpoint surveillance that regulates
T-lymphocyte activation. By suppressing T-cell activation, tumors that express PD-
1/PD-L1 are able to hide from the immune system. This PD-L1 protein expression has
changed the way we think about tumor biology, and it has been identified in lymphoma,
non–small cell lung cancer, glioblastoma, melanoma, and malignancies in the kidney,
breast, and GI tract, among others.103 By targeting PD-1 or PD-L1 protein on these
tumor cells, a growing number of monoclonal antibody therapies allow restoration of
the body’s T-cell antitumor function. This therapy has been so effective that in late 2017
the US Food and Drug Administration (FDA) granted accelerated approval to
pembrolizumab (KEYTRUDA®, Merck & Co., Inc.) for patients with recurrent locally
advanced or metastatic gastric or GEJ adenocarcinoma whose tumors express PD-L1 as
determined by FDA-approved testing. At the time of this publication, only PD-L1 IHC
22C3 pharmDx (Dako) is approved, but this information is rapidly evolving, and up-to-
date information can be found at http://www.fda.gov/CompanionDiagnostics. For more
information about PD-1/PD-L1 function and immunohistochemical testing and
interpretation, see “Colon” chapter.

PEARLS & PITFALLS: PD-1 and PD-L1 Negative Results on Archived Material
If PD-L1 expression is not detected in an older archived gastric cancer
specimen, the FDA recommends assessing the feasibility of a fresh tumor
biopsy for repeat testing.

Human Epidermal Growth Factor Receptor 2 (HER2)

Human epidermal growth factor receptor 2 (HER2), also known as CerB-2 and ERBB2,
is a proto-oncogene located on chromosome 17q21 that encodes a transmembrane
protein with tyrosine kinase activity, is a member of the HER receptor family, and is
involved in signal transduction pathways leading to cell growth and differentiation.104
Quick fact: Proto-oncogenes are the normal inactive precursors of oncogenes.
Recall, tumors may arise owing to oncogene activation (turning on) or tumor suppressor
gene inactivation (turning off).
Amplification of the HER2 gene and overexpression of its product were first
discovered in breast cancer, in which it is significantly associated with worse
outcomes.105 Other carcinomas also demonstrate HER2 overexpression, including
colorectal, ovarian, prostatic, lung, gastroesophageal, and gastric, although a direct link
to outcomes is not as clear in these sites.106 In 2010, a landmark study demonstrating the
efficacy of trastuzumab in gastric and gastroesophageal carcinoma was published:
“Trastuzumab in combination with chemotherapy versus chemotherapy alone for
treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer
(ToGA): a phase 3, open-label, randomized controlled trial.”
Quick fact: Open-label or open trial is when both researchers and participants
know which treatment is being administered. In a randomized controlled trial, patients
are randomly assigned to either an intervention (in this case trastuzumab in combination
with standard chemotherapy) or control group (in this case, chemotherapy alone).
Clinical trials are conducted in “phases,” and each phase has a different purpose:

Phase I: Experimental treatment is tested on a small group of people (20 to 80) for the
first time. The purpose is to evaluate safety and identify side effects.
Phase II: Experimental treatment is tested on a larger group of people (100 to 300) to
determine effectiveness and further evaluate safety.
Phase III: Experimental treatment is administered to large groups of people (1000 to
3000) and compared against standard treatment. Effectiveness, side effects, and
safety are further assessed.
Phase IV: This phase follows FDA approval and availability of the treatment for the
public. Researchers continue to track safety and monitor information about risks,
benefits, and optimal use.

Trastuzumab is a monoclonal antibody directed against HER2 (for more information

about monoclonal antibodies, see “Colon”, chapter). It was one of the first molecular-
targeted drugs developed and was introduced for treatment of HER2-positive advanced
breast cancer. As stated in the study’s title, the ToGA trial demonstrated that the addition
of trastuzumab to standard chemotherapy results in increased overall survival among
patients with HER2-expressing unresectable gastric and gastroesophageal tumors.107
The benefit was modest (2.1 months improvement in progression-free survival and 2.7
months improvement in median overall survival) but was statistically significant and
resulted in early termination of the study. As a result of the ToGA trial, trastuzumab is
the first molecular-targeted agent approved by the FDA for treatment in gastric cancer
and other molecular HER2 agents continue to be tested (e.g., pertuzumab, lapatinib).
Combined guidelines from the College of American Pathologists (CAP), American
Society of Clinical Pathology (ASCP), and American Society of Clinical Oncology
(ASCO) recommend that patients with HER2-positive tumors be offered combination
chemotherapy and HER2-targeted therapy as the initial treatment of either primary or
metastatic disease.108 Thus, it is imperative to determine the HER2 status in advanced
gastric or gastroesophageal adenocarcinoma to select patients who may benefit from
treatment. These CAP/ASCP/ASCO guidelines endorse immunohistochemistry for
HER2 as the initial test of choice. However, it is not as simple as applying one’s
knowledge of breast cancer HER2 testing to these gastric/gastroesophageal cancers
(Figs. 2.203–2.212). Interpretation of HER2 IHC relies on three fundamental factors:
stain location, extent, and intensity, and there are several key differences in the
interpretation as compared with breast carcinoma (see Table 2.5). First, and perhaps
most importantly, stain location does not require complete circumferential membranous
staining. Instead, basal, basolateral, or lateral staining is adequate (Figs. 2.205 and
2.206). The exception is isolated luminal membrane staining, which is considered
negative (Fig. 2.211). Second, stain extent requires a minimum amount of tumor cell
reactivity: ≥10% in resection specimens or at least 1 cancer cell cluster, defined as ≥5
cells, in biopsy specimens. Finally, if these two previous criteria are met, a score may
be applied based on stain intensity: 1+ for faint or barely perceptible (seen only on 40x
objective, Figs. 2.203 and2.204), 2+ for weak to moderate (visible at 10 to 20x, Figs.
2.205 and 2.206), or 3+ for strong (seen at 2 to 4x, Figs. 2.207 and 2.208). If the tumor
fails to meet minimum thresholds for location or extent, it is scored as 0. Scores of 0
and 1+ are considered negative results, and no additional testing is required, as these
patients will not benefit from trastuzumab therapy. A score of 2+ is equivocal and
requires additional testing by fluorescence in situ hybridization (FISH) analysis or other
in situ hybridization method (e.g., silver in situ hybridization, chromogenic in situ
hybridization, and dual-color dual-hapten in situ hybridization). Criteria for in situ
hybridization results are listed in Table 2.6.

PEARLS & PITFALLS: Beware False-Positive Staining Patterns!

IM and dysplastic epithelium commonly stain with partial membranous reactivity
(Fig. 2.209). Stay alert to this pitfall and inspect well-preserved areas of
invasive tumor only (Fig. 2.210). Recall, positive staining is linear membranous
reactivity, either complete or incomplete lateral/basolateral, in 10% of tumor
cells for resections or a single cancer cell cluster (≥5 cells) in biopsy
specimens. Avoid calling false-positive results in the following scenarios:
• Cytoplasmic or nuclear reactivity only (Fig. 2.212)
• Isolated luminal membrane staining (Fig. 2.211)
• Granular or pericellular pattern
• Metaplasia or regenerative changes (e.g., near ulceration)
• Edge or crush artifact
• IM or dysplasia (Figs. 2.209 and 2.210)
• Signet ring cells with marginated cytoplasmic staining

PEARLS & PITFALLS: Select the Tissue Block With the Lowest-Grade Tumor
Morphology!
As mentioned earlier, HER2 overexpression is seen more frequently in the
intestinal phenotype (24%) and mixed tumors (20%) as compared with diffuse
signet ring cell type (0% to 6%).109,110 Low-grade tumors show greater
frequency of HER2 positivity compared with high grade,111 and so block
selection should target the lowest-grade tumor and areas showing intestinal
differentiation to have the highest prospect of positive HER2 result. Other rare
morphologic variants of gastric adenocarcinoma (e.g., adenosquamous,
papillary, hepatoid) lack sufficient data for comment. More than one tissue
block may be tested if the tumor is morphologically heterogeneous.

Figure 2-203. HER2 negative, score 1+. A sufficient number of tumor cells are staining in this
biopsy specimen (one cancer cell cluster of ≥5 cells), and membranous staining is present.
However, the intensity of the stain is 1+, as it is barely perceptible, even at high magnification
(40x). A score of 1+ is considered negative, and there is no need to reflex to FISH testing.
Figure 2-204. HER2 negative, score 1+. In this biopsy example, the staining criteria are met,
with basal, lateral, and basolateral staining present in at least ≥5 cells of this tumor cell cluster.
However, the stain is faint and requires 40x magnification to be perceptible. This 1+ staining
intensity is considered negative, and there is no need to reflex to FISH testing.
Figure 2-205. HER2 equivocal, score 2+. Sufficient tumor cells are staining (at least one cancer
cell cluster defined as ≥5 cells in a biopsy) in a basolateral membranous pattern. However, the
staining intensity is weak to moderate and is best seen at 10–20x. This is scored as 2+ and
requires follow-up with FISH studies.
Figure 2-206. HER2 equivocal, score 2+. Complete or incomplete basolateral staining is
present in a sufficient number of tumor cells (at least one cancer cell cluster defined as ≥5 cells
in a biopsy). Compared with breast carcinoma, which requires circumferential membranous
staining, discontinuous basal/lateral membranous staining is acceptable with gastric cancers.
However, the staining intensity is weak to moderate and is best seen at 10–20x. This is scored
as 2+ and requires follow-up with FISH studies to determine whether the patient can benefit
from trastuzumab therapy.
Figure 2-207. HER2 positive, score 3+. Sufficient tumor cells are staining (at least one cancer
cell cluster defined as ≥5 cells in a biopsy) in a complete membranous pattern with strong
reactivity visible at low magnification (2–4x). A 3+ interpretation is positive and requires no
additional FISH testing. Although this example shows complete membranous reactivity, gastric
HER2 interpretation requires only basal, lateral, or basolateral staining.
Figure 2-208. HER2 positive, score 3+. At least 10% of tumor cells in a resection specimen
must show reactivity to be properly graded. This example shows strong and complete
membranous reactivity in the tumor cells visible at low magnification (2–4x). A score of 3+ is
considered positive, and patients may benefit from trastuzumab therapy.
Figure 2-209. HER2 negative, score 0. Beware of strong staining in areas of overlying dysplasia
(pictured) or intestinal metaplasia. Areas of metaplasia or regenerative changes near ulcers
may also show false-positive reactivity. Interpretation of HER2 should be limited to the invasive
carcinoma only, and it is helpful to have the H&E handy for correlation.
Figure 2-210. HER2 negative, score 0. The invasive cancer (bracket) is negative for HER2, but
the overlying dysplastic epithelium shows strong reactivity. This is a common false-positive
pitfall to avoid; always inspect well-preserved areas of invasive tumor only.
Figure 2-211. HER2 negative, score 0. Although there appears to be linear membranous
reactivity in at least five tumor cells, this shows an isolated luminal staining pattern (arrowhead
). True positive staining requires basal, lateral, or basolateral membranous reactivity.
 Figure 2-212. HER2 negative, score 0. There is moderate to strong reactivity in at least five
tumor cells, but the staining patterns are nuclear, cytoplasmic, granular, and extracellular. True
linear membranous reactivity in a lateral/basolateral distribution is absent. This HER2 should be
interpreted as negative.

TABLE 2.5: Evaluation of HER2 Immunohistochemistry in Gastric

Adenocarcinoma
 Membranous
Complete or Objective Surgical Biopsy
Basolateral With Visible Specimen Specimen Interpretation Next
Staining Staining Criteria Criteria Score by IHC Step

None or any <10% of No cells 0 Negative None

membranous cancer staining
cells

Faint or barely 40x ≥10% of 1 cancer 1+ Negative None

perceptible cancer cell cluster
cells (≥5 cells)
Weak to moderate 10–20x 2+ Equivocal HER2
FISH
or
other
ISH
study

Strong 2–4x 3+ Positive None

FISH, fluorescence in situ hybridization; IHC, immunohistochemistry; ISH, in situ hybridization.

TABLE 2.6: Comparison of HER2 Scoring Between Gastroesophageal and

Breast Carcinomas
 Gastroesophageal
Adenocarcinoma Breast Carcinoma

IHC criteria Extent Biopsy: 1 tumor cluster >10% tumor cells

(≥5 cells)
Resection: ≥10%
tumor cells

Circumferential No. Basal, basolateral, Yes

membrane staining lateral, or complete
required acceptable.

Fixation requirements None 6–72 hours; 10%

neutral buffered
formalin

ISH criteria Single probe HER2 ≥6.0 signals per cell Same
copy number

Dual probe Ratio ≥2.0 with ≥4.0 Same

HER2/CEP17 ratio signals/cell
and HER2 copy Ratio ≥2.0 with <4.0
number signals/cell
Ratio <2.0 with ≥6.0
signals/cell

Positive HER2 Tumor types 25% Intestinal 25% Ductal

Features 10% Mixed
5% Diffuse (signet ring
cell)

Tumor location 30% GE junction Not applicable

15% Gastric

GE, gastroesophageal; IHC, immunohistochemistry; ISH, in situ hybridization.

FAQ: Should HER2 testing be performed on biopsy or resection specimens?

What about metastases?

Any tumor tissue may be tested, but preferably treatment-naive tumor.

Biopsy or resection specimens are both acceptable, and either primary or
metastatic tumor is suitable. HER2 rates in the ToGA trial were similar between
biopsy and surgical specimens (23.2% and 19.7%, respectively).107,112 Other
studies confirm that HER2 amplification is similar between paired resection and
biopsy specimens, as well as their metastatic samples (>90%
concordance).112-116 Given the overall high degree of concordance, HER2
testing on neoplastic tissue from primary or metastatic tumor in either biopsy or
resections specimens is appropriate. Selection of tumor tissue acquired before
the initiation of trastuzumab therapy is preferred if such specimens are
available and adequate.

FAQ: Are there requirements for tissue fixation?

