2019 IEEE International Ultrasonics Symposium (IUS)
Nonalcoholic fatty liver disease (NAFLD) affects ~25% of the world population. Confounder-correct... more Nonalcoholic fatty liver disease (NAFLD) affects ~25% of the world population. Confounder-corrected chemical-shift-encoded MRI-derived proton density fat fraction (MRI-PDFF) is an established quantitative noninvasive biomarker of hepatic steatosis but has limited availability. There is a clinical need for more practical and accessible methods to noninvasively assess hepatic steatosis. Previous work has shown that two quantitative ultrasound (QUS) biomarkers - attenuation coefficient (AC) and backscatter coefficient (BSC) - are correlated with hepatic steatosis. Examining a broad range of QUS biomarkers, this study aims to develop an improved, multi-parametric QUS-based approach to diagnose NAFLD and quantify hepatic fat, with MRI-PDFF as the reference standard. 102 participants recruited from the UCSD NAFLD Research Center underwent QUS exams on the right liver lobe with an Acuson S3000 ultrasound scanner and the 4C1 and 6C1HD transducers. Seven QUS biomarkers - AC, BSC, three Lizzi-Feleppa parameters (slope, intercept, midband), and two envelope parameters (k and μ) - were derived from ultrasound radiofrequency data. Two multivariable models were developed based on QUS biomarkers: a generalized linear regression model to predict hepatic PDFF using stepwise regression for biomarker selection and a regularized logistic regression model to classify NAFLD (MRI-PDFF>5%, N=78/102) versus no NAFLD (MRI-PDFF≤5%) using LASSO regularization for biomarker selection. Leave-one-out cross-validation was performed. The final regression model selected the midband and k-parameter. The cross-validated predicted PDFF values were correlated with the reference MRI-PDFF values (Spearman ρ = 0.82 and Pearson’s r = 0.76). In comparison, Pearson’s r was 0.59 between AC and MRI-PDFF and 0.58 between BSC and MRI-PDFF. The final classifier model selected the midband, k-parameter and μ-parameter, achieving an area under the receiver operating characteristic curve (AUROC) of 0.88. In comparison, AUROC was 0.83 using AC and 0.84 using BSC. The results suggest that multi-parametric QUS can improve the quantification of hepatic steatosis and diagnosis of NAFLD.
Objectives To compare the diagnostic accuracy of US shear wave elastography (SWE) and magnetic re... more Objectives To compare the diagnostic accuracy of US shear wave elastography (SWE) and magnetic resonance elastography (MRE) for classifying fibrosis stage in patients with nonalcoholic fatty liver disease (NAFLD). Methods Patients from a prospective single-center cohort with clinical liver biopsy for known or suspected NAFLD underwent contemporaneous SWE and MRE. AUCs for classifying biopsy-determined liver fibrosis stages ≥ 1, ≥ 2, ≥ 3, and = 4, and their respective performance parameters at cutoffs providing ≥ 90% sensitivity or specificity were compared between SWE and MRE. Results In total, 100 patients (mean age, 51.8 ± 12.9 years; 46% males; mean BMI 31.6 ± 4.7 kg/m2) with fibrosis stage distribution (stage 0/1/2/3/4) of 43, 36, 5, 10, and 6%, respectively, were included. AUCs (and 95% CIs) for SWE and MRE were 0.65 (0.54–0.76) and 0.81 (0.72–0.89), 0.81 (0.71–0.91) and 0.94 (0.89–1.00), 0.85 (0.74–0.96) and 0.95 (0.89–1.00), and 0.91 (0.79–1.00) and 0.92 (0.83–1.00), for dete...
