Liver shear wave elastography using a mechanical index exceeding regulatory limits is safe and effective

Abstract

Background

Metabolic dysfunction-associated steatotic liver disease affects 1 in 3 people worldwide. Ultrasound shear wave elastography (SWE) in obese patients, the target population for testing, is hampered by beam attenuation, leading to unreliable liver fibrosis quantification.

Purpose

We assess the safety and efficacy of increased push mechanical index (IPMI) above U.S. Food and Drug Administration limits to improve SWE.

Materials and methods

This single-center prospective trial (July 2023-April 2024) (NCT05792423) enrolled healthy adults stratified by body mass index (BMI). Participants underwent conventional push pulse (mechanical index [MI] 1.4) and IPMI (MI 2.5) SWE (GE Healthcare LOGIQ E10) performed by 1 of 3 sonographers and serial liver function testing (LFT) before and up to 7 days after imaging. Liver injury was defined as increased serum alanine transaminase (ALT), aspartate aminotransferase (AST), or alkaline phosphatase (ALP) (non-inferiority margins: AST 7.5 U/L, ALT 12 U/L, ALP 17.5 U/L). Secondary endpoints included SWE variability and measurement number.

Results

Twenty-two analyzable participants (mean age 39.6 ± 16.3 years; 15 women) had normal BMI (6), overweight (6), class 1 obesity (7), and class 2 obesity (3). Conventional shear wave speed was 1.34 ± 0.21 m/s, and IPMI yielded 1.36 ± 0.20 m/s (velocities ≤ 1.34 m/s indicate minimal or no fibrosis). The mean [95% CI] LFT change from baseline to day 1 was 1) AST: −0.86 [−2.34, 0.61], P = .24, 2) ALT: 0.32 [−1.04, 1.68], P = .63, 3) ALP: 1.73 [−1.02, 4.47], P = .21. The upper 95% CI for all biomarkers met non-inferiority criteria. Mean IPMI interquartile range (IQR) to median ratio decreased 0.019 (29.2% relative reduction) (P = .01) with 0.68 [IQR: 0.0.75] (P = .058), fewer average attempts.

Conclusion

IPMI SWE in healthy volunteers did not cause injury and reduced measurement variability. IPMI SWE should be developed to improve examination quality and reliability in obese patients.

Abbreviations

ALP = Alkaline phosphatase; ALT = Alanine transaminase; AST = Aspartate aminotransferase; BMI = Body Mass Index; FDA = Food and Drug Administration; LFT = Liver Function Test; IPMI = Increased Push Mechanical Index; IQRM = Interquartile range to median ratio; I_SPTA = spatial peak temporal average intensity; MASLD = Metabolic dysfunction-associated steatotic liver disease; MI = Mechanical Index; NAFLD = Nonalcoholic fatty liver disease; SWE = Shear Wave Elastography; TI = Thermal Index

Summary

Acoustic output above current U.S. Food and Drug Administration limits improves ultrasound shear wave elastography measurement quality without causing liver injury.

Key Results

• Liver ultrasound shear wave elastography (SWE) with an increased push mechanical index (IPMI) above current U.S. Food and Drug Administration limits did not cause liver injury measured by blood-based liver function testing.

• IPMI improved image quality, with a 29.2% [95% confidence interval: 7.7%-50.9%, P = .01] reduction in measurement variability, the most important clinical marker of liver SWE quality.

• IPMI elastography required 0.68 (P = .058) fewer attempts per participant than conventional elastography.

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as nonalcoholic fatty liver disease (NAFLD),¹ is the most common chronic liver disorder in the United States, with a prevalence of 30%.^2,3 MASLD includes simple steatosis, characterized by excess liver fat, and metabolic dysfunction-associated steatohepatitis (MASH), where the surplus fat is accompanied by inflammation and/or fibrosis. Over time, tissue injury and scarring may cause cirrhosis,^4–6 leading to portal hypertension, synthetic liver failure, and hepatocellular carcinoma.^7,8 MASH has substantial morbidity and mortality and is the second leading indication for liver transplantation in the United States.⁹

Early and accurate liver fibrosis diagnosis and staging are critical to optimizing MASLD patient management. The gold standard, tissue biopsy, is costly, invasive, semi-quantitative, prone to sampling error, and associated with procedural complications. Ultrasound (US) shear wave elastography (SWE) is a safe, widely available, and relatively low-cost technique commonly used for liver fibrosis diagnosis and staging in at-risk populations.¹⁰ Unfortunately, liver SWE fails at much higher rates in patients with MASLD than in patients with viral hepatitis due to higher rates of obesity.^11,12 Increased subcutaneous fat thickness in obesity causes US beam distortion (aberration) and weakening (attenuation), significantly reducing measurement quality and the ability to quantify fibrosis.^13–18 As a result, SWE failure rates and diagnostic accuracy are worse in obese patients, the population most likely to have MASLD.¹⁹