No. Unlike breast carcinoma, there are no specific fixation requirements for
HER2 testing of gastric/gastroesophageal adenocarcinoma (see Table 2.6).

PEARLS & PITFALLS: Scoring of HER2 IHC in Gastric Cancer Is Different

From Scoring in Breast Cancer
The Ruschoff/Hofmann scoring method for gastric cancer, adapted from the
ToGA trial, has a lower threshold as compared with that of breast cancer.
Complete membranous staining is not required. Any linear membranous (not
cytoplasmic) basolateral, lateral, or complete staining is considered sufficient
for further scoring, except in the case of isolated luminal membrane staining.
The scoring method then takes into account two additional factors: intensity of
staining and minimum amount of tumor cells positive. Intensity of staining is
scored as 1+ faint or barely perceptible, visible only on 40x objective; 2+ weak
to moderate, visible at 10 to 20x; and 3+ strong, visible at 2 to 4x. This staining
must be seen in at least 10% of tumor cells in resections, or in a single cancer
cell cluster (≥5 cells) in a biopsy specimen.117 See Table 2.5 and Figs.
2.203–2.208.

FAQ: How many biopsy fragments are required?

At minimum, five biopsies should be taken at the time of endoscopy.

However, clinicians should be advised that six to eight biopsy fragments are
optimal owing to tumor heterogeneity.118 Make sure to communicate any
concerns for adequacy, in particular if the HER2 result is negative or
uninterpretable.
It is never a bad idea to have a generous sample at first pass. If
subsequent genomic testing is requested (there is currently not sufficient
evidence to recommend for or against), this archived tissue expedites the
process and obviates the need for rebiopsy.
FAQ: Are HER2 results reliable on fine needle aspiration (FNA) specimens? Is
it appropriate to use an FNA specimen if it is the only material available?

HER2 testing on FNA specimens (cell blocks) is an acceptable alternative

it is the only material available, as per CAP/ASCP/ASCO guidelines. However,
this recommendation is extrapolated primarily from breast carcinoma data, and
limited conclusive data are available for gastric/gastroesophageal cancers.
Given the wide variations in tissue handling and processing of cytologic
samples (e.g., use of decalcification and fixative selection: proprietary vs.
alcohol based vs. direct formalin vs. alternative), laboratories should confirm
test performance of HER2 assays on these types of specimens before
reporting patient results. There are no set regulations guiding this practice, and
testing paradigms may be determined by the local laboratory director.

FAQ: Should clinicians wait for HER2 results before initiating therapy?

No. Many centers send out testing to reference laboratories, which may
increase turnaround time for HER2 results. Although most oncologists will be
anxious to receive these results, there is no need to know HER2 status before
starting chemotherapy. If HER2 results are subsequently positive, clinicians
may later add trastuzumab to the treatment plan. In addition, most patients with
gastric cancer are symptomatic at the time of diagnosis and will benefit from
immediate oncologic management. For pathologists, the turnaround time
benchmark is 90% of reports within 10 working days from the date of
procedure. For send outs, 90% of specimens should be sent within 3 days of
tissue processing.

FAQ: Why do scoring criteria for HER2 IHC differ between gastric and breast
cancers?

Interpretation of gastric HER2 IHC is based on criteria used in the ToGA

trial. Recall, this trial demonstrated survival benefit among patients with HER2-
positive tumors when trastuzumab was added to their chemotherapy regimens.
The HER2 scoring system used in this study differed from breast scoring,
hence the variation in criteria. For gastric/gastroesophageal cancers, complete
circumferential membranous staining is not required. Instead, any basal,
basolateral, lateral, or complete membranous staining may be scored (Figs.
2.203–2.208). Note, however, that isolated luminal membrane staining is
negative (Fig. 2.211), as is isolated nuclear or cytoplasmic staining (Fig. 2.212).
Scoring then relies on additional factors, such as extent of tumor cells staining
and intensity of staining, as detailed in Table 2.6.

FAQ: Why not just perform FISH on all patients?

In gastric cancers, HER2 testing by FISH and IHC show only moderate
fidelity, as up to 20% of gastric cancers with negative IHC interpretation (0 or
1+) show positive amplification by FISH. However, no significant survival benefit
from trastuzumab is seen in these patients and, as a result, IHC is the first-line
test for HER2 overexpression.119 This is in contrast to the experience with
breast carcinoma whereby an extremely low threshold to perform FISH is
maintained and every attempt is sought to achieve a positive HER2 result.
Partially driving this practice is the high efficacy of trastuzumab in HER2+
breast carcinoma along with the low toxicity profile of the drug. This makes
trastuzumab low risk and potentially high yield for clinicians to include in their
arsenal against breast cancer. However, this does not hold true for gastric
cancers, and thus the testing algorithms diverge. Unlike in breast cancer, there
is no significant survival benefit for trastuzumab if IHC is 0 or 1+ even when
FISH shows HER2 amplification. On the other hand, if there is uncertainty over
IHC scoring of 1+ versus 2+, the best approach is to simply reflex to FISH.

FAQ: Should I ever repeat HER2 testing?

No. There are no data to support repeat HER2 testing if initial testing is
negative, although some clinical colleagues may request this.

PEARLS & PITFALLS: The Path Report Should Specify the Antibodies and
Probes Used
Several HER2 antibodies are offered through various vendors, and there is no
specific recommended antibody. The ToGA trial used HercepTest, whereas
other studies have applied Ventana 4B5 or Thermo Fischer Scientific CB11, and
even more variations are available on the market. Although concordance among
antibodies is moderate to good, the ASCO/CAP/ASCP guidelines strongly
recommend laboratories specify the antibodies and probes used for the test.120
An example of this standard verbiage is included later. The guidelines also
emphasize that assays should be appropriately validated for HER2 IHC and
ISH on gastroesophageal adenocarcinoma specimens, although this is standard
practice for any laboratory test. As with any other test, 20 positive and 20
negative gastroesophageal adenocarcinomas should be verified for FDA-
approved tests, and 40 samples for laboratory-developed tests.
Example verbiage for inclusion in report:
Method: Testing is performed using FDA-approved Ventana Pathway HER2
(4B5) rabbit monoclonal primary antibody and a proprietary detection system.
No expression (HER2 score of 0), low expression (HER2 score of 1+), and high
expression (HER2 score of 3+) controls are used. All controls show
appropriate reactivity.
Scoring: Scoring is performed according to the following article: Ruschoff J,
Dietel M, Baretton G, et al. HER2 diagnostics in gastric cancer-guideline
validation and development of standardized immunohistochemical testing.
Virchows Arch. 2010; 457(3):299-307.

CHECKLIST: Steps to Reporting HER2 in Gastric/Esophageal Carcinoma

□ Advise clinicians to begin other chemotherapy while waiting for HER2 results;
trastuzumab may be added later if HER2 results are positive
□ IHC is the first test of choice; reflex to ISH only if results are 2+ equivocal
□ No special fixation constraints required
□ Use biopsy, resection, or metastatic sample from treatment-naive tumor
□ Cytology cell block is appropriate if no other tissue is available
□ Select the tissue block with lowest-grade or intestinal-type carcinoma
□ Score with three factors:
○ Stain location: Complete or basolateral membranous
○ Stain extent: >10% in resection; 1 cluster of ≥5 cells in biopsy
○ Stain intensity: 1+ weak (40x); 2+ moderate (10 to 20x); 3+ strong (2 to 4x)
 □ Report:
○ Negative: 0 or 1+
○ Equivocal: 2+
○ Positive: 3+
□ Take action:
○ Negative: 0 or 1+ → None
○ Equivocal: 2+ → Send for ISH HER2
○ Positive: 3+ → None
□ Specify antibodies and probes used in path report

WELL-DIFFERENTIATED NEUROENDOCRINE
TUMORS (FORMERLY “CARCINOID”)
Based on WHO 2017 classification, well-differentiated neuroendocrine tumors (WD-
NETs) are distinguished from NECs by their morphologic features, independent of
mitotic count or Ki-67 proliferative index. This is a departure from previous WHO
2010 classification for neuroendocrine neoplasms, in which tumors were stratified
based solely on mitoses and Ki-67. The updated system not only more reliably indicates
prognosis and response to therapy but is also more intuitive to pathologists. For
example, tumors that are morphologically uniform are classified as WD-NETs, whereas
NECs are classified by their poorly differentiated high-grade cytology and marked
nuclear atypia. NEC is discussed separately in the gastric adenocarcinoma section
“Tumor Classification”, as its prognosis and staging more closely reflect those of
gastric adenocarcinoma.
Most gastric WD-NETs are composed of ECL cells, typically in the corpus and
fundus (90%),121 which express chromogranin A or synaptophysin by
immunohistochemistry. Endoscopically they appear as submucosal nodules or polyps
(Fig. 2.213), sometimes with ulcerations. Gastric WD-NETs are classified into three
groups, each arising in different clinical contexts (see Table 2.7), and each with
divergent prognoses and treatment protocols. In isolation, the tumors are histologically
indistinguishable and are composed of nests or trabeculae of small, uniform, polygonal,
or cuboidal cells with lightly eosinophilic and finely granular cytoplasm. The nuclei are
round or oval with smooth nuclear borders and stippled chromatin with indistinct
nucleoli. The key to differentiating the three types of gastric WD-NETs from one another
requires examination of the nonpolypoid background mucosa, yet again underscoring the
importance of background gastric biopsies with all gastric lesions.

Figure 2-213. Gastric well-differentiated neuroendocrine tumor. This pale endoscopic nodule
appears submucosal. The background gastric mucosa also shows a mosaic pattern and
patchy atrophy.

TABLE 2.7: Comparison of Gastric Well-Differentiated Neuroendocrine

Tumors
 Type 1 Type 2 Type 3

Frequency 70%–80% of gastric NETs Rare 10%–15% of gastric

NETs

Focality Multifocal Multifocal Solitary lesion

Size 0.5–1.0 cm ≤1.5 cm Variable; one third >2

cm

Location Body/Fundus Body/Fundus Anywhere in

stomach

Associated with Hypergastrinemia Hypergastrinemia Sporadic

Hypochlorhydria/achlorhydria MEN1 No clinical
AMAG Zollinger-Ellison associations
ECL-cell hyperplasia syndrome
Pernicious anemia

Clinical behavior Usually benign 30% metastasize Dependent on size

Rare metastases and depth of
invasion

Demographics F≫M M= F M> F

70%-80% of patients are Mean age 50 Mean age 65
50’s–60’s

GASTRIC WELL-DIFFERENTIATED NEUROENDOCRINE

TUMOR, TYPE I
Type I gastric WD-NETs are the most common and arise in the setting of AMAG. These
occur in middle-aged women (70% to 80%) and are the result of ECL-cell hyperplasia.
In AMAG, the autoimmune destruction of parietal cells leads to reduced gastric acid
production and loss of feedback inhibition of gastrin secretion in the antral G cells (i.e.,
the G cells cannot turn off gastrin secretion). The resulting hypergastrinemia stimulates
ECL cells to proliferate, which appear as multiple small nodules in the body/fundus of
the stomach. Technically, this early change represents a reversible hyperplasia, but may
progress to malignancy, especially as tumors enlarge. As compared with type II and type
III gastric WD-NETs, type I lesions have an excellent prognosis with exceedingly low
rates of metastatic disease.122 EMR of any visible lesions and close endoscopic follow-
up is prudent, but there are no existing guidelines for surveillance. Antrectomy to
remove the stimulatory G cells has also proven useful as long-term therapy, and
treatment of underlying pernicious anemia is required because the absence of parietal
cells also results in a deficit of intrinsic factor, the transporter for vitamin B12.123
Differentiating this well-performing tumor from other gastric WD-NETs requires
examination of the background flat mucosa (Figs. 2.214–2.229). Look for the
constellation body/fundus-predominant characteristics of AMAG: (1) IM, (2) atrophy of
oxyntic glands, (3) inactive chronic gastritis (typically low-lying and lymphocytic, in
contrast to the superficial plasmacytic inflammation of H. pylori), (4) linear or nodular
ECL-cell hyperplasia (≥5 cells in a row or cluster as highlighted by chromogranin or
other neuroendocrine immunohistochemistry), and (5) pyloric metaplasia (and
sometimes pancreatic acinar cell metaplasia).