The Journal of the Acoustical Society of America, 2019
Quantitative ultrasound (QUS) aims to improve diagnostic ultrasound imaging by extracting objecti... more Quantitative ultrasound (QUS) aims to improve diagnostic ultrasound imaging by extracting objective tissue parameters from backscattered signals. Deep learning can facilitate this process because of its ability to extract high-level information from the raw data. We have demonstrated that deep learning approaches applied to backscattered signals can accurately quantify liver fat noninvasively. We will discuss three deep learning approaches herein: (1) a one-dimensional convolutional neural network (1-D-CNN) for uncalibrated radiofrequency (RF) data; (2) a two-dimensional CNN (2-D-CNN) for uncalibrated spectrograms; and (3) a 2-D-CNN for calibrated spectrograms. Approaches (1) and (2) do not require system calibration with a physical phantom, whereas Approach (3) obtains calibrated spectrograms with a physical phantom. Each approach is evaluated using three datasets that contain ultrasound RF data acquired from the right liver lobe and contemporaneous MRI-PDFF (reference standard) in participants with and without nonalcoholic fatty liver disease: 204 scanned by a Siemens S2000 ultrasound scanner (4C1), 70 scanned by a Siemens S3000 scanner (4C1 and/or 6C1HD), and 104 scanned by both the S3000 and a GE Logiq e9 scanner (C1-6). We will discuss the advantages and disadvantages of each approach evaluated using data acquired across multiple ultrasound scanner platforms. [Supported by R01DK106419.]Quantitative ultrasound (QUS) aims to improve diagnostic ultrasound imaging by extracting objective tissue parameters from backscattered signals. Deep learning can facilitate this process because of its ability to extract high-level information from the raw data. We have demonstrated that deep learning approaches applied to backscattered signals can accurately quantify liver fat noninvasively. We will discuss three deep learning approaches herein: (1) a one-dimensional convolutional neural network (1-D-CNN) for uncalibrated radiofrequency (RF) data; (2) a two-dimensional CNN (2-D-CNN) for uncalibrated spectrograms; and (3) a 2-D-CNN for calibrated spectrograms. Approaches (1) and (2) do not require system calibration with a physical phantom, whereas Approach (3) obtains calibrated spectrograms with a physical phantom. Each approach is evaluated using three datasets that contain ultrasound RF data acquired from the right liver lobe and contemporaneous MRI-PDFF (reference standard) in participants with and ...
The Journal of the Acoustical Society of America, 2019
Nonalcoholic fatty liver disease (NAFLD) is a leading cause of chronic liver disease worldwide. C... more Nonalcoholic fatty liver disease (NAFLD) is a leading cause of chronic liver disease worldwide. Current diagnosis relies on invasive biopsy. Quantitative ultrasound (QUS) may provide a noninvasive alternative, using parameters such as attenuation (AC) and backscatter (BSC) coefficients. This study examines how exclusion of larger blood vessels from fields of interest (FOI) affects AC and BSC measurements in NAFLD patient livers. 201 total participants with known or suspected NAFLD underwent liver right lobe QUS. Of these, 84 underwent a second QUS exam on the same day and 172 patients had contemporaneous MRI with computation of proton density fat fraction (MRI-PDFF). AC and BSC were calculated from FOIs drawn on reconstructed B-mode images in two ways: FOIs that included as much of the liver as possible including blood vessels and FOIs modified manually to exclude larger blood vessels. Compared to blood vessel-exclusive FOIs, blood vessel-inclusive FOIs provided higher (<0.005) between-exam ICCs (AC, 0.71...
Liver fibrosis is a histological hallmark of most chronic liver diseases, can progress to cirrhos... more Liver fibrosis is a histological hallmark of most chronic liver diseases, can progress to cirrhosis and liver failure, and predisposes to hepatocellular carcinoma. Accurate diagnosis of liver fibrosis is necessary for prognosis, risk-stratification, and treatment decision-making. Liver biopsy, the reference standard for assessing liver fibrosis, is invasive, costly, and impractical for surveillance and treatment response monitoring. Elastography offers a noninvasive, objective, and quantitative alternative to liver biopsy. This article discusses the need for noninvasive assessment of liver fibrosis and reviews the comparative advantages and limitations of US and MR elastography techniques with respect to their basic concepts, acquisition, processing, and diagnostic performance. Variations in clinical contexts of use and common pitfalls associated with each technique are considered. In addition, current challenges and future directions to improve the diagnostic accuracy and clinical utility of elastography techniques are discussed.