Increasing US beam energy or pressure for push pulses and/or tracking pulses has the potential to mitigate these limitations. However, U.S. Food and Drug Administration (FDA) guidance limits “Track 3” devices to derated spatial peak temporal average intensity (I_SPTA.3)  ≤720 mW/cm², mechanical index (MI)  ≤1.9, thermal index (TI)  ≤6.0, and probe patient-contact surface temperature ≤43 °C.²⁰ The acoustic techniques currently used in SWE approach these limits.^10,21,22 Therefore, increasing acoustic output above current FDA limits may increase the risks of thermal and non-thermal bioeffects, including cavitation. The FDA limitation of MI (a measure of the ratio of the rarefactional pressure to the frequency that quantifies the likelihood of generating cavitation) to 1.9 aims to minimize the risk of mechanical (ie, non-thermal) bioeffects; however, this limitation is conservative for in vivo conditions.²³ Initial human studies have evaluated elevated MI values for push pulses (MI 2.2) and tracking pulses (MI 2.4) without signs of adverse effects, though safety was not formally assessed.^24,25

Our study assessed the safety and efficacy of increased push mechanical index (IPMI) to improve SWE quality in normal-weight and obese patients. The acoustic pressures, and therefore MI values, of the SWE push pulses were raised above FDA Track 3 levels. The I_SPTA.3, TI, and patient-contact surface temperature, as well as the MI values of the SWE tracking pulses, were kept below FDA limits. The primary aim of this study was to evaluate for liver injury after IPMI SWE through serial liver function testing (LFTs). The secondary aim quantified SWE measurement variability reduction. These endpoints are clinically relevant, as safe and more reliable SWE has the potential to detect more patients at risk of poor long-term outcomes and may reduce the need for costly additional testing in MASLD.

Materials and methods

Study cohort

The local institutional review board approved this (clinical trial number NCT05792423) prospective HIPAA-compliant pre/post single-arm non-inferiority study (July 2023-April 2024). An investigational device exemption was not required, as the IPMI SWE US device was determined to be a non-significant risk device. The Helsinki declaration of 1964 was followed, and written informed consent was obtained from all participants.

Healthy participants were recruited via an electronic flyer on a publicly available clinical trial recruitment website. Inclusion criteria were age 18-65 years, body mass index (BMI) between 18.5 and 39.9 kg/m², stable medications for 30 days, and ability to undergo study procedures. Participants with a BMI ≥ 40 were excluded. Participants were excluded for excessive alcohol consumption (>7 units/week for women or >14 units/week for men); drug-induced liver injury; liver transplantation; viral hepatitis; implanted electronic device; planned medication changes; known or suspected pregnancy; recent abnormal LFTs, including alanine transaminase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP); recent US contrast administration; or any of the following within 30 days: pharmaceutical trial participant, systemic chemotherapy, or acute illness. We attempted to stratify recruitment evenly by BMI group: 18.5-24.9 kg/m², 25.0-29.9 kg/m², 30.0-34.9 kg/m², and 35.0-39.9 kg/m². A prior technical publication analyzed the same cohort (22 participants) and focused on individual measurement signal improvement, while this analysis reports safety and examination-level quality.²⁶

Clinical data

Participants underwent focused clinical history, serum beta-human chorionic gonadotropin (if applicable), and LFTs within 48 hours before undergoing US imaging. Serial LFTs were performed 1, 2, and 7 days after US imaging at a single clinical laboratory. Study staff confirmed consistent alcohol intake and medication use and assessed for adverse events during the study and via phone call within 1 week of the final blood test. Adverse events were defined as any untoward or unfavorable medical occurrence, including any abnormal physical examination or unexpected abnormal laboratory finding, symptom, or disease, temporally associated with a subject’s participation in the research. Specifically, participants were assessed for pain and any other patient-reported symptoms. Serious adverse events were defined as death, life-threatening, resulting in hospitalization or prolongation of existing hospitalization, resulting in a persistent or significant disability/incapacitation, or jeopardizing the participant’s health requiring medical or surgical intervention to prevent one of the other outcomes listed above.

Ultrasound shear wave elastography

After fasting for at least 4 hours, conventional energy and IPMI SWE of the liver were performed during 1 imaging visit using a GE HealthCare LOGIQ E10 modified to permit IPMI SWE with a C1-6 probe (GE HealthCare, Chicago, IL, United States). Other US pulse settings, such as tracking pulses, were within FDA limits. Image acquisition was performed by 3 certified sonographers with 7-18 years of experience. For conventional SWE with a push pulse MI of 1.4, we performed up to 20 attempts to obtain 10 diagnostic shear wave speed measurements (in meters per second) in the right lobe of the liver via an intercostal window during suspended respiration. The same general region was assessed for each measurement, though images were not acquired from the exact same location due to slight patient and operator motion during the study. Measurement adequacy was determined by the commercially available on-machine SWE signal analysis algorithm that displays non-diagnostic measurements as red regions of interest. SWE regions of interest were placed approximately 2 cm deep to the liver capsule and preferentially located toward the image midline while avoiding vessels or focal lesions. The median shear wave speed and interquartile range to median (IQRM) ratio were calculated.