Figure 2-214. Gastric WD-NET, histology of the previous figure. At low magnification, this is an
expansile lesion composed of sheets of uniform cells with a slightly trabecular architecture. The
base of the lesion is invading into the muscularis mucosae. The prognosis of this lesion
depends upon the etiopathogenesis, which can be deduced from the changes found in the
background nonlesional sample.
Figure 2-215. Autoimmune metaplastic atrophic gastritis, background mucosa of the previous
figure. Background nonlesional gastric tissue is important when encountering a WD-NET in the
stomach. Type 1 WD-NETs have an excellent prognosis and arise in the setting of AMAG,
which shows a constellation of features. The oxyntic glands are absent in this image because
of complete oxyntic gland atrophy, and this is accompanied by background chronic
inflammation, intestinal metaplasia (arrowhead ), and pyloric metaplasia (arrow).
Figure 2-216. Autoimmune metaplastic atrophic gastritis. At lower magnification, one can
appreciate the complete absence of oxyntic glands, the presence of a low-lying lymphocytic
infiltrate, intestinal metaplasia (arrowhead ), and pyloric gland metaplasia (arrow). These
features, in conjunction with ECL-cell hyperplasia (not pictured), indicate AMAG.
Figure 2-217. Autoimmune metaplastic atrophic gastritis, nodular ECL-cell hyperplasia. ECL-
cell hyperplasia is not always visible on H&E stain, but these small nodular aggregates
(arrowheads) stain with chromogranin. To diagnose AMAG, one must find linear or nodular
ECL-cell hyperplasia (defined as at least five adjacent cells) in combination with oxyntic gland
atrophy (note the absence of the typical pink parietal cell and blue chief cells), low-lying
lymphocytic inflammation, intestinal metaplasia, and pyloric gland metaplasia.
Figure 2-218. Autoimmune metaplastic atrophic gastritis, nodular ECL-cell hyperplasia. Nodular
ECL-cell hyperplasia (arrowhead ) can be confirmed by chromogranin immunostain (not
pictured). The mucosa additionally shows a combination of complete atrophy of oxyntic glands,
low-lying lymphocytic inflammation, intestinal metaplasia, and pyloric metaplasia. The absence
of oxyntic glands leads to loss of acid secretion, which normally inhibits gastrin secretion. In the
absence of acid, gastrin levels increase and lead to ECL-cell hyperplasia, seen here.
Figure 2-219. Autoimmune metaplastic atrophic gastritis, nodular ECL-cell hyperplasia.
Uninhibited gastrin secretion results in nodular hyperplasia of ECL cells (arrowhead ), which
can become neuroendocrine tumors. The size cutoff varies by publication (0.5 vs. 0.5 cm).
However, because the metastatic rate of small lesions is negligible, one practical approach is to
report all small ECL-cell nodules as nodular hyperplasia and reserve the term WD-NET for
endoscopically visible lesions submitted as nodules or polyps.
Figure 2-220. WD-NET arising in AMAG, endoscopy. This endoscopically visible lesion is a
WD-NET, and the background gastric mucosa is nodular with atrophy. Biopsy samples of the
background mucosa (previous figures) show histologic features of AMAG, including numerous
areas containing nodular ECL-cell hyperplasia. Size cutoffs for differentiating hyperplasias from
NETs are arbitrary, and the authors take a practical approach in calling lesions WD-NETs only if
they correlate with an endoscopically visible lesion (arrow).
Figure 2-221. Quick tutorial on the interpretation of AMAG. AMAG is a body-predominant
disease. Biopsies of the gastric antrum (top row) are essentially unremarkable or may show
changes of chemical/reactive gastropathy. Gastrin and chromogranin stains in the antrum
highlight the normal band of stimulatory “G” cells that secrete gastrin. By comparison, the
gastric body and fundus (bottom row) show a constellation of features that can be identified by
H&E, including (1) partial or complete atrophy of oxyntic glands; (2) lymphocytic inflammation,
often low lying; (3) intestinal metaplasia; and (4) pyloric gland metaplasia. The features are
almost indistinguishable from gastric antral tissue containing intestinal metaplasia. Therefore, a
gastrin stain can be performed to confirm the absence of G cells (which reside only in the true
antrum), thus identifying the tissue as true body/fundus with atrophy. A chromogranin stain
highlights the final diagnostic feature of ECL-cell hyperplasia, either linear (arrowhead ) or
nodular (arrow).
Figure 2-222. Quick tutorial on the interpretation of AMAG. By H&E, the gastric antrum (top row)
is unremarkable or has chemical/reactive gastropathy, as seen here. The gastric body (bottom
row) contains red flags to further pursue an AMAG workup. The easiest red flag to spot is
intestinal metaplasia in a biopsy labeled as body/fundus. Other features include the absence of
normal oxyntic glands, a low-lying lymphocytic infiltrate, and pyloric gland metaplasia. Any
combination of these should prompt a basic AMAG workup, which includes gastrin and
chromogranin immunostains. In the antrum, gastrin and chromogranin both highlight the
horizontal band of gastrin-secreting G cells. In the body/fundus, the widespread absence of G
cells is expected and confirms the tissue source. Chromogranin highlights linear (arrow) and
nodular (arrowhead ) ECL-cell hyperplasia.
Figure 2-223. WD-NET arising in AMAG. At low magnification, this tumor is expansile and
composed of small nests of cells. The background mucosa appears atrophic with intestinal
metaplasia and pyloric-type glands.

Figure 2-224. WD-NET arising in AMAG. A chromogranin immunostain of the previous case
highlights the tumor cells, confirming neuroendocrine differentiation.
Figure 2-225. WD-NET arising in AMAG, higher magnification of the previous figure. The tumor
cells are uniform and arranged in small nests and cords.
Figure 2-226. WD-NET arising in AMAG. The mucosa adjacent to the tumor (right) provides
clues to the pathogenesis and prognosis. A specific combination of features offers a telltale
story of AMAG: complete atrophy of oxyntic glands, chronic inflammation, intestinal metaplasia,
and pyloric metaplasia. ECL-cell hyperplasia can be found on the chromogranin stain.
Figure 2-227. AMAG, background mucosa of the previous figure. The H&E features show a
complete absence of oxyntic glands (no chief cells or parietal cells), chronic inflammation,
intestinal metaplasia, and pyloric metaplasia. The last feature of AMAG (ECL-cell hyperplasia)
can be confirmed by chromogranin immunostain.
Figure 2-228. AMAG, linear and nodular ECL-cell hyperplasia, chromogranin stain. Linear and
nodular ECL-cell hyperplasia is defined as five or more adjacent chromogranin reactive cells.

Figure 2-229. AMAG, linear and nodular ECL-cell hyperplasia, chromogranin stain. Linear and
nodular ECL-cell hyperplasia is defined as five or more adjacent chromogranin reactive cells.
GASTRIC WELL-DIFFERENTIATED NEUROENDOCRINE
TUMOR, TYPE II
Type II NETs (Fig. 2.230) are rare and arise in the setting of ZE syndrome due to MEN1
syndrome or a gastrin-secreting tumor elsewhere in the GI tract. Similar to the
mechanism in AMAG, the uninhibited gastrin secretion stimulates ECL cells to
proliferate, resulting in WD-NETs (often multiple). These type II tumors have worse
prognosis than type I, with metastasis in about 30% of cases.122 However, type II tumors
behave distinctly better than type III tumors, again underscoring the importance of
differentiating WD-NETs, which can be achieved by reviewing tandem biopsies of the
nonpolypoid mucosa (Figs. 2.230–2.237). Biopsies of the background mucosa in ZE
syndrome show oxyntic gland hyperplasia (Fig. 2.231) (as compared with atrophy in
AMAG), and diffuse endocrine cell hyperplasia can be identified by
immunohistochemistry. Local resection of the NET, evaluation for metastatic disease,
and resection of the stimulatory gastrin-secreting tumor (usually found in the small
bowel) is the mainstay of therapy.122

Figure 2-230. Zollinger-Ellison syndrome in a patient with MEN1, endoscopic view. The gastric
folds are hypertrophic.
Figure 2-231. Gastric oxyntic mucosa of a patient with Zollinger-Ellison syndrome. These
patients have a gastrin-secreting tumor (gastrinoma), often found in the small bowel, and the
resulting hypergastrinemia causes direct stimulation of oxyntic mucosa to secrete copious
amounts of acid. Biopsies show hyperplastic oxyntic mucosa with proton pump inhibitor effect
(pictured); PPIs are prescribed to suppress acid secretion. Curative treatment relies on
identification and resection of the gastrinoma.
Figure 2-232. WD-NET arising in gastric oxyntic mucosa of a patient with Zollinger-Ellison
syndrome. Multiple nests of uniform cells are present between the oxyntic glands.
Figure 2-233. WD-NET arising in gastric oxyntic mucosa of a patient with Zollinger-Ellison
syndrome. Higher magnification of the previous figure shows uniform tumor cells without
prominent nucleoli.
Figure 2-234. WD-NET arising in gastric oxyntic mucosa of a patient with Zollinger-Ellison
syndrome. Biopsies submitted as gastric “nodules” show extensive involvement of the gastric
mucosa by WD-NET.
Figure 2-235. WD-NET arising in gastric oxyntic mucosa of a patient with Zollinger-Ellison
syndrome, chromogranin immunostain. The tumor cells are reactive for chromogranin.
Figure 2-236. WD-NET arising in gastric oxyntic mucosa of a patient with Zollinger-Ellison
syndrome. Another look at the tumor cells, which appear remarkably uniform.
 Figure 2-237. WD-NET arising in gastric oxyntic mucosa of a patient with Zollinger-Ellison
syndrome, chromogranin immunostain.

GASTRIC WELL-DIFFERENTIATED NEUROENDOCRINE

TUMOR, TYPE III
Type III gastric WD-NET (Figs. 2.238–2.241), the second most frequent type, has no
associated clinical syndrome or context; rather, these lesions are sporadic. Type I and
type II tumors arise predominantly in the gastric body and are multiple, whereas type III
tumors can arise anywhere in the stomach and are typically solitary. In contrast to the
excellent prognosis of type I and type II tumors, lymph node metastasis is found in 71%
of type III tumors measuring >2 cm.124 Small <1 cm tumors rarely metastasize, but in all
cases of type III tumors, surgical resection is advised. Type III WD-NETs are not
associated with any specific background mucosal changes.
Figure 2-238. Gastric WD-NET. These tumors are staged by both size and depth of invasion.
For example, this tumor, which invades into the submucosa would be staged as pT1, but only if
the tumor is ≤1 cm. Anything >1 cm is automatically pT2 even if only superficially invasive.
Figure 2-239. Gastric WD-NET chromogranin stain of the previous figure. A chromogranin stain
highlights the tumor cells, confirming their neuroendocrine differentiation.
Figure 2-240. Gastric WD-NET. Differentiation (well differentiated vs poorly differentiated) is
determined by morphology. Uniform cells such as seen here define this tumor as a well-
differentiated NET. Grading, by comparison, is determined by mitotic count and Ki-67
proliferation.
 Figure 2-241. Gastric WD-NET. Well-differentiated NETs have uniform round nuclei that are
similar in size and shape (pictured). The chromatin is “salt and pepper” with indistinct nucleoli.
If there is significant pleomorphism (variation in shape) and anisonucleosis (variation in nuclear
size), consider a diagnosis of poorly differentiated neuroendocrine carcinoma, instead.

SAMPLE NOTE: Well-Differentiated NET Arising in

AMAG
Stomach, Body, Nodule, Biopsy
• Well-differentiated NET (G1), type I, arising in a backdrop of AMAG, see Comment.

Comment
Type I gastric WD-NETs arise in the setting of hypergastrinemia owing to AMAG and
are well-performing tumors. The rate of metastatic disease to lymph nodes or distant
sites is negligible and, if the lesion is amenable, conservative EMR is adequate
treatment. Patients with AMAG are at risk for pernicious anemia, dysplasia, and gastric
adenocarcinoma. Continued endoscopic surveillance is suggested, if clinically
appropriate.
FAQ: What is the size cutoff for a NET versus hyperplasia, and what is the
significance?

Size cutoffs for NETs vary based on published sources. For example:
College of American
Pathologists

≥0.5 mm Neuroendocrine tumor

<0.5 mm In situ, neuroendocrine dysplasia or

hyperplasia

World Health Organization

≥0.5 cm Neuroendocrine tumor

>0.5 mm–<0.5 cm Microcarcinoid

≤0.5 mm Endocrine cell hyperplasia

The discrepancy in size criteria is a nonissue for type III WD-NETs, which
are typically bulky masses at presentation, but it can cause some confusion for
type I and II lesions, which are stimulated by excess gastrin and can range
from minute endocrine cell aggregates (technically reversible hyperplasias) to
larger neoplasms that persist even following removal of the gastrin stimulus.
Because the metastatic rate of small lesions is negligible, the authors exercise
a more practical approach over measuring endocrine cell aggregates: the term
WD-NET is reserved for endoscopically visible nodules or polyps (Figs. 2.213
and 2.220), whereas endocrine aggregates found on random samples are
described as hyperplasia (Figs. 2.221, 2.222, 2.228, and 2.229). Far more
important than splitting hairs over this nomenclature, identifying the features in
the background mucosa (AMAG vs. ZE vs normal/nonspecific) provides
information for subtyping, prognostication, and treatment.

PEARLS & PITFALLS: Grading of WD-NET Requires Both Mitotic Count and
Ki-67 Proliferative Index
WD-NETs are graded by both mitotic index and Ki-67 proliferation index (Figs.
2.242–2.244). The Ki-67 index frequently results in a higher grade than mitotic
count, and studies have shown these grade-discordant tumors more likely to
have metastases to lymph nodes and distant sites, perineural invasion, small
vessel invasion, and overall survival.125 In cases for which grade results are
discordant, assign the higher grade.
WHO 2017 Classification of
Neuroendocrine Neoplasms

WD-NET G1 <3% Ki-67 <2 mitoses/10

HPF

WD-NET G2 3% to 20% 2–20

Ki-67 mitoses/10
HPF

WD-NET G3 >20% Ki- >20

67 mitoses/10
HPF
Figure 2-242. Gastric WD-NET, grade 3. Tumor differentiation and tumor grade are independent
assessments and a well differentiated tumor can be high grade, even if that sounds
counterintuitive. Differentiation is dependent on cell morphology; this case is well differentiated
because the tumor cells are uniform. Grading, however, is based on mitotic activity and Ki-67
proliferation index. Any tumor with >20 mitoses in 10 HPF is G3, even if well differentiated. This
example shows three mitoses (arrowheads) within a single HPF.
Figure 2-243. Gastric WD-NET, grade 3, higher magnification of the previous figure. This
mitotically active (arrowhead ) tumor reaches the threshold for >20 mitoses in 10 HPF and is
considered high grade (G3). However, the tumor is well differentiated because the cells are
uniform, have smooth nuclear borders, and nucleoli are absent. Differentiation is morphologic,
whereas tumor grade is based on mitoses and Ki-67.
Figure 2-244. Digital image analysis for Ki-67 proliferation index. Digital image analysis software
can automate the counting of cells, reducing burden on pathologists. The top image is a
photomicrograph of Ki-67 immunohistochemistry in a WD-NET submitted to the software
program. The bottom image displays what the program identifies as tumor cells (marked blue)
and Ki-67 positive cells (marked gold) with an automated 1% count in this example.
Note of Caution
Performing Ki-67 labeling on type 1 WD-NETs can be misleading because a sizable
minority of them (up to about a third) have the proliferation of a G2 tumor but behave
like grade 1 lesions.126 If this testing is performed (some do not add it), the results may
benefit from a disclaimer.

FAQ: Can I eyeball the Ki-67 or does this require a manual count?