2019 IEEE International Ultrasonics Symposium (IUS)
Nonalcoholic fatty liver disease (NAFLD) affects ~25% of the world population. Confounder-correct... more Nonalcoholic fatty liver disease (NAFLD) affects ~25% of the world population. Confounder-corrected chemical-shift-encoded MRI-derived proton density fat fraction (MRI-PDFF) is an established quantitative noninvasive biomarker of hepatic steatosis but has limited availability. There is a clinical need for more practical and accessible methods to noninvasively assess hepatic steatosis. Previous work has shown that two quantitative ultrasound (QUS) biomarkers - attenuation coefficient (AC) and backscatter coefficient (BSC) - are correlated with hepatic steatosis. Examining a broad range of QUS biomarkers, this study aims to develop an improved, multi-parametric QUS-based approach to diagnose NAFLD and quantify hepatic fat, with MRI-PDFF as the reference standard. 102 participants recruited from the UCSD NAFLD Research Center underwent QUS exams on the right liver lobe with an Acuson S3000 ultrasound scanner and the 4C1 and 6C1HD transducers. Seven QUS biomarkers - AC, BSC, three Lizzi-Feleppa parameters (slope, intercept, midband), and two envelope parameters (k and μ) - were derived from ultrasound radiofrequency data. Two multivariable models were developed based on QUS biomarkers: a generalized linear regression model to predict hepatic PDFF using stepwise regression for biomarker selection and a regularized logistic regression model to classify NAFLD (MRI-PDFF>5%, N=78/102) versus no NAFLD (MRI-PDFF≤5%) using LASSO regularization for biomarker selection. Leave-one-out cross-validation was performed. The final regression model selected the midband and k-parameter. The cross-validated predicted PDFF values were correlated with the reference MRI-PDFF values (Spearman ρ = 0.82 and Pearson’s r = 0.76). In comparison, Pearson’s r was 0.59 between AC and MRI-PDFF and 0.58 between BSC and MRI-PDFF. The final classifier model selected the midband, k-parameter and μ-parameter, achieving an area under the receiver operating characteristic curve (AUROC) of 0.88. In comparison, AUROC was 0.83 using AC and 0.84 using BSC. The results suggest that multi-parametric QUS can improve the quantification of hepatic steatosis and diagnosis of NAFLD.
Objectives To compare the diagnostic accuracy of US shear wave elastography (SWE) and magnetic re... more Objectives To compare the diagnostic accuracy of US shear wave elastography (SWE) and magnetic resonance elastography (MRE) for classifying fibrosis stage in patients with nonalcoholic fatty liver disease (NAFLD). Methods Patients from a prospective single-center cohort with clinical liver biopsy for known or suspected NAFLD underwent contemporaneous SWE and MRE. AUCs for classifying biopsy-determined liver fibrosis stages ≥ 1, ≥ 2, ≥ 3, and = 4, and their respective performance parameters at cutoffs providing ≥ 90% sensitivity or specificity were compared between SWE and MRE. Results In total, 100 patients (mean age, 51.8 ± 12.9 years; 46% males; mean BMI 31.6 ± 4.7 kg/m2) with fibrosis stage distribution (stage 0/1/2/3/4) of 43, 36, 5, 10, and 6%, respectively, were included. AUCs (and 95% CIs) for SWE and MRE were 0.65 (0.54–0.76) and 0.81 (0.72–0.89), 0.81 (0.71–0.91) and 0.94 (0.89–1.00), 0.85 (0.74–0.96) and 0.95 (0.89–1.00), and 0.91 (0.79–1.00) and 0.92 (0.83–1.00), for dete...
