Quality Assurance of Ultrasound Imaging

Systems

IPEM Report No 102

A Report produced by the Institute of Physics and Engineering in

Medicine

Editor
Stephen Russell, Manchester

Authors
Nick Dudley, Lincolnshire
Tony Evans, Leeds
Peter Hoskins, Edinburgh
Amanda Watson, Glasgow
Hazel Starritt, Bath
 © Institute of Physics and Engineering in Medicine 2010
Fairmount House, 230 Tadcaster Road
York YO24 1ES
ISBN 978 1 903613 43 6

All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic, mechanical, photocopying,
recording or otherwise, without the prior permission of the publisher.

Published by the Institute of Physics and Engineering in Medicine

Fairmount House, 230 Tadcaster Road, York YO24 1ES

Legal Notice

This report was prepared and published on behalf of the Institute of Physics and
Engineering in Medicine (IPEM). Whilst every attempt is made to provide accurate and
useful information, neither the IPEM, the members of IPEM or other persons contributing
to the formation of the report make any warranty, express or implied, with regard to
accuracy, omissions and usefulness of the information contained herein. Furthermore, the
same parties do not assume any liability with respect to the use, or subsequent damages
resulting from the use, of the information contained in this report.

Prepared and printed by:

The Charlesworth Group, Wakefield, UK. [Link]
Contents
Acknowledgements vii

Preface viii

Chapter 1: Introduction 1

1.1 Background 1

1.2 Overview 1

1.3 Image quality 2

1.4 New imaging technologies 3

1.5 Test structure 3

1.6 Display screen equipment tests 4

1.7 Record keeping 4

Chapter 2: B-Mode imaging 6

2.1 Table of measurements 7

2.2 Test equipment 7

2.2.1 Tissue equivalent test object 7
2.2.2 Open topped test object 8

2.3 Acceptance tests 9

2.3.1 Calliper accuracy 9
2.3.2 Image uniformity 10
2.3.3 Functional checks 11
2.3.4 Comparison with demonstration scanners 12

2.4 Baseline tests 12

2.4.1 Sensitivity 12
 2.4.2 Noise 14
2.4.3 Low contrast penetration 15

2.5 User tests 16

2.5.1 Image uniformity 16
2.5.2 Sensitivity 17
2.5.3 Noise 18

Chapter 3: Doppler 19

3.1 Tables of measurements 19

3.1.1 Spectral Doppler 19
3.1.2 Colour/power Doppler 20

3.2 Test equipment 21

3.2.1 Simple string test object 21

3.3 Spectral Doppler preparation and positioning 23

3.4 Spectral Doppler minimum system performance tests 24

3.4.1 Maximum velocity estimation accuracy at 50 cm s–1
(or similar) set velocity 24
3.4.2 Intrinsic spectral broadening at 50 cm s–1 (or similar)
set velocity 26
3.4.3 Mean velocity estimation 26
3.4.4 Doppler noise 27

3.5 Spectral Doppler basic functional checks 27

3.5.1 Flow direction at 50 cm s–1 (or similar) set velocity 27
3.5.2 Axial range gate registration and dimensions at 50 cm s–1
(or similar) set velocity 28
3.5.3 Doppler gain control function response 28
3.5.4 Pulse repetition frequency (PRF)/velocity range control
function response 29
3.5.5 Baseline shift control response 30
3.5.6 Angle correction accuracy 30

iv
3.6 Colour and power Doppler basic functional checks 30
3.6.1 Colour flow direction at 50 cm s–1 (or similar) set velocity 31
3.6.2 Colour/power Doppler gain control function response 31
3.6.3 Colour PRF/velocity range control function response 32
3.6.4 Colour baseline shift control response 32
3.6.5 Colour/power box positioning and sizing 32

Chapter 4: Safety 34

4.1 Table of measurements 35

4.2 Acoustics safety test equipment 36

4.2.1 Acoustic force balance 36
4.2.2 Radiometer 36
4.2.3 Hydrophones 37

4.3 Acoustics safety acceptance measurements 37

4.3.1 Visual check of mechanical index (MI) 37
4.3.2 Visual check of thermal index (TI) 39
4.3.3 Measurement of TI 40
4.3.4 Measurement of surface temperature of transducer 42

4.4 Acoustics safety baseline/routine measurements 44

4.4.1 Visual check of MI 44
4.4.2 Visual check of TI 45
4.4.3 Measurement of TI 45
4.4.4 Surface temperature of transducer 46
4.4.5 Acoustic power 46

4.5 Electrical safety and physical integrity acceptance,

routine and baseline testing 48
4.5.1 Physical integrity 49
4.5.2 Electrical integrity of system 51

Chapter 5: Extended testing 58

5.1 B-Mode imaging 58

5.1.1 Test equipment 58
5.1.2 Preparation 59

v
 5.1.3 Dead zone 60
5.1.4 Axial resolution, lateral resolution and slice thickness 60
5.1.5 Grey-scale performance 61
5.1.6 Anechoic target detection 62
5.1.7 Novel methods 62

5.2 Doppler extended testing 63

5.3 Extended acoustic and thermal safety testing 66

5.3.1 Table of measurements 66
5.3.2 Equipment for acceptance and routine testing 67
5.3.3 Evaluation of MI safety index 71
5.3.4 Spatial peak temporal average intensity 73
5.3.5 In situ temperature rise 75

Appendix A: B-Mode test objects 77

Appendix B: Features and calibration of a string phantom 80

Appendix C: Ultrasound safety test equipment 83

Appendix D: References and Standards 92

vi
Acknowledgements
This publication is the result of a collaborative effort by the authors Nick Dudley, Tony
Evans, Peter Hoskins, Hazel Starritt and Amanda Watson. Without the considerable time
and effort that they contributed to this project it would not have been possible.
Thanks are also due to the reviewers whose valuable input significantly improved the
document.

vii
Preface
The Authors of this group were brought together by the IPEM Ultrasound and Non
Ionising Radiation Special Interest group to review and recommend action with respect
to ultrasound quality assurance (QA) and the current guidance given in IPEM Reports 70
and 71. After much deliberation, it was decided that an update to the protocols and methods
given in these earlier reports would be useful.
It has been over 15 years since IPEM Reports 70 and 71 were produced to give guidance
on the measurement of ultrasound system quality. Over this period, there has been a con-
tinuous significant improvement in the quality, capability and application of ultrasound
imaging systems. The complexity and range of facilities on the most advanced scanners
have expanded and the devices have moved from being mostly analogue in design to
almost entirely digital. New ranges of scanners have been introduced aimed at medical
specialties outside of the traditional ultrasound diagnostic imaging, vascular and cardiol-
ogy applications. It is hoped that this document which both simplifies and updates the
QA techniques will find a wide audience within both the newer and traditional areas of
ultrasound application.
The tests identified are aimed at a wide skills base with some only applicable to physicists
specialising in ultrasound or to service engineers maintaining systems, but many will be
suitable for clinicians, radiographers, clinical technologists, nurses and midwives who use
ultrasound systems.
Stephen Russell (Editor),The Christie NHS Foundation Trust, July 2009

viii
1 Introduction
1.1 Background
The measurement of ultrasound system performance has been a difficult subject from the
inception of medical ultrasound scanning. There are few, if any, measurements which are
universally accepted as quantitative, reproducible and representative of clinical perform-
ance. The nature of the ultrasound image is somewhat unique in that even for parameters
which can be measured, many vary significantly depending on the location in the image,
the settings on the system and the particular transducer employed. The range of possible
settings and transducers is a major hurdle in characterising the performance of a system.
The purpose of this report is not therefore to measure system performance but to establish
a group of tests monitoring aspects of the performance which are considered likely to
change or deteriorate and which will probably impact on the clinical efficacy of the system.
Since the earlier IPEM reports on ultrasound testing, many aspects of the ultrasound
system have moved from analogue-based electronics to digital. The effects previously
tested relating to component changes (for example, drifting reference voltages or changing
frequencies) no longer apply. However, there are still some aspects of the machine subject
to such changes mostly related to the transducer and display monitor.
The newest systems are software driven and implemented on flexible platforms. These
may be subject to frequent software changes either to enable system improvements or to
fix bugs. The software changes are a new potential source of problems, with users often
unaware of which software version their system is operating on and when and whether any
software updates have taken place.
The routine monitoring of systems against a baseline group of tests provides reassurance
that they retain a consistent level of performance. They can both monitor the impact of
components which age or wear and potentially identify issues arising from changes in the
system resulting from software updates.
On the arrival of a new piece of medical equipment, the application of a fairly rigorous
group of tests, often referred to as acceptance tests, is already well established. At the
time of this report, it is recommended in the UK by the Medical and Healthcare Products
Regulatory Agency in their Device Bulletin DB2006(5) (MHRA, 2006). It is now possible
for radical changes to occur in a system from software updates alone, and such updated
equipment should be treated as new for a number if not all aspects of its performance.

1.2 Overview
The tests within this report have been selected on a criterion of merit based on a number
of factors including: likelihood of change, cost and portability of test equipment, skill
required by the operator and the reproducibility, effectiveness and clinical relevance of the

1
Quality Assurance of Ultrasound Imaging Systems

tests. Many of the tests rely on skilled and experienced operators with considerable under-
standing of the scanner controls but some may be performed by less experienced operators
with basic ultrasound training. Effective quality assurance (QA) for mainstream diagnostic
systems does require considerable experience and skill. It is achieved by having operators
with a thorough understanding of ultrasound, responsible for large numbers of imaging
systems, such that they are familiar with the subtle changes which indicate early system
problems. However, systems used by non-imaging medical specialties may have signifi-
cant problems which go unnoticed by their users and for these systems, regular checks
by operators with specific training in ultrasound QA can be effective and ensure that the
devices continue to be fit for purpose. It is recommended that all systems should be checked
at least annually by a well trained physicist or technician to ensure that the level of QA for
the system is effective and to perform the regular tests which are outside of the scope of
the user or local QA operator. This provides a regular audit of QA procedures, skills and
system maintenance levels.
The measurements identified in this report relate to one of three criteria:
System acceptance tests: to determine in so far as possible that the performance of a
system is of a required standard and at a level that is expected.
System baseline quality assurance tests: to make a reference measurement of perform-
ance on a system. These measurements are identical to those performed as routine quality
assurance checks but are repeated several times to establish a baseline average.
System routine quality assurance tests: to determine whether the system continues
to perform to the same standard as at the reference measurement. They may also be
performed reactively if a query on performance is raised.

1.3 Image quality

The term image quality is used throughout this report and it refers to the aspects of the
image which are expected to impact the diagnostic capability of the system. There are a
number of traditional measurable image performance parameters such as resolution (in
ultrasound, this is usually split into three components: axial, lateral and slice resolution),
contrast, noise, and penetration which may affect diagnostic capability. There are also
other aspects related to the real time nature of the image, temporal effects, such as frame
rate and temporal resolution.
The fact that ultrasound systems are used in an exceptionally wide range of disciplines
means that the diagnostic significance of image performance parameters will vary. With
the present range of knowledge and tests, it is not possible to attribute particular levels of
significance to a measured parameter. Only by further research recording the values of
measurable imaging parameters and monitoring the clinical efficacy of systems will the
significance of changes in imaging performance parameters become clear. In the mean-
time, monitoring changes and deciding on sensible investigation or recommended
actions based on comparisons with performance at or near acceptance are the best options
available.

2
 Introduction

It is important to remember that, as a diagnostic ultrasound scanner is used in real time,

it is the user of the system at the time who must be able to justify the suitability of the
equipment if challenged. Static imaging records made at the time showing the efficacy of
the scan are essential. In addition, a background of regular local testing and professional
annual inspection can also strongly support this requirement.

1.4 New imaging technologies

The rapid development of computing technology in recent years has had a huge impact
on ultrasound imaging with new methods for encoding pulses, imaging with harmonics
and processing raw echo information. These methods which continue to evolve need
investigating both to establish their efficacy in clinical applications and their ongoing
performance. There is a requirement to have rational objective methods for evaluating
manufacturers’ claims made for the introduction or evolution of a new technology. It is
possible that methods could be developed from the extended testing identified in this
document. The monitoring of changes in clinical outcomes following the implementation
of new technologies may also prove to be an appropriate method in some circumstances.
However, this does require rigorous monitoring of the outcomes before the new technol-
ogy is implemented which is rarely undertaken. Overall, there is a strong case for more
research into measuring the benefits of ‘improved technologies’.

1.5 Test structure

This report focuses on acceptance testing, baseline measurements and quality assurance
measurements which may be implemented in a routine testing programme or as a reactive
response to performance queries. The measurements have been separated into three areas
of standard tests for B-mode imaging, Doppler imaging and Safety (which includes ultra-
sound safety, electrical safety and physical integrity). A table at the front of each section
identifies when each of the tests described should be performed and whether it applies to
all of the transducers on the system (transducer related test) or just one (system related
test). These sections set the core testing requirements. Additional tests which may be use-
ful in researching system performance or form the basis of more extensive testing regimes
are discussed in Chapter 5. The range of standard tests described in Chapters 2 to 4 should
be considered as the minimum to be performed on any clinical diagnostic ultrasound
system with an appropriate subset applied to task-specific ultrasound systems. An
implementation of these tests will require a mix of QA trained users and medical physics
specialists.
Where possible, limits or recommended actions are described for each test but, in many
cases, these will have to be established using local criteria. Factors such as clinical applica-
tion, operator skills and range of operators will determine appropriate interventions or
escalation of investigation to minimise risk. The recommended action levels are most often
the point at which further investigation over and above the basic testing is required. It is

3
Quality Assurance of Ultrasound Imaging Systems

not possible to be more specific because of the extremely wide range of systems and
applications. The action levels quoted are based on recommendations from the literature
and the experience of the authors. As more data are reported and become available
then national rather than local limits should become established. The recent National
Physical Laboratory report AC2 (Shaw and Hekkenberg, 2007) gives extensive details and
recommendations on testing programmes.

1.6 Display screen equipment tests

There are no specific tests in this report relating to display equipment (DSE) or visual
display units (VDU). There are several reasons for this, the principal difficulty being that
there is no readily identifiable method for testing across all or even a wide range of ultra-
sound systems. Almost all methods require test patterns generated from stored images and
these are not universally available on ultrasound systems. The tests usually consist of
visual assessment of elements within the test pattern and may be augmented by measuring
luminance at various points within the pattern using a calibrated luminance meter. Modern
ultrasound systems use computer displays and if stored test images are available then test-
ing using a range of different standards is feasible. Images are also available in a DICOM
format and it should be possible to upload these onto compatible systems.
It is anticipated that the future use of imaging stations will encompass an increasing range
of modalities where the distinctions applicable will principally be between 2D/3D and
B&W/colour environments. The testing regime for ultrasound system displays used for
diagnosis should therefore be the same as that applied to any picture archival and com-
munication system (PACS) primary diagnostic monitor or modality workstation monitor.
The tests should be applied wherever possible, for example on DICOM compliant systems
by importing the appropriate DICOM test patterns. The standards for display monitor
testing described in IPEM Report 91 (IPEM, 2005) may be appropriate.
There are some systems in use, particularly portable devices and low end systems, for
which there is no practicable method for testing the monitor other than visual checks of a
clinical or test object image. It is expected that the situation will improve over time as all
diagnostic imaging systems become DICOM compliant and manufacturers become aware
of the need to provide display test patterns.

1.7 Record keeping

During the life of an ultrasound scanner, there is a substantial amount of paperwork
generated; the initial delivery documentation and user manuals, the acceptance test report,
service reports from the agent maintaining the system, quality control (QC) reports from
physics staff, user QC records, repairs, replacements and software updates. All of this
documentation should ideally be kept in a single location such that the history and any
issues with the system can easily be referenced. For example, knowing when a fault with

4
 Introduction

a transducer was identified, what actions were taken and when a replacement was installed
and tested may be very important if there is a query on the efficacy of an examination made
at that time.
Ultrasound systems are not fixed installations, are used across many applications and may
be moved between units and assigned to different clinical applications during their work-
ing life. When this occurs, it is important that both the new and previous users continue
to have access to the system records. Often in such moves, records may be lost which
becomes a serious issue if that system is subsequently implicated in legal proceedings.
It is important in departments with several machines of similar types and with interchange-
able probes to keep records of when each probe is used on a particular system. Ideally,
probes should not be swapped between systems so that a complete record of the equipment
as used can be kept.
If an issue with a system is identified either from QA or an incidental finding and a deci-
sion is taken to continue using the system, then a record must be kept of that decision. This
should include the reasoning and control measures in place to minimise risk. In some
cases, it would be prudent to record a full risk assessment but for others where there is, for
example, no impact on the image quality, a simple memo or temporary change in proce-
dure may be adequate. Whichever method is used, a copy of this information should also
be kept with the main records as it is important to have contemporary records of decisions
along with details of the faults they relate to.

5
2 B-Mode imaging
There is a limited range of evidence for the efficacy of tests using tissue equivalent test
objects (TETOs). As a result, this guidance focuses as far as possible on simple inherently
stable tests that reflect system accuracy and performance changes that would be expected
to impact on image quality. Some equivalent tests can also be performed with a TETO
but there is no evidence to show that there is any advantage to this. The use of a TETO
invariably slows the process and additional checks on test object stability are required.
The functional tests do require a TETO but these are qualitative and do not rely on the
consistency of the TETO. There are many additional tests not listed here which may be
undertaken to provide the evidence required to show effective methods of B-mode quality
assurance (QA). A number of these additional tests are outlined in Chapter 5. As part of
the frequencies for checks indicated in the table, it is recommended that at least one of
these is made an annual check and audit of the user testing programme, performed by a
well trained physicist or physics technician. If regular user tests are not undertaken, these
annual checks are still advised although this would not comply with the recommendations
of this report.
Current national guidance in IPEM Report 71 (IPEM, 1995) and international guidance
including that of the AAPM (Goodsitt et al., 1998) recommend a range of routine tests
using a TETO. Tests were recommended based on expert opinion and no evidence of
efficacy was referenced. Kollmann (1995) found changes in resolution or calliper accuracy
in 11 of 17 probes over a period of 18 months; however, the accuracy of the measurements
was not discussed and no remedial action was taken as a result of the changes. Kanal et al.
(1998) found that the 95% confidence interval on resolution measurements was ±40%,
suggesting that only gross changes in performance may be demonstrated with confidence.
Limited negative evidence for the value of these methods was provided by Dudley et al.
(2001) who found no imaging faults on 17 scanners over a period of 3 years.
Since the publication of IPEM Report 71, harmonic imaging, high frequency imaging
(>10 MHz), 3D and 4D imaging, compounding and speckle reduction imaging have
become widely used. Few specific tests have been developed, although TETOs are avail-
able for 3D imaging (e.g. Kollmann et al., 2001). The tests detailed here will work at high
frequencies and are capable of demonstrating general faults that will affect harmonic
imaging, 3D and 4D imaging. Freehand elastography is now available on a number of
mainstream systems and there are some which give quantitative strain measurements.
These are outside the scope of this report but will need to be considered in the future.
The user tests identified for B-mode can easily be performed in a few minutes and are
ideally suited to a simple frequent local testing programme. Minimum frequencies of
monthly tests are suggested but checks could readily be performed weekly due to their low
time overhead.

6
 B-Mode imaging

2.1 Table of measurements

Test Methods Whena Notes
Axial and lateral calliper Open topped test object A Use more than one
accuracy or tissue equivalent test scale setting and
object zoom setting
Circumference and area Open topped test object A
measuring facility (and
others)
Sensitivity/penetration Reverberation in air U All frequencies at
Tissue equivalent test R baseline
object
Spatial uniformity Air & paperclip A/U
Tissue equivalent test A
object
Electronic noise Gain threshold U
Functional checks, e.g. Tissue equivalent test A
focal zones, time-gain object
compensation (TGC)
a
A, acceptance (new scanner or probe, or repair/service affecting this parameter);
U, user (minimum frequency monthly); R, reactive; baseline tests are required for
user and reactive tests; these baseline tests may be included as part of the acceptance.
Where resources allow, reactive tests may be performed routinely as planned tests on
a biannual or annual basis; this may be required where there is no user testing
undertaken or where external audit of the user testing programme is required.
All except the functional checks are transducer related tests.

2.2 Test equipment

2.2.1 Tissue equivalent test object
When using tissue equivalent test objects (TETOs), it is important to understand their
limitations. The authors have noted differences in image characteristics between TETOs
of the same model and it is well known that TETOs can deteriorate with age, e.g. due to
desiccation. TETOs should be stored, used and maintained according to the manufacturers’
instructions; some manufacturers offer ‘rejuvenation’ services. It is good practice to
monitor the condition of TETOs by, for example, weighing, checking target separation
with a scanner calibrated by other means and observing images for changes such as cracks
in the internal medium. Units with large numbers of ultrasound systems may be able to use
statistical trend monitoring across their systems to establish the stability of the TETO.

7
Quality Assurance of Ultrasound Imaging Systems

In smaller units, collaboration with others including audit, exchanging TETO and
cross-calibration checks, for example, with fluid filled calibration objects, may be suitable
methods for maintaining standards. Small changes in TETO performance will not affect
qualitative tests, such as those recommended for scanner acceptance, but may influence
the results of quantitative tests, in particular calliper accuracy. The speed of sound and
attenuation in TETOs can have a significant temperature dependence (Browne et al.,
2003).
At the time of writing, traceable measurements of attenuation and speed of sound in
material samples can be made by the National Physical Laboratory (NPL). Measurements
of backscatter coefficient and non-linearity parameter (B/A) are planned. The tests, both
available and proposed, will not be possible on an intact test object.
The TETO should be of uniform tissue mimicking material, with targets to allow a
functional check of focal zones, e.g. column of nylon filaments, and, if necessary, targets
for testing calliper accuracy in lateral and axial directions. Cyst or contrast targets can
be useful in general system evaluation and should ideally be spherical although most
commercial TETOs use cylinders which do not differentiate slice resolution issues.
The speed of sound should be 1540 m s–1, unless an open topped test object (OTTO, see
below) is to be used for calliper checks in which case alternatives such as urethane may be
considered. It should be noted that due to the potential impact of the velocity changes on
the image, urethane and other similar objects should only be used for ongoing consistency
checks (Goldstein, 2000; Dudley et al., 2002).
Attenuation and backscatter should be appropriate to the generation of speckle throughout
the normal clinical depth of imaging. Attenuation of 0.5 to 0.7 dB cm–1 MHz–1 is likely
to be suitable for most applications; the higher attenuation may be more suitable for
penetration measurements at low frequencies.