A minimum of 500 cells (suggested range 500 to 2000) is counted to

determine the Ki-67 index, which is reported as the percent of positive tumor
cells. Manual counting is time consuming, and a number of studies have
examined different techniques for Ki-67 index, including digitized automatic
counting and eyeballing. Digital image analysis software is not widely available
and requires software modification to prevent inaccurate counting, for example,
intratumoral or peritumoral lymphocytes (Fig. 2.244). Eyeballing in areas of
highest density “hot spots” is accurate in higher-grade lesions, but when tumors
are close to grade cutoffs, it is best to perform manual count. Then again,
anyone who has attempted to tally up 500 to 2000 cells under the microscope
knows how quickly one can lose track of which cells have been counted. A
simple and economical approach is to print out a photo or screenshot of a hot
spot area and manually cross off each cell as it is counted (Fig. 2.245). Take
care to exclude lymphocytes, which will skew the Ki-67 count higher.

PEARLS & PITFALLS: Grading by Mitotic Rate Requires Counting 50 HPF But
Is Reported as per 10 HPF
The mitoses in 50 HPF should be counted to accurately grade the tumor. This
number is divided by five to report mitoses per 10 HPF. These minimum
requirements for grading (50 HPF for mitotic count and 5000 to 2000 cells for
Ki-67 index) presume there is enough tissue for accurate grading, but small
biopsy material may be insufficient in many cases. See the following sample
note.

SAMPLE NOTE: When Small Biopsies Contain

Insufficient WD-NET for Mitotic Count or Ki-67
Index Grading
Stomach, Antrum, Polyp, Biopsy
• Well-differentiated neuroendocrine tumor, see Comment.

Comment
There is insufficient tumor cell quantity to accurately grade this WD-NET (minimum
requirement 50 HPF for mitotic count and 500 cells for Ki-67 proliferation index).
Based on the available material, this tumor appears to be (G1, G2, G3). Final grading
will be revised following review of a larger sample (e.g., excision specimen).

PEARLS & PITFALLS: Both Size and Depth of Invasion Are Considered in
Staging of Gastric WD-NETs
Although the depth of invasion defines most staging criteria, gastric WD-NETs
are among the few that also take into account tumor size at the lower stages.
For example, staging criteria by depth is fairly typical, with invasion into the
lamina propria and submucosa staged as pT1, but only if the tumor is ≤1 cm
Any tumor >1 cm is automatically pT2 or higher, even if superficially invasive.
Once tumors invade at least into the muscularis propria, the tumors are staged
by depth regardless of size: pT2, involvement of muscularis propria, pT3,
involvement of subserosa, pT4, invasion of serosa or adjacent tissue/organs.
 Figure 2-245. Manual count for Ki-67 proliferation index. A simple computer printout is quick and
economical if digital image analysis software is not available. Each cell can be marked off
during the 500–2000 cell count to avoid duplicate counts. To facilitate turnaround time, ancillary
staff can be trained in this method for reporting Ki-67 proliferation index.

MALT LYMPHOMA
GI lymphomas are challenging to recognize because the GI tract serves many
immunologic functions, and there is considerable histologic overlap between benign
inflammatory conditions and malignant lymphomas. In-depth coverage of lymphomas is
beyond the scope of this text and is left to our subspecialty hematopathology colleagues.
We encourage a low threshold to liberally share cases with such experts, but any
pathologist reviewing GI biopsies will be faced with gastric MALT extranodal marginal
zone lymphoma. The tools given in the following text are intended to ensure readers are
comfortable triaging these cases and are confident in recognizing features requiring
additional workup and consultation with hematopathology colleagues.
Gastric MALT lymphoma is driven by H. pylori infection, and eradication of the
organism is the first-line treatment of MALT lymphoma, resulting in remission in nearly
80% of cases.127 At low magnification, a robust and expansile deep chronic
inflammatory infiltrate is the first red flag to evaluate further for MALT lymphoma
(Figs. 2.246–2.250). At higher magnification, the infiltrate is typically composed of
monomorphic small lymphocytes with pericellular clearing or “halos” (Figs.
2.251–2.254). Features that serve as red flags to differentiate a malignant infiltrate from
benign gastritis include glandular destruction, in which lymphocytes (usually three or
more) invade the glandular epithelium and disrupt normal architecture (i.e.,
lymphoepithelial lesions) (Figs. 2.251–2.257) and the presence of dense lymphoid
infiltrates involving the muscularis mucosae (Figs. 2.246–2.248).128 These features
should trigger immunohistochemical workup, including, at minimum, CD3, CD20, and
CD43. This limited and economical panel can identify about half of MALT lymphomas,
which will show aberrant coexpression of CD43+ in the predominantly CD20+ B-cell
infiltrate (Fig. 2.258). The CD3 immunostain will provide a contrast by highlighting any
T cells, which normally express CD43. If this panel is insufficient to make a diagnosis,
expansion to a more comprehensive immunohistochemical panel will show the
following pattern in MALT lymphoma: CD20+, CD79a+, BCL2+, CD5−, CD10−,
cyclin D1−, CD23− CD43± (Figs. 2.259–2.264). This fundamental panel will also
differentiate other mature B-cell neoplasms, such as chronic lymphocytic
leukemia/small lymphocytic lymphoma (CD5+), follicular lymphoma (CD10+), and
mantle cell lymphoma (cyclin D1/BCL1+).

Figure 2-246. Gastric MALT lymphoma. Features that should trigger a MALT lymphoma workup
are seen here, including a deep monotonous lymphoid infiltrate that is gland destructive and
splaying out the muscularis mucosae.
Figure 2-247. Gastric MALT lymphoma. The infiltrate at low power is far more robust than
expected for a simple chronic gastritis. The lymphoid infiltrate is densely packed, gland
destructive, deep, and expanding the muscularis mucosae.
Figure 2-248. Gastric MALT lymphoma, higher magnification of the previous figure. The infiltrate
is deep, dense, and composed of a monotonous population of lymphocytes. The lymphoid cells
not only cross the muscularis mucosae but also spread the muscle bundles (arrow) apart.
Figure 2-249. Gastric MALT lymphoma. The infiltrate is more dense and monotonous than the
usual chronic gastritis. At low magnification, the expansile and gland-destructive quality should
trigger further workup for lymphoma.

Figure 2-250. Gastric MALT lymphoma, higher magnification of the previous figure. Gland
destruction is seen in the center of the field, as lymphocytes invade the glandular epithelium.
Figure 2-251. Gastric MALT lymphoma, higher magnification of the previous figure.
Lymphoepithelial lesions are characterized by lymphocytes (usually ≥3) invading the glandular
epithelium and disrupting the normal architecture (arrowhead ). Destroyed glands (arrow) leave
areas of drop-out which are filled in by the monotonous lymphocytes.
Figure 2-252. Lymphoepithelial lesion of gastric MALT lymphoma. The malignant lymphoid cells
invade the glandular epithelium and disrupt the normal architecture. A feature of the malignant
cells is the pericellular clearing or halo around each cell.
Figure 2-253. Gastric MALT lymphoma. These malignant lymphocytes are destroying areas of
glandular epithelium (arrowhead ) and muscularis mucosae (arrow). Features of glandular
destruction and muscularis mucosae abnormality are not seen in benign gastritis and should
prompt further workup for lymphoma.
Figure 2-254. Lymphoepithelial lesion of gastric MALT lymphoma. These lymphocytes have a
characteristic pericellular halo seen in MALT lymphoma cells. The presence of three or more
lymphocytes invading the glandular epithelium (arrow) is called a lymphoepithelial lesion (LEL).
Compared with benign intraepithelial lymphocytosis, which are T cells, these LELs are
composed of malignant B cells.
Figure 2-255. Gastric MALT lymphoma. At low magnification, this infiltrate differs from a benign
chronic gastritis because it is deep, dense, and monotonous with gland destruction.
Figure 2-256. Lymphoepithelial lesions in gastric MALT lymphoma, higher magnification of the
previous figure. Lymphoepithelial lesions are seen in various stages (arrows). On the far left,
the gland structure is still visible. The far right shows marked disruption of the glandular
architecture, but remnant epithelial cells are still visible. The center lesion is a nearly destroyed
gland and is barely visible.
Figure 2-257. Gastric MALT lymphoma, pancytokeratin stain. A pancytokeratin stain can be
helpful in highlighting residual glands and areas of lymphoepithelial lesions, which may be
obscured by the dense lymphocytic infiltrate on H&E.
Figure 2-258. Gastric MALT lymphoma, CD20+ with coexpression of CD43+. About half of
gastric MALT lymphomas can be identified by a limited immunopanel of CD3, CD20, and CD43.
These tumors show aberrant coexpression of CD43+ (normally found in T cells) in the
predominantly CD20+ B-cell infiltrate. The CD3 immunostain provides a contrast by highlighting
the T cells. Should this panel fail to solidify a diagnosis, an extended immunopanel can be
performed.
Figure 2-259. Gastric MALT lymphoma. This infiltrate is just a little too dense and too
monotonous to consider chronic gastritis. Some areas appear expansile, whereas others
appear gland destructive. In these instances, it is best to rule out lymphoma.
Figure 2-260. Gastric MALT lymphoma, CD43−, CD20+ with coexpression of BCL2. As noted
earlier, about half of MALT lymphomas do not express CD43. An extended panel of
immunostains will show reactivity for BCL2+ in the CD20-positive B cells.
Figure 2-261. Lymphoepithelial lesions in gastric MALT lymphoma. Lymphoepithelial lesions can
be found in various stages. Early lesions show ≥3 tumor lymphocytes invading the glandular
epithelium (arrow). More mature lesions show disruption of glandular architecture with
degenerating epithelial cells (arrowhead ).
Figure 2-262. Plasmacytoid variant of gastric MALT lymphoma. At low power, the architecture of
this gastric biopsy is abnormal. The lamina propria appears expanded and cellular, whereas the
glands are irregularly distributed. The pigment is incidental hemosiderin.
Figure 2-263. Plasmacytoid variant of gastric MALT lymphoma, higher magnification of the
previous figure. Do not be falsely reassured by the presence of a plasmacytic infiltrate in this
case. The plasma cells in the lamina propria are atypical with binucleate forms and marked
variation in size. A plasmacytic clone is found in 30% of MALT lymphomas. The pigment is
incidental hemosiderin.
 Figure 2-264. Plasmacytoid variant of gastric MALT lymphoma, higher magnification of the
previous figure. These subtle MALT lymphomas can be CD20 negative but should express
CD79a and show kappa or lambda restriction. This example shows CD20−, CD79a+, and
lambda restriction.

PEARLS & PITFALLS: Always Report H. pylori Status

H. pylori treatment and eradication is effective and results in remission in up to
80% of gastric MALT lymphomas. However, a subset of tumors contains
molecular genetic changes of t(11;18)(q21;q21) that have been associated with
the failure of MALT lymphoma to regress after H. pylori eradication therapy and
may arise in the absence of H. pylori infection.129 Reporting of H. pylori status
stratifies patients into prognostic groups and directs treatment (e.g., the
addition of radiotherapy or chemotherapy).

PEARLS & PITFALLS: Plasmacytoid MALT Lymphoma Variant as a Pitfall

Do not be falsely reassured by the presence of a plasmacytic infiltrate and
assume it represents H. pylori gastritis. Although H. pylori gastritis is
associated with a plasmacytic infiltrate compared with the typical lymphoid
infiltrate of MALT lymphoma, a plasmacytic clone is present in about 30% of
MALT lymphomas, and the plasmacytoid variant of MALT lymphoma can be
subtle and tricky (Figs. 2.262 and 2.263).130 Do not forget to always assess for
a deep, expansile, or destructive pattern, even if the cells appear predominantly
plasmacytic. Immunostains for CD20 can also be negative in the plasmacytoid
variant, another pitfall to diagnosis, but the variant will stain for CD79a and
show either kappa or lambda restriction (Figs. 2.264 and 2.265).

FAQ: How does one handle biopsies for posttreatment assessment of MALT
lymphoma?

Biopsies from treated patients largely show regression of lymphoma, with

sparse residual lymphoid cells that are insufficient for further workup, and yet
cannot be cleared as complete remission (Figs. 2.266 and 2.267). In these
cases, the first step is immunohistochemistry for H. pylori to ensure eradication.
Following this, should sufficient lymphoid cells be available, a limited
immunohistochemical panel mirroring the original immunoprofile of the tumor
cells (e.g., if the tumor was CD43+, then the limited panel of CD20/CD43/CD3
is sufficient) may be useful in confirming residual disease. However,
immunohistochemistry is not necessary, and a histological grading system for
posttreatment evaluation of gastric MALT lymphoma includes the following
categories131:
Complete remission: absent or scattered lymphoid/plasma cells
Probable Minimal Residual Disease (pMRD): aggregates of lymphoid
cells or lymphoid nodules with empty lamina propria
Responding Residual Disease: Dense, diffuse, nodular lymphoid infiltrate
with or without lymphoepithelial lesions, with partially empty lamina propria
No Change: Dense, diffuse, nodular lymphoid infiltrate with lymphoepithelial
lesions and no change in lamina propria
The differentiation of complete remission from pMRD is not always
histologically clear-cut, but this distinction is not too important because a
diagnosis of pMRD is not an indication for further treatment and clinicians
manage this as a state of remission.
SAMPLE NOTE: Probable Minimal Residual Disease
of MALT Lymphoma
Although the presence of patchy basal lymphoid aggregates is consistent with pMRD,
this finding is not necessarily an indication for further treatment and could be managed
as a state of remission with appropriate follow-up.

Reference:
Copie-Bergman C, Gaulard P, Lavergne-Slove A, et al. Proposal for a new histological
grading system for post-treatment evaluation of gastric MALT lymphoma. Gut.
2003;52(11):1656. PMID:14570741; PMCID:PMC1773845.