The Journal of the Acoustical Society of America, 2019
Quantitative ultrasound (QUS) aims to improve diagnostic ultrasound imaging by extracting objecti... more Quantitative ultrasound (QUS) aims to improve diagnostic ultrasound imaging by extracting objective tissue parameters from backscattered signals. Deep learning can facilitate this process because of its ability to extract high-level information from the raw data. We have demonstrated that deep learning approaches applied to backscattered signals can accurately quantify liver fat noninvasively. We will discuss three deep learning approaches herein: (1) a one-dimensional convolutional neural network (1-D-CNN) for uncalibrated radiofrequency (RF) data; (2) a two-dimensional CNN (2-D-CNN) for uncalibrated spectrograms; and (3) a 2-D-CNN for calibrated spectrograms. Approaches (1) and (2) do not require system calibration with a physical phantom, whereas Approach (3) obtains calibrated spectrograms with a physical phantom. Each approach is evaluated using three datasets that contain ultrasound RF data acquired from the right liver lobe and contemporaneous MRI-PDFF (reference standard) in participants with and without nonalcoholic fatty liver disease: 204 scanned by a Siemens S2000 ultrasound scanner (4C1), 70 scanned by a Siemens S3000 scanner (4C1 and/or 6C1HD), and 104 scanned by both the S3000 and a GE Logiq e9 scanner (C1-6). We will discuss the advantages and disadvantages of each approach evaluated using data acquired across multiple ultrasound scanner platforms. [Supported by R01DK106419.]Quantitative ultrasound (QUS) aims to improve diagnostic ultrasound imaging by extracting objective tissue parameters from backscattered signals. Deep learning can facilitate this process because of its ability to extract high-level information from the raw data. We have demonstrated that deep learning approaches applied to backscattered signals can accurately quantify liver fat noninvasively. We will discuss three deep learning approaches herein: (1) a one-dimensional convolutional neural network (1-D-CNN) for uncalibrated radiofrequency (RF) data; (2) a two-dimensional CNN (2-D-CNN) for uncalibrated spectrograms; and (3) a 2-D-CNN for calibrated spectrograms. Approaches (1) and (2) do not require system calibration with a physical phantom, whereas Approach (3) obtains calibrated spectrograms with a physical phantom. Each approach is evaluated using three datasets that contain ultrasound RF data acquired from the right liver lobe and contemporaneous MRI-PDFF (reference standard) in participants with and ...
The Journal of the Acoustical Society of America, 2019
Nonalcoholic fatty liver disease (NAFLD) is a leading cause of chronic liver disease worldwide. C... more Nonalcoholic fatty liver disease (NAFLD) is a leading cause of chronic liver disease worldwide. Current diagnosis relies on invasive biopsy. Quantitative ultrasound (QUS) may provide a noninvasive alternative, using parameters such as attenuation (AC) and backscatter (BSC) coefficients. This study examines how exclusion of larger blood vessels from fields of interest (FOI) affects AC and BSC measurements in NAFLD patient livers. 201 total participants with known or suspected NAFLD underwent liver right lobe QUS. Of these, 84 underwent a second QUS exam on the same day and 172 patients had contemporaneous MRI with computation of proton density fat fraction (MRI-PDFF). AC and BSC were calculated from FOIs drawn on reconstructed B-mode images in two ways: FOIs that included as much of the liver as possible including blood vessels and FOIs modified manually to exclude larger blood vessels. Compared to blood vessel-exclusive FOIs, blood vessel-inclusive FOIs provided higher (<0.005) between-exam ICCs (AC, 0.71...
Liver fibrosis is a histological hallmark of most chronic liver diseases, can progress to cirrhos... more Liver fibrosis is a histological hallmark of most chronic liver diseases, can progress to cirrhosis and liver failure, and predisposes to hepatocellular carcinoma. Accurate diagnosis of liver fibrosis is necessary for prognosis, risk-stratification, and treatment decision-making. Liver biopsy, the reference standard for assessing liver fibrosis, is invasive, costly, and impractical for surveillance and treatment response monitoring. Elastography offers a noninvasive, objective, and quantitative alternative to liver biopsy. This article discusses the need for noninvasive assessment of liver fibrosis and reviews the comparative advantages and limitations of US and MR elastography techniques with respect to their basic concepts, acquisition, processing, and diagnostic performance. Variations in clinical contexts of use and common pitfalls associated with each technique are considered. In addition, current challenges and future directions to improve the diagnostic accuracy and clinical utility of elastography techniques are discussed.
Uploads
Papers by Vivian Montes