2.2.2 Open topped test object

An open topped test object (OTTO) is used to check calliper accuracy or image linearity.
The test object should contain target patterns of similar shapes to those measured in
clinical practice. For some applications where only linear measurements are performed,
the targets found in current TETOs (lateral and axial bands of uniformly spaced wires,
e.g. Gammex 403GS) are suitable (bearing in mind the limitations outlined above).
For checking non-linear measurements, a test object with targets arranged in a regular
two-dimensional array or in circles or ellipses may be appropriate (Dudley and Griffith,
1996). This is likely to be an OTTO containing a liquid with a sound transmission speed
of 1540 m s–1. A number of options are available, e.g. 9.5% ethanol by volume in distilled
water at 20°C (Martin and Spinks, 2001), water at 50°C; the speed of sound in all such
liquids has a significant temperature dependence and volatile liquids will be subject to
evaporation losses. A container which can be sealed when not in use will minimise the
effects of evaporation. It is recommended that the temperature is monitored and recorded
when using such devices. A significant advantage of the OTTO is the ability to readily
change target configurations should the need arise.

8
 B-Mode imaging

2.3 Acceptance tests

The principal role of acceptance testing is to establish that a system is operating correctly
and safely. These tests may be used subsequently as additional reactive tests, to
confirm issues identified from user tests comparing the results with those of the original
acceptance.
For all tests, it is suggested that either a custom test preset or a factory default is used as a
starting point for system settings. It is important that the details of this setting and any
adjustments made for each test are clearly recorded at acceptance or the first instance of
testing. Physically recording the screen settings using printers or PACS or saving the preset
to media stored away from the system are all suitable. Any manual adjustments of these
base settings for particular measurements should be recorded as part of the protocol.

2.3.1 Calliper accuracy

Reproducibility and absolute accuracy are both important to the clinical use of
measurement functions.

[Link] Set-up and assessment

The test object (OTTO or TETO as appropriate to the measurement function under test)
should be imaged at an appropriate magnification to allow measurements to be made with
each method in clinical use, e.g. linear, trace and ellipse. Measurements should be made
between targets within the test object at clinically realistic distances and scanning depths
for the transducer concerned, e.g. a few millimetres to replicate nuchal translucency meas-
urement. If using an OTTO, it is likely that measurements with high frequency transducers
will be possible at depths which would not be available in the clinical setting.
When imaging the test object, adjust overall gain and time-gain compensation (TGC) to
reduce the level of speckle and noise around targets, so that they are clearly distinguished
and do not saturate the grey-scale.
A series of measurements (say, 10) should be made for each method and the mean and
coefficient of variation (CoV) should be calculated.
It is possible that measurements may be affected by scale setting, as pixel sizes will be
rescaled by the scanner’s software. It is impractical to test callipers on every possible scale
setting, but a subset of measurements should be repeated on a range of clinically realistic
scale settings. Due to the nature of current digital systems, these tests should be repeated
whenever software is updated.

[Link] Recorded values and recommended actions

The tolerance for each measurement and its CoV should be set to reflect the clinical
accuracy required (Dudley, 2003), illustrated by the following examples. For measure-
ments of fetal biparietal diameter (in the range 20 mm to 95 mm), accuracy of ±1 mm is

9
Quality Assurance of Ultrasound Imaging Systems

required (the absolute accuracy of a linear measurement in a test object should not be
affected by the distance); the CoV should be less than 3% at 20 mm and less than 1% at
95 mm. For measurements of the fetal abdominal circumference (in the range 150 mm to
400 mm), an accuracy of ±2.5% is required, however the accuracy of fitting methods, e.g.
ellipse fitting, should be ±1%; the CoV should be less than 2% for traced measurements
and less than 1% for fitting methods.
Where a TETO shows measurements outside acceptable limits, a repeat test using
an OTTO may be advisable before reporting the issue. If it is shown in both cases that
a measurement error is present, then the manufacturer should be informed and the issue
investigated further.

2.3.2 Image uniformity

[Link] Set-up and assessment
This test applies to probes which operate as linear or curvilinear arrays for phased arrays.
It is useful to simply record the in-air reverberation for future reference. Operate the probe
dry in air and increase the overall gain to maximum; move a single focus as close to
the probe face as possible. It is important to switch off advanced imaging functions,
especially spatial compounding, spatial smoothing and speckle reduction imaging as these
may mask fundamental problems with the probe. Inspect the reverberation pattern seen
for axial banding, which represents a localised transmission/reception fault for linear or
curvilinear probes. There are a variety of possible causes. By running a narrow, smooth
metal object (such as a paperclip or the back of a thin key) along the probe over a very thin
film of coupling gel looking for reduced echo levels at the defective area, the magnitude
of the fault may be assessed (Figure 2.1). There may be intrinsic banding evident on a
brand new transducer for which there are no faults. These can be the result of minor

Figure 2.1 Illustration of the ‘paperclip test’ where a reduction in reverberation amplitude
is shown as a thin metal object is moved over the probe face. The effects of drop-out can
clearly be seen

10
 B-Mode imaging

imperfections in the lens surface or as a product of the beam forming, the latter evidenced
by symmetry in the banding around the centre line. These effects are slight compared with
crystal faults involving more than one element and will not be evident within a TETO
image. Faults may occur within the system which result in structured bands which
are clearly evident both for the reverberations in air and for the TETO. These are usually
system rather than probe related.
Image a TETO using a scanner preset that provides a uniform TGC slope, e.g. abdominal
preset, as this should be well matched to the attenuation properties of the TETO. If such a
preset is not available, TGC will require optimisation during imaging. An optimised image
should show uniform grey-scale in the tissue mimicking material throughout the useful
penetration depth. Moving the probe along the scanning axis whilst imaging in real time
will often emphasise banding effects; it may also be necessary to reduce time smoothing
(frame averaging) for this test.

[Link] Recorded values and recommended actions

Images should be taken of both the in-air reverberation and of a uniform area of the
TETO.
Axial banding (perpendicular to the probe face) usually indicates a localised transmission/
reception fault. Any axial banding due to such a fault is unacceptable and should be
notified to the supplier.
Lateral banding (parallel to the probe face) may indicate inappropriate TGC or a beam
forming issue; the supplier’s clinical applications specialist should be able to assist in
distinguishing between these causes.

2.3.3 Functional checks

[Link] Set-up and assessment
Whilst imaging a TETO, check the function of those controls used routinely during clinical
imaging, e.g. TGC, focusing, image processing, frequency, harmonic imaging and com-
pound imaging. The effects for each system type quickly become familiar and the relative
stability of the TETO enables qualitative comparisons. Where digital image archives
are available then reference images with particular settings can be used for side-by-side
comparisons. Some functions on the system will have been offered as options; these often
require a licence key. It is common for suppliers to provide temporary evaluation keys for
all of the system options such that they can be assessed during the early months of use.
It is important to ensure that permanent licence keys are provided for all of the options
purchased. Subsequent loss of a function several months after installation may not be
noted by the users and is unlikely to be checked again after the initial acceptance.

[Link] Recorded values and recommended actions

Record reference images with particular settings for future side-by-side comparisons.
A sequence of images recorded from the standard images for each of the base user presets

11
Quality Assurance of Ultrasound Imaging Systems

will also be useful for future comparisons, especially on queries where users suggest there
have been changes in settings.
Any control that fails to operate or that produces an unexpected result should be brought
to the attention of the supplier or service agent.

2.3.4 Comparison with demonstration scanners

A TETO may also be useful in making qualitative comparisons with the demonstration
scanner on which the decision to purchase was based. Images of the available targets taken
with standard settings, e.g. a factory preset, using the same storage or hardcopy medium,
may be visually compared. This can be a useful indication that the scanner is performing
similarly to the demonstration unit and may be helpful if clinical pressures require that the
scanner is used before a full acceptance test can be undertaken. This situation is far from
ideal and not recommended but is sometimes unavoidable.

2.4 Baseline tests

These tests provide tolerance values for the results of subsequent reactive and user tests.
For broadband probes with multiple transmit modes, repeat the tests for each frequency
or setting. They may be performed as an adjunct or as part of the acceptance tests. Should
an operator suspect a problem then these tests are used as reactive tests and the results
compared to the baseline measurements.

2.4.1 Sensitivity
The following test provides a measurement that is suitable for assessment of change rather
than absolute measurement as it reflects scanner sensitivity but is not actually a direct
measure of sensitivity.

[Link] Set-up and assessment

Note that on some instruments it may not be possible to set controls exactly as described
here. Adaptation of the procedure is allowed but any departure from the standard setting
methods described here must be recorded sufficiently for settings and results to be
reproduced.
Achieve standard settings by, for example, selecting a preset appropriate to the normal
clinical use of the probe; a copy of this preset may be made and renamed QC or Test to reduce
the risk of changes. Ensure that the probe is clean and dry and is imaging in air. Reduce
the light level in the room to a normal clinical scanning level (ideally perform the
measurement when a few minutes have passed to allow the eyes to adjust).
If you have a broadband probe, start with the lowest frequency or the setting that is expected
to give the deepest penetration. Experience has shown this test to be less reproducible for
harmonic frequencies and so the test should be performed only for fundamental frequencies

12
 B-Mode imaging

(other image processing such as compounding or speckle reduction may also make the test
less reliable as may modes yet to be introduced; it is advisable to disable all such advanced
processing). Turn overall gain to maximum. Turn all TGC controls to mid-range, unless
there is little or no reverberation and noise, or the sliders do not click into position at
mid-range, in which case set TGC controls to maximum and record this for future reference.
Use a single focus set to the most superficial position. Select a scale/depth setting that
allows measurement of the full depth of probe reverberation, allowing a few centimetres
extra for possible changes, and record this setting.
Freeze the image and measure vertically from the probe surface to the deepest visible
reverberation plane in the middle third of the image (ignore reverberations at the edge of
the image) (Figure 2.2). Two factors may make this test more difficult:
(i) ignore any deeper reappearance of reverberations after they have initially
faded to background
(ii) for phased arrays, there is often a haze, rather than distinct reverberation
planes – measure to the end of this haze, observing the advice in (i).

Figure 2.2 Measurement of the depth of ‘in-air’ reverberations as a proxy for sensitivity

[Link] Recorded values and recommended actions

Record the distance from the probe surface to the deepest visible reverberation.

13
Quality Assurance of Ultrasound Imaging Systems

This measurement is the baseline for subsequent tests. The acceptable range (tolerance) for
future routine tests is ±distance to the adjacent reverberation plane.

2.4.2 Noise
[Link] Set-up and assessment
Reduce TGC to minimum over the region of reverberation in the image, so that only noise
in the distal image (where TGC should remain as set for the sensitivity test) is clearly
visible (Figure 2.3). Turn off any time smoothing/frame averaging/persistence in the
image. Reduce the light level in the room to a normal clinical scanning level (ideally
perform the measurement when a few minutes have passed to allow the eyes to adjust).
Ensuring that overall gain is initially at maximum, reduce it to the point where all noise in
the distal part of the image is eliminated (this often persists longest in the bottom corners).
Note the overall gain setting, usually shown on the display. If this is taken from the position
of the control knob, estimate how far it is (in tenths of a setting) between marked settings,
e.g. if halfway between markings 2 and 3 record 2.5, or count the number of clicks if
unmarked.

Figure 2.3 Reduction of gain to measure the noise threshold

Perform this measurement several times to establish an acceptable range. If any isolated
values are very different from the others, discard them and repeat.

14
 B-Mode imaging

An alternative method may be used where there is no ready indication of the gain setting.
Set the TGC to maximum and the gain to maximum (note that if this saturates the image,
the TGC may be reduced to a midpoint click). Set the scanning depth to maximum and turn
off frame averaging. Use a single focus set to the most superficial depth. The background
should be a haze of noise which increases with depth to a plateau. (A piece of card with a
2 cm square cut in the middle can aid in the measurement; the card is moved slowly down
the image until the noise across the square is uniform; the top of the card is then the depth
to the top of the plateau.) Measure the distance from the probe face to the plateau. Repeat
the measurement as for the previous method.
Harmonic imaging will often significantly increase apparent noise levels as the amplifier
gain settings are often much higher to compensate for significantly weaker signals.
Independent measurements may be made with and without harmonics.

[Link] Recorded values and recommended actions

Record the range of measurements (smallest and largest) as the acceptable range for future
routine tests. If the measurements are all the same, record the range as the result ±2 gain
increments, e.g. 72±2, 2.5±0.2.
For the alternative method, again record the range of measurements as the acceptable range
for routine tests. If all of the measurements differ by less than ±10% then use ±10% of the
mean measured distance.
If either of these measurements are outside these limits, then report the issue of increased
system noise to the manufacturer or service agent. Some systems have variable noise levels
as a result of local factors and may require adjustment of acceptable ranges. Such factors
can be external or mains-borne interference and local variations in mains power quality.
Interventions such as mains smoothing filters or even relocation of the system may be
required if the effects are significant. If such interventions are not possible, then the
acceptable range will need adjusting and the problem tolerated.

2.4.3 Low contrast penetration

[Link] Set-up and assessment
Image a TETO using a scanner preset that provides a uniform TGC slope, e.g. abdominal
preset, as this should be well matched to the attenuation properties of the TETO. Increase
output to maximum, turn all TGC to mid-range (if there is no mid-range click then
maximum settings may be a suitable reproducible alternative, otherwise a physical marker
may be required) and adjust overall gain to achieve a mid-grey speckle level in as much of
the TETO as possible; note the gain setting. Before freezing the image, make an assessment
of the position of the boundary between speckle and electronic noise, ignoring isolated
regions of speckle and areas under low contrast targets. Freeze the image and use the
scanner’s callipers to measure the distance from the top of the TETO to the boundary
between speckle and electronic noise only (Figure 2.4).

15
Quality Assurance of Ultrasound Imaging Systems

Figure 2.4 Low contrast penetration measurement in tissue mimicking material. Note that
the proximal cursor is placed above the top of the image at the same level as the top of the
TETO, as the first few millimetres of the image have been blanked by the scanner

[Link] Recorded values and recommended actions

Using the system callipers, record the depth of the speckle/noise boundary (it may be
necessary to freeze the image at which point the boundary may be less distinct; if this is the
case, then fix the spot for measurement by eye before freezing the image and place the first
cursor on this spot immediately). This measurement is the baseline for subsequent tests.
The acceptable reduction in low contrast penetration for reactive tests is 5% or 5 mm,
whichever is greater.

2.5 User tests

2.5.1 Image uniformity
[Link] Set-up and assessment
Operate the probe dry in air and increase the overall gain to maximum and move a single
focus to the most superficial setting. Inspect the reverberation pattern seen for axial
banding, which represents a localised transmission/reception fault. There are a variety of

16
 B-Mode imaging

possible causes. By running a narrow, smooth metal object (such as a paperclip or the back
of a key) along the probe over a very thin film of coupling gel looking for reduced echo
levels at the defective area, the magnitude of the fault may be assessed. If a defect is found
then it is useful to image a TETO or a human subject focusing on the suspect area to assist
in determining appropriate action. Note that although observation of in-air reverberation
can be useful for phased array transducers, the banding evidence will not be seen because
of the mode of operation in phased arrays. In the case of a phased array, it is an observation
of overall changes in the in-air reverberation that should raise suspicions for further
investigation.
Common causes of banding within an image are cable faults and crystal detachment.
The cable fault can sometimes be confirmed by flexing the cable and seeing the banding
shadows change. Faults may occur within the system which result in strong bands in
a uniform pattern evident both for the reverberations in air and for the TETO. These are
usually due to a fault on an image processing board within the beam former.

[Link] Recorded values and recommended actions

Record the position of any axial banding and whether this resulted in reduced echo levels
with the metal object, TETO or human subject. If axial banding is present in the image of
a TETO or human subject, the probe should be removed from use for technical assessment.
If axial banding is present, but not visibly affecting the clinical image and TETO, the
probe may be retained in use subject to frequent monitoring; users should be warned
that the Doppler beam from the affected area may produce unreliable results. Further
technical assessment of the Doppler accuracy is advised on any system used for flow
measurement.

2.5.2 Sensitivity
[Link] Set-up and assessment
Achieve standard settings by selecting the preset, TGC and scale established at baseline.
Ensure that the probe is clean and dry. Turn overall gain to maximum. Check previous
measurements and determine a frequency and setting for measurement from the range
tested at baseline. On subsequent occasions, test a different frequency/setting, so that, over
a period, all frequencies/settings are tested.
Freeze the image and measure vertically from the probe surface to the deepest visible
reverberation plane in the middle third of the image taking into account the factors detailed
previously.

[Link] Recorded values and recommended actions

Record the frequency/setting used and the distance to the deepest reverberation plane.
If the measurement is out of tolerance (determined from baseline measurements), check
the preset, TGC and overall gain settings and, if correct, perform a reactive repeat test for

17
Quality Assurance of Ultrasound Imaging Systems

all frequencies/settings. If two or more results are out of tolerance, no upgrade has been
performed and the monitor has been correctly set, a fault is likely. If the measurements
have increased, this may be due to damage to or delamination of the probe. If the
measurements have decreased, there may be an amplifier fault. In either case, contact the
manufacturer or service agent and report the fault.

2.5.3 Noise
[Link] Set-up and assessment
For the frequency/setting tested for sensitivity, ensure that overall gain is initially at
maximum. Reduce TGC to minimum over the region of reverberation in the image, so that
only noise in the distal image (where TGC should remain as set for the sensitivity test)
is clearly visible. Reduce overall gain to the point where all noise in the distal part of the
image is eliminated (this often persists longest in the bottom corners).
An alternative method may be used where there is no ready indication of the gain setting.
Set the TGC to maximum and the gain to maximum (note that if this saturates the image,
the TGC may be reduced to a mid-point click). Set the scanning depth to maximum
and turn off frame averaging. Use a single focus set to the most superficial depth. The
background should be a haze of noise which increases with depth to a plateau. (A piece of
card with a 2 cm square cut in the middle can aid in the measurement; the card is moved
slowly down the image until the noise across the square is uniform; the top of the card is
then the depth to the top of the plateau.) Measure the distance from the probe face to the
plateau.

[Link] Recorded values and recommended actions

Record the overall gain setting where noise is eliminated or the distance to the noise
plateau for the alternative test. If the measurement is out of tolerance, check the preset,
TGC and overall gain settings and, if correct, repeat the test for all frequencies/settings.
From these tests, if two or more results are out of tolerance, no upgrade has been performed
and the monitor has been correctly set, a fault is likely. A decrease in the result indicates
increased noise in the system. If the results have increased, this may be due to a reduction
in sensitivity of the scanner itself, in which case the sensitivity result will probably have
decreased. In either case, contact the manufacturer or service agent and report the fault.

18
3 Doppler
The limited availability and high price of suitable portable, commercial test-objects
designed for Doppler testing combined with a poor evidence base for the validity of such
testing have made the proscribing of appropriate tests for Doppler difficult. However, since
the publication of the previous IPEM ultrasound and Doppler quality assurance (QA)
documents (IPEM Reports 71 and 70), the majority of diagnostic ultrasound systems now
feature both spectral and colour Doppler modes. It was therefore deemed important that a
minimum set of spectral and colour Doppler tests based on the use of a simple string test
object should be included to comply with guidance on best practice for the acceptance of
new medical equipment. These tests would then also form a baseline for reference should
an investigation of Doppler performance be required. It is hoped that this will also encour-
age the incorporation of Doppler testing into QA routines and build up a base of evidence
for the future.
The basic system performance tests and basic functional checks listed in Section 3.1 below
can be carried out using a simple string test object with the requirements as described in
Section 3.2. A number of the tests may also be performed with a simple fluid flow phan-
tom. Performance tests give an indication of the system’s capabilities when measuring or
imaging flow. Functional checks are designed to simply assess whether a control is active
or not or to identify faults which should be addressed before clinical use.
It is expected that most Doppler faults will be as a result of software upgrades or from
probe faults. It is recommended that these tests are repeated after software updates and
where probes with identified minor problems in B-mode continue to be used clinically.

3.1 Tables of measurements

3.1.1 Spectral Doppler

Method Whena Notes

Basic system performance tests
Maximum velocity estimation accuracy at Simple string test ABR
50 cm s–1 (or similar) set velocity object
Spectral broadening at 50 cm s–1 (or similar) Simple string test ABR
set velocity object
Mean velocity estimation accuracy at Simple string test ABR
50 cm s–1 (or similar) set velocity object
Basic functional checks
Flow direction at 50 cm s–1 (or similar) set Simple string test ABR
velocity object

19
Quality Assurance of Ultrasound Imaging Systems

Axial range gate registration and dimensions Simple string test ABR
at 50 cm s–1 (or similar) set velocity object
Doppler gain control function response Simple string test ABR
object/free air
Pulse repetition frequency (PRF)/velocity On-screen scale ABR
range control function response display
Baseline shift control response On-screen scale ABR
display
Angle correction accuracy Simple string test ABR
object
a
A, acceptance (new scanner or probe, or repair/service affecting this parameter); B,
baseline; U, user (minimum frequency monthly); R, reactive test to evaluate system
where a fault is suspected.
All are system related tests although they may be performed on a range of transducers
particularly if different types are attached.

3.1.2 Colour/power Doppler

Basic functional checks Methods Whena Notes

Colour flow direction at Simple string test object ABR Not relevant for
50 cm s–1 (or similar) set power Doppler
velocity
Colour/power Doppler gain Simple string test object/ ABR
control function response free air
Colour PRF/velocity range On-screen colour scale ABR Not relevant for
control function response display power Doppler
Colour scale baseline shift On-screen colour scale ABR Not relevant for
control response display power Doppler
Colour/power Duplex or Simple string test object/ ABR
Triplex priority control free air
function
Colour/power box On-screen colour scale ABR
positioning and sizing display
a
A, acceptance (new scanner or probe, or repair/service affecting this parameter); B,
baseline; U, user (minimum frequency monthly); R, reactive test to evaluate system
where a fault is suspected.
All are system related tests although they may be performed on a range of transducers
particularly if different types are attached.