CHECKLIST: Features that Trigger MALT Lymphoma Workup or Consult

□ Predominantly lymphocytic infiltrate
□ Deep monotonous lymphoid infiltrates
□ Expansile lymphoid infiltrate
□ Lymphocytes with pericellular clearing or halos
□ Lymphoid infiltrate correlating with endoscopic nodule or mass
□ Destructive lymphoid infiltrates invading glands (“lymphoepithelial lesions”)
□ Dense lymphoid infiltrates traversing the muscularis mucosae
Figure 2-265. Lymphoepithelial lesion in plasmacytoid variant of gastric MALT lymphoma. These
malignant plasma cells are invading the glandular epithelium and destroying the normal
architecture.
Figure 2-266. Probable minimal residual disease. Some residual lymphoid cells are common
following treatment of MALT lymphoma. So long as they are not dense, diffuse, nodular, or
accompanied by lymphoepithelial lesions, these are considered probable minimal residual
disease. This designation is not an indication for further treatment but should be managed as a
state of remission.

Figure 2-267. Probable minimal residual disease. Sometimes residual lymphoid cells are
present in posttreatment biopsies yet are insufficient in size to immunophenotype. These can
simply be reported as probable minimal residual disease, which is managed clinically as a
state of remission.
MESENCHYMAL LESIONS
Mesenchymal lesions cover a broad spectrum of mesodermally derived tumors, which
are covered more completely in “Mesenchymal Tumors” chapter. Select mesenchymal
polyps common to the stomach are briefly covered herein, including the inflammatory
fibroid polyp (IFP), gastrointestinal stromal tumor (GIST), leiomyoma, and granular
cell tumor (GCT).

INFLAMMATORY FIBROID POLYP

IFP was first described in 1949 by Vanek as “gastric submucosal granulomas with
eosinophilic infiltration.”132 IFPs can occur in all ages but are most common in age 50
to 60 years and have a slightly higher incidence in women.121 They are rare lesions with
an estimated relative prevalence of 0.09%.133,134 IFPs present most often as a solitary
polyp or submucosal nodule in the gastric pylorus or distal antrum (Fig. 2.268) and are
typically small (<1.5 cm) and sessile.135 They are characterized by CD34
immunoreactive spindle and stellate stromal cells mixed with inflammatory cells
(predominantly eosinophils) in a myxoid or edematous stroma and thin-walled vessels
(Figs. 2.269–2.274). The spindle cells are sometimes seen swirling or forming an
“onion skin” pattern around vessels (Figs. 2.271–2.273). Although this was once
believed a reactive lesion, activating mutations have been identified in the platelet-
derived growth factor receptor alpha (PDGFRα) gene. This mutation is also found in a
subset of GISTs, typically the gastric benign epithelioid variant that does not have a KIT
mutation. IFPs are now viewed as PDGFRα-driven benign neoplasms.135 These tumors
rarely cause clinical symptoms; however, a few cases of large gastric IFPs causing
gastric outlet obstruction have been reported.136 They are believed to have no malignant
potential, thus no endoscopic follow-up is recommended after initial histologic
confirmation, unless symptomatic, in which case complete resection is recommended.137

KEY FEATURES: IFP

• PDGFRα-mutation-driven benign neoplasms
• Mean age 50 to 60 years with female predominance
• Solitary, sessile submucosal nodule in the antrum/pylorus
• Stellate and spindle cells mixed with inflammatory cells, predominantly eosinophils
• Myxoid or edematous stroma with thin-walled vessels
• Spindle cells are sometimes seen swirling or forming an onion skin pattern around
vessels
• CD34+ immunohistochemistry highlights spindle/stellate cells
• Excision required only if symptomatic

Figure 2-268. Inflammatory fibroid polyp. The most common location for IFP is the gastric
antrum, as seen here. The epicenter of these lesions is submucosal, and they appear
endoscopically as a nodule.
Figure 2-269. Inflammatory fibroid polyp, CD34 immunohistochemistry of the previous figure. A
CD34 highlights the scope of this benign lesion, which is surprisingly more extensive than
appreciated on H&E.
Figure 2-270. Inflammatory fibroid polyp. The spindle cells extend from the submucosa and
percolate through the lamina propria toward the surface. However, the findings are subtle and
one can appreciate how a superficial biopsy might be challenging to interpret. Diagnostic clues
include unexplained bland spindle cells and the presence of eosinophils.
Figure 2-271. Inflammatory fibroid polyp. The spindle cells may swirl concentrically around
vessels in an onion skin pattern, as seen here. As the spindle cells extend upward, they
traverse the muscularis mucosae and splay the muscle fibers.
Figure 2-272. Inflammatory fibroid polyp, CD34 immunohistochemistry of the previous figure.
CD34 highlights the spindle cells of the IFP, which are more abundant than appreciated by
H&E.
Figure 2-273. Inflammatory fibroid polyp. Higher magnification shows the bland spindle cells
forming a concentric pattern around the artery. Intimately admixed are frequent eosinophils, an
extremely helpful diagnostic clue to this entity.
 Figure 2-274. Inflammatory fibroid polyp. Some IFPs appear more edematous or myxoid. In
these areas, the spindle cells (arrows) may be sparse. Often the first clue to diagnosis is the
eye-catching eosinophils.

GASTROINTESTINAL STROMAL TUMOR

GISTs are rare mesenchymal tumors derived from the interstitial cells of Cajal (the
pacemaker cells of the GI tract), which reside between the inner circular and outer
longitudinal layers of the muscularis propria (Figs. 2.275–2.276). GISTs account for
1% to 3% of all malignant GI tumors,138 or about 5,000 new cases per year in the
United States.139,140 Although these tumors may arise anywhere along the luminal GI
tract, the most common site is the stomach, and they are often found incidentally during
upper endoscopy for indications unrelated to the tumor.141 Most GISTs contain a
mutation of the protooncogenes KIT (75%) or PDGFRA (5%) with known positivity for
CD117 (c-KIT) in 95% of the tumors.142 GISTs lacking cytoplasmic immunoreactivity
for CD117 show reactivity for DOG-1 immunohistochemistry (overall 97%), an equally
sensitive and specific marker.140,143,144 The majority of GISTs exhibit stereotypical
features of monotonous bland spindled or epithelioid cells with pale eosinophilic
cytoplasm and oval nuclei with vesicular chromatin (Figs. 2.277–2.280). The most
reliable prognostic factors are site, size of primary tumor, and mitotic index.
Endoscopic ultrasound and CT scans are important to determine local and metastatic
spread.145-147 If the tumor is metastatic or unresectable, imatinib (a tyrosine-kinase
inhibitor) is the first-line chemotherapeutic agent of choice in tumors expressing c-kit
mutation.148-150 Be attentive to any epithelioid GISTs with multinodular or plexiform
growth, or lymphovascular invasion, as these features are red flags for the imatinib-
resistant GISTs found in Carney-Stratakis syndrome, an autosomal dominant familial
syndrome characterized by paraganglioma and GIST with germline mutations in
succinate dehydrogenase genes SDHB, SDHC, or SDHD.151 These GISTs are found
almost exclusively in the stomach, have absent SDHB immunohistochemistry, and have
higher risk of metastatic disease irrespective of the usual GIST prognostic
predictors.151-153 Other SDHB-negative GISTs can be found in Carney triad and
sporadic pediatric SDHB-deficient tumors.

KEY FEATURES: GIST

• Arise from interstitial cells of Cajal found in the muscularis propria
• Stomach is the most common site, although can arise anywhere along GI tract
• Small fascicles of monotonous pale eosinophilic cells, spindled (70%), epithelioid
(20%), or mixed (10%)
• Immunohistochemistry: CD117+ (95%), DOG-1+ (97%), CD-34+ (60%)
• Prognosis of most GISTs rely on site, size, and mitotic activity
• Most GISTs have mutation-specific response to tyrosine kinase inhibitors, such as
imatinib
• Red flags: Multinodular or plexiform growth, lymphovascular invasion indicate
SDH-deficient GISTs, a feature of Carney-Stratakis syndrome
• SDH-deficient GISTs are found almost exclusively in the stomach, are more likely to
metastasize, and are resistant to imatinib therapy

LEIOMYOMA
Leiomyomas are benign smooth muscle tumors, typically asymptomatic and found
incidentally.154-156 Endoscopically they appear as rounded submucosal lesions with
intact overlying mucosa (Fig. 2.281) and range in size from 0.5 to 20 cm.157 Both
leiomyomas and GISTs can grow inwardly and outwardly to form a dumbbell shape,
although leiomyomas are more likely to grow intraluminally (vs. GIST, which expands
predominantly in an extramural fashion). Histologically, the tumor is composed of
intersecting bundles of smooth muscle without atypia, frequent mitotic activity, or
necrosis (Figs. 2.282–2.284). The tumor can be differentiated from GIST, which is
CD117 or DOG-1 positive, whereas leiomyoma is smooth muscle actin and desmin
positive and negative for CD117/DOG-1.
Figure 2-275. Gastrointestinal stromal tumor (GIST). These lesions derive from the interstitial
cells of Cajal and arise almost exclusively from the myenteric (Auerbach) plexus, which is
located between the inner circular and outer longitudinal layers of the muscularis propria
(pictured). For this reason, a spindle cell lesion arising from the muscularis mucosae cannot be
a GIST.
Figure 2-276. Gastrointestinal stromal tumor (GIST), CD117 immunohistochemistry of the
previous figure. CD117 immunoreactivity confirms the diagnosis in 95% of cases. CD117-
negative GISTs can be stained for DOG1.
Figure 2-277. GIST, epithelioid type. The cells are round and fairly uniform. There is no
prognostic significance to the morphologic variant.
Figure 2-278. GIST, spindled type. These cells are elongated with cigar-shaped nuclei.
Figure 2-279. GIST. These tumors can take on many morphologic variations and are wonderful
mimickers of other tumors. This region shows features similar to Verocay bodies found in
schwannomas.
Figure 2-280. GIST. Perpendicular fascicles of spindle cells raise the differential for leiomyoma.
GISTs are excellent mimickers of other tumors.
Figure 2-281. Leiomyoma. This spindle cell neoplasm arises from the muscularis mucosae.
Because of this location, the diagnosis cannot be GIST.
Figure 2-282. Leiomyoma, higher magnification of the previous figure. The tumor arises from
the muscularis mucosae, and the cells are bland and spindled.
Figure 2-283. Leiomyoma, smooth muscle actin (SMA) immunohistochemistry. SMA confirms
the smooth muscle differentiation of this tumor.
 Figure 2-284. Leiomyoma. The smooth muscle bundles characteristically intersect at
perpendicular angles.

GRANULAR CELL TUMOR

For unknown reasons, about half of gastric GCTs occur synchronously with esophageal
GCTs.158 These lesions tend to occur equally in both sexes, most frequently in the fourth
to sixth decade of life.158 Endoscopically, they are found incidentally in the proximal
stomach, range from a few millimeters in size up to 7 cm, and appear as a yellow
subepithelial mass or nodule that resembles a lipoma.158 Histologically, the submucosal
tumor is composed of sheets of uniform polygonal tumor cells with abundant
eosinophilic granular cytoplasm and small hyperchromatic nuclei (Figs. 2.285 and
2.286). The cytoplasmic accumulation of secondary lysosomes is PAS positive and
diastase resistant, whereas immunohistochemistry shows the tumors are positive for
S100 and NSE and negative for HMB45, keratins, and desmin. The vast majority of
GCTs are benign, but malignant and metastatic cases have been reported.159

KEY FEATURES: GCT

• About 50% of gastric GCTs will have accompanying esophageal GCT
• Mean age 50 years, M = F
• Submucosal epicenter, with sheets of polygonal cells containing abundant granular
eosinophilic cytoplasm
• Cytoplasmic accumulation of PAS-D+ secondary lysosomes
• Immunohistochemistry positive: S100+, NSE+
• Immunohistochemistry negative: HMB45−, cytokeratin−, desmin−

Figure 2-285. Granular cell tumor. The tumor cells have indistinct cell borders and slightly
atypical angulated nuclei. There is abundant eosinophilic and granular cytoplasm.
 Figure 2-286. Granular cell tumor. Oil immersion shows the granular quality of the cytoplasm,
which by electron microscopy is filled with lysosomes (not pictured).

NEAR MISS
METASTATIC LOBULAR BREAST CARCINOMA
New diagnoses of diffuse-type gastric cancer in women should include at least one
immunohistochemical marker to exclude metastatic lobular breast carcinoma, such as
GATA3. Although breast cancer metastasis to the GI tract is a rare occurrence, the
stomach is the most common location aside from the liver, and the discohesive
infiltrating cells of lobular breast carcinoma can be easily mistaken for a primary
diffuse-type gastric cancer (Figs. 2.287–2.294). Differentiation of these tumors relies
almost entirely upon immunohistochemical confirmation, an important step because the
treatment of these tumors diverge. A diagnosis of primary gastric cancer leads to
surgical management, whereas metastatic breast cancer may benefit from
chemotherapeutic options depending on the hormone receptor status. Histologically,
metastatic lobular breast cancer cells infiltrate the lamina propria individually or in
single-file cords (Fig. 2.288). These uniform small discohesive cells characteristically
lack E-cadherin (Fig. 2.289), a surface cohesion molecule. Because diffuse-type gastric
cancer lacks a precursor lesion, this metastatic pattern is nearly indistinguishable from a
primary gastric tumor. A PAS stain is negative for cytoplasmic PAS staining or
intracytoplasmic mucin (Fig. 2.290), whereas most gastric signet ring cell carcinomas
will show PAS staining. One helpful clue to metastatic lobular breast carcinoma is the
presence of intracytoplasmic lumina, which have a sharply demarcated edge and may
contain a hyaline globule imparting a targetoid appearance (Fig. 2.294).