20
 Doppler

3.2 Test equipment

3.2.1 Simple string test object
A simple string test object should consist of a moving filament submerged in a fluid with
a speed of sound of 1540 m s–1. A full description of the rationale behind the characteristics
with some suggestions for additional functionality is given in Appendix B. The character-
istics of a simple phantom designed to allow the minimum tests and functional checks to
be carried out are as follows.
The string test object, motor assembly and tank should be designed to ensure that a
variety of transducers may be positioned to allow easy insonation of the string in the centre
of the scan plane. The filament assembly should be large enough to minimise insonation
of objects other than the filament itself and small enough to minimise the tank size and
consequent volume of fluid required. The design must ensure that the string is submerged
in the fluid at all times. Figure 3.1 illustrates a commercial string test object configuration.
Whilst the driving string is not fully submerged as recommended, the imaged string

Figure 3.1 Diagram of commercial string test object (note that although the drive belt
is not submerged as recommended, the imaged belt is, which although not ideal is an
acceptable compromise)

21
Quality Assurance of Ultrasound Imaging Systems

is. This represents a commercial compromise between the ideal of not drawing in air
bubbles from the surface and the difficulty and cost in making a watertight motor or drive
bearing.
Ideally, the filament should be in the centre of the default field of view for each transducer
type. In practice, if it is not possible to have depth adjustment, the depth should be fixed
such that the filament will appear in the deepest third of the field of the default field of view
of the highest frequency transducer likely to be tested. The filament should be constructed
from commercially available o-ring rubber with a diameter no more than 1 mm (see
Appendix B).
The filament should be mounted at 30º to the horizontal so that a vertical beam will have
a 60º beam–filament angle (Figure 3.2).

Figure 3.2 Beam filament angle

As described in further detail in Appendix B, the fluid in which the filament is suspended
should be clean, degassed and ideally have a speed of sound of 1540 m s–1. The tank should
be physically clean and lined with acoustic absorber and the position of the filament should
remain fixed with the motor and drive wheel mounted securely to minimise vibrations of
the filament velocity.
There should be the facility to drive the filament at a steady speed of around 50 cm s–1.
Other speeds and velocity profiles may allow for extended tests but are not required
within this procedure.
Where possible, the tank should be positioned on a vibration-free, horizontal surface. The
transducer should be positioned in the tank with the aid of a retort stand and clamp. Care

22
 Doppler

should be taken to avoid over-immersion of the transducer, i.e. it is not necessary to

immerse beyond the active transducer element area. Manufacturers will usually indicate
within the documentation a maximum immersion depth often relating to disinfection or
electrical safety testing.

3.3 Spectral Doppler preparation and positioning

It is recommended that the tank and filament assembly be set up and filled with the
degassed water–glycerol fluid at the start of the QA session. This will give time for any air
bubbles in the system to escape while the B-mode and other tests are being carried out. If
the string is completely submerged then driving the string during this time will help to
speed up this process.
The methods described in Section 3.4 onwards assume that the transducer is positioned
such that the filament is in the B-mode image plane at an angle of 60º to the central vertical
beam, where possible in the centre of the default field of view.
All of the visible filament should be in the centre of the elevation plane of the B-mode
image. If this is not the case then the intensity of the filament will vary along its length.
Adjust the position of the transducer relative to the filament to obtain an even, peak inten-
sity along the visible string length. If the intensity of the filament is saturating, reduce the
image gain and then acoustic power to minimise this as far as possible (Figure 3.3).

Figure 3.3 Beam filament image; (a) saturated; (b) suitable; (c) low

23
Quality Assurance of Ultrasound Imaging Systems

3.4 Spectral Doppler minimum system performance tests

Where scanners are used with spectral Doppler to provide measurements of velocity and
their associated calculations to assist with diagnosis, it is important that these measure-
ments can be made with a degree of confidence. The maximum velocity estimation
accuracy and intrinsic spectral broadening tests provide an indication of the machine
accuracy in making such measurements. Provided that the measurement has been set up
correctly, these errors should be small compared with errors due to operator positioning
and physiological variation. It is recommended that, as for B-mode testing, a preset is
selected for the tests and a copy or record of that preset made. Ideally the preset chosen
should be appropriate to the range of clinical scans the system will be used for. Any system
settings adjusted from the preset base should be recorded for each of the measurements
made.

3.4.1 Maximum velocity estimation accuracy at 50 cm s–1 (or similar) set velocity
[Link] Set-up and assessment
1. Switch on the spectral (or pulsed-wave (PW)) Doppler control and ensure
that the B-mode image is frozen.
2. Adjust the sample volume (SV) or range gate size to its minimum setting.
(Please note that on some systems sub-millimetre range gates have been
shown to give unreliable readings on string test objects and these should
also be checked on slightly larger gates.)
3. Keep the Doppler beam indicator as near vertical as possible and position
the sample volume over the image of the filament.
4. Make any further minor adjustments and stop at the position which appears
to give the highest intensity Doppler spectral signal.
5. Apply angle correction and line up with the image of the filament.
6. Adjust the spectral Doppler scale and baseline to avoid aliasing but avoid
setting the scale too high as shown in Figure 3.4a–c.
7. If the Doppler trace is saturating, reduce first the Doppler gain, then the
acoustic power to minimise saturation and display a range of greys on the
spectrum. See Figure 3.5a–c.
8. Freeze the spectrum once satisfied with its appearance and use the velocity
cursors to measure maximum and minimum velocities (the minimum
velocity values will be used in the estimation of intrinsic spectral
broadening).

[Link] Recorded values and recommended actions

Record the minimum range gate values and the gate depth and angle correction indicated
when positioned over the filament. Record five maximum velocity values (peak of the trace

24
 Doppler

Figure 3.4 Doppler scale setting; (a) scale too low, signal aliased; (b) suitable; (c) scale
too high

Figure 3.5 Doppler gain and power settings; (a) signal level too low; (b) suitable; (c)
signal saturated

25
Quality Assurance of Ultrasound Imaging Systems

ignoring noise spikes). Obtain an average maximum velocity and standard deviation from
these values.
The percentage error involved in the estimation of the maximum velocity can be
calculated using
  EstimatedMaxV elocity  
Error (%) =    − 1 ×100 (3.1)
  FilamentV elocity  
If the error indicated is greater than the manufacturer’s tolerance for the system then this
should be drawn to the manufacturer’s attention.

3.4.2 Intrinsic spectral broadening at 50 cm s–1 (or similar) set velocity

[Link] Set-up and assessment
Use the same methods as for maximum velocity estimation accuracy (Section 3.4.1).

[Link] Recorded values and recommended actions

Record the measured values, and the percentage intrinsic spectral broadening (ISB) can be
calculated using Equation 3.2.
The angle indicated is the recorded angle correction value, so for 60º, this would be ISB60.
Use the five values to calculate and record the mean and standard deviation. Note that
in fluid-based test objects, there may be additional spectral broadening from the flow
velocity profile encompassed by the sample volume.

 EstimatedMaxV elocity − EstimatedMinV elocity 

ISB(angle ) (%) =   ×100 (3.2)
 EstimatedMaxV elocity + EstimatedMinV elocity 

3.4.3 Mean velocity estimation

[Link] Set-up and assessment
Use the same general method as for maximum velocity estimation (Section 3.4.1). Activate
the mean Doppler frequency trace (e.g. Auto Doppler) on the machine, and obtain an
average value of mean Doppler frequency over a time period of about 1 second. Note
that not all scanners will have the mean velocity function in which case a visual calliper
measurement of the mid-point of the trace may be used.

[Link] Recorded values and recommended actions

Record the measured value and the type of measurement made and calculate the error in
mean velocity from Equation 3.3:
  EstimatedMeanV elocity  
Error (%) =    − 1 ×100 (3.3)
  FilamentV elocity  

26
 Doppler

If the error indicated is greater than the manufacturer’s tolerance for the system then this
should be drawn to the manufacturer’s attention.

3.4.4 Doppler noise

[Link] Set-up and assessment
Clean and dry the face of the transducer. Set the Doppler spectral trace to full screen and
with the transducer operating in free air, increase the gain to maximum. There should be
visible noise within the trace area. Reduce the gain gradually to the point where all noise
in the image is eliminated. Note the overall gain setting, usually shown on the display.
If this is taken from the position of the control knob, estimate how far it is (in tenths of a
setting) between marked settings, e.g. if halfway between markings 2 and 3, record 2.5, or
count the number of clicks if unmarked.

[Link] Recorded values and recommended actions

Record the overall gain setting where noise is eliminated. Repeat the measurement
approximately 10 times and record the minimum and maximum values as the expected
range for future checks. If there is no variation then take either ±2 steps or ±5% of the range
of gain (whichever is the greater) instead. For reactive checks, if the measurement is
outside these values then check all of the system settings carefully. If the results are out of
tolerance, no upgrade has been performed and the monitor has been correctly set, a fault is
likely. A decrease in the result indicates increased noise in the system. If the results have
increased, this may be due to a reduction in sensitivity of the scanner itself, in which case
the sensitivity result will have probably decreased. Such faults should be investigated
further in conjunction with the manufacturer or service agent.

3.5 Spectral Doppler basic functional checks

3.5.1 Flow direction at 50 cm s–1 (or similar) set velocity
[Link] Set-up and assessment
Use the same set-up as for maximum velocity estimation accuracy (Section 3.4.1). Set the
velocity scale so that the measured velocity is one-fifth of the full scale (i.e. ±125 cm s–1
for a measured velocity of 50 cm s–1) and the scale is symmetrical around the baseline.

[Link] Recorded values and recommended actions

Increase the Doppler gain and note the presence of any signal appearing in the opposite
direction to that applied. This is a crosstalk signal and should only be observed at
relatively high gain settings.
If a strong crosstalk signal is seen across a range of settings this should be reported as a
fault and the supplier or manufacturer notified.

27
Quality Assurance of Ultrasound Imaging Systems

3.5.2 Axial range gate registration and dimensions at 50 cm s–1 (or similar) set
velocity
[Link] Set-up and assessment
Use the same set-up as for maximum velocity estimation accuracy (Section 3.4.1). Set the
sample volume to a mid-point size (usually around 3 to 8 mm) and record this value. On
the B-mode image, measure the depth from the probe face along the Doppler sample line
to the point of maximum intensity of the string. Adjust the sample volume position in an
axial (vertical) direction around the B-mode image of the filament until the strength or
intensity of the detected Doppler signal is maximised, and measure the distance from the
probe face to the centre of the range gate (the angle correction line can be used to indicate
this point).
Adjust the sample volume position in an axial (vertical) direction around the B-mode
image of the filament and note the maximum and minimum range gate depths at which the
Doppler signal is still present.

[Link] Recorded values and recommended actions

The measured depth for the maximum intensity of the B-mode image of the filament
should coincide with the measured depth for the centre of the sample volume at peak
Doppler intensity (see Figure 3.6).
If this position does not coincide with the maximum intensity of the B-mode image, further
investigations should be made at several depths and with a number of range gate sizes. The
difference between the two depths should be recorded for each case. If the distance is
consistently more than one-quarter of the range gate width, this should be reported as a
fault to the manufacturer or service agent. Otherwise users should be made aware of the
error and monitor any clinical impact. (The strong echo from a string test object is an
atypical Doppler signal and it is recommended that smaller errors are investigated for
clinical impact before reporting a fault.)
The difference between the two extreme gate depths should be recorded as an estimation
of the axial sample volume dimensions. The magnitude of the axial sample volume dimen-
sion should be recorded as an indication of the spatial resolution of the spectral Doppler.
It is not unusual for this measurement to produce values for range gate dimensions which
are generally 20–30% larger than indicated due to the strength of the Doppler signal.
The size indicated should therefore be considered as a proportional rather than absolute
measurement. The recorded value should be used as a reference for queries with regard to
Doppler performance.

3.5.3 Doppler gain control function response

[Link] Set-up and assessment
Adjustment of the spectral Doppler gain control should produce noticeable even stepped
changes in the intensity of the spectrum or waveform from the moving filament. Increasing

28
 Doppler

Figure 3.6 Sample volume positioning

the gain with the transducer operating in free air should produce a noise pattern of
increasing quantity and intensity.

[Link] Recorded values and recommended actions

If the operator detects erratic steps in gain or non-response of the control, then this should
be reported as a fault to the supplier or manufacturer.

3.5.4 Pulse repetition frequency (PRF)/velocity range control function response

[Link] Set-up and assessment
The PRF or velocity scale control should be adjusted up and down through its full range.
This should be repeated with and without the B-mode image frozen as the highest velocity
scales are usually only possible in full spectral Doppler mode not duplex.

[Link] Recorded values and recommended actions

The upper and lower limits of the scale maximum should be recorded both with and without
B-mode duplex imaging.

29
Quality Assurance of Ultrasound Imaging Systems

The control should operate smoothly within the normal upper and lower limits. If it does
not then this should be regarded as a fault.

3.5.5 Baseline shift control response

[Link] Set-up and assessment
Select a mid-point velocity range with a spectral Doppler image from the string. Adjust the
baseline of the spectral display to its maximum and minimum values. It should be possible
to adjust the baseline of the spectral display to the maximum and minimum of any of the
PRF or velocity scale settings. Often this control is not continuous but changes in steps of
around 10% of the scale.

[Link] Recorded values and recommended actions

The control should operate smoothly between the maximum and minimum limits of the
scale set. If it does not then this should be regarded as a fault.

3.5.6 Angle correction accuracy

[Link] Set-up and assessment
The on-screen angle correction indicator should be adjusted such that it is aligned with the
filament and the image is then recorded either by printing or digital storage. Measure the
angle on this image.
Change the angle correction through the full range available; the indicator should move in
small, evenly spaced steps. Some systems have a fixed number of larger steps (usually
small portable or low end systems) such that only an approximation to the actual angle is
possible.

[Link] Recorded values and recommended actions

The measured beam–filament angle on the recorded image should be recorded and com-
pared with the set angle indicated on the screen and the difference recorded. If an angle
difference of greater than ±2º is seen, then the position of the probe should be checked
before reporting this as a fault.
Any change from the normal operation or increase in the angle difference outside of estab-
lished normal ranges of the angle correction control should be reported as a fault to the
supplier or manufacturer.

3.6 Colour and power Doppler basic functional checks

The string test object is not ideal for colour and power Doppler testing as the signal from
the filament is several orders of magnitude higher than the low echo-level signal from
blood (or from blood mimics as used in a flow phantom). This can result in excessive

30
 Doppler

Figure 3.7 Colour Doppler signals; (a) signal loss and bleeding; (b) acceptable

colour ‘bleeding’ and/or signal loss at and around the filament (Figure 3.7a). However, it
may be possible to minimise gain (B-mode and colour) and acoustic power controls to
obtain an acceptable colour map over the moving string (Figure 3.7b) which will allow
some basic functional checks to be carried out.

3.6.1 Colour flow direction at 50 cm s–1 (or similar) set velocity

[Link] Set-up and assessment
Set the colour scale to around 10 to 20 cm s–1 higher than the set velocity. Depending on
the set-up of the scanner pre-set, the filament should appear with either red or blue tones.
Reverse the filament direction if possible or reverse the orientation of the transducer. The
colour should also be reversed. Also check the function of the ‘invert’ control.

[Link] Recorded values and recommended actions

If the filament fails to show the correct colour for the flow direction then this should be
regarded as a fault and reported to the supplier or manufacturer.
Note that this test does not apply for normal power Doppler with no flow direction.

3.6.2 Colour/power Doppler gain control function response

[Link] Set-up and assessment
Increasing the colour Doppler gain control should produce a noticeable increase in
the displayed intensity of the colour map of the moving filament with a ‘bleeding-over’ of
the colour and increasing random colour noise in the image. Increasing the gain with the
transducer operating in free air will produce a colour noise pattern of increasing quantity
and intensity.

31
Quality Assurance of Ultrasound Imaging Systems

[Link] Recorded values and recommended actions

The control should operate smoothly with a gradual increase in intensity. Noise spikes
which appear as the control is moved may indicate a problem as will erratic steps in gain.
These should be reported to the supplier or manufacturer.

3.6.3 Colour PRF/velocity range control function response

[Link] Set-up and assessment

The colour velocity scale control should be adjusted up and down with the bottom of the
colour box at a set depth and the upper and lower limits of the scale maximum noted.

[Link] Recorded values and recommended actions

The depth of the base of the colour box, and upper and lower limits of the scale maximum
should be recorded.

The control should operate smoothly with a gradual increase in velocity scale. On many
systems, this will be in steps usually increasing the scale limit by 20–30% for each step.

3.6.4 Colour baseline shift control response

[Link] Set-up and assessment

Adjust the baseline of the colour bar to its maximum and minimum value. It should be
possible to adjust the baseline of the colour bar to the maximum and minimum of any of
the colour velocity scale settings.

[Link] Recorded values and recommended actions

The control should operate smoothly between the maximum and minimum limits of the
scale set. If it does not then this should be regarded as a fault.

3.6.5 Colour/power box positioning and sizing

[Link] Set-up and assessment

For the chosen preset, use the default field of view and check that the colour box may
be expanded to cover the whole of that field of view. For linear array probes, it is often
possible to have the box sides either parallel to the imaging beams or steered at an angle
to them. With any steering turned off (i.e. a ‘centred’ box), minimise the box size and note
the axial and lateral dimensions with the box positioned both at the probe face and at
maximum depth.

32
 Doppler

[Link] Recorded values and recommended actions

Measure and record the horizontal and vertical dimensions of the minimised box when it
is positioned at the top and bottom of the field of view. Note any ‘dead-zone’ areas where
the box cannot be positioned.
If the box positioning is abnormally restricted or the control fails to move or change the
box size in smooth and even steps, then there may be a fault with the system and this should
be reported to the supplier or manufacturer.

33
4 Safety
There are three aspects of safety considered in this section: the safety of the ultrasound
field exposure, the electrical safety of the system and the physical safety of the system.
Ultrasound used prudently has been identified as a generally safe method of providing
diagnostic image information, although potential risks have been identified and an ‘as low
as reasonable achievable’ (ALARA) principle is recommended when using ultrasound
(WFUMB, 1998). The recent report by the Health Protection Agency independent
Advisory Group on Non-ionising Radiation on Health Effects of Exposure to Ultrasound
and Infrasound (AGNIR, 2010) continues to support the approach taken by the WFUMB.
The current consensus is that, at high levels, thermal and cavitation effects will cause harm
and for medical uses of ultrasound, measures are in place to ensure that these harmful
levels do not occur. The British Medical Ultrasound Society (BMUS) has issued guidelines
for the safe application of diagnostic ultrasound (BMUS/BIR, 2009). In this guidance, two
safety indices are identified, which as a result of Food and Drug Administration (FDA)
regulations are displayed on the majority of systems. The BMUS guidance refers to many
factors which contribute to the safe use of ultrasound equipment but specifically refers to
the use of these indices to regulate and monitor exposure and to apply constraints where
significant risks have been reported in the literature. The risks are identified in the guid-
ance and include: Potential for cavitation at mechanical index (MI) values of 0.7 and above
with a lower limit applied for sensitive gas-containing organs of 0.3; Potential for thermal
damage at temperature rises of 4°C for over 5 minutes which may occur from thermal
index (TI) values of 2 upwards. Higher temperature rises cause damage in a shorter time
interval; for each 1ºC rise in temperature, the advice is to reduce the exposure time by a
factor of four. Other more specific constraints apply for sensitive organs and tissue groups.
These factors clearly show the importance that is placed on the indices displayed on ultra-
sound systems. Checking these indices to ensure that they are displayed, seem to behave
appropriately and to identify the maximum values is an important process in testing an
ultrasound system.
The guidance given in this chapter focuses on tests that are relatively straightforward to
carry out and where the results are obtained directly or require only simple computation.
More complex testing methods are described in a later chapter. In assessing a system
against the recommended actions, allowance should be made for local measurement
uncertainties which will depend on the specific test equipment used.

34
 Safety

4.1 Table of measurements

Test Method Whena Notes
Mechanical Visual assessment A/R
index (MI)
Thermal index Visual assessment A/R
(TI)
Acoustic Measurements of A/R Equipment operating at higher
power acoustic power acoustic output levels
TI (imaging Calculation from A/R Centres with appropriate
mode) acoustic power facilities
measurement
Probe in air Using radiometer or A/R Routinely on probes and at
temperature thermocouple settings producing significant
heating
Physical Visual inspection A/B/P/U Normally this should be recorded
integrity (probes) weekly but informal inspections
(probes) should be more frequent
Physical Visual inspection A/B/P/U Normally this should be
integrity (console) recorded quarterly but informal
(console) inspections should be more
frequent
Electrical Measurement of A/P According to local medical
safety checks electrical safety equipment testing policy
(console) according to local
testing procedure
Electrical Measurement of A/P When identified by visual
safety checks electrical safety inspection for all probes. As
(probes as currents and specified by the manufacturer for
patient applied insulation intra-cavity or intra-
part) operative probes
Electrical Measurement of A/P According to local medical
safety checks electrical safety equipment testing policy
(ECG applied currents and
part) insulation
a
A, acceptance (new scanner or probe); B, baseline; R, reactive (this should be
included any time a software update is implemented); P, planned (minimum frequency
biannual); U, user (minimum frequency monthly).
All except the physical integrity (console), Electrical safety checks (console) and
electrical safety checks (ECG applied parts) are transducer related tests.

35
Quality Assurance of Ultrasound Imaging Systems

Tests type P are the reactive tests performed as part of a planned preventative maintenance
visit.
Electrical tests may additionally be required at the completion of certain types of service
work on the system or where physical problems with the system have been observed.

4.2 Acoustic safety test equipment

4.2.1 Acoustic force balance
Acoustic power measurement is based on the principle that an acoustic beam exerts a
radiation force on a target when it is absorbed or reflected. This effect has led to the devel-
opment of acoustic force balances where a suitable target is insonated by the acoustic beam
under investigation. The force exerted by a beam of ultrasound is actually very small and
a highly sensitive device is required to measure it.
Over the years, many designs for acoustic force balances have been published in the
literature and a number of highly sensitive devices have successfully been constructed.
Currently there are three devices available commercially which allow measurement of
acoustic power in the diagnostic range. One device has a reflecting target and the others
absorbing targets and each has strengths and weaknesses. It is doubtful whether these
devices have the sensitivity required to measure acoustic power accurately at the lowest
end of the diagnostic power range, but each may be used with ease at powers from several
tens of mW upwards. Further details on what to look for in an acoustic force balance are
given in Appendix C.
The National Physical Laboratory has produced a Best Practice Guide to Measurement
of Acoustic Output Power (NPU, 2008). Additional information on radiation force
balances can be found in IPEM Report 84 (IPEM, 2001) which deals primarily with
power measurement in physiotherapy ultrasound beams.
An aperture limiting mask or slit is required for use with an acoustic force balance to
measure bounded acoustic power (power per unit length). An appropriate design is recom-
mended in IEC 62359 (IEC, 2005a) and is summarised in Appendix C. Bounded acoustic
power is required in the calculation of the thermal index (TI) for imaging beams.
Use of a force balance will often require the following items of additional equipment:
aperture limiting slit (see Appendix C), clamp stand, coupling gel.

4.2.2 Radiometer
A radiometer is a non-contact device that detects and measures infrared radiation emitted
by an object. The technique provides a direct measurement of the surface temperature
of an ultrasound transducer. Although the measurements are simple to make, there are
several potential small sources of error.

36
 Safety

[Link] Spatial averaging

A radiometer measures the average temperature within its field of view and the smaller the
detector diameter, the smaller the area over which the temperature is averaged. There are
two issues associated with spatial averaging. First, the diameter of the radiometer field of
view (FOV) must not exceed the smallest dimension of the transducer. If it exceeds the size
of the transducer then the measured temperature will be inaccurate because of the inclusion
of lower temperature background regions. Second, the field of view of the radiometer
needs to be smaller than the active area of the transducer to measure the maximum tem-
perature accurately. Thermal conduction effects mean that the temperature distribution
rapidly becomes fairly uniform over the active elements of the transducer and measure-
ments made within this area should accurately reflect the average maximum surface
temperature.

[Link] The transducer surface

The theory of infrared thermometry applies to black-body radiators. The surface of an
ultrasound transducer, although matt, is not a black-body equivalent. This results in an
error of <5% in temperature estimation (°C), for which an emissivity correction can be
applied. Emissivity data are often supplied and can be found in published literature.

4.2.3 Hydrophones
Hydrophone measurements are not included in this chapter due to their increased level of
complexity. However, a number of safety related measurements require a hydrophone, for
example, MI and TI in pulsed Doppler mode and M-mode. With a hydrophone it is also
possible to check the local intensity of the beam against upper limits set by the Food
and Drug Administration (FDA) in the USA. For this reason, centres with appropriate
resources may wish to carry out hydrophone measurements. Chapter 5 describes a range
of appropriate tests.

4.3 Acoustic safety acceptance measurements

4.3.1 Visual check of mechanical index (MI)
Figure 4.1 shows an ultrasound scanner displaying the safety indices MI and TI. Where
these indices are displayed, they should provide valid information to the user about the
ultrasonic exposure received by the patient (European Medical Device Directive). Tests of
varying levels of complexity may be carried out to check this. In addition, comparison
should be made between maximum displayed values and published safety guidelines
(BMUS/BIR, 2000). Users should be aware that any item of equipment for which the
safety indices never exceed limits defined by international standards is not required to
display the values.
A visual check of the performance of the on-screen indication of the MI is carried out
for each mode of scanning and each transducer. The aim of the test is to ensure that the
displayed MI tracks changes in the acoustic beam and is within acceptable limits.

37
Quality Assurance of Ultrasound Imaging Systems

Figure 4.1 Values of MI and TI displayed on an ultrasound scanner

When MI is displayed, it is important to:

• check that displayed MI varies as the equipment settings are altered
• determine the maximum MI value and the settings associated with this
value
• inform the user of any settings producing excessive values of MI.

[Link] Set-up and assessment

For each transducer and mode in turn:
1. Check whether MI is displayed on the screen.
2. Vary the equipment settings and note that MI varies (it will not necessarily
vary for every change in each parameter value).
3. Find the maximum displayed value of MI by varying the operating condi-
tions systematically until the maximum value is identified.
4. Note this value, together with the corresponding equipment settings.
Note that some systems may adjust parameters other than the one being changed by the
operator to constrain levels to those indicated by the FDA or for image optimisation. This

38
 Safety

form of linked setting adjustment is becoming increasingly common in complex software

driven digital scanners.

[Link] Recorded values and recommended actions

For each transducer, the maximum value of MI displayed is recorded together with
full details of the operating conditions. A similar set of data should be recorded for each
clinical setting used.
The displayed MI should be in the range 0 to 1.9 for all machine settings.
If the maximum displayed MI > 1.9 at any point, then the user and manufacturer or
supplier should be notified. The machine can continue to be used as long as this setting is
avoided.

4.3.2 Visual check of thermal index (TI)

A visual check of the performance of the on-screen indication of TI is carried out for each
mode of scanning and each transducer. The aim of the test is to ensure that the displayed
TI tracks changes in the acoustic beam and is within acceptable limits. For ultrasound
equipment on which TI does not exceed the limit set by international standards, an
on-screen display is not required.
Where TI is displayed, it is important to:
• check that the displayed TI varies appropriately as the equipment settings
are altered
• determine the maximum TI value and the settings associated with this
value
• inform the user of any settings producing excessive values of TI.

[Link] Set-up and assessment

For each transducer and operating mode in turn:
1. Check whether TI is displayed on the screen.
2. Vary the equipment settings and check that the displayed TI changes in an
appropriate way, e.g. it increases with increasing acoustic power, focal
depth and frame rate.
3. Find the maximum displayed value of TI by varying the operating condi-
tions systematically until the maximum value is identified.
Note that as for MI values, some systems may adjust parameters other than the one being
changed.

39
Quality Assurance of Ultrasound Imaging Systems

[Link] Recorded values and recommended actions

For each transducer, frequency and mode, the maximum value of TI displayed is recorded
together with full details of the operating conditions giving rise to this value. If desired, a
similar set of data may be recorded for clinical settings. The user should be informed of
conditions resulting in high values of TI.
The displayed TI should be in the range 0 to 3 for all machine settings.
If the maximum displayed TI > 3 at any point, then the user and manufacturer or supplier
should be notified. The machine can continue to be used as long as this setting is
avoided.

4.3.3 Measurement of TI
Clinical applications for which TI should be measured at acceptance and reactively include
obstetric and neonatal scanning and ophthalmology. The case for measuring TI for scanned
ultrasound beams is a strong one because of the high number of routine obstetric scans
carried out in the UK. For scanned modes, the location of maximum heating is always near
to the surface and the following equation is used to calculate the index:
w 01f c
TI = (4.1)
210
Here w01 is the bounded acoustic power (expressed in mW) at the source, i.e. the power
emitted from the central 1 cm of the active aperture and fc is the centre frequency (acoustic
working frequency in MHz).
TI can also be calculated for other modes using more complex equations (AIUM/NEMA,
1992, 2004a,b; IEC, 2005a). The TI calculation depends on the location of maximum heat-
ing, which in turn generally depends on the operating mode and the target. As a result, the
on-screen display on an ultrasound scanner may give values of TIS (soft tissue), TIB (bone
at focus) or TIC (bone at surface). Different equations are used for calculating TI in differ-
ent modes of operation (IEC, 2005a). Only imaging will be considered here. Imaging of
TIS and TIB are both given by Equation 4.1. The result of the calculation of TIS can be
used to check the accuracy of the value displayed on the screen for a range of settings.
The equipment required for measurement of TIS is an acoustic force balance with a 1 cm
aperture limiting mask. Appendix C gives detailed information about acoustic force
balances available commercially. Manufacturers of these balances claim a sensitivity of
10 mW, which makes such devices suitable for higher acoustic power conditions. At
3.5 MHz, an output level resulting in TI>0.2 should be measurable.
The nominal acoustic working frequency of the transducer (in MHz) must also be known.
On some equipment, the acoustic working frequency is displayed on the screen and may
alter with mode or be directly selectable by the user. On simpler systems, the frequency is
sometimes marked on the probe or the connector. If there is no indication of the working

40
 Safety

frequency on the equipment, the manufacturers’ technical specification or supplied

acoustic output information should be consulted.

[Link] Set-up and assessment

1. The acoustic balance should be set up according to the manufacturer’s
instructions, with the addition of an aperture limiting mask. Locate the
balance on a firm, level surface. Allow sufficient time before use for fluid
movement to cease. Ensure environmental draughts and vibrations are kept
to a minimum.
2. Mount the transducer so that it is aligned with the target and coupled to the
balance. Coupling may be into water for an open topped balance or through
an acoustic window using coupling gel. The transducer alignment should
be checked and adjusted carefully by imaging the target. Small errors in
alignment will result in a significant reduction in measured acoustic power
with a device employing a conical target.
3. For each transducer and mode, select ultrasound equipment settings which
generate the maximum TIS. This process is complex and time consuming
and requires experience. There are some controls such as power setting
which are easy to determine but as stated earlier, some systems link
controls, and settings which may be presumed to increase TI, will not, as
a result of these linked adjustments. Experimentation with the general
controls does enable a systematic approach to be developed although new
modes and controls may require continual modification of the process.
4. Measure bounded acoustic power and use the nominal centre frequency to
calculate TI.
5. Repeat measurements if necessary, after optimising equipment settings to
obtain maximum acoustic output.
6. Measurements may also be made using clinical settings.

[Link] Recorded values and recommended actions

Measured TI values are recorded and compared with displayed TI values for a range of
settings allowing for the accuracy of power measurement – radiation force balances should
achieve an accuracy of ±10% and this should be verified locally.
If the calculated TI>3, then the user should be informed and the conditions recorded. Users
should be made aware of the setting conditions where these levels may occur such that they
can adjust their clinical practice to comply with safety guidelines.
If any measured TI corresponds to a system displayed TI outside of the uncertainty of the
method, then the manufacturer should be informed.

41
Quality Assurance of Ultrasound Imaging Systems

4.3.4 Measurement of surface temperature of transducer

The surface temperature of a transducer is of particular concern for ophthalmology,
neonatal ultrasound, interstitial transducers and any clinical applications involving tissue
where the sensitivity to temperature is compromised. The images in Figure 4.2 were taken
using a thermal imager and show that the size of the heated area on the transducer face
depends on the operating conditions.

Figure 4.2 Thermal images of a transducer face showing the heating pattern generated
when operated in air under different focal conditions

International standards (IEC, 2007a) limit the surface temperature of an ultrasound trans-
ducer operating in air by requiring that the temperature of the transducer face does not
exceed 50°C after 30 minutes and that the temperature rise after 30 minutes is less than
27°C. A pragmatic approach is proposed here which allows a series of measurements to be
made in a reasonable time. Experience has shown that a temperature measurement made
after 3 minutes gives an acceptable indication of the maximum temperature, since the
initial heating rate is very rapid. If the surface temperature exceeds 42°C or the tempera-
ture rise is greater than 20°C, after 3 minutes, then a full 30-minute test is carried out. This
requires a repeat of the measurements with the temperature recorded after 30 minutes. It
is beneficial during this to record the values at 3-minute intervals such that the heating
profile may be established and appropriate guidance on safe use given.

[Link] Set-up and assessment

An appropriate radiometer is selected. The radiometer is positioned 1 cm in front of
the centre of the transducer face in a draught-free environment. It should be supported
horizontally in a clamp as shown in Figure 4.3.
Measurements are made at 3-minute intervals and a suitable clock or stopwatch is required.
Carry out the following steps for each transducer and each mode in turn.

42
 Safety

Figure 4.3 Measurement of transducer surface temperature using a radiometer

1. Select the equipment settings which give maximum power or maximum

value of displayed TI.
2. Turn the transducer output off and allow the transducer to cool to ambient
temperature.
3. When the transducer has cooled, record the starting temperature.
4. Turn the transducer output on and start the timer.
5. Record the temperature of the transducer 3 minutes after switch-on.
6. Turn the transducer output off.
7. Calculate the temperature rise during the 3-minute period.
8. Allow the transducer to cool to ambient temperature between consecutive
measurements (this may take a long time).
9. Each time the equipment settings are altered, ensure that the transducer
output is turned off and the transducer is allowed to cool.

43
Quality Assurance of Ultrasound Imaging Systems

[Link] Recorded values and recommended actions

Record initial and final temperatures and calculate the temperature rise.
If the temperature rise 3 minutes after switching on a transducer exceeds 20°C or the sur-
face temperature is greater than 42°C, 3 minutes after switch on, then a 30-minute test
should be carried out. This requires a repeat of the measurements with the temperature
recorded after 30 minutes. It is beneficial during this to record the values at 3-minute
intervals such that the heating profile may be established and appropriate guidance on safe
use given.
A surface temperature in excess of 50°C, 30 minutes after switching on the transducer or
a temperature rise of greater than 27°C, 30 minutes after switch on exceeds IEC standards
(IEC 2007a). This information should be reported to the user and manufacturer, together
with information about the settings resulting in these conditions. The transducer should be
removed from use or appropriate precautions applied to prevent such a temperature rise
before scanning.

4.4 Acoustic safety routine/reactive measurements

The tests listed below are performed to obtain baseline values and repeated for routine
assessment. Many of these tests are similar to those carried out at acceptance. For these
tests, detailed descriptions of test methods are given in Section 4.3.

4.4.1 Visual check of MI

[Link] Set-up and assessment
Routine tests are carried out using the method described in Section 4.3.1 and using identi-
cal equipment settings. A change in displayed MI may arise if the software is upgraded and
this cause should be eliminated before further investigations are carried out. After an
upgrade, it is appropriate to recheck selected transducers under the operating conditions
determined in Section 4.3.1 above. If a change in MI is found, the new maximum MI
values and settings should be determined and recorded for all transducers and modes.

[Link] Recorded values and recommended actions

For each transducer, the maximum value of MI displayed is recorded together with full
details of the operating conditions.
If the maximum displayed MI>1.9 at any point, then the user and manufacturer or
supplier should be notified. The machine can continue to be used as long as this setting is
avoided.
If the maximum displayed MI has increased then the user should be informed of these
conditions and of the new maximum.

44
 Safety

4.4.2 Visual check of TI

[Link] Set-up and assessment
Routine tests are carried out using the method described in Section 4.3.2 and using identi-
cal equipment settings. A change in displayed TI may arise if the software is upgraded and
this cause should be eliminated before further investigations are carried out. After an
upgrade, it is appropriate to recheck selected transducers under the operating conditions
determined in Section 4.3.2 above. If a change in TI is observed, the new maximum TI
values and settings should be determined and recorded for all transducers and modes.

[Link] Recorded values and recommended actions

For each transducer, the maximum value of TI displayed is recorded together with full
details of the operating conditions.
If the maximum displayed TI>3 at any point, then the user and manufacturer or
supplier should be notified. The machine can continue to be used as long as this setting is
avoided.
If the maximum displayed TI has increased, then the user should be informed of these
conditions and of the new maximum.

4.4.3 Measurement of TI
[Link] Set-up and assessment
Using the set-up described in Section 4.3.3, routine quantitative measurements of TI can
be made for each transducer under a variety of scanner settings such as those that produce
maximum TI or the settings that are used clinically. A range of standard settings should be
established as a baseline for which TI is evaluated. In general, scanner settings which
produce maximum acoustic power will also give rise to maximum TI. Hence an iterative
process can be used to identify these settings using an acoustic force balance.
For each transducer operating in each scanning mode and each set of baseline operating
conditions, the following two-stage procedure is followed:
1. The acoustic power is measured (see Section 4.4.5) through an appropri-
ately designed 1 cm slit.
2. Using knowledge of the nominal frequency, a TI value is calculated for each
set of operating conditions selected.
It may be decided that this measurement is not required routinely. However, if a change in
displayed TI is found during a visual check, it is appropriate to recheck the measured TI
under the standard operating conditions. After upgrade, the TI should be measured using
equipment settings which produce maximum acoustic power and maximum displayed TI
to obtain new baseline values.

45
Quality Assurance of Ultrasound Imaging Systems

[Link] Recorded values and recommended actions

Measured TI values are recorded and compared with displayed TI values for a range of
settings allowing for the accuracy of power measurement – radiation force balances should
achieve accuracy of ±10% and this should be verified locally.
If the calculated TI>3, then the user should be informed and the conditions recorded.
If any measured TI corresponds to a system displayed TI outside of the tolerance of the
method, then the manufacturer should be informed.

4.4.4 Surface temperature of transducer

[Link] Set-up and assessment
The measurement equipment and method of measurement are described in Section 4.3.4.
The equipment settings required for this measurement are those which produce maximum
acoustic power (if measured), or the maximum value of displayed TI. Hence, the baseline
temperature values are measured under worst-case conditions. An increased temperature
compared with baseline should be investigated as it could arise from a fault condition or a
software upgrade resulting in changed output.

[Link] Recorded values and recommended actions

Baseline: Record the starting temperature and the final temperature for each transducer,
mode and set of conditions. Calculate and record the temperature rise in each case.
Routine: Record the starting temperature and the final temperature for selected transducers
and modes under standard conditions. Calculate and record the temperature rise in each
case.
If after 3 minutes, the surface temperature is greater than 42°C or the temperature rise
exceeds 20°C, then, as for the acceptance test (Section 4.3.4), repeat the measurements
with the temperature recorded at 3-minute intervals up to 30 minutes.
A change in maximum temperature or a temperature rise of more than 3°C may arise as a
result of a software upgrade. If an upgrade has occurred since previous measurements,
record new baseline values for the transducers affected.
If no software upgrade has occurred, the change in heating may indicate a fault condition
and should be investigated further.

4.4.5 Acoustic power

Measurement of acoustic power provides fundamental information about an acoustic beam
producing an image, and so is to be encouraged (Figure 4.4). Detailed information about
appropriate measurement equipment is given in Appendix C.

46
 Safety

Figure 4.4 Example of a commercial power balance employing an absorbing target

(courtesy of Precision Acoustics, Dorchester, UK)

[Link] Set-up and assessment

The first issue to be considered is the baseline settings which will be used for the test. The
user can decide whether to select the acoustic power for typical clinical settings or the
maximum acoustic output power available as the baseline value. Baseline values of acous-
tic power should be measured for each transducer in each operating mode under known
conditions (where it is within the capability of the measurement device to do so). In each
case, it is imperative to record the details of the scanner settings associated with each
measurement. Subsequent measurements are compared with baseline values and signifi-
cant changes in acoustic output are reported. The machine settings required for baseline
values of acoustic power have to be determined. The conditions selected may be either
those preset for clinical use or the conditions producing maximum acoustic output.
The following procedure should be followed to determine the equipment settings that
result in maximum acoustic power output.
1. For all modes, the transmit power should be set to maximum, i.e. 0 dB or
100% as appropriate. In general, the equipment should initially be set up

47
Quality Assurance of Ultrasound Imaging Systems

with the maximum frame rate selected, the maximum image depth and the
focal zone located as far from the transducer as possible.
2. The effect on measured acoustic power of altering the various set-up
controls is then observed and an iterative process followed to maximise
the acoustic output. For pulsed Doppler mode, the effect of varying the
sample volume and gate (or pulse repetition frequency) should also be
investigated.
3. The settings producing maximum acoustic power are recorded. The impor-
tance of recording all of the set-up conditions affecting the acoustic output
cannot be over-emphasised.
Once the baseline conditions have been determined for a particular transducer and mode,
a precise measurement can be made:
1. Select all of the required equipment settings.
2. Freeze the image, ensuring that the ultrasound is off, and zero the balance.
3. Turn the ultrasound on and allow the reading to reach a maximum. Around
10 seconds should be sufficient (do not allow slow drift upwards to continue
if observed).
4. Switch ultrasound off and check the zero again.
5. Repeat if necessary.
Repeat for each transducer, frequency and mode.

[Link] Recorded values and recommended actions

For baseline values, the measured acoustic power is recorded for all transducers, frequen-
cies and modes together with the associated scanner settings.
For routine testing, the acoustic power is measured and recorded for selected transducers,
frequencies and modes using the baseline settings. Values are compared with baseline. A
variation of more than 20% may arise if the software is upgraded in the interim and this
cause would need to be excluded before further action is taken.
A variation from baseline of greater than 20% without software upgrade should be consid-
ered suspicious and confirmed by further tests and the supplier or manufacturer should be
notified.

4.5 Electrical safety and physical integrity acceptance, routine and

baseline testing
The electrical and physical integrity tests on systems do not vary between acceptance,
routine and baseline testing. In fact, the baseline tests recorded will be the tests at the

48
 Safety

first instance of checking the system which should normally be the acceptance test but
may on occasion be the first routine test. The electrical tests are such that repetition
may cause damage to the system and baseline values should be established from a single
measurement.

4.5.1 Physical integrity

The inspection of the physical integrity of an ultrasound system is an important part of a
recorded quality assurance programme as many systems are large and complicated with a
number of cables, connectors, panels, plugs and switches. Identifying early deterioration
for any of these items may prevent an expensive repair bill or more importantly ensure the
system is available for scanning when required.