PEARLS & PITFALLS: Metastatic Lobular Breast Carcinoma

• Metastatic lobular breast carcinoma and primary diffuse-type gastric

cancer can be indistinguishable by H&E
• Both lesions are unassociated with background gastritis or precursor
lesion
• Make it a habit to perform a GATA3 on all new diagnoses of diffuse-type
gastric adenocarcinoma; GATA3 should be negative in primary gastric
cancer
• Clear intracytoplasmic lumina with hyaline globules imparting a targetoid
appearance is a clue to lobular breast cancer
Figure 2-287. Metastatic lobular breast carcinoma. Single infiltrating hyperchromatic cells
invade through the lamina propria. At low magnification, one might consider inactive chronic
gastritis or diffuse-type gastric adenocarcinoma.
Figure 2-288. Metastatic lobular breast carcinoma, higher magnification of the previous figure.
Single-file cords (arrows) are highly characteristic of invasive lobular carcinoma of the breast.
Figure 2-289. Metastatic lobular breast carcinoma, E-cadherin immunostain. Lobular
carcinoma of the breast has a characteristic loss of cell adhesion molecule E-cadherin. This
feature can also be seen in primary gastric carcinomas of diffuse type and therefore should not
be considered a reliable marker for breast origin.
Figure 2-290. Metastatic lobular breast carcinoma, PAS stain. The tumor cells of lobular breast
carcinoma do not show cytoplasmic reactivity for PAS or contain mucin.
Figure 2-291. Metastatic lobular breast carcinoma, GATA3 immunostain. GATA3 is a reliable
marker that highlights tumors of breast origin and is negative in gastric cancer.
Figure 2-292. Metastatic lobular breast carcinoma. This example is extremely subtle,
underscoring the importance of always reviewing gastric biopsies at high magnification and
accounting for each cell.
Figure 2-293. Metastatic lobular breast carcinoma, higher magnification of the previous figure.
The metastatic cells are uniform and bland with slightly eosinophilic cytoplasm that blends in
with the benign oxyntic glands. In all gastric biopsies, take a moment to point out individual cells
and try to place a name to them, rather than grouping them together in one’s mind.
 Figure 2-294. Metastatic lobular breast carcinoma. Although at first glance the tumor cells
resemble signet ring cells, note the intracytoplasmic lumina with distinct borders and the single
hyaline inclusion. These targetoid cells (arrow) are characteristic of lobular breast carcinoma.

GASTRIC XANTHOMA
These subepithelial aggregates of histiocytes are submitted as endoscopic nodules,
polyps, or plaques and cause no diagnostic difficulty for pathologists when encountered
in the gallbladder (i.e., cholesterolosis) but can prove tricky when seen in the stomach
(Figs. 2.295–2.302). The most common concern among extramural consultations is
exclusion of diffuse-type gastric cancer (Figs. 2.295–2.297). At low magnification, an
area of pallor is eye-catching as the foamy histiocytes expand the lamina propria (Fig.
2.298). Collections of bland macrophages with abundant foamy cytoplasm are usually
subepithelial but can be found anywhere within the tissue (Fig. 2.299 and 2.301). Small
and crushed biopsies provide the most challenging material, but application of CD68
immunohistochemistry is almost always helpful (Figs. 2.297 and 2.302). PAS stain is
negative for intracytoplasmic mucin (Fig. 2.300).
Figure 2-295. Gastric xanthoma in a crushed and suboptimal biopsy. By H&E, several cells in
this biopsy are hard to name (arrows). They are subepithelial in the lamina propria and contain
clear cytoplasm. Small and crushed biopsies are always difficult to interpret, and diffuse-type
gastric cancer raises the stakes even further. Do not hesitate to request repeat biopsy if the
diagnosis is unclear.
Figure 2-296. Gastric xanthoma, pancytokeratin immunostain of the previous figure. A
pancytokeratin stain can highlight the gastric foveolar and glandular architecture and provide
reassurance that the cells are not invasive carcinoma cells.
Figure 2-297. Gastric xanthoma, CD68 immunostain. A CD68 immunostain in this example
highlights the scant crushed cells, confirming they are foamy macrophages and not signet ring
cells.
Figure 2-298. Gastric xanthoma. Gastric xanthoma cells can mimic a diffuse-type gastric
cancer at low magnification. All areas of increased cellularity or pallor (arrow) should be
reviewed more closely to exclude a sneaky diffuse-type gastric cancer.
Figure 2-299. Gastric xanthoma, higher magnification of the previous figure. Gastric xanthomas
are composed of collections of foamy macrophages. At high magnification, they have bland
uniform nuclei and abundant foamy cytoplasm that makes them, in most cases, easy to
distinguish from diffuse gastric cancer by H&E alone.
Figure 2-300. Gastric xanthoma, PAS stain. The cytoplasm of foamy macrophages is not PAS
positive and does not contain mucin.
Figure 2-301. Gastric xanthoma. This example is more challenging, as the foamy macrophages
(arrow) are embedded within the muscularis mucosa and raise concern for a sneaky invasive
diffuse-type gastric cancer.
 Figure 2-302. Gastric xanthoma, CD68 immunostain of the previous figure.
Immunohistochemistry confirms histiocytic differentiation and reassures that these are benign
cells.

GASTRITIS CYSTICA PROFUNDA

Do not mistake this rare benign lesion for adenocarcinoma. Gastritis cystica profunda
(GCP) is characterized by downgrowth of cystically dilated gastric glands through the
muscularis mucosae into the submucosa (Figs. 2.303–2.308), with or without polypoid
hyperplasia (gastritis cystica polyposa), and typically has a backdrop of chronic
inflammation.160 The proliferation of these benign displaced glands should not be
interpreted as malignant. Some have postulated that inflammation causes mucosal
erosion and migration of epithelial cells into the submucosa, followed by cystic
dilation. Mucosal prolapse and herniation of glands into the submucosa is thought to be
the pathogenesis in cases of prior instrumentation (65%), but this lesion can also be
found in unoperated stomachs.161-163 These lesions are associated with older age, male
gender, and proximal location, and rare examples are found in combination with true
gastric cancer (0.7% of gastric cancers).164 A helpful history of prior surgery and prior
biopsy can be key to avoiding overdiagnosis, and histologically these submucosal
glands are lined by bland nonneoplastic epithelial cells (Figs. 2.304, 2.306, and 2.308).
Figure 2-303. Gastritis cystica profunda. Herniating into the submucosa are collections of
gastric glands (arrow). At scanning magnification, these submucosal glands appear
disorganized and raise concern for an invasive adenocarcinoma.
Figure 2-304. Gastritis cystica profunda, higher magnification of the previous figure. Upon
closer review, these submucosal glands are bland and lined by benign epithelial cells.
Figure 2-305. Gastritis cystica profunda. An endoscopic mucosal resection shows a gastric
adenoma with surface epithelial hyperchromasia. Deep to this are glands that extend into the
submucosa (arrow). The association with overlying dysplasia certainly raises concern for an
invasive adenocarcinoma.
Figure 2-306. Gastritis cystica profunda, higher magnification of the previous figure. A closer
examination of the submucosal glands shows they are lined by bland, benign epithelial cells
and are cystically dilated. A dive into the electronic medical record reveals that this lesion was
previously biopsied. The submucosal glands are likely herniated owing to prior instrumentation,
a characteristic history for GCP.
Figure 2-307. Gastritis cystica profunda. Cystically dilated glands are present in the submucosa
of this gastric biopsy (arrow). A background of inactive chronic gastritis is present, which is
characteristic for GCP. Even at this magnification, the glands appear benign and misplaced.
 Figure 2-308. Gastritis cystica profunda, higher magnification of the previous figure. The
epithelial cells lining the herniated glands are uniform and show no pleomorphism, stratification,
or atypia.

INACTIVE CHRONIC GASTRITIS HIDES TUMORS

A theme throughout this chapter has been the emphasis on not diagnosing a chronic
gastritis and moving on too quickly. Lymphoid aggregates and dense collections of
chronic inflammatory cells can appear hyperchromatic and busy, masking other
important findings. To avoid overlooking sneaky tumor cells, always take a moment to
review gastric biopsies at higher magnification and make sure each cell in the biopsy is
acknowledged. Avoid mentally grouping cells together as inactive chronic gastritis or
lymphoid aggregates; instead, pause and put a name to each cell. Once this becomes a
habit, it happens automatically and swiftly. This attention to each cell prevents
overlooking sneaky signet ring cells, gastric MALT lymphomas, metastatic lobular
breast carcinomas, and other subtle tumors, as well as nonneoplastic findings, such as
viral cytopathic effect (Figs. 2.309 and 2.310).
Figure 2-309. Tumor hiding in a lymphoid aggregate. Pattern recognition is key in pathology, and
pathologists quickly group together like findings, such as clusters of lymphocytes in inactive
chronic gastritis. However, always take a moment to look at higher magnification and
acknowledge each cell in case there is tumor hiding there (arrow).
 Figure 2-310. Tumor hiding in a lymphoid aggregate, higher magnification of the previous figure.
These intralymphatic tumor cells were nearly obscured by the surrounding lymphoid infiltrate
but are clearly visible at high magnification.