[Link] Set-up and assessment

The system should be inspected ideally in a room with good access to all sides and with
natural light. Alternatively, the room must have a good source of even, bright light. For
some parts of the inspection, a torch may be required to see effectively (particularly for
inspections of wheels and brakes).
The console visual inspection should be performed in the following stages to ensure that
all parts have been covered. Any gel on the system should be cleaned and dried before
further tests are undertaken. The controls and switches should be checked for signs of abra-
sion or wear. Non-sealed knobs, sliders or switches are also inspected for ingress of scan-
ning gel. The panels and cover plates should be soundly fixed with no signs of physical
damage. The ergonomic adjustments, for example, monitor height and lateral movement,
keyboard height and orientation, should be checked to ensure that movements are smooth,
easy and stable when fixed in place. Digital photographs of damaged areas should be used
to record status for future reference.
The probe usually consists of a hard plastic body and a softer membrane material for the
active scanning face. If there is any gel on the probe then this should be cleaned and dried
before further tests are undertaken. Both of these should be inspected for damage. The
membrane is prone to splits and cuts and to lifting or bubbling. The splits and cuts may be
difficult to see if they have been caused by a sharp object such as a scalpel or syringe.
Particular attention needs to be paid to the edges of the membrane where it is bonded
to the hard probe casing usually by a filler or adhesive material which may peel or split.
Applying a light pressure to the probe face and observing the bounce back may reveal
bubbles or lifting of the membrane. The hard casing should be inspected carefully
for hairline cracks in the body. If physical damage is seen which may compromise
the electrical integrity, then this should be tested as a patient applied part, see Section
[Link].
The cables attaching the main console and any attached peripherals should be inspected
for signs of damage. The cables should be examined visually for cuts, abrasions, twisting
or deformation of outer sleeves and for signs of stress (often seen as a change in shade or

49
Quality Assurance of Ultrasound Imaging Systems

colour of the outer sleeve). A tactile inspection of the cable may reveal damage or stress
internal to the cable and is also advised. Particular attention should be paid to the area
around the strain relief where the cable enters the probe or connector (see Figure 4.5). At
this point, kinks or damage to the external sheath often result in signal cable damage soon
after. When probes are removed from the system (the probe should not be removed purely
for inspection), the connectors should be checked for physical damage. The pins in the
connectors should be visually checked for alignment. Please note probes should never be
removed when active.

Figure 4.5 Probe cable strain relief and early signs of cable sheathing failure

If the console unit is moved routinely, then the brakes should be tested for functionality.
Each brake should be operated in turn and each should prevent movement of the system.
Where the wheels swivel and/or have direction locks, these should also be checked for
smooth and effective operation.

[Link] Recorded values and recommended actions

When performing acceptance testing, systems should show no signs of physical damage
or wear and all physical movements and adjustments should operate as expected. Any
defects should be reported to both the user and the manufacturer or supplier.
There is no specific difference between these acceptance tests and tests for baseline unless
the baseline tests are being performed on an older system where the initial state may
involve recording existing flaws. It is recommended that a digital camera is used to record
the initial state of the system and subsequent changes.
When performing subsequent routine testing, any damage or functional problems with the
system are recorded. New problems or existing ones showing further deterioration should
be recorded and, where possible, digital images taken.
There are no specific recommended actions for physical integrity problems as their impact
will very much depend on the use made of the system and the extent of the damage. In
general, the assessment should be performed in several steps:

50
 Safety

1. Identify the problem and record details of the extent of the damage.
2. If electrical integrity tests are indicated, then these should be completed
before continuing to use the device.
3. Perform a risk assessment considering the likelihood of harm in the short,
medium and long term and the likelihood of the condition further deteriorat-
ing.
4. Apply the control measures determined from the risk assessment and ensure
that all users are fully aware of the issues. These may involve redefining the
range of tasks the system is suitable for or taking the system out of use until
repaired.
As an example, the recommended action for systems which are used as a static shared
services machine with many different users may be to take it out of service if the function-
ality of the ergonomic adjustments is compromised. Users would be unable to set the
system for their own physique and application. Where such problems exist, there is a
risk that people will work around them potentially compromising their own safety.
Alternatively, users may agree on a temporary change in practice whilst the problem is
rectified. It is important that any changes in working practice to cope with problems are
documented, the risks considered and the changes applied by all users.
When considering the state of a system where there is no compromise in functionality and
the issue is principally cosmetic, common sense must prevail. Signs of bashed panels may
be expected on a 5-year-old system which has been regularly moved but on a 6-month-old
system, they may be signs that a change of practice is needed. Similarly, signs of stress or
damage to probe cable sheathing and/or strain relief on a probe which is less than a few
years old may mean that operators are twisting or bending the cable excessively or it could
indicate a design fault in the way the probe cable is routed or supported. Simple modifica-
tions to working practice or cable support to relieve the stress could extend the working
life of the probe significantly.

4.5.2 Electrical integrity of system

The specific electrical safety tests undertaken in most environments are determined
according to local policies and procedures across a wide range of medical devices and
there should be no need to modify these significantly for ultrasound systems. An IPEM
document is available giving guidance on electrical safety testing (IPEM, 2009). The
purpose of this section is not to give a full prescriptive testing regime but to identify the
specific issues which apply when attempting electrical tests on ultrasound systems. As a
matter of good practice, any protocols in place should be subject to regular review to
monitor associated risks. Specific attention should be given to early implementations of
new generations of systems where there is no previous history. Examples are given for
electrical tests which are commonly performed on ultrasound systems such as insulation,
earth leakage and touch current. Values or limits for these tests are not quoted because they
will depend on the standards applied within the local protocol.

51
Quality Assurance of Ultrasound Imaging Systems

It should be noted that the electrical safety regulations identified within this document may
change with changes in European or national legislation; care must be taken to confirm that
the standards being applied are current. Historically, the IEC 60601 standard has been
the most appropriate standard for establishing electrical safety measurement checks in
Europe. This standard was first published in 1977 and is now on its third revision, IEC
60601-1:2005 (IEC, 2005b). This is a type approval standard and the tests are intended to
stress the device and may result in the destruction of the device under test. However, there
are subsets of tests within this standard which have been identified as safe to use in the field
and are regularly applied to establish ongoing safety. It is still possible to damage systems
with some of these tests and any operator should check that they are appropriate for the
system before undertaking the tests. A new standard, IEC 62353:2007 (IEC, 2007b), is
now in place which is specifically designed for in-field testing, however, the measurements
in this testing process deviate from those in the type approval standard such that they are
not comparable. Since the manufacturers will be under no obligation to publish results for
this ‘in-field’ test, it is difficult to see where results could be compared and how measured
changes may be raised as an issue. For this reason and because established practice and
risk assessments will meet legal requirements, the recommendation of IPEM Report 97
(IPEM, 2009) is to continue with the subset tests from IEC 60601-1:2005 (IEC, 2005b).
The group or individual responsible for ultrasound system quality will often not be the
same as those responsible for electrical safety across a site or organisation. Where this is
the case, care should be taken to ensure that electrical safety tests are performed in accord-
ance with the manufacturers’ guidance and that the testers are aware of the specific issues
with ultrasound systems. Invariably, this will mean that those responsible for ultrasound
system quality will need to oversee electrical safety testing.
It is fairly common for additional peripherals to be added to ultrasound systems subsequent
to the original installation (common examples are video recorders, thermal printers and
auxiliary display monitors). If these peripherals are powered from a single socket along
with the main system, then this has changed the constituent parts; the person installing the
additional peripheral has modified the system and is responsible for the electrical integrity
of the whole. If the peripherals are powered from a separate wall socket, the items are
separate devices and can be tested individually. The definition of what constitutes a system
when multiple electrical devices are connected via a multi-way distribution block is more
complex; if the block is fixed behind a panel in the system requiring tools for access then
it is treated as a single system. If the block is accessible and used in place of having enough
wall sockets, then in principle, the devices are separate but the actual definition remains
unclear. The generally accepted consensus is that multi-way mains distribution blocks
should not be used to power medical devices. Again, local policies will identify acceptable
configurations and testing criteria.
For all electrical safety tests, repeat measurements are not advised because of the stress
such tests place on the equipment. It is useful if, during the purchasing phase, a request is
made to the supplier for details of the testing protocols for the system, particularly with
regard to probes as patient applied parts and any special conditions which may apply. (For

52
 Safety

example, trans-oesophageal probes (TOE) often need frequent testing and some may
include a user testing kit.)
Please note: The same standards are available from BSI. The year of the BS EN and IEC
standard may not be identical with the BS EN designated standard sometimes published in
the following year. (For example, for the IEC 62353:2007 document (IEC, 2007b), the BSI
version is BS EN 62353:2008 (BSI, 2008).)

[Link] Electrical integrity of main console

A range of typical measurements is included in this section, principally to identify the
measures which can be undertaken for ultrasound systems. The reader is advised to refer
to local policies and procedures before undertaking electrical safety testing. Any visual
inspections or panel removal/replacement should have been completed before undertaking
electrical tests. Care should always be taken to ensure that, for any power up or down
processes, the correct procedure is followed for the scanner concerned. Reference should
be made to the service manuals to confirm that there are no specific precautions required
during the electrical safety testing. Electrical testing should only ever be carried out
by suitably qualified staff, which in this case would mean staff that are trained both in
electrical safety testing and the basic operation of the ultrasound system.
It is important to identify the class and type of equipment to ensure that appropriate limits
apply to the measurement. The equipment class and type will be identified usually by an
appropriate symbol (Figure 4.6) on the system chassis. It is becoming more common for
ultrasound machines to identify the system type on the probe, with probes of type BF
and/or CF connected to the system (Figure 4.7). If this is the case, then the system type is
according to the probe in use. It is important to make sure that the appropriate limits for
measurement are applied and probes should always be checked for classification symbols.
Battery powered equipment should be tested according to the appropriate class only if it is
capable of being operated when connected to the mains. Systems which disable operation
during charging do not require testing as mains connected medical equipment.
It is important to note that there is an issue with using automatic safety testers and the shut-
down requirements of ultrasound systems which use a soft key power down sequence. The
sequence warns that the mains should not be disconnected until the shutdown is complete.
Automatic testers cannot be allowed to cycle through their test sequence as most discon-
nect the mains power and this may damage the ultrasound scanner. Manually operated
testers are safest although automatic devices used in a manual mode or programmable
testers appropriately configured may be suitable. Care should be taken to check manufac-
turer’s advice on testing or usable configurations for the system as some tests which are
usually safe may prove destructive. (For example, high current earth bonding tests which
path through functional earth.) Some systems may have filters or protective circuits in
place which give anomalous readings for tests; often the only way to resolve these is by
consulting the manufacturer or supplier to ascertain the cause.

53
Quality Assurance of Ultrasound Imaging Systems

Figure 4.6 Electrical class and earth symbols (note there is no specific symbol for Class
1; equipment which carries an earth or protective earth will be Class 1)

Figure 4.7 System type symbols

Often the most powerful method of ensuring electrical safety is to inspect the system visu-
ally regularly; if the slightest doubt exists over the integrity of the system, then it should
be checked by appropriate measurements.

Set-up and assessment

Before starting any measurements of electrical integrity, a full visual inspection of the
system should be undertaken as described in Section 4.5.1. If this is not appropriate,

54
 Safety

then as a minimum the panels, power switches and mains cables should be tested as
described.
The system to be tested is plugged into the safety tester; at this point, the system should
be switched off (including all connected peripherals powered from the same socket).
Separately powered peripherals connected by signal leads should be disconnected from the
mains. Network connections should be removed from the wall socket during testing, it is
worthwhile on at least one occasion to measure the leakage with this connection remade.
If the network is connected correctly, there should be no difference between these two
measurements.
When the safety tester is switched between measurement types, the ultrasound system
should always be powered down and then restarted for the next test. This may take some
time as the power up–power down sequences for ultrasound systems may take several
minutes. Any attempt to circumvent this cycle may result in severe damage to the
system.
The earth continuity tests are performed first using the procedure applicable to the spe-
cific electrical safety tester. Ensuring that the earth bonding is sound before continuing
with other tests reduces risks to the individual performing the tests. (The tests are usually
performed with a current from 0.1 A to 10 A generated by the safety tester and do not
require power to be applied to the system. The standard requires a current of 25 A to be
applied for several seconds but testing at this level is known to damage some systems and
may adversely stress others and is not advised. Even a lower 10 A test has been shown to
damage some systems.)
Measurements should be made of earth continuity, insulation (resistance) and leakage
currents in that order as advised in IPEM Report 97. If faults are identified at any of these
stages, then they should be rectified before continuing with the other tests. This ensures
that, for example, the earthing is sound before continuing with the tests which increases
the safety of the tester.
The leakage current measurements should be made for both the system leakage and Touch
Current (often identified as enclosure leakage on testers). The measurements should be
performed under normal conditions apart from Touch Current where it is usual to use the
single fault condition (SFC) earth disconnected.

Recorded values and recommended actions

All of the measurements made should be recorded and checked against the limits in the
applicable standard and against the previous values for the system where they exist. If the
system is outside the tolerance stipulated in the standard, then the system should be taken
out of use until the cause can be determined and rectified. Where a system consists of
multiple devices, it may be that one of these is the cause and removal rectifies the problem
allowing the system to continue in use whilst the fault on the device is repaired. Small
changes over time may be expected in any electrical system as components age, however,

55
Quality Assurance of Ultrasound Imaging Systems

a significant stepped increase in the values from previous measurement usually indicates
a fault. At this point, a decision will have to be made as to whether the system is safe to
continue in use. There are too many variables to provide specific guidance as to when a
system which is still within the tolerance of the standard should be removed from use. The
decision will depend on how close to the limit for the standard the system is, how large a
change has occurred and what the suspected cause of the problem is.

[Link] Electrical integrity of probe as patient applied part

For particular probe types on some systems, a specific test may be indicated in the manu-
facturer’s service specification. If this is the case, then these should be undertaken at
the recommended intervals. There are some cases where specific constraints apply to the
depth to which the probe is immersed; these should be checked in the manufacturer’s
documentation before carrying out any tests.
If a visual inspection reveals a defect which would clearly allow ingress of water into the
transducer, then it should be assumed that the probe is unsafe for use and it should not be
tested. For less obvious problems such as hairline cracks or peeling or splitting of seals
where the defect may not penetrate through to the electrical connections, then it will be
necessary to check the electrical integrity of the probe.

Set-up and assessment

A saline solution (1 g/litre for example) is placed in an appropriate container (probe
dependent) and the probe face immersed and held in the solution such that all areas that
could be in contact with gel during normal scanning are below the surface. Allow to stand
for 10–15 minutes before making measurements to allow gel which may be dried in
the damaged areas to rehydrate. Ensure that any damaged/repaired area to be checked is
covered by the saline solution.
The system to be tested is plugged into the safety tester; at this point, the system should
be switched off (including all connected peripherals powered from the same socket).
Separately powered peripherals connected by signal leads should be disconnected from the
mains. The system is then powered up according to the manufacturer’s instructions.
The electrical test probe provided is plugged into the applied parts socket on the safety
tester. The test probe is then clipped onto an aluminium strip (for example, folded
aluminium foil, 60×100 mm) and the strip is immersed in the saline solution.
The electrical leakage tests are then performed using the procedure applicable to the
specific electrical safety tester.

Recorded values and recommended actions

All values should be recorded and checked against the limits applied in the applicable
standard and against the previous values for the system where they exist. If the probe is
outside of the tolerance stipulated in the standard then it should be taken out of use until

56
 Safety

the cause can be determined and rectified. Where the values obtained show a significant
increase from those obtained previously but the probe remains within tolerance for the
applicable standard, then a decision will have to be made as to whether the probe may
continue in use. In general, this will be influenced by a number of factors including: what
the probe is used for, how close to the limits the probe is, how different the measurement
is from the normal and what the suspected cause of the change is. It is also noted in IPEM
Report 97, Chapter 6 that some manufacturers’ limits for probe leakage are higher than
those indicated in the standard. If that is the case, then these should be the limits that are
applied as they have been through the type approval and accepted. For example, for some
trans-oesophageal probes, a limit of 180 µA is applied.

57
5 Extended testing
5.1 B-Mode imaging
Currently, there is no good published evidence to support the use of resolution measure-
ments and the assessment of dead zone and of grey-scale and anechoic target detection
performance in routine testing. However, individual centres may wish to include the tests
where there is local evidence of efficacy or where there is the capacity and capability to
provide this evidence.
The following tests may be performed using either manual/visual methods or software
methods (Gibson et al., 2001; van Wijk and Thijssen, 2002):

Table 5.1 Table of measurements

Test Method When Notes
Dead zone/black-out zone TETO Baseline and as locally
determined
Axial and lateral resolution/ TETO Baseline and as locally Two methods
beam profile determined available
Slice thickness TETO Baseline and as locally Two methods
determined available
Grey-scale performance TETO Baseline and as locally
(contrast targets) determined
Anechoic target detection TETO Baseline and as locally
determined
TETO, tissue equivalent test object.
All are transducer related tests.

5.1.1 Test equipment

A tissue equivalent test object (TETO) is required for these tests. This should be of uniform
tissue mimicking material, with targets for assessment of resolution (a column or pairs
of nylon filaments), targets for measurement of dead zone (nylon filaments close to the
surface), anechoic targets of different sizes at a number of depths, and targets with different
amounts of backscatter. Targets must be within the low contrast penetration depth of the
probe under test and, ideally, anechoic and resolution targets should be placed at intervals
throughout this depth (Figure 5.1).
Attenuation and backscatter should be appropriate to the generation of speckle throughout
the normal clinical depth of imaging. Attenuation of 0.5 dB cm–1 MHz–1 is likely to be
suitable for most applications (in some instances, a test object with a higher attenuation

58
 Extended testing

Figure 5.1 Example composite image of a TETO, showing resolution targets, anechoic
targets and grey-scale targets in a tissue mimicking background
may be preferred, 0.7 dB cm–1 MHz–1). A speed of sound of 1540 m s–1 is desirable but not
essential (defocusing of the beam will occur at other speeds, but measurements should be
made relative to a baseline established with the same TETO).
Inclined plane TETOs are available, as an alternative to conventional TETOs, for
measuring slice thickness.

5.1.2 Preparation
Scanner settings for serial performance testing must be easily reproducible. Making use of
the presets available on many modern scanners can facilitate this. TETOs have uniform
attenuation and so selection of a preset for a uniformly attenuating body part is advised;
abdominal settings may be appropriate for general purpose scanners and breast or thyroid
settings may be appropriate for small parts scanners. A suitable preset may be selected
by imaging the test object and finding the setting that gives the most uniform mid-grey
speckle image with time-gain compensation (TGC) set as follows. TGC may be easily
reproduced by setting to mid-scale if slider controls click into place here, or pushing

59
Quality Assurance of Ultrasound Imaging Systems

slider controls to the maximum position. In centres where there is a possibility that presets
will be altered, it is advisable to copy the chosen settings to a new, suitably named,
preset.
For the individual tests, only the scale and magnification settings need then be changed
and recorded for future testing. Other settings may change with scale, e.g. focal depth,
but should remain consistent for a given scale setting.
Image measurements should be made using the scanner’s calliper system unless software
is available for automated measurement.
Artefacts are common when imaging test objects. The bottom of the test object generates
a strong echo, which may appear as a range artefact within the tissue mimicking material
when the image scale is set at less than the test object depth. In urethane test objects, this
can be reduced by smearing the bottom of the test object with coupling gel, thus reducing
the amplitude of the echo. In boxed test objects, the artefact may be moved out of critical
areas by changing the scale setting. When imaging through water with curved arrays,
reflections from the water surface generate hazy artefacts within the tissue mimicking
material. Placing absorbent paper into the water at the ends of the probe can reduce
these.

5.1.3 Dead zone

[Link] Set-up and assessment
Image the test object on high magnification to show targets close to the surface. Measure
the distance from the probe face to the first clearly defined target in the image. Note
that on some scanners there may be a ‘black-out’ zone, eliminating the dead zone from the
image.

[Link] Recorded values and recommended actions

Record the depth of the first visible target, with careful reference to the TETO and not
to the apparent imaged TETO surface, which may be affected by a ‘black-out’ zone.
Any change in this depth, allowing for random measurement errors, should be regarded as
a fault.

5.1.4 Axial resolution, lateral resolution and slice thickness

[Link] Set-up and assessment
There are two ways of measuring resolution.
Primary method: Measuring the spread on an image of a line target imaged in cross-
section. A small target (less than the resolving power of the scanner) is imaged. The
depth and width of the target image (usually a nylon filament) give the axial and lateral
resolution, respectively. This method is subject to significant inter- and intra-observer
variation. Slice thickness may be measured in a similar fashion, by measuring the width
of the line target image with the probe rotated to 45° (Skolnick, 1991).

60
 Extended testing

A centred and vertical image of a column of targets should be obtained, with the scale set
to just beyond the penetration depth. Target images should be optimised for brightness by
slight rocking of the probe. It is good practice to utilise read zoom to magnify the targets
for accuracy of calliper placement. Note that the use of write zoom will have the beneficial
effect of reducing pixel size, but, as the zoom box is panned, may cause changes in
focusing that are difficult to record and reproduce. After freezing the image, each target
should be measured axially and laterally, taking care not to include speckle. Callipers
should be placed at points with approximately 50% of maximum brightness, i.e. avoid
measuring weak ‘tails’ of the target profile.
To measure slice thickness, the probe should be rotated to 45°; use of a jig or template
may facilitate this (the simplest template is a folded piece of paper or card). It is important
to centre the image of the target column and to ensure that is vertical, otherwise image
profiles may not be symmetrical. Each target image should be measured as for lateral
resolution.
Secondary method: Observing the closest pair of objects that may be distinguished.
Test objects will generally contain groups of closely spaced line targets for the assessment
of lateral and axial resolution. These may be available at several depths. This method is
quantised, i.e. only separations provided in the test object may be assessed. There is no
equivalent method for slice thickness.
The target group should be laterally centred in the image and the smallest visible target
separation in axial and lateral directions noted. Again, read zoom should be used. It is
important to understand and take account of the fact that, where targets are staggered
rather than horizontally and vertically aligned, they may apparently be resolved laterally
due to good axial resolution.

[Link] Recorded values and recommended actions

For the primary method, record the axial and lateral dimensions of each target, or of
a sample of targets. An acceptable range for routine measurements may be established
by acquiring and measuring further images; note that this may generate an artificially
small tolerance and a more realistic range may be established by repeated measurements
on separate occasions. An acceptable initial tolerance for this test is ±40% (Kanal et al.,
1998).
For the secondary method, record the smallest visible target separations in the axial and
lateral directions. The smallest detectable change, and therefore the tolerance on the test,
is the increment in target spacing.

5.1.5 Grey-scale performance

[Link] Set-up and assessment
A number of TETOs contain a series of cylindrical or triangular prism targets with
different amounts of backscatter, aligned at the same depth. These should be imaged and
the visibility of each target recorded on a 3-point scale:

61
Quality Assurance of Ultrasound Imaging Systems

• not seen
• seen but indistinct
• clearly seen.

[Link] Recorded values and recommended actions

Record the level of visibility of each target. This is a rather coarse measure of grey-scale
performance. The tolerance for the test should be established by observing further images.
An appropriate initial tolerance may be an increment on the visibility scale.

5.1.6 Anechoic target detection

[Link] Set-up and assessment
Most TETOs contain rows of scatter-free targets with different diameters, often at a number
of depths. The targets should be imaged with the scale set to just beyond the penetration
depth. Target images should be optimised by slight rocking of the probe; any adjacent
nylon filaments can be used to aid this adjustment (as for resolution). If there is a large
matrix of targets, the size of the smallest target visible at each depth should be recorded.
If the number of targets is small, the visibility of each target may be recorded on a 4-point
scale:
• not seen
• poorly seen, where the target is ill-defined
• fairly clear, where the target is well-defined but has internal noise
• clearly seen, where the target is well-defined and black internally.

[Link] Recorded values and recommended actions

Record the smallest target visible in each row or the level of visibility of each target. The
tolerance for the test should be established by observing further images. An appropriate
initial tolerance may be an increment in the size visible or an increment on the visibility
scale.

5.1.7 Novel methods

A number of novel devices have been developed. The efficacy of these devices for
B-mode quality assurance (QA) has yet to be established, but their use is becoming more
widespread and evidence is being produced.
Kofler and Madsen (2001) proposed the use of signal-to-noise ratio analysis of images
using a vertical plane of different sized spherical anechoic targets (simulated lesions) to
determine resolution zones. The results obtained depend on resolution in three dimensions,
providing an overall performance metric.

62
 Extended testing

Satrapa et al. (2002) extended this concept, developing a test object with layers of tissue
mimicking material. Alternate layers contain vertical cylindrical voids. The ultrasound
probe is fixed in a jig and scanned across the top of the test object. During the scan, a series
of images is captured and a three-dimensional reconstruction generated. The signal-
to-noise ratio (SNR) of each void is calculated and a plot of SNR against depth is pro-
duced. A curve is fitted to the plot to show the functional range (SNR>2.5) of the probe.
The system also produces a three-dimensional grey-scale representation of the SNR
function, a set of images through each plane of voids (C-images) and the corresponding
planar SNR images, as an adjunct to the SNR plot, allowing the observer to visually
correlate the functional range with the images.
Pye et al. (2004) introduced the concept of a Resolution Integral and implemented
this using a novel test object containing anechoic pipe structures embedded in tissue
mimicking material. As in the methods outlined above, this accounts for resolution in three
dimensions with the advantage that the targets are continuous, rather than discrete, and a
wider range of target sizes is included. Images may be assessed visually or computation-
ally, with comparable results. The results are used to provide further performance metrics
– depth of field and characteristic resolution – and can be used to discriminate between
transducers of different ages, frequencies and clinical performance.
Moore et al. (2005) described the FirstCall2000 (Sonora Medical Systems, Longmont,
CO, USA) probe testing device. This tests the functionality of the probe in isolation from
the scanner by pulsing each individual element and testing for sensitivity, capacitance,
frequency and bandwidth. This is a sophisticated version of the ‘paperclip test’ described
in Chapter 2, isolating the fault to probe component level. Sonora Medical Systems has
also developed a hand-held probe testing device – the Magic NickelTM. This device has
a very small transmit/receive element, which is placed on the probe face and detects
any transmission. If a transmission is detected, the device then sends a sequence of pulses
back into the probe and these are shown on the scanner display. The device thus tests the
function of the probe elements on transmission and reception.
All of the above devices deserve rigorous evaluation to determine their efficacy as part of
a QA programme.

5.2 Doppler extended testing

There are a number of limitations to the level of tests which can be performed with a
basic string-type Doppler test object, mainly arising from the single-velocity nature of the
device. Various other test systems exist including flow phantoms, belt and rotating test
objects, vibrating test objects and electronic injection systems. The information in this
section is a short review of the literature which may be used as a basis for expanding or
evaluating a Doppler testing regime. There is still a great deal of work required if we are
to improve the currently sketchy evidence base and range of Doppler testing. Only articles

63
Quality Assurance of Ultrasound Imaging Systems

published since IPEM Report 70 (IPEM, 1994) are included as many aspects of that report
are still relevant.
In terms of standardisation of phantom design, there is an IEC Standard 61685 (IEC, 2001)
in place for flow test objects but at present, the authors are not aware of any commercially
available test object based on this standard. The design of test objects is discussed in
the IEC technical specification document 61895 (IEC, 1999) on the testing of pulsed
wave Doppler systems. There are a few flow and string test objects available to purchase,
and other Doppler test objects have arisen from within research institutions. It is however
unclear which if any of these may be of further use and as already stated, even the limited
use of a stable string test object has not been validated. An AAPM (American Association
of Physicists in Medicine) committee with a working title of Developing Quantitative
QC tests for Doppler Ultrasound was instigated in 2005 but has not yet reported.
For those interested in flow phantom use, components such as the pump and blood mimic
and C-flex tubing may be purchased. These generally require assembly in the lab, and pos-
sibly in-house manufacture of tissue mimic and blood mimic. The in-house construction
of flow phantoms is well described in the literature. Tierlinck et al. (1998) report on testing
five similar Doppler test objects constructed to the IEC 1685 draft standard. They found
that, for a number of parameters, variations are too large between phantoms or observers
to produce useful results but there are a range of parameters which can be measured
effectively with such objects: penetration depth, sample volume position, minimum
velocity, Doppler angle. Tubing material can cause distortion of the Doppler beam and the
received Doppler spectrum, by refraction of the beam at the vessel interface, and attenua-
tion within the vessel. A rubber-based vessel, C-flex, provides the best choice of commer-
cially available tubing. Alternatively, a vessel based on PVA-cryogel may be manufactured
in-house which has good acoustic matching to real arteries (Dineley et al., 2006). To
overcome the problem of spectral distortion, wall-less phantoms have been developed by
several authors including Rickey et al. (1995) and Ramnarine et al. (2001). The wall-less
models are substantially artifact-free which contrasts with the heavy shadowing seen for
the others. Ramnarine et al. (2001) developed a wall-less stenosis phantom where they
presented a novel method for sealing the junction between the wall-less vessels and the
external tubing. The object was found to have a stable geometry when tested under high
flow rates. Poepping et al. (2004) describe a flow phantom using a vessel mimicking
material and a wall-less version formed into anatomically correct models of the carotid
bifurcation. Subsequent work by Watts et al. (2007) and Meagher et al. (2007) further
develop the wall-less anatomical model to investigate the accuracy of flow measurements
made in the normal and diseased carotid.
A specification for the acoustic and viscous properties of blood mimicking fluids (BMF)
is provided in IEC 61685 (IEC, 2001). The recipe for BMF production in IEC 61685 is
based on that described in the papers by Ramnarine et al. (1998, 1999). The BMF is based
on suspended nylon particles and was shown to be easy to prepare, and provides stable
acoustic backscatter over a 3-month period. Samavat and Evans (2006) also confirmed
this BMF to be an effective mimic.

64
 Extended testing

Fleming et al. (1994) assess tissue Doppler imaging using a rotating phantom consisting
of reticulated foam to determine the velocity estimation accuracy and velocity resolution,
and a sliding gel phantom to measure the spatial resolution. The test phantoms were
effective for the measurement of velocity and spatial resolution but the velocity resolution
was found to depend on backscatter intensity and target velocity.
Dineley et al. (2006) use PVA-cryogel in the design of a wall motion phantom.
Measurements of wall motion are often used to derive quantities such as elasticity but the
accuracy and reproducibility of the techniques are difficult to determine. The phantom was
found to produce physiological distension waveforms.
Li et al. (1997) describe the use of an acoustic grid around the tubing of a flow phantom
to measure colour spatial resolution. The grid effectively creates a sequence of closely
spaced moving targets; by varying the grid, the window of separation within which the
spatial resolution falls can be established. Two moving line sources have been described
by Novario et al. (1994) using a flow phantom and Lange and Loupas (1996) using a string
phantom.
Browne et al. (2004) look at the measurement of Doppler sensitivity and in particular
validating a ‘single figure of merit’ sensitivity performance index and the ability to use this
to discriminate between scanners. The measurement of Doppler sensitivity is a complex
problem because there are many aspects which may be attributed to the parameter includ-
ing: flow velocity, vessel size, penetration depth and clutter filter. Findings show that the
method is valid although an additional parameter for vessel size would be needed to utilise
the method with phantoms other than the one used.
Rickley and Fenster (1996) describe methods to evaluate the effectiveness of the wall filter.
These are based on the use of a partially reflective mirror which splits the Doppler beam
enabling the Doppler sample volume to be placed simultaneously within two moving
regions, one which simulates blood flow and one which simulates wall motion. They
describe two phantoms with the ability to produce flow signals with and without clutter to
measure the effectiveness of the wall filter. The first phantom consists of two belts, one for
flow and one for clutter, and the second phantom has a fluid component for the vascular
flow and a belt for the clutter component.
Li et al. (1998) describe the use of an acoustic injection test object to test ultrasound
scanners. These devices receive the pulse from the scanner and generate a synthesised
echo with an appropriate modulation of the frequency. Since they have no moving parts or
tissue equivalent material, they are stable and easy to use. (The Magic NickelTM from
Sonora Medical mentioned earlier, can be used to produce a simple signal injection for
Doppler tests.)
Cathignol et al. (1994) and Hoskins (1994a,b) provide detailed information on acoustic
properties and selection of suitable filaments for use in string-type test objects.
Both electronic injection and vibrating target phantoms have been described, e.g. Evans
et al. (1989), McAleavey et al. (2001). These are non-physiologic: in the electronic

65
Quality Assurance of Ultrasound Imaging Systems

injection system, the beam forming characteristics and spectral broadening are not
accounted for, and vibrating systems produce a Doppler signal which is mirror-imaged.
In the review article by Hoskins (2008), these Doppler test objects are compared in detail.
It is noted that there are now tissue-equivalent recipes for soft tissue, blood mimic and the
vessel wall. The flow phantom will remain the principal device for use in scientific studies.
String and belt test objects offer simplicity of design, especially for evaluation of velocity
estimators, and have a role as portable test objects. Electronic injection and vibrating test
objects produce non-physiologic Doppler signals and their role is limited.

5.3 Extended acoustic and thermal safety testing

Useful quantitative safety information about an ultrasound beam can be obtained from
measurement of the acoustic pulse. It is recognised that while hydrophone measurements
are not straightforward, centres with the capacity and capability may wish to carry them
out. A review of metrology for ultrasonic applications by Zeqiri (2007) provides a detailed
background for measurements of ultrasonic fields including the use of hydrophones.
The difficulty and cost involved in making accurate measurements of ultrasound fields are
the principal reasons why details of these and other extended safety measurements are
given here rather than under the routine testing.
In comparing equipment performance against recommended limits, typical values
have been applied and an additional allowance may be necessary for local measurement
uncertainties dependent on the technology used.

5.3.1 Table of measurements

Test Method Whena Notes

Mechanical index Measurement of acoustic A/R Centres with appropriate
(MI) pressure and centre facilities to input type
frequency specific maximum values
into a national database
Intensity Hydrophone and A/R
integrator
In-situ temperature Using appropriate R Centres with appropriate
thermal test object facilities to input type
specific maximum values
into a national database
a
A, acceptance (new scanner or probe); R, routine or reactive (minimum frequency
biannual).
All are transducer related tests

66
 Extended testing

5.3.2 Equipment for acceptance and routine testing

[Link] Hydrophone with signal capture and analysis
Several methods may be used to obtain information about acoustic pulses. Each one
uses a calibrated hydrophone as the basic measurement device. At the simplest level,
a calibrated hydrophone in water is used to detect the acoustic waveform and from this
signal, acoustic pressure and acoustic frequency may be measured. The next level of
complexity involves a system which performs integration of the digitised waveform
to obtain a measurement of the pulse intensity integral and the spatial peak intensity. A
digitised waveform of the acoustic pulse from a calibrated hydrophone in water may be
displayed, for example, using a digital oscilloscope and a personal computer and further
analysis carried out. An example of an acoustic pulse measured using a calibrated
hydrophone is shown in Figure 5.2.

Figure 5.2 Acoustic pulse showing non-linear distortion

A hydrophone may be of needle or membrane (Figure 5.3) design with a sensitive element
which is small compared to the acoustic wavelength (IEC 1994, 2007c) and small com-
pared to the lateral dimensions of the acoustic beam. Generally, a sensitive element of
0.2 mm may be used, but for high frequency beams, an element size of less than 0.1 mm

67
Quality Assurance of Ultrasound Imaging Systems

Figure 5.3 Needle and membrane hydrophone (courtesy of Precision Acoustics)

is required to sample the beam width adequately. Factors such as element sensitivity, signal
amplification, bandwidth and calibration also need to be considered. Hydrophones must be
calibrated up to at least the acoustic working frequency; the standard 1–20 MHz available
from the National Physical Laboratory, for example, will be sufficient for the majority of
systems.
Commonly, a hydrophone is held in position inside a glass or plastic tank filled with
degassed water. If the beam is vertically aligned, the acoustic transducer is coupled into
the water above the hydrophone. Alternatively, a polyethylene window may be inserted
into a side or end of a tank to permit a horizontal measurement arrangement. In each case,
a method of accurately aligning the hydrophone and transducer is required, which allows

68
 Extended testing

adjustments of the order of 0.1 mm to be made. An automated positioning system may be

used but is not essential.
This method appears initially to be straightforward. However, it is not easy to implement
for the majority of modern ultrasound scanners in modes other than Doppler and M-mode
because of the complex beams employed and difficulties with triggering.
Details of equipment suppliers are given in Appendix C.

[Link] Acoustic intensity integrator

The use of an in-line integrator (BMUS/BIR, 2000; Whittingham et al., 2004) can over-
come many of the problems of the methods described above for measurement of intensity.
It enables the temporal-average acoustic intensity to be measured easily for all modes. This
relatively simple, portable measurement system is based around the use of a hydrophone
in water and a radio frequency (RF) power meter. As shown in Figure 5.4, the signal from
a hydrophone is passed via an amplifier to the RF power meter, which acts as an integrator.
A positioning system is required that permits precise adjustments to be made, either to
the location of the transducer or to the location of the hydrophone. For intensity measure-
ments, the system requires calibration before use by exposure to a range of known
intensity sources. Details of suppliers of the component parts of the system can be found
in Appendix C.

Figure 5.4 Diagram of components

[Link] Thermal test object

Use of a thermal test object allows a local measurement of the temperature rise generated
by the absorption of an acoustic beam. The design requirements for thermal test objects are

69
Quality Assurance of Ultrasound Imaging Systems

Figure 5.5 Schematic of thermal test object (courtesy of National Physical Laboratory)

set out in IEC TS 62306 (IEC, 2006). At the present time, there is only one commercial
supplier of thermal test objects (TTO). The basic design is shown in Figure 5.5 and a
picture in Figure 5.6. The thermal sensor is a thin film thermocouple embedded in a
sandwich of tissue mimicking material. The TTO is connected to a sensitive voltmeter

Figure 5.6 Image of thermal test object (courtesy of National Physical Laboratory)

70
 Extended testing

(resolution 1 µV). Further details may be found in Appendix C. Use of the TTO allows
a worst-case temperature rise to be estimated under standard set-up conditions. The tem-
perature in-situ may be obtained to a first approximation by de-rating the temperature rise,
i.e. making an assumption that it increases linearly with acoustic power. Two types of
test object are available, a soft tissue type and a bone-soft tissue type. The bone-soft tissue
type is probably the most suitable model for a third trimester pregnancy. If required,
polyethylene attenuators could be inserted between the transducer and the TTO to simulate
overlying tissue but this is not recommended for routine measurement.

5.3.3 Evaluation of MI safety index

Where safety indices are displayed on ultrasound scanners they should provide valid
information to the user about the ultrasonic exposure to the patient. In addition to a visual
check on MI described in Section 4.3.1, quantitative measurements may be carried out
under a variety of standard conditions to validate the on-screen displayed values. Measure-
ments should be made at the settings producing the maximum MI in each mode and may
be made at other clinical settings.
MI is defined in IEC 62359 (IEC, 2005) as the ratio of the spatial peak of the peak rarefac-
tion pressure (derated) to the square root of the centre acoustic frequency. It is measured
at the point on the beam axis where the pulse intensity integral (PII) is a maximum.
Equation 5.1 is used to calculate the index.
pr
MI = (5.1)
fc

Here fc is the centre frequency (acoustic working frequency) and pr is the derated
peak rarefaction pressure. A derating factor of 0.3 dB cm–1 MHz–1 is used to convert from
acoustic pressure in water to in-situ rarefaction pressure. BS EN 61157:2007 (BSI, 2007)
recommends the evaluation of arithmetic mean acoustic working frequency for pulsed
fields. The zero-crossing acoustic frequency method, whilst easy to apply, is not strictly
applicable to pulsed beams. However, it may be used for pulses containing three or
more cycles, e.g. Doppler and colour Doppler beams. Measurement of acoustic frequency
should be carried out using low amplitude pulses to avoid pulse distortion due to
non-linear propagation in water.
The arithmetic mean acoustic working frequency (BSI, 2007) is defined as the mean of
the most widely separated frequencies f1 and f2, at which the magnitude of the acoustic
pressure spectrum is 3 dB below the peak magnitude.
The purpose of the test at acceptance is to allow the following checks to be carried out:
• confirming the accuracy of the value displayed on the screen
• ensuring that the maximum system MI does not exceed current safety
guidelines and regulatory limits, and
• confirming that the MI is below the level requiring on-screen display.

71
Quality Assurance of Ultrasound Imaging Systems

Routine/reactive testing may also be carried out in the following situations:

1. Where no change is observed in the displayed value of MI generated under
standard conditions, the aim of this test is to ensure that no change has
occurred in the acoustic beam. If a change has occurred in either peak rar-
efaction pressure or centre frequency, the effect of the change on calculated
MI is evaluated.
2. Where a change is observed in the displayed value of MI generated under
standard conditions, the aim of this test is to determine whether this reflects
a real change in the acoustic beam.
For routine/reactive quantitative measurements of MI, it is vital that the settings are
identical to the settings used at baseline/acceptance testing (i.e. standard conditions). If a
change in MI has occurred as a result of a software upgrade, new baseline measurements
are required.

[Link] Set-up and assessment

These measurements are intrinsically difficult and need considerable patience and practice
before reliable values are obtained. Ideally, the measurements should be made calculating
the derated PII values to determine the location of the maximum derated PII, however,
to make the measurements more practicable, the maximum PII or maximum peak
negative pressure is suggested. This may not be in the same location as the measured
maxima indicated but is acceptable.
For each transducer and mode in turn:
1. Couple the acoustic transducer to the measurement tank.
2. Obtain a trigger signal (external or self-trigger).
3. Align the transducer and the hydrophone carefully to maximise the signal.
4. Align the beam axis as close as possible with the movement axis of the tank
by making adjustments and finding the peak signal at several depths.
5. For systems displaying MI on the screen, set the ultrasound scanner to
produce the maximum displayed value of MI by varying the operating
conditions systematically.
6. Move the transducer on axis to obtain the maximum value of PII from the
digitised hydrophone signal.
7. Record the distance from the transducer where this occurs.
8. Measure the acoustic working frequency and the peak acoustic rarefaction
pressure at this location using a calibrated hydrophone.
9. Record these results together with full details of the equipment settings,
including displayed MI.

72
 Extended testing

10. Use measured acoustic data to calculate MI for comparison with display.
11. Now, vary the equipment settings systematically to ensure that the
maximum PII (and hence MI) has been located. If necessary, repeat steps
6–9 for these new equipment settings.
12. Repeat the measurement at clinical settings if required.
13. For systems not displaying MI, the equipment settings that result in
maximum PII on axis need to be discovered. Measurements may also be
made at clinical settings if desired.
Finding the correct location for the measurement of acoustic frequency and acoustic pres-
sure can be a lengthy procedure. An alternative is to make the measurement at the location
of the peak negative pressure rather than locating the maximum PII.

[Link] Recorded values and recommended actions

Measured MI values, calculated from measurement of acoustic pressure and centre
frequency, are recorded together with full details of the operating conditions including
displayed MI. If desired, a similar set of conditions may be recorded for each clinical
setting.
The MI should not be greater than 1.9 for all machine settings.
Recommended action levels for acceptance testing:
• maximum displayed MI>1.9
• measured MI>1 in B-mode and MI not displayed
• displayed MI differs from calculated MI by >±0.3.
In addition, for routine/reactive testing:
• displayed MI or calculated MI differ from baseline by >±0.3
• calculated MI exceeds 1.9.
For any of these, the user should be made aware of the issue and the manufacturer or
supplier informed. Further investigations will be required.
If the maximum displayed MI>1.9 or calculated MI>1.9 at any point, then the user and
manufacturer or supplier should be notified. The machine can continue to be used as long
as this setting is avoided.

5.3.4 Spatial peak temporal average intensity

Manufacturers of ultrasound equipment may provide the user with information about
maximum Ispta levels and measurements can be made to confirm that the equipment
is operating to the type standard. If this information is not available, maximum Ispta is

73
Quality Assurance of Ultrasound Imaging Systems

measured and checked against the regulatory limits applied in the USA (US FDA 510k,
2008) (US FDA, 2008). Remember the limits are applied to measurements derated by
0.3 dB MHz–1 cm–1.
In principle, any of the systems using a hydrophone described in Section 5.2.1 can be
extended to allow calculation of the spatial peak temporal average intensity. The process
is fairly straightforward for pulsed Doppler and M-mode beams, where the Ispta can be
calculated from the pulse average intensity, pulse duration and pulse repetition period.
For complex beams, it becomes increasingly difficult and an alternative approach has
been developed which uses a RF power meter to integrate the analogue signal from a
hydrophone. This system overcomes the difficulty which can arise with triggering, for
example, in complex scanned beams having multiple focal zones (Section [Link]).