References
1. Carmack SW, Genta RM, Schuler CM, Saboorian MH. The current spectrum of
gastric polyps: a 1-year national study of over 120,000 patients. Am J
Gastroenterol. 2009;104(6):1524-1532.
2. Jhala NC, Montemor M, Jhala D, et al. Pancreatic acinar cell metaplasia in
autoimmune gastritis. Arch Pathol Lab Med. 2003;127(7):854-857.
3. Lam-Himlin D, Park JY, Cornish TC, Shi C, Montgomery E. Morphologic
characterization of syndromic gastric polyps. Am J Surg Pathol.
2010;34(11):1656-1662.
4. Beggs AD, Latchford AR, Vasen HF, et al. Peutz-Jeghers syndrome: a systematic
review and recommendations for management. Gut. 2010;59(7):975-986.
5. Hemminki A, Markie D, Tomlinson I, et al. A serine/threonine kinase gene
defective in Peutz-Jeghers syndrome. Nature. 1998;391(6663):184-187.
6. Volikos E, Robinson J, Aittomaki K, et al. LKB1 exonic and whole gene deletions
are a common cause of Peutz-Jeghers syndrome. J Med Genet. 2006;43(5):e18.
 7. Giardiello FM, Welsh SB, Hamilton SR, et al. Increased risk of cancer in the
Peutz-Jeghers syndrome. N Engl J Med. 1987;316(24):1511-1514.
8. Hearle N, Schumacher V, Menko FH, et al. Frequency and spectrum of cancers in
the Peutz-Jeghers syndrome. Clin Cancer Res. 2006;12(10):3209-3215.
9. Jeghers H, Mc KV, Katz KH. Generalized intestinal polyposis and melanin spots of
the oral mucosa, lips and digits; a syndrome of diagnostic significance. N Engl J
Med. 1949;241(26):1031-1036.
10. Offerhaus GJA BM, Gruber SB. Peutz-Jeghers syndrome. In: Bosman FT, Carneiro
F, Hruban RH, et al, eds. Tumors of the Digestive System. Lyon, France: IACR;
2010:160-170.
11. Latchford AR, Phillips RK. Gastrointestinal polyps and cancer in Peutz-Jeghers
syndrome: clinical aspects. Fam Cancer. 2011;10(3):455-461.
12. Tse JY, Wu S, Shinagare SA, et al. Peutz-Jeghers syndrome: a critical look at
colonic Peutz-Jeghers polyps. Mod Pathol. 2013;26(9):1235-1240.
13. Giardiello FM, Brensinger JD, Tersmette AC, et al. Very high risk of cancer in
familial Peutz-Jeghers syndrome. Gastroenterology. 2000;119(6):1447-1453.
14. Burkart AL, Sheridan T, Lewin M, Fenton H, Ali NJ, Montgomery E. Do sporadic
Peutz-Jeghers polyps exist? Experience of a large teaching hospital. Am J Surg
Pathol. 2007;31(8):1209-1214.
15. Burt RW, Bishop DT, Lynch HT, Rozen P, Winawer SJ. Risk and surveillance of
individuals with heritable factors for colorectal cancer. WHO Collaborating
Centre for the Prevention of Colorectal Cancer. Bull World Health Organ.
1990;68(5):655-665.
16. Howe JR, Roth S, Ringold JC, et al. Mutations in the SMAD4/DPC4 gene in
juvenile polyposis. Science. 1998;280(5366):1086-1088.
17. Sweet K, Willis J, Zhou XP, et al. Molecular classification of patients with
unexplained hamartomatous and hyperplastic polyposis. JAMA.
2005;294(19):2465-2473.
18. Howe JR, Mitros FA, Summers RW. The risk of gastrointestinal carcinoma in
familial juvenile polyposis. Ann Surg Oncol. 1998;5(8):751-756.
19. Giardiello FM, Hamilton SR, Kern SE, et al. Colorectal neoplasia in juvenile
polyposis or juvenile polyps. Arch Dis Child. 1991;66(8):971-975.
20. Offerhaus GJA HJ. Juvenile polyposis. In: Bosman FT, Carneiro F, Hruban RH, et
al, eds. WHO Classification of Tumors of the Digestive System. Lyon, France:
IARC; 2010:166-167.
21. Cronkhite LW Jr, Canada WJ. Generalized gastrointestinal polyposis; an unusual
 syndrome of polyposis, pigmentation, alopecia and onychotrophia. N Engl J Med.
1955;252(24):1011-1015.
22. Nakamura M, Kobashikawa K, Tamura J, et al. Cronkhite-Canada syndrome. Intern
Med. 2009;48(17):1561-1562.
23. Daniel ES, Ludwig SL, Lewin KJ, Ruprecht RM, Rajacich GM, Schwabe AD. The
Cronkhite-Canada syndrome. An analysis of clinical and pathologic features and
therapy in 55 patients. Medicine (Baltim). 1982;61(5):293-309.
24. Heald B, Mester J, Rybicki L, Orloff MS, Burke CA, Eng C. Frequent
gastrointestinal polyps and colorectal adenocarcinomas in a prospective series of
PTEN mutation carriers. Gastroenterology. 2010;139(6):1927-1933.
25. Kato M, Mizuki A, Hayashi T, et al. Cowden’s disease diagnosed through
mucocutaneous lesions and gastrointestinal polyposis with recurrent hematochezia,
unrevealed by initial diagnosis. Intern Med. 2000;39(7):559-563.
26. Levi Z, Baris HN, Kedar I, et al. Upper and lower gastrointestinal findings in
PTEN mutation-positive Cowden syndrome patients participating in an active
surveillance program. Clin Transl Gastroenterol. 2011;2:e5.
27. Tan MH, Mester JL, Ngeow J, Rybicki LA, Orloff MS, Eng C. Lifetime cancer
risks in individuals with germline PTEN mutations. Clin Canc Res.
2012;18(2):400-407.
28. Abraham SC, Singh VK, Yardley JH, Wu TT. Hyperplastic polyps of the stomach:
associations with histologic patterns of gastritis and gastric atrophy. Am J Surg
Pathol. 2001;25(4):500-507.
29. Koch HK, Lesch R, Cremer M, Oehlert W. Polyps and polypoid foveolar
hyperplasia in gastric biopsy specimens and their precancerous prevalence. Front
Gastrointest Res. 1979;4:183-191.
30. Koga S, Watanabe H, Enjoji M. Stomal polypoid hypertrophic gastritis: a polypoid
gastric lesion at gastroenterostomy site. Cancer. 1979;43(2):647-657.
31. Zea-Iriarte WL, Sekine I, Itsuno M, et al. Carcinoma in gastric hyperplastic polyps.
A phenotypic study. Dig Dis Sci. 1996;41(2):377-386.
32. Ginsberg GG, Al-Kawas FH, Fleischer DE, Reilly HF, Benjamin SB. Gastric
polyps: relationship of size and histology to cancer risk. Am J Gastroenterol.
1996;91(4):714-717.
33. Orlowska J, Jarosz D, Pachlewski J, Butruk E. Malignant transformation of benign
epithelial gastric polyps. Am J Gastroenterol. 1995;90(12):2152-2159.
34. Genta RM, Schuler CM, Robiou CI, Lash RH. No association between gastric
fundic gland polyps and gastrointestinal neoplasia in a study of over 100,000
 patients. Clin Gastroenterol Hepatol. 2009;7(8):849-854.
35. Zelter A, Fernandez JL, Bilder C, et al. Fundic gland polyps and association with
proton pump inhibitor intake: a prospective study in 1,780 endoscopies. Dig Dis
Sci. 2011;56(6):1743-1748.
36. Vieth M, Stolte M. Fundic gland polyps are not induced by proton pump inhibitor
therapy. Am J Clin Pathol. 2001;116(5):716-720.
37. Bianchi LK, Burke CA, Bennett AE, Lopez R, Hasson H, Church JM. Fundic gland
polyp dysplasia is common in familial adenomatous polyposis. Clin Gastroenterol
Hepatol. 2008;6(2):180-185.
38. Garrean S, Hering J, Saied A, Jani J, Espat NJ. Gastric adenocarcinoma arising
from fundic gland polyps in a patient with familial adenomatous polyposis
syndrome. Am Surg. 2008;74(1):79-83.
39. Zwick A, Munir M, Ryan CK, et al. Gastric adenocarcinoma and dysplasia in
fundic gland polyps of a patient with attenuated adenomatous polyposis coli.
Gastroenterology. 1997;113(2):659-663.
40. Wu TT, Kornacki S, Rashid A, Yardley JH, Hamilton SR. Dysplasia and
dysregulation of proliferation in foveolar and surface epithelia of fundic gland
polyps from patients with familial adenomatous polyposis. Am J Surg Pathol.
1998;22(3):293-298.
41. Offerhaus GJ, Giardiello FM, Krush AJ, et al. The risk of upper gastrointestinal
cancer in familial adenomatous polyposis. Gastroenterology. 1992;102(6):1980-
1982.
42. Torbenson M, Lee JH, Cruz-Correa M, et al. Sporadic fundic gland polyposis: a
clinical, histological, and molecular analysis. Mod Pathol. 2002;15(7):718-723.
43. Li J, Woods SL, Healey S, et al. Point mutations in exon 1B of APC reveal gastric
adenocarcinoma and proximal polyposis of the stomach as a familial adenomatous
polyposis variant. Am J Hum Genet. 2016;98(5):830-842.
44. Abraham SC, Nobukawa B, Giardiello FM, Hamilton SR, Wu TT. Sporadic fundic
gland polyps: common gastric polyps arising through activating mutations in the
beta-catenin gene. Am J Pathol. 2001;158(3):1005-1010.
45. Lauwers GY, Shimizu M, Correa P, et al. Evaluation of gastric biopsies for
neoplasia: differences between Japanese and Western pathologists. Am J Surg
Pathol. 1999;23(5):511-518.
46. Schlemper RJ, Riddell RH, Kato Y, et al. The Vienna classification of
gastrointestinal epithelial neoplasia. Gut. 2000;47(2):251-255.
47. Singhi AD, Lazenby AJ, Montgomery EA. Gastric adenocarcinoma with chief cell
 differentiation: a proposal for reclassification as oxyntic gland polyp/adenoma.
Am J Surg Pathol. 2012;36(7):1030-1035.
48. Vieth M, Kushima R, Borchard F, Stolte M. Pyloric gland adenoma: a clinico-
pathological analysis of 90 cases. Virchows Arch. 2003;442(4):317-321.
49. Abraham SC, Montgomery EA, Singh VK, Yardley JH, Wu TT. Gastric adenomas:
intestinal-type and gastric-type adenomas differ in the risk of adenocarcinoma and
presence of background mucosal pathology. Am J Surg Pathol. 2002;26(10):1276-
1285.
50. Chen ZM, Scudiere JR, Abraham SC, Montgomery E. Pyloric gland adenoma: an
entity distinct from gastric foveolar type adenoma. Am J Surg Pathol.
2009;33(2):186-193.
51. Abraham SC, Park SJ, Lee JH, Mugartegui L, Wu TT. Genetic alterations in gastric
adenomas of intestinal and foveolar phenotypes. Mod Pathol. 2003;16(8):786-795.
52. Spigelman AD, Williams CB, Talbot IC, Domizio P, Phillips RK. Upper
gastrointestinal cancer in patients with familial adenomatous polyposis. Lancet.
1989;2(8666):783-785.
53. Bakotic BW, Robinson MJ, Sturm PD, Hruban RH, Offerhaus GJ, Albores-
Saavedra J. Pyloric gland adenoma of the main pancreatic duct. Am J Surg Pathol.
1999;23(2):227-231.
54. Hackeng WM, Montgomery EA, Giardiello FM, et al. Morphology and genetics of
pyloric gland adenomas in familial adenomatous polyposis. Histopathology.
2017;70(4):549-557.
55. Hashimoto T, Ogawa R, Matsubara A, et al. Familial adenomatous polyposis-
associated and sporadic pyloric gland adenomas of the upper gastrointestinal tract
share common genetic features. Histopathology. 2015;67(5):689-698.
56. Choi WT, Brown I, Ushiku T, et al. Gastric pyloric gland adenoma: a multicentre
clinicopathological study of 67 cases. Histopathology. 2018;72(6):1007-1014.
57. Wood LD, Salaria SN, Cruise MW, Giardiello FM, Montgomery EA. Upper GI
tract lesions in familial adenomatous polyposis (FAP): enrichment of pyloric gland
adenomas and other gastric and duodenal neoplasms. Am J Surg Pathol.
2014;38(3):389-393.
58. Ueyama H, Yao T, Nakashima Y, et al. Gastric adenocarcinoma of fundic gland
type (chief cell predominant type): proposal for a new entity of gastric
adenocarcinoma. Am J Surg Pathol. 2010;34(5):609-619.
59. Matsukawa A, Kurano R, Takemoto T, Kagayama M, Ito T. Chief cell hyperplasia
with structural and nuclear atypia: a variant of fundic gland polyp. Pathol Res
 Pract. 2005;200(11-12):817-821.
60. Muller-Hocker J, Rellecke P. Chief cell proliferation of the gastric mucosa
mimicking early gastric cancer: an unusual variant of fundic gland polyp. Virchows
Arch. 2003;442(5):496-500.
61. Kushima R, Sekine S, Matsubara A, Taniguchi H, Ikegami M, Tsuda H. Gastric
adenocarcinoma of the fundic gland type shares common genetic and phenotypic
features with pyloric gland adenoma. Pathol Int. 2013;63(6):318-325.
62. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J
Clin. 2005;55(2):74-108.
63. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer
statistics. CA Cancer J Clin. 2011;61(2):69-90.
64. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin.
2017;67(1):7-30.
65. Ahn HS, Lee HJ, Yoo MW, et al. Changes in clinicopathological features and
survival after gastrectomy for gastric cancer over a 20-year period. Br J Surg.
2011;98(2):255-260.
66. Brown LM, Devesa SS. Epidemiologic trends in esophageal and gastric cancer in
the United States. Surg Oncol Clin N Am. 2002;11(2):235-256.
67. Amin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging
Manual: continuing to build a bridge from a population-based to a more
“personalized” approach to cancer staging. CA Cancer J Clin. 2017;67(2):93-99.
68. Park JC, Lee YK, Kim SY, et al. Long-term outcomes of endoscopic submucosal
dissection in comparison to surgery in undifferentiated-type intramucosal gastric
cancer using propensity score analysis. Surg Endosc. 2018;32(4):2046-2057.
69. Lauren P. The two histological main types of gastric carcinoma: diffuse and so-
called intestinal-type carcinoma. An attempt at a histo-clinical classification. Acta
Pathol Microbiol Scand. 1965;64:31-49.
70. Cimerman M, Repse S, Jelenc F, Omejc M, Bitenc M, Lamovec J. Comparison of
Lauren’s, Ming’s and WHO histological classifications of gastric cancer as a
prognostic factor for operated patients. Int Surg. 1994;79(1):27-32.
71. Bosman FT, Carneiro F, Hruban RH, et al., ed WHO Classification of Tumours of
the Digestive System. 4th ed. Lyon: WHO Press; 2010.
72. Talamonti MS, Kim SP, Yao KA, et al. Surgical outcomes of patients with gastric
carcinoma: the importance of primary tumor location and microvessel invasion.
Surgery. 2003;134(4):720-727; discussion 727-729.
73. Luebke T, Baldus SE, Grass G, et al. Histological grading in gastric cancer by
 Ming classification: correlation with histopathological subtypes, metastasis, and
prognosis. World J Surg. 2005;29(11):1422-1427; discussion 1428.
74. van Beek J, zur Hausen A, Klein Kranenbarg E, et al. EBV-positive gastric
adenocarcinomas: a distinct clinicopathologic entity with a low frequency of
lymph node involvement. J Clin Oncol. 2004;22(4):664-670.
75. Zur Hausen A, van Rees BP, van Beek J, et al. Epstein-Barr virus in gastric
carcinomas and gastric stump carcinomas: a late event in gastric carcinogenesis. J
Clin Pathol. 2004;57(5):487-491.
76. Ushiku T, Shinozaki A, Shibahara J, et al. SALL4 represents fetal gut
differentiation of gastric cancer, and is diagnostically useful in distinguishing
hepatoid gastric carcinoma from hepatocellular carcinoma. Am J Surg Pathol.
2010;34(4):533-540.
77. Yong KJ, Gao C, Lim JS, et al. Oncofetal gene SALL4 in aggressive hepatocellular
carcinoma. N Engl J Med. 2013;368(24):2266-2276.
78. Tang LH, Untch BR, Reidy DL, et al. Well-differentiated neuroendocrine tumors
with a morphologically apparent high-grade component: a pathway distinct from
poorly differentiated neuroendocrine carcinomas. Clin Cancer Res.
2016;22(4):1011-1017.
79. Yachida S, Vakiani E, White CM, et al. Small cell and large cell neuroendocrine
carcinomas of the pancreas are genetically similar and distinct from well-
differentiated pancreatic neuroendocrine tumors. Am J Surg Pathol.