[Link] Set-up and assessment

For each transducer in each mode:
1. Couple the acoustic transducer to the measurement tank so that the acoustic
beam is aligned with the hydrophone.
2. Obtain a trigger signal if required (external or self-trigger).
3. Make fine adjustments to the alignment of the transducer and the
hydrophone to maximise the signal.
4. Select the equipment settings which produce maximum acoustic output
(e.g. maximise TI) and vary the axial location of the hydrophone to find the
location of maximum Ispta in the beam. Note: the maximum Ispta will be
located at or near the minimum slice thickness (Dixon and Duck, 1998).
5. Measure and calculate derated Ispta (derating factor 0.3 dB MHz–1 cm–1).
6. Repeat the measurement at clinical settings if required.

[Link] Recorded values and recommended actions

The maximum Ispta is recorded for each transducer in each mode together with full details
of scanner settings and the location of measurement.
Recommended actions for acceptance testing:
• Maximum Ispta >±30% of manufacturers quoted values
• Maximum derated Ispta >720 mW cm2 (systems with safety indices
displayed)
• Maximum derated Ispta >50 mW cm2 (ophthalmic systems).
If the systems exceed any of these conditions, then the user and manufacturer or supplier
should be informed of the problem. If the values are significantly higher, then additional
measures should be put in place to ensure the settings involved are not applied clinically.

74
 Extended testing

If on a routine or reactive retest of these values, the maximum Ispta >±30% of baseline
value without software upgrade, then the user and manufacturer or agent should be
informed.

5.3.5 In-situ temperature rise

In-situ temperature measurements may be made at acceptance in all operating modes, for
both clinical and worst-case conditions, using appropriate thermal test objects (TTOs).
Comparison can be made between measured values and the displayed TI and this
comparison informs advice given to the user on interpretation of TI.
It is not necessary to make routine measurements of the in-situ temperature rise. However,
a significant change in output power or Ispta will result in a change in the temperature rise
produced by absorption of the beam. It is appropriate to reassess in-situ temperature under
these conditions.

[Link] Set-up and assessment

For each transducer and each operating mode:
1. Position the TTO in the water.
2. Position the transducer so that the acoustic beam is coupled into the test tank
and aligned with the TTO.
3. Adjust the ultrasound equipment settings to maximise the on-screen display
of TI.
4. Carefully adjust the position of the transducer to maximise the temperature
rise measured in 30 seconds. The test object should be allowed to cool
between measurements by turning the acoustic beam off (pressing freeze
will normally but not always do this). Note: this is a lengthy, iterative
process.
5. Once the maximum location is found, allow the TTO to cool to ambient
temperature and record this temperature.
6. Switch the ultrasound on and measure the temperature after 5 minutes.
7. Calculate the temperature rise and apply the derating factor based on
centre frequency (f) and distance from transducer (z) (derating factor is
I0.3 = l10−0.3f.z where 0.3 refers to an attenuation coefficient of 0.3 dB cm−1
MHz−1) (AGNIR, 2010) or use acoustic attenuation (TMM).
8. The process may be repeated using equipment settings used clinically if
desired.

[Link] Recorded values and recommended actions

For each transducer and mode, record the maximum temperature rise, the measurement
location in the beam and the ultrasound equipment settings.

75
Quality Assurance of Ultrasound Imaging Systems

For obstetric imaging, a temperature increase of greater than 4°C in 5 minutes in an

appropriate TTO would be of serious concern (WFUMB, 1998). Such results should be
drawn to the attention of the users immediately.
If the displayed TI differs from the measured temperature rise by >±50%, then the issue is
reported to the users so that they can make appropriate temporary adjustments to their
practice. Further investigations involving repeat measurements should then be undertaken
to establish findings before reporting the issue to the manufacturers or agent. These will
also establish if the changes to user practice need to be applied for a longer period.

76
Appendix A B-mode test objects
No test objects are required for the sensitivity, noise and basic uniformity tests described
in Chapter 2.
The non-linear calliper checks in Chapter 2 require an open topped test object. There are
none commercially available but construction is fairly straightforward if mechanical work-
shop facilities are available (see Dudley and Griffith (1996) for an example). Two plastic
sheets approximately 12 cm × 12 cm should be drilled (0.5 mm drill) in the required pattern
and held at least 3 cm apart by pillars. The holes should be threaded with 0.1 mm diameter
nylon filament (fishing line is suitable and is readily available), which should be pulled
tight and secured. This construction should then be placed into a rigid plastic container and
submerged in liquid with the correct speed of sound (1540 m s–1) – a number of options are
available, e.g. 9.5% ethanol by volume in distilled water at 20ºC (Martin and Spinks,
2001), or water at 50ºC; the speed of sound in all such liquids has a significant temperature
dependence.
Linear calliper checks, follow-up uniformity checks and low contrast penetration measure-
ments described in Chapter 2 require a simple tissue mimicking test object. This should
contain tissue mimicking material and appropriate targets (e.g. nylon filaments). The speed
of sound should be 1540 m s–1, unless an open topped test object is to be used for all
calliper checks in which case alternatives such as urethane (1450 m s–1) may be considered
for consistency checks. Attenuation and backscatter should be appropriate to the genera-
tion of speckle throughout the normal clinical depth of imaging. Attenuation of 0.5 to
0.7 dB cm–1 MHz–1 is likely to be suitable for most applications; the higher attenuation may
be more suitable for penetration measurements at low frequencies. Note that in some test
objects, the dependence of attenuation coefficient on frequency is not linear.
The extended tests described in Chapter 5 require a tissue mimicking test object with
additional targets: anechoic targets of sizes sufficiently small to challenge the resolution of
the scanners and grey-scale targets with increased and decreased backscatter compared to
background at levels close enough to background to challenge the contrast capabilities of
the scanner. Such test objects are also likely to be suitable for the tests in Chapter 2. A
speed of sound of 1540 m s–1 is desirable, although urethane may have advantages in terms
of durability and lifetime; the lower speed in urethane affects resolution measurements
and causes mis-registration artefacts where spatial compounding is used to form the
image. Again, an attenuation of 0.5 to 0.7 dB cm–1 MHz–1 is likely to be suitable for most
applications.
Table A1 shows examples of commercially available test objects; most manufacturers offer
other, more specialised, test objects. Browne et al. (2003) have investigated the properties
of a number of tissue mimicking materials.

77
78
Table A1 Manufacturers, website addresses and the models of test object available for basic (Chapter 2) and extended (Chapter
5) testing
Manufacturer Materials HQ and UK distributor Basic tests Extended tests
websites
ATS Urethane unless [Link] 535 (general) 539 (general)
otherwise stated [Link] 549 (general and small parts)
551 (small parts)
514 (gel, general and small parts)
CIRS Zerdine unless [Link] 054 040
otherwise stated [Link]. 042 (urethane)
[Link] 044 (grey-scale and anechoic targets)
Quality Assurance of Ultrasound Imaging Systems

Precision Gel [Link] Edinburgh Pipe Phantom Resolution

Acoustics Integral measurement
Gammex Gel [Link] 411 (general) 403 GS (general)
No UK website 404 (small parts) 404 GS (small parts)
421 (slice thickness)
Tissue Foam and saline No manufacturer’s website Signal-to-noise ratio measurement
Characterisation [Link]
Consulting
 Appendix A

Test objects for general use are suitable for lower frequencies, where penetration is greater
than 10 cm, and small parts test objects are more suitable for higher frequencies.

References
Browne, J.E., Ramnarine, K.V., Watson, A.J. and Hoskins, P.R. (2003) Assessment of the
acoustic properties of common tissue-mimicking test phantoms. Ultrasound in Medicine &
Biology, Vol. 29, 1053–1060.
Dudley, N.J. and Griffith, K. (1996) The importance of rigorous testing of circumference
measuring calipers. Ultrasound in Medicine & Biology, Vol. 22, 1117–1119.
Martin, K. and Spinks, D. (2001) Measurement of the speed of sound in ethanol/water
mixtures. Ultrasound in Medicine & Biology, Vol. 27(2), 289–291.

79
Appendix B Features and calibration of a
string phantom
Overall design
The string phantom consists of a moving filament. Generally, the filament is threaded
around several wheels, one of which is the drive wheel, and the others are free wheels. The
drive wheel is connected to a motor whose rotational rate is controlled by an external drive
box or computer. The main design criterion which affects performance is whether the
filament loop is completely submerged or only partially submerged. Mounting of the drive
wheel out of the water may be considered desirable by some manufacturers, however,
as the filament enters the water, it drags bubbles into the water which can appear on the
Doppler spectrum. Ideally, the entire filament should be submerged, hence the drive wheel
and motor should also be submerged. Some designs go halfway to this ideal by utilising a
drive pulley which drives a belt from above the surface of the water to move the imaged
filament which remains on submerged pulleys. This reduces the air bubbles which may be
dragged around the imaged filament and avoids the need to submerge a motor.

Filament composition and size

The acoustic backscatter characteristics of the filament are important. Blood has a rela-
tively uniform backscatter with angle whereas it is known that spiral wound filaments such
as silk or cotton exhibit preferential scattering of sound at particular angles (Cathignol
et al., 1994; Hoskins, 1994a). The Doppler spectra from such filaments have an unusual
appearance which may give the impression of reduced spectral broadening. When used to
check velocity measured using Doppler ultrasound, this can create false readings. Hence
the use of spiral wound filaments is not advised. Published work (Hoskins, 1994b) sug-
gests that the backscatter variation from O-ring rubber varies smoothly with angle, and that
this is a suitable target.
Ideally, the filament should have minimal diameter in order that measurements of sample
volume and range gate registration can be made. Modern Doppler systems have a mini-
mum sample volume of 0.5–1 mm, hence it is desirable that range gate registration can be
measured to an accuracy of 0.5 mm. Ideally, the filament diameter would be less than this,
however, commercial O-ring rubber is produced which is typically 1 mm in diameter. This
represents a testing limitation if a single filament is to be used.

Fluid characteristics
The fluid in which the filament is suspended should have a speed of sound of 1540 m s–1.
This is usually achieved using a 9% solution by volume of glycerol and water although if
it is to be used for a prolonged period, an antibacterial agent is required. The fluid should
be physically clean and degassed as air bubbles and dirt will be dragged along by the
moving filament and detected on the Doppler spectrum.

80
 Appendix B

Tank
The tank should be physically clean. Lining of the bottom of the tank with acoustic
absorber should always be performed to prevent the ultrasound beam bouncing around the
tank and creating false Doppler signals. For some string phantoms, false Doppler signals
can be generated from the wheels, through insonation by beam sidelobes, which may be
avoided by shielding with acoustic absorber.

Overall movement
The position of the filament should remain fixed even when the filament is moving. Such
movement might occur, for example, if one of the wheels was mounted offset, or there
may be some transient deviation if the filament is knotted, though this is not generally of
concern.

Control of filament speed

There should be the facility to drive the filament at a steady speed within a physiological
range with low speed variability (i.e. <1%) from about 1 to 100 cm s–1. If filament speeds
much beyond this are used, then there may be considerable movement, though some string
phantoms may be specified with very high filament speeds up to 10 m s–1. The system
should also allow the operator to drive the filament in a physiologic manner, with control
over both the waveform and the cyclic rate. This is best done using a programmable system
or by computer rather than by manual adjustment of knobs and other hand controls as these
are less reproducible. It is desirable that waveforms with different pulsatilities are available
from high pulsatile femoral waveforms to low pulsatile carotid waveforms.

Display and output

The minimum information which the string phantom must display is the filament speed
(cm s–1). For pulsatile flow, a real-time display of filament speed is not useful as this varies
too quickly. Ideally, there should be some facility to display the intended time–velocity
waveform, and to provide details of the intended variation in speed with time in a quantita-
tive form, for example, as data files on an accompanying disc, to allow comparison of
Doppler data with the intended filament velocity waveform.

Calibration
It is generally assumed that the filament is in contact with the drive wheel, i.e. no slip,
hence the speed of the drive wheel where it contacts the filament is equal to the speed of
the filament. The display of speed on the string phantom may in the first instance simply
be the speed of the drive wheel, assuming that the filament is infinitesimally thin. This is
not the case for O-ring rubber, hence the speed of the O-ring may be slightly higher than
that of the drive wheel by a few percent. For this reason, it is generally advisable to check
the calibration of the displayed velocity. It is possible to perform this in-house using the

81
Quality Assurance of Ultrasound Imaging Systems

following procedure (Hoskins, 1996). This relies on measuring the length of the filament
and the time for the filament to traverse a single loop.
1. Nylon target. Knot a thin nylon loop on the O-ring with 2 mm or so of nylon
sticking out. This acts as the target which produces a regular series of clicks
as it passes through the beam, especially if the O-ring is misaligned slightly
so that the beam mainly intercepts the nylon rather than the O-ring.
2. Measurement of loop time. Set a speed of about 20 cm/s and note what this
is. Use the time measurement facility on the Doppler display to measure the
time T taken for the nylon target to pass several times and calculate the time
for a single passage.
3. Measurement of loop distance. The filament is stretched when it is mounted
on the string phantom, so the increase in length needs to be measured.
Switch off the string phantom and remove from the tank. Dry the filament
and attach a second nylon marker to the O-ring. Measure the distance L1
between the two nylon markers with the O-ring mounted. Take the O-ring
off and measure the distance L2 between the two nylon markers. The
stretch factor S is L1/L2. Measure the total length L3 of the O-ring without
stretching. The final estimated length L4 of the O-ring is S*L3.
4. Calculation of calibration factor. The true speed is L4/T. This should be
compared with the displayed speed to derive a calibration factor. For a drive
wheel which employs a 90o groove, the calibration factor will be about
1.05.

References
Cathignol, D., Dickerson, K., Newhouse, V.L., Faure, P. and Chapelon, J.Y. (1994) On
the spectral properties of Doppler thread phantoms. Ultrasound in Medicine & Biology,
Vol. 20, 601–610.
Hoskins, P.R. (1994a) Choice of moving target for a string phantom. I. Backscattered
power characteristics. Ultrasound in Medicine & Biology, Vol. 20, 773–780.
Hoskins, P.R. (1994b) Choice of moving target for a string phantom. II. On the testing of
Doppler ultrasound devices. Ultrasound in Medicine & Biology, Vol. 20, 781–789.
Hoskins, P.R. (1996) Accuracy of maximum velocity estimates made using Doppler
ultrasound systems. British Journal of Radiology, Vol. 69, 172–177.

82
Appendix C Ultrasound safety test
equipment
Acoustic force balances
The following pages are intended to act as a guide to acoustic power balances for diagnos-
tic ultrasound. The measurement devices described here were commercially available at
the time of writing. Other devices may become available in the future and this general
guidance about the principles of acoustic measurement will still apply.

Sensitivity
Sensitivity is of fundamental importance in selecting an acoustic force balance for a given
purpose. Ideally for diagnostic measurements, an acoustic force device would be able to
measure as low as 1 mW of acoustic power. However, this level of sensitivity has yet to
be achieved in a commercial design. Some force balances currently on the market claim
sensitivity down to 10 mW but this requires ideal measurement conditions, i.e. draught
and vibration free. One balance claims to achieve a sensitivity of 1 mW by employing
measurement averaging. At acoustic powers of several tens of mW, the measurements
become considerably easier to make and more reliable. For therapy level ultrasound, an
acoustic balance needs to be able to measure up to at least 10 W and to handle the heat
generated by absorption of the acoustic beam.

Physical aspects
1) General
• The size and shape of the measurement container must to be sufficient to
accommodate the desired range of transducers, e.g. longer linear arrays and
fetal heart monitors.
• Weight may be important if the balance is to be transported regularly.
• The measured result may be viewed as a LCD display or logged to a
computer.
• Environmental factors such as heat and air conditioning will affect the
performance of the balance.
• An acoustic force balance requires a solid, level surface as a base and will
take a time to stabilise after moving.
2) Target
• The size of the target will determine the size of the beam that can be
measured. Even though the chamber is sufficiently large to accommodate

83
Quality Assurance of Ultrasound Imaging Systems

the transducer, the total power will not be measured if the target does not
‘see’ the whole beam.
• Power cannot be measured accurately in a sector beam using either a
reflecting or an absorbing target.
• Alignment errors are greater for reflecting than for absorbing targets
and may lead to a significant reduction in measured power. The geometry
of focused fields may also give rise to unpredictable errors with conical
reflecting targets.
• An absorbing target may heat up as acoustic energy is absorbed, resulting
in errors in measurement. Measurements may also be frequency-dependent,
requiring correction. Reflecting target devices are not immune to thermal
effects as it is necessary to absorb the reflected beam in the measurement
chamber and hence heat can be generated giving secondary effects.
• In order to measure the thermal index (TI) in scanned mode, the area of
the target that is insonated by the beam must be restricted.
3) Transmission fluid
• Convection currents can arise in the fluid due to environmental effects.
An open topped system is vulnerable and will need to be shielded from
draughts when measuring low powers.
• Convection currents arise in the fluid due to thermal effects as the
device absorbs acoustic energy. This is more likely for measurements at
high powers and after a sustained period of use at moderate powers.
• A short transmission path is required since transmission losses in water will
lead to inaccurate results, particularly at high frequencies.
4) Cost
The cost of an acoustic force balance in 2010 was around £4600 with calibration at
around £1000.
5) Commercial equipment (in alphabetical order)
i) NPL Acoustic Force Balance
Manufacturers’ specifications
Absorbing target
Measurement range 10 mW–10 W
Recommended frequency range 1 MHz–10 MHz
Measurement uncertainties (range 50 mW to 5 W) less than ±7% at 95%
confidence
Accommodates transducer diameters up to 5 cm (on standard model)
Digital display and can be connected to PC
Traceable to UK primary standard.

84
 Appendix C

Suppliers of NPL Balance

Precision Acoustics, Hampton Farm Business Park, Higher Bockhampton,
Dorchester, Dorset, DT2 8QH, UK
Tel: 01305 264669; Fax: 01305 260866
[Link]
ii) Ohmic UPM-DT 1AV ultrasound power meter
Manufacturers’ specifications
Reflecting target (45° air-backed cone)
Measurement range 0–30 W
Frequency 0.5 MHz–10 MHz
Measurement uncertainties (at 10 mW) ±3%
Accommodates transducer diameters up to 3 inches (approximately
7.5 cm)
Weight 6.4 kg
Digital display
Suppliers of Ohmic Balance
Available from: Ohmic Instruments Co., 508 August Street, Easton, MD
21601, USA
Tel: +1 410 820 5111; Fax: +1 410 822 9633
Email: ohmic@[Link]
[Link]
iii) Onda Corporation RFB Radiation Force Balance
Manufacturers’ specifications
Conical target
Measurement range 1 mW (with extended averaging)–2 W
Low frequency limit 1 MHz
Measurement uncertainty <5%
Maximum beam diameter 5 cm
Noise <3 mW without averaging
Calibration certificate supplied
USB interface and Windows XP compatible
Suppliers of Onda Balance
Available from: Onda Corporation, 592 Weddell Drive, Suite 7, Sunnyvale,
CA 94089, USA
Tel: +1 408 745 0383; Fax: +1 408 745 0956
Email: info@[Link]
[Link]
European distributor: Dahn International ApS, Denmark

85
Quality Assurance of Ultrasound Imaging Systems

6) Build your own power balance

A number of designs for successful power balances can be found in the literature.
It is beyond the scope of this publication to compare different designs. Further
information can be found in a useful review published by Perkins (1989). Methods
for measuring acoustic power are described in detail in IEC 61161 (IEC, 2006) which
has become a British Standard BS EN 61161:2007 (BSI, 2007).
7) Check source
Precision Acoustics also market two versions of a continuous wave check source
which is a sound source which can be calibrated to a traceable standard and serves to
test radiation force balance calibration.
Manufacturers’ specifications
Excitation Two continuous-wave outputs at specified frequencies
Transducers Two 19 mm diameter, unfocused at specified frequencies
Nominal power output 10 mW, 100 mW and 1 W – front-panel selected, inter-
mediate ‘off’
Stability Power output stable to better than ±2.5% during
1000 hours of use
Calibration accuracy ±7% as standard (see Option 1 or Option 2 below)
Electrical input Supply voltage 80 V to 250 V, 50 Hz to 60 Hz, universal
Cost around £6000
Option 1 Calibration of power output against a secondary standard
radiation force balance – accuracy = ±7% (95% con-
fidence)
Option 2 Calibration of power output against the NPL primary
standard radiation force balance – accuracy = ±4% (95%
confidence).
8) Making an aperture limiting mask for acoustic power measurement for TI
In order to measure the acoustic power generated by the central 1 cm of a transducer,
as required to determine the scanning TI, an aperture limiting mask is required for the
force balance.
The mask is constructed of an attenuating material in which there is a 1 cm wide slit
which has a lateral dimension exceeding the width of the transducer. The slit is posi-
tioned at the centre of the long axis of the transducer, with the 1 cm width aligned with
the long axis. In this way, only the acoustic power emitted from a 1 cm strip of the
ultrasound transducer is detected. A material with a high attenuation coefficient and
with minimum impedance match with water is recommended and additional attenua-
tion can be obtained by sandwiching a low impedance reflector between two layers of
the attenuating material. The design is described in IEC 62359:2005 (IEC, 2005).

86
 Appendix C

Temperature measurement
Radiometer/Infrared pyrometer
A small field of view is required. The temperature range covered should extend from room
temperature to above 50°C. A resolution of 0.1°C is desirable for accurate measurement.
Suggested suppliers (this is not intended as a complete list):
Linear Laboratories Model C-1600MP supplied by Calex, PO Box 2, Leighton Buzzard,
Beds, LU7 1WZ, UK.
Tel: 01525 373178
Fluke Precision Measurement Ltd, Hurricane Way, Norwich, Norfolk NR6 6JB, UK.
Tel: 01603 256758; Fax: 01603 256759, e-mail: fpm@[Link]

Thermal test object

Thermal test objects with embedded thermocouples are available from the National
Physical Laboratory and are connected to a micro-voltmeter in use. The dimensions of the
test objects are 150 mm diameter × 55 mm deep. The test objects comprise a thin film
thermocouple embedded in a tissue mimicking material.
A soft tissue model and/or a bone-soft tissue model can be purchased. The design is shown
in Figure 5.5 in Chapter 5 and reproduced below. In addition, a positioning system is
required, which allows transducers of various sizes and shapes to be accurately aligned,
such as that shown in Figure C1.