2012;36(2):173-184.
80. Basturk O, Yang Z, Tang LH, et al. The high-grade (WHO G3) pancreatic
neuroendocrine tumor category is morphologically and biologically heterogenous
and includes both well differentiated and poorly differentiated neoplasms. Am J
Surg Pathol. 2015;39(5):683-690.
81. Raj N, Valentino E, Capanu M, et al. Treatment response and outcomes of grade 3
pancreatic neuroendocrine neoplasms based on morphology: well differentiated
versus poorly differentiated. Pancreas. 2017;46(3):296-301.
82. Sorbye H, Welin S, Langer SW, et al. Predictive and prognostic factors for
treatment and survival in 305 patients with advanced gastrointestinal
neuroendocrine carcinoma (WHO G3): the NORDIC NEC study. Ann Oncol.
2013;24(1):152-160.
83. Lloyd RV OR, Kloppel G, Rosai J. WHO Classification of Tumours of Endocrine
Organs. Vol 10. 4th ed. Lyon: IARC; 2017.
84. Fennerty MB. Gastric intestinal metaplasia on routine endoscopic biopsy.
 Gastroenterology. 2003;125(2):586-590.
85. Correa P. Human gastric carcinogenesis: a multistep and multifactorial process–
first American Cancer Society Award Lecture on cancer epidemiology and
prevention. Cancer Res. 1992;52(24):6735-6740.
86. Correa P, Haenszel W, Cuello C, Tannenbaum S, Archer M. A model for gastric
cancer epidemiology. Lancet. 1975;2(7924):58-60.
87. Park JY, Cornish TC, Lam-Himlin D, Shi C, Montgomery E. Gastric lesions in
patients with autoimmune metaplastic atrophic gastritis (AMAG) in a tertiary care
setting. Am J Surg Pathol. 2010;34(11):1591-1598.
88. Coggon D, Barker DJ, Cole RB, Nelson M. Stomach cancer and food storage. J
Natl Cancer Inst. 1989;81(15):1178-1182.
89. La Vecchia C, Negri E, D’Avanzo B, Franceschi S. Electric refrigerator use and
gastric cancer risk. Br J Cancer. 1990;62(1):136-137.
90. Tricker AR. N-nitroso compounds and man: sources of exposure, endogenous
formation and occurrence in body fluids. Eur J Cancer Prev. 1997;6(3):226-268.
91. Song P, Wu L, Guan W. Dietary nitrates, nitrites, and nitrosamines intake and the
risk of gastric cancer: a meta-analysis. Nutrients. 2015;7(12):9872-9895.
92. La Vecchia C, Ferraroni M, D’Avanzo B, Decarli A, Franceschi S. Selected
micronutrient intake and the risk of gastric cancer. Cancer Epidemiol Biomarkers
Prev. 1994;3(5):393-398.
93. Palli D, Russo A, Decarli A. Dietary patterns, nutrient intake and gastric cancer in
a high-risk area of Italy. Cancer Causes Control. 2001;12(2):163-172.
94. Hansford S, Kaurah P, Li-Chang H, et al. Hereditary diffuse gastric cancer
syndrome: CDH1 mutations and beyond. JAMA Oncol. 2015;1(1):23-32.
95. van der Post RS, Vogelaar IP, Carneiro F, et al. Hereditary diffuse gastric cancer:
updated clinical guidelines with an emphasis on germline CDH1 mutation carriers.
J Med Genet. 2015;52(6):361-374.
96. Fujita H, Lennerz JK, Chung DC, et al. Endoscopic surveillance of patients with
hereditary diffuse gastric cancer: biopsy recommendations after topographic
distribution of cancer foci in a series of 10 CDH1-mutated gastrectomies. Am J
Surg Pathol. 2012;36(11):1709-1717.
97. Lee AF, Rees H, Owen DA, Huntsman DG. Periodic acid-schiff is superior to
hematoxylin and eosin for screening prophylactic gastrectomies from CDH1
mutation carriers. Am J Surg Pathol. 2010;34(7):1007-1013.
98. Barber ME, Save V, Carneiro F, et al. Histopathological and molecular analysis of
gastrectomy specimens from hereditary diffuse gastric cancer patients has
 implications for endoscopic surveillance of individuals at risk. J Pathol.
2008;216(3):286-294.
99. Worthley DL, Phillips KD, Wayte N, et al. Gastric adenocarcinoma and proximal
polyposis of the stomach (GAPPS): a new autosomal dominant syndrome. Gut.
2012;61(5):774-779.
100. Yanaru-Fujisawa R, Nakamura S, Moriyama T, et al. Familial fundic gland
polyposis with gastric cancer. Gut. 2012;61(7):1103-1104.
101. McDuffie LA, Sabesan A, Allgaeuer M, et al. beta-Catenin activation in fundic
gland polyps, gastric cancer and colonic polyps in families afflicted by ‘gastric
adenocarcinoma and proximal polyposis of the stomach’ (GAPPS). J Clin Pathol.
2016;69(9):826-833.
102. Repak R, Kohoutova D, Podhola M, et al. The first European family with gastric
adenocarcinoma and proximal polyposis of the stomach: case report and review of
the literature. Gastrointest Endosc. 2016;84(4):718-725.
103. Guan J, Lim KS, Mekhail T, Chang CC. Programmed death Ligand-1 (PD-L1)
expression in the programmed death receptor-1 (PD-1)/PD-L1 blockade: a key
player against various cancers. Arch Pathol Lab Med. 2017;141(6):851-861.
104. Akiyama T, Sudo C, Ogawara H, Toyoshima K, Yamamoto T. The product of the
human c-erbB-2 gene: a 185-kilodalton glycoprotein with tyrosine kinase activity.
Science. 1986;232(4758):1644-1646.
105. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human
breast cancer: correlation of relapse and survival with amplification of the HER-
2/neu oncogene. Science. 1987;235(4785):177-182.
106. Yan M, Schwaederle M, Arguello D, Millis SZ, Gatalica Z, Kurzrock R. HER2
expression status in diverse cancers: review of results from 37,992 patients.
Cancer Metastasis Rev. 2015;34(1):157-164.
107. Bang YJ, Van Cutsem E, Feyereislova A, et al. Trastuzumab in combination with
chemotherapy versus chemotherapy alone for treatment of HER2-positive
advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-
label, randomised controlled trial. Lancet. 2010;376(9742):687-697.
108. Bartley AN, Washington MK, Ventura CB, et al. HER2 testing and clinical decision
making in gastroesophageal adenocarcinoma: guideline from the College of
American Pathologists, American Society for Clinical Pathology, and American
Society of Clinical Oncology. Arch Pathol Lab Med. 2016;140(12):1345-1363.
109. Cruz-Reyes C, Gamboa-Dominguez A. HER2 amplification in gastric cancer is a
rare event restricted to the intestinal phenotype. Int J Surg Pathol. 2013;21(3):240-
 246.
110. Kunz PL, Mojtahed A, Fisher GA, et al. HER2 expression in gastric and
gastroesophageal junction adenocarcinoma in a US population: clinicopathologic
analysis with proposed approach to HER2 assessment. Appl Immunohistochem
Mol Morphol. 2012;20(1):13-24.
111. Park JS, Rha SY, Chung HC, et al. Clinicopathological features and prognostic
significance of HER2 expression in gastric cancer. Oncology. 2015;88(3):147-
156.
112. Okines AF, Thompson LC, Cunningham D, et al. Effect of HER2 on prognosis and
benefit from peri-operative chemotherapy in early oesophago-gastric
adenocarcinoma in the MAGIC trial. Ann Oncol. 2013;24(5):1253-1261.
113. Janjigian YY, Werner D, Pauligk C, et al. Prognosis of metastatic gastric and
gastroesophageal junction cancer by HER2 status: a European and USA
International collaborative analysis. Ann Oncol. 2012;23(10):2656-2662.
114. Pirrelli M, Caruso ML, Di Maggio M, Armentano R, Valentini AM. Are biopsy
specimens predictive of HER2 status in gastric cancer patients? Dig Dis Sci.
2013;58(2):397-404.
115. Qiu Z, Sun W, Zhou C, Zhang J. HER2 expression variability between primary
gastric cancers and corresponding lymph node metastases.
Hepatogastroenterology. 2015;62(137):231-233.
116. Bozzetti C, Negri FV, Lagrasta CA, et al. Comparison of HER2 status in primary
and paired metastatic sites of gastric carcinoma. Br J Cancer. 2011;104(9):1372-
1376.
117. Ruschoff J, Dietel M, Baretton G, et al. HER2 diagnostics in gastric cancer-
guideline validation and development of standardized immunohistochemical
testing. Virchows Arch. 2010;457(3):299-307.
118. Gullo I, Grillo F, Molinaro L, et al. Minimum biopsy set for HER2 evaluation in
gastric and gastro-esophageal junction cancer. Endosc Int Open. 2015;3(2):E165-
E170.
119. Bang YJ. Advances in the management of HER2-positive advanced gastric and
gastroesophageal junction cancer. J Clin Gastroenterol. 2012;46(8):637-648.
120. Abrahao-Machado LF, Jacome AA, Wohnrath DR, et al. HER2 in gastric cancer:
comparative analysis of three different antibodies using whole-tissue sections and
tissue microarrays. World J Gastroenterol. 2013;19(38):6438-6446.
121. Stolte M, Sticht T, Eidt S, Ebert D, Finkenzeller G. Frequency, location, and age
and sex distribution of various types of gastric polyp. Endoscopy. 1994;26(8):659-
 665.
122. Borch K, Ahren B, Ahlman H, Falkmer S, Granerus G, Grimelius L. Gastric
carcinoids: biologic behavior and prognosis after differentiated treatment in
relation to type. Ann Surg. 2005;242(1):64-73.
123. Burkitt MD, Pritchard DM. Review article: pathogenesis and management of
gastric carcinoid tumours. Aliment Pharmacol Ther. 2006;24(9):1305-1320.
124. Rorstad O. Prognostic indicators for carcinoid neuroendocrine tumors of the
gastrointestinal tract. J Surg Oncol. 2005;89(3):151-160.
125. McCall CM, Shi C, Cornish TC, et al. Grading of well-differentiated pancreatic
neuroendocrine tumors is improved by the inclusion of both Ki67 proliferative
index and mitotic rate. Am J Surg Pathol. 2013;37(11):1671-1677.
126. Chen WC, Warner RR, Ward SC, et al. Management and disease outcome of type I
gastric neuroendocrine tumors: the Mount Sinai experience. Dig Dis Sci.
2015;60(4):996-1003.
127. Raderer M, Kiesewetter B, Ferreri AJ. Clinicopathologic characteristics and
treatment of marginal zone lymphoma of mucosa-associated lymphoid tissue
(MALT lymphoma). CA Cancer J Clin. 2016;66(2):153-171.
128. Isaacson PG, Spencer J. Malignant lymphoma of mucosa-associated lymphoid
tissue. Histopathology. 1987;11(5):445-462.
129. Ye H, Liu H, Raderer M, et al. High incidence of t(11;18)(q21;q21) in
Helicobacter pylori-negative gastric MALT lymphoma. Blood. 2003;101(7):2547-
2550.
130. Bacon CM, Du MQ, Dogan A. Mucosa-associated lymphoid tissue (MALT)
lymphoma: a practical guide for pathologists. J Clin Pathol. 2007;60(4):361-372.
131. Copie-Bergman C, Gaulard P, Lavergne-Slove A, et al. Proposal for a new
histological grading system for post-treatment evaluation of gastric MALT
lymphoma. Gut. 2003;52(11):1656.
132. Vanek J. Gastric submucosal granuloma with eosinophilic infiltration. Am J Pathol.
1949;25(3):397-411.
133. Carmack SW, Vemulapalli R, Spechler SJ, Genta RM. Esophagitis dissecans
superficialis (“sloughing esophagitis”): a clinicopathologic study of 12 cases. Am
J Surg Pathol. 2009;33(12):1789-1794.
134. Rossi P, Montuori M, Balassone V, et al. Inflammatory fibroid polyp. A case report
and review of the literature. Ann Ital Chir. 2012;83(4):347-351.
135. Rittershaus AC, Appelman HD. Benign gastrointestinal mesenchymal BUMPS: a
brief review of some spindle cell polyps with published names. Arch Pathol Lab
 Med. 2011;135(10):1311-1319.
136. Morandi E, Pisoni L, Castoldi M, Tavani E, Trabucchi E. Gastric outlet obstruction
due to inflammatory fibroid polyp. Ann Ital Chir. 2006;77(1):59-61.
137. Stolte M. Clinical consequences of the endoscopic diagnosis of gastric polyps.
Endoscopy. 1995;27(1):32-37; discussion 59-60.
138. Cassier PA, Ducimetiere F, Lurkin A, et al. A prospective epidemiological study of
new incident GISTs during two consecutive years in Rhone Alpes region:
incidence and molecular distribution of GIST in a European region. Br J Cancer.
2010;103(2):165-170.
139. Fletcher CD, Berman JJ, Corless C, et al. Diagnosis of gastrointestinal stromal
tumors: a consensus approach. Hum Pathol. 2002;33(5):459-465.
140. Miettinen M, El-Rifai W, H L Sobin L, Lasota J. Evaluation of malignancy and
prognosis of gastrointestinal stromal tumors: a review. Hum Pathol.
2002;33(5):478-483.
141. Miettinen M, Sarlomo-Rikala M, Lasota J. Gastrointestinal stromal tumors: recent
advances in understanding of their biology. Hum Pathol. 1999;30(10):1213-1220.
142. Goddard AF, Badreldin R, Pritchard DM, Walker MM, Warren B. The management
of gastric polyps. Gut. 2010;59(9):1270-1276.
143. Dei Tos AP, Laurino L, Bearzi I, Messerini L, Farinati F. Gastrointestinal stromal
tumors: the histology report. Dig Liver Dis. 2011;43 Suppl 4:S304-S309.
144. Liegl-Atzwanger B, Fletcher JA, Fletcher CD. Gastrointestinal stromal tumors.
Virchows Arch. 2010;456(2):111-127.
145. Caletti G, Deviere J, Fockens P, et al. Guidelines of the European Society of
Gastrointestinal Endoscopy (ESGE) Part II: retroperitoneum and large bowel,
training. The European Endosonography Club working party. Endoscopy.
1996;28(7):626-628.
146. Van Dam J, Brady PG, Freeman M, et al. Guidelines for training in electronic
ultrasound: guidelines for clinical application. From the ASGE. American Society
for Gastrointestinal Endoscopy. Gastrointest Endosc. 1999;49(6):829-833.
147. Nesje LB, Laerum OD, Svanes K, Odegaard S. Subepithelial masses of the
gastrointestinal tract evaluated by endoscopic ultrasonography. Eur J Ultrasound.
2002;15(1-2):45-54.
148. Casali PG, Jost L, Reichardt P, Schlemmer M, Blay JY. Gastrointestinal stromal
tumours: ESMO clinical recommendations for diagnosis, treatment and follow-up.
Ann Oncol. 2009;20 Suppl 4:64-67.
149. Casali PG VJ, Kotasek D, LeCesne A, Reichardt P, Blay JY. Imatinib mesylate in
 advanced gastrointestinal stromal tumors (GIST): survival analysis of the
intergroup EORTC/ISG/AGITG randomized trial in 946 patients. Eur J Cancer.
2005;3:201; abstract 711.
150. Debiec-Rychter M, Sciot R, Le Cesne A, et al. KIT mutations and dose selection
for imatinib in patients with advanced gastrointestinal stromal tumours. Eur J
Cancer. 2006;42(8):1093-1103.
151. Carney JA, Stratakis CA. Familial paraganglioma and gastric stromal sarcoma: a
new syndrome distinct from the Carney triad. Am J Med Genet. 2002;108(2):132-
139.
152. Pasini B, McWhinney SR, Bei T, et al. Clinical and molecular genetics of patients
with the Carney-Stratakis syndrome and germline mutations of the genes coding for
the succinate dehydrogenase subunits SDHB, SDHC, and SDHD. Eur J Hum Genet.
2008;16(1):79-88.
153. Gaal J, Stratakis CA, Carney JA, et al. SDHB immunohistochemistry: a useful tool
in the diagnosis of Carney-Stratakis and Carney triad gastrointestinal stromal
tumors. Mod Pathol. 2011;24(1):147-151.
154. Miettinen M, Sarlomo-Rikala M, Sobin LH, Lasota J. Esophageal stromal tumors:
a clinicopathologic, immunohistochemical, and molecular genetic study of 17 cases
and comparison with esophageal leiomyomas and leiomyosarcomas. Am J Surg
Pathol. 2000;24(2):211-222.
155. Hyun JH, Jeen YT, C