Figure 5.5 Schematic of thermal test object (courtesy of National Physical Laboratory)

87
Quality Assurance of Ultrasound Imaging Systems

Figure C1 Commercially available thermal test object in test rig (courtesy of National
Physical Laboratory)

Supplier for TTO

National Physical Laboratory, Hampton Road, Teddington, Middlesex TW11 0LW, UK.
Tel: 020 8977 3222; Fax: 020 8614 0466
[Link]
objects

Hydrophone measurements
1) The following items (i–iv) are required to build a system for measurement of Ispta
using an RF power meter.

88
 Appendix C

i) Measurement tank/container
A glass or Perspex aquarium may be obtained from any pet supplier. Various
sizes are available and the main consideration is ease of use and the size of
components to be inserted. For a vertical measurement system, a bucket can
be used.
ii) Positioning system
A positioning system which allows for fine adjustment of the order of 50 µm
over a minimum volume extending 150 mm × 100 mm × 100 mm. Suitable
systems may be obtained from suppliers of laboratory equipment. Alterna-
tively, Precision Acoustics supply a fully automated scanning system which
provides spatial resolution of 6.25 µm over a scanning volume of 600 mm
× 250 mm × 250 mm (see below for contact details).
A suitable transducer mount is also required, capable of securely supporting
a range of transducer sizes and shapes varying from 1.5 cm diameter up to
2.0 cm × 20 cm.
iii) Hydrophone
Needle hydrophones
Manufacturer’s specification (Precision Acoustics)
PVDF element
Element diameter 0.04 mm to 1 mm
Element thickness 9 µm to 28 µm
100 MHz wide-band amplifier
In-house calibration over the range 1 MHz to 20 MHz
Cost: Hydrophone £1080 (1 mm element) to £1860 (0.04 mm element) plus
amplifier and dc coupler £1080
Manufacturer’s specification (Onda Corporation)
Polymer or ceramic elements (model HNP and HNC)
Electrode diameter 0.2 mm to 1.5 mm
Frequency response 1 dB–20 dB (HNP), 1 dB–10 dB (HNC)
Capsule hydrophones
Manufacture’s specification (Onda Corporation, Model HGL)
Electrode size 0.085 mm to 1 mm
Frequency response ±3 dB over range 0.25 MHz to 40 MHz
Sensitivity (nV Pa–1) 8 (0.85 mm) to 45 (1 mm)
Acceptance angle >150° (0.85 mm) to 100° (1 mm)
Membrane hydrophones
Manufacturer’s specification (Precision Acoustics)
PVDF film
9, 13, 16 microns thick
Element diameter 0.2 mm or 0.4 mm

89
Quality Assurance of Ultrasound Imaging Systems

100 MHz wide-band amplifier

In-house calibration over the range 1 MHz to 20 MHz
Cost: Hydrophone £7355 plus signal compensator £1830
Manufacture’s specification (Onda Corporation, HMA series)
Frequency response 0.5 MHz to 45 MHz
Sensitivity (nV Pa–1) 100 to 250
Manufacture’s specification (Sonora Medical Systems, Model 805)
PVDF film
Frequency response 1 MHz to 20 MHz
Suppliers
1. Precision Acoustics Ltd, Hampton Farm Business Park, Higher Bock-
hamton, Dorchester, Dorset, DT2 8QH, UK
Tel: 01305 264669; Fax 01305 260866
[Link]
2. Onda Corporation, 592 Weddell Drive, Suite 7, Sunnyvale, CA 94089,
USA
Tel: +1 408 745-0383; Fax: +1 408 745-0956
Email: info@[Link]
[Link]
European distributor for Onda: Dahn International ApS, Denmark
3. Sonora Medical Systems Inc., 1751 S. Fordham St, Suite 100,
Longmont, CO 80503, USA
Tel: +1 888 476 6672
Email: sales@[Link]
[Link]
iv) RF Power meter
The following power meter may be used:
Rohde and Schwarz power meter, Model NRVS, together with Rohde and
Schwarz insertion unit (50 Ω).
Supplier
Rohde and Schwarz UK Ltd, 1 Ancells Business Park, Fleet, Hants GU51
2UZ, UK.
Tel: 01252 818888; Fax: 01252 811447
[Link]
2) Items i–iii above may be used to capture the acoustic waveform for off-line analysis
by the addition of a PC or laptop with appropriate software. A suitable range of
PC-based oscilloscopes is available from Pico Technology Ltd (for example). Any
high specification digital oscilloscopes may be used.

90
 Appendix C

Cost: From £600 for high specification PC-based oscilloscope with signal analysis
capability. Plus cost of standard PC or laptop computer
Suppliers (this list is not intended to be inclusive)
Pico Technology Limited, The Mill House, Cambridge Street, St Neots, Cambridge-
shire PE19 1QB, UK
Tel: 01480 396395; Fax: 01480 396296
[Link]
Tektonics UK Limited, Western Peninsula, Western Road, Bracknell, Berks RG12
1RF, UK
Tel: 01344 392 000; Fax: 01344 392401
[Link]@[Link]
[Link]
For Gould oscilloscopes: TICS International Ltd, Unit 7, Industrial Estate, Derby
Road, Hounslow, Middx TW3 3UH, UK
Tel: 0870 787 4910; Fax: 0870 787 4920
E-mail: sales@[Link]
LeCroy Ltd, 27 Blacklands Way, Abingdon Business Park, Abingdon, Oxon OX14
1DY, UK
Tel: 01235 536973; Fax: 01235 528796
Email: [Link]@[Link]
[Link]

References
BSI (2007) BS EN 61161 Ultrasonics – Power Measurement – Radiation Force Balances
and Performance Requirements. BSI, London, UK.
IEC (2005) International Standard IEC 62359:2005 Ultrasonics – Field Characterisation
– Test Methods for Determination of Thermal and Mechanical Indices Related to Medical
Diagnostic Ultrasonic Fields. International Electrotechnical Commission, Geneva,
Switzerland.
IEC (2006) International Standard IEC 61161:2006 Ultrasonics – Power Measurement
– Radiation Force Balances and Performance Requirements. International Electrotechni-
cal Commission, Geneva, Switzerland.
Perkins, M.A. (1989) A versatile force balance for ultrasound power measurement.
Physics in Medicine and Biology, Vol. 34, 1645–1652.

91
Appendix D
References
AGNIR (2010) Health Effects of Exposure to Ultrasound and Infrasound. RCA 14. HPA,
London, UK.

BMUS/BIR (2000) The Safe Use of Ultrasound in Medical Diagnosis, edited by G. ter
Haar and F. A. Duck. BMUS/British Institute of Radiology, London, UK.

BMUS/BIR (2009) Guidelines for the Safe Use of Diagnostic Ultrasound Equipment.
BMUS/BIR, London. [Link]
[Link]. [accessed on 18/06/2010]

Browne, J.E., Ramnarine, K.V., Watson, A.J. and Hoskins, P.R. (2003) Assessment of the
acoustic properties of common tissue-mimicking test phantoms. Ultrasound in Medicine
& Biology, Vol. 29, 1053–1060.

Browne, J.E., Watson, A.J., Hoskins, P.R. and Elliott, A.T. (2004) Validation of a sensitivity
performance index test protocol and evaluation of colour Doppler sensitivity for a range of
ultrasound scanners. Ultrasound in Medicine & Biology, Vol. 30, Issue 11, 1475–1483.

Cathignol, D., Dickerson, K., Newhouse, V.L., Faure, P. and Chapelon, J.-Y. (1994) On
the spectral properties of Doppler thread phantoms. Ultrasound in Medicine & Biology,
Vol. 20, Issue 7, 601–610.

Dineley, J., Meagher, S., Poepping, T.L., McDicken, W.N. and Hoskins, P.R. (2006)
Design and characterisation of a wall motion phantom. Ultrasound in Medicine & Biology,
Vol. 32, Issue 9, 1349–1357.

Dixon, K.L. and Duck, F.A. (1998) Determining the conditions for measurement of
spatial-peak temporal-averaged intensity in scanned ultrasound beams. British Journal of
Radiology, Vol. 71, 968–971.

Dudley, N.J. (2003) B-mode measurements. In: Diagnostic Ultrasound Physics and Equip-
ment, edited by P. R. Hoskins, A. Thrush, K. Martin and T. A. Whittingham. Greenwich
Medical Media Limited, London, UK.

Dudley, N.J. and Griffith K. (1996) The importance of rigorous testing of circumference
measuring calipers. Ultrasound in Medicine & Biology, Vol. 22, 1117–1119.

Dudley, N.J., Griffith, K., Houldsworth, G., Holloway, M. and Dunn, M.A. (2001) A
review of two alternative ultrasound quality assurance programmes. European Journal of
Ultrasound, Vol. 12, 233–245.

92
 Appendix D

Dudley, N.J., Gibson, N.M., Felckney, M.J. and Clark, P.D. (2002) The effect of speed of
sound in ultrasound test objects on lateral resolution. Ultrasound in Medicine & Biology,
Vol. 28, 1561–1564.

Evans, J.A., Price, R. and Luhana, F. (1989) A novel testing device for Doppler ultrasound
equipment. Physics in Medicine and Biology, Vol. 34, 1701–1708.

Fleming, A.D., McDicken, W.N., Sutherland, G.R. and Hoskins, P.R. (1994) Assessment
of colour Doppler tissue imaging using test-phantoms. Ultrasound in Medicine & Biology,
Vol. 20, Issue 9, 937–951.

Gibson, N.M., Dudley, N.J. and Griffith, K. (2001) A computerised ultrasound quality
control testing system. Ultrasound in Medicine & Biology, Vol. 27, 1697–1711.

Goldstein, A. (2000) The effect of acoustic velocity on phantom measurements. Ultrasound

in Medicine & Biology, Vol. 26, 1133–1143.

Goodsitt, M.M., Carson, P.L., Witt, S., Hykes, D.L. and Kofler, J.M. Jr. (1998) Real-time
B-mode ultrasound quality control test procedures. Report of AAPM Ultrasound Task
Group No. 1. Medical Physics, Vol. 25, 1385–1406.

Hoskins, P.R. (1994a) Choice of moving target for a string phantom: I. Measurement of
filament backscatter characteristics. Ultrasound in Medicine & Biology, Vol. 20, Issue 8,
773–780.

Hoskins, P.R. (1994b) Choice of moving target for a string phantom: II. On the perfor-
mance testing of Doppler ultrasound systems. Ultrasound in Medicine & Biology, Vol. 20,
Issue 8, 781–789.

Hoskins, P.R. (2008) Simulation and validation of arterial ultrasound imaging and blood
flow. Ultrasound in Medicine & Biology, Vol. 34, 693–717.

IPEM (1994) Testing of Doppler Ultrasound Equipment. IPEM Report 70. IPEM, York,
UK

IPEM (1995) Routine Quality Assurance of Ultrasound Imaging Systems, edited by R.

Price, M. Docker, K. Martin, D. McHugh and J. Pelmore. IPEM Report 71. IPEM, York,
UK.

IPEM (2001) Guidelines for Testing and Calibration of Physiotherapy Ultrasound

Machines, edited by S. Pye and B. Zeqiri. IPEM Report 84. IPEM, York, UK.

93
Quality Assurance of Ultrasound Imaging Systems

IPEM (2005) Recommended Standards for the Routine Performance Testing of Diagnostic
X-Ray Imaging Systems. IPEM Report 91. IPEM, York, UK.

IPEM (2009) Guide to Electrical Safety Testing of Medical Equipment: the Why and the
How, edited by S. Wentworth. IPEM Report 97. IPEM, York, UK.

Kanal, K.M., Kofler, J.M. and Groth, D.S. (1998) Comparison of selected ultrasound per-
formance tests with varying overall receiver gain and dynamic range, using conventional
and magnified field of view. Medical Physics, Vol. 25, 642–647.

Kofler, J.M. Jr and Madsen, E.L. (2001) Improved method for determining resolution
zones in ultrasound phantoms with spherical simulated lesions. Ultrasound in Medicine &
Biology, Vol. 27, 1667–1676.

Kollmann, C. (1995) Results of a study of quality control of diagnostic ultrasound equip-

ment. Ultraschall in der Medizin, Vol. 16, 206–209.

Kollmann, C., Bergmann, H., Trabold, T. and Zotz, R. (2001) A testing device for quality
assurance of 3D-ultrasound equipment. Zeitschrift für medizinische Physik, Vol. 11, 45–
52.

Lange, G.J. and Loupas, T. (1996) Spectral and color Doppler sonographic applications
of a new test object with adjustable moving target spacing. Journal of Ultrasound in
Medicine, Vol. 15, 775–784.

Li, S., Hoskins, P.R. and McDicken, W.N. (1997) Rapid measurement of the spatial resolu-
tion of colour flow scanners. Ultrasound in Medicine & Biology, Vol. 23, Issue 4, 591–
596.

Li, S.F., Hoskins, P.R., Anderson, T. and McDicken, W.N. (1998) An acoustic injection test
object for colour flow imaging systems. Ultrasound in Medicine & Biology, Vol. 24, Issue
1, 161–164.

Martin, K. and Spinks, D. (2001) Measurement of the speed of sound in ethanol/water

mixtures. Ultrasound in Medicine & Biology, Vol. 27(2), 289–291.

McAleavey, S., Hah, Z. and Parker, K. (2001) A thin film phantom for blood flow simula-
tion and Doppler test. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency
Control, Vol. 48, 737–742.

Meagher, S., Poepping, T.L., Ramnarine, K.V., Black, R.A. and Hoskins, P.R. (2007)
Anatomical flow phantoms of the nonplanar carotid bifurcation, Part II: Experimental
validation with Doppler ultrasound. Ultrasound in Medicine & Biology, Vol. 33, Issue 2,
303–310.

94
 Appendix D

MHRA (2006) Device Bulletin. Managing Medical Devices. Guidance for Healthcare and
Social Services Organisations. DB2006(5). Medical and Healthcare Products Regulatory
Agency, London, UK.

Moore, G.W., Gessert, A. and Scafer, M. (2005) The need for evidence based quality
assurance in the modern ultrasound clinical laboratory. Ultrasound, Vol. 13, 158–162.

Novario, R., Goddi, A., Crespi, A. and Conte, L. (1994) A new phantom for quality
assurance of color-coded ultrasound flow equipment. Physica Medica, Vol. 10, 101–106.

NPL (2008) Good Practice Guide to Measurement of Acoustic Output Power. National
Physical Laboratory, Teddington, Middlesex, UK. [Link]
ConWebDoc.2282 [accessed on 18/06/2010]

Poepping, T.L., Nikolov, H.N., Thorne, M.L. and Holdsworth, D.W. (2004) A thin-walled
carotid vessel phantom for Doppler ultrasound flow studies. Ultrasound in Medicine &
Biology, Vol. 30, Issue 8, 1067–1078.

Pye, S.D., Ellis, W. and MacGillivray, T. (2004) Medical ultrasound: a new metric of
performance for grey-scale imaging. Journal of Physics: Conference Series, Vol. 1, 187–
192.

Ramnarine, K.V., Nassiri, D.K., Hoskins, P.R. and Lubbers, J. (1998) Validation of a new
blood-mimicking fluid for use in Doppler flow test objects. Ultrasound in Medicine &
Biology, Vol. 24, Issue 3, 451–459.

Ramnarine, K.V., Hoskins, P.R., Routh, H.F. and Davidson, F. (1999) Doppler backscatter
properties of a blood-mimicking fluid for Doppler performance assessment. Ultrasound in
Medicine & Biology, Vol. 25, Issue 1, 105–110.

Ramnarine, K.V., Anderson, T. and Hoskins, P.R. (2001) Construction and geometrical
stability of physiological flow rate wall-less stenosis phantoms. Ultrasound in Medicine &
Biology, Vol. 27, Issue 2, 245–250.

Rickey, D.W., Picot, P.A., Christopher, D.A. and Fenster, A. (1995) A wall-less vessel
phantom for Doppler ultrasound studies. Ultrasound in Medicine & Biology, Vol. 21, Issue
9, 1163–1176.

Rickey, D.W. and Fenster, A. (1996) A Doppler ultrasound clutter phantom. Ultrasound in
Medicine & Biology, Vol. 22, Issue 6, 747–766.

Samavat, H. and Evans, J.A. (2006) An ideal blood mimicking fluid for Doppler
ultrasound phantoms. Journal of Medical Physics, Vol. 31, Issue 4, 275–278.

95
Quality Assurance of Ultrasound Imaging Systems

Satrapa, J., Doblhoff, G. and Schultz, H.J. (2002) Automated quality control of diagnostic
ultrasound appliances. Ultraschall in der Medizin, Vol. 23, 123–128.

Shaw, A. and Hekkenberg, R. (2007) Standards to Support Performance Evaluation

for Diagnostic Ultrasound Imaging Equipment. AC2. National Physical Laboratory,
Teddington, Middlesex, UK.

Skolnick, M.L. (1991) Estimation of ultrasound beam width in the elevation (slice thickness)
plane. Radiology, Vol. 180, 286–288.

Teirlinck, C.J.P.M., Bezemer, R.A., Kollmann, C., Lubbers, J., Hoskins, P.R., Fish, P.,
Fredfeldt, K.-E. and Schaarschmidt, U.G. (1998) Development of an example flow test
object and comparison of five of these test objects, constructed in various laboratories.
Ultrasonics, Vol. 36, Issues 1–5, 653–660.

US FDA (2008) Guidance for Industry and FDA Staff – Information for Manufacturers
Seeking Marketing Clearance of Diagnostic Ultrasound Systems and Transducers. http://
[Link]/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/
[Link]. [accessed on 06/2010]

van Wijk, M.C. and Thijssen, J.M. (2002) Performance testing of medical ultrasound
equipment: fundamental vs. harmonic mode. Ultrasonics, Vol. 40, 585–591.

Watts, D.M., Sutcliffe, C.J., Morgan, R.H., Meagher, S., Wardlaw, J., Connell, M., Bastin,
M.E., Marshall, I., Ramnarine, K.V., Hoskins, P.R. and Black, R.A. (2007) Anatomical
flow phantoms of the nonplanar carotid bifurcation, Part I: Computer-aided design and
fabrication. Ultrasound in Medicine & Biology, Vol. 33, Issue 2, 296–302.

WFUMB (1998) Conclusion and recommendations on thermal and non-thermal mecha-

nisms for biological effects of ultrasound, in: World Federation for Ultrasound in
Medicine and Biology Symposium on Safety of Ultrasound in Medicine, edited by S. B.
Barnet. Ultrasound in Medicine & Biology, Vol. 24, 1–55.

Whittingham, T.A., Mitchell, G., Tong, J. and Feeney, M. (2004) Towards a portable
system for the measurement of thermal and mechanical indices. Journal of Physics:
Conference Series, Vol. 1, 64–71.

Zeqiri, B. (2007) Metrology for ultrasonic applications. Progress in Biophysics and

Molecular Biology, Vol. 93, 138–152.

Standards
AIUM/NEMA (1992) The Standard for Real-Time Display of Thermal and Mechanical
Acoustic Output Indices on Diagnostic Ultrasound Equipment. AIUM/NEMA, Laurel,
MD.

96
 Appendix D

AIUM/NEMA (2004a) UD 3-2004 Revision 2 Standard for Real-Time Display of Thermal

and Mechanical Acoustic Output Indices on Diagnostic Ultrasound Equipment. AIUM,
Rockville Pike, MD.

AIUM/NEMA (2004b) UD 2-2004 Revision 3 Acoustic Output Measurement Standard for

Diagnostic Ultrasound Equipment. AIUM, Rockville Pike, MD.

BSI (2007) BS 61157:2007 Standard Means for the Reporting of the Acoustic Output of
Medical Diagnostic Ultrasonic Equipment. BSI, London, UK.

BSI (2008) BS EN 62353:2008 Medical Electrical Equipment. Recurrent Test and Test
After Repair of Medical Electrical Equipment. BSI, London, UK.

IEC (1994) International Standard IEC 61102 Specification for Measurement and Char-
acterisation of Ultrasound Fields Using Hydrophones in the Frequency Range 0.5 MHz
to 15 MHz. International Electrotechnical Commission, Geneva, Switzerland (historical
document now withdrawn).

IEC (1999) IEC 61895 Ultrasonics – Pulsed Doppler Diagnostic Systems – Test Proce-
dures to Determine Performance. International Electrotechnical Commission, Geneva,
Switzerland.

IEC (2001) IEC 61685 Ultrasonics – Flow Measurement Systems – Flow Test Object.
International Electrotechnical Commission, Geneva, Switzerland.

IEC (2005a) International Standard IEC 62359:2005 Ultrasonics – Field Characterisa-

tion – Test Methods for Determination of Thermal and Mechanical Indices Related
to Medical Diagnostic Ultrasonic Fields. International Electrotechnical Commission,
Geneva, Switzerland.

IEC (2005b) International Standard IEC 60601-1:2005 Medical Electrical Equipment.

General Requirements for Basic Safety and Essential Performance. International Electro-
technical Commission, Geneva, Switzerland.

IEC (2006) International Standard IEC TS 62306 Ultrasonics – Field Characterisation –

Test Objects for Determining Temperature Elevation in Diagnostic Ultrasound Fields.
International Electrotechnical Commission, Geneva, Switzerland.

IEC (2007a) International Standard IEC 60601-2-37:2007 Medical Electrical

Equipment – Part 2-37: Particular Requirements for the Safety of Ultrasound Medical
Diagnostic and Monitoring Equipment. International Electrotechnical Commission,
Geneva, Switzerland.

97
Quality Assurance of Ultrasound Imaging Systems

IEC (2007b) International Standard IEC 62353:2007 Medical Electrical Equipment.

Recurrent Test and Test after Repair of Medical Electrical Equipment. International
Electrotechnical Commission, Geneva, Switzerland.

IEC (2007c) International Standard IEC 62127-1:2007 Ultrasonics: Hydrophones Part 1:

Measurement and Characterisation of Medical Ultrasonic Fields up to 40 MHz. Interna-
tional Electrotechnical Commission, Geneva, Switzerland.

98