IPMI SWE measurements followed the same protocol with a push pulse MI of 2.5, a value selected based on vendor equipment capability. For the higher MI measurements, the imaging frame rate was decreased such that the soft tissue TI (quantifying the total expected rise in temperature) was the same (0.98) for both acquisition types; that is, the maximum expected temperature rise was expected to be on the order of 1 °C. For both the conventional and IPMI techniques, the push frequency was 2.3 MHz, which corresponds to a wavelength of roughly 670 µm. In the linear regime, there is no change in wavelength as the amplitude is varied. While higher amplitudes might yield some nonlinear behavior (eg, harmonic generation) or exacerbate dispersive effects, these would be second-order and unlikely to affect shear wave speed estimates. Participants were evaluated for signs of pain and discomfort before, during, and immediately after the study visit.

To ensure consistent hardware functionality, a custom electronic probe assessment tool assessed the US transducer channel sensitivity profile within 14 days before each participant’s imaging visit and immediately after each imaging visit. Ten malfunctioning elements were determined a priori as the threshold for hardware replacement.

All of the collected data were entered into a secure REDCap database by one of the study staff and confirmed by a board-certified radiologist. GE Healthcare provided financial support and the US device. Case report forms, data management systems, and data analyses were only accessible or performed by study authors not affiliated with GE Healthcare.

Statistical analysis

The study was powered to show that day 7 LFTs (AST, ALT, and ALP) did not significantly increase following IPMI SWE. LFTs are routinely used in clinical trials to assess the severity of liver injury.²⁷ This corresponds to a non-inferiority study design where success is defined by the upper 95% confidence limit of a detected LFT increase being less than a pre-defined clinically relevant safety margin. The non-inferiority margin was defined as the LFT standard deviation (σ), which was less than one-quarter of the upper limit of normal. This was extrapolated from drug trials where LFT elevation of 3x-8x the upper limit of normal serves as the criterion for drug discontinuation due to hepatotoxicity.²⁸ Our non-inferiority margin was selected to be much more stringent, given that US is an otherwise noninvasive diagnostic test. LFT σ was estimated by assuming that the laboratory normal range defines the population mean value ± 2σ. Here, σ was estimated as 7.5 U/L for AST (normal range: 10-40 U/L), 12 U/L for ALT (normal range: 7-55 U/L), and 17.5 U/L for ALP (normal range: 45-115 U/L). Paired non-inferiority testing assuming no mean difference between pre- and post-imaging LFTs, a non-inferiority margin equal to the standard deviation, α = .05, and target power = .9 required 18 participants. Assuming a 25% dropout rate, we aimed to recruit 24 participants.

Secondary endpoints assessed LFT change at day 1 and day 2, reduction in IQRM by paired t-test, and number of imaging attempts by Wilcoxon signed rank test. Detailed technical analyses of US signal improvement are reported separately.²⁶ Pairwise cross-correlation between LFTs performed on different days was quantified by Pearson Correlation (Figure S1). We constructed separate linear mixed effects models to characterize the association between each LFT measure and time (day 0, 1, 2, and 7). Subject-specific random intercepts were included to account for the dependent data. Linear combinations of coefficients were computed along with their 95% confidence intervals to summarize temporal LFT comparisons (eg, day 1 vs day 0). Significance was assessed at P < .05. Analyses were performed in R 4.5.1. Statistical analysis code can be made available upon reasonable request.

Results

Cohort description

Twenty-four participants were enrolled (Figure 1); 2 withdrew prior to imaging and were excluded from analysis. Twenty-two participants (39.6 ± 16.3 years; 15 women) had normal BMI (6), overweight (6), class 1 obesity (7), and class 2 obesity (3). Additional baseline demographics are shown in Table 1. Race and ethnicity were self-reported by the participants.

Figure 1.

Participant recruitment. BMI = body mass index.

Table 1.

Baseline demographics.

Body mass index category (kg/m2)	18.5-24.9 (n = 6)	25.0-29.9 (n = 6)	30.0-34.9 (n = 7)	35.0-39.9 (n = 3)	Total (n = 22)
Age (years)	34.3 (12.9)	48 (16.4)	37.3 (16.9)	39 (22.6)	39.6 (16.3)
Female	5 (83.3%)	2 (33.3%)	6 (85.7%)	2 (66.7%)	15 (68.2%)
Racea
Asian	2 (33.3%)	1 (16.7%)	0 (0%)	0 (0%)	3 (13.6%)
Black	0 (0%)	0 (0%)	2 (28.6%)	0 (0%)	2 (9.1%)
White	3 (50%)	4 (66.7%)	3 (42.9%)	3 (100%)	13 (59%)
Other	1 (16.7%)	1 (16.7%)	2 (28.6%)	0 (0%)	4 (18.2%)
Hispanic or Latino	0 (0%)	0 (0%)	2 (28.6%)	0 (0%)	2 (9%)
Albumin (g/dL)	4.7 (0.2)	4.8 (0.3)	4.5 (0.2)	4.1 (0.3)	4.6 (0.3)
Total bilirubin (mg/dL)	0.55 (0.24)	0.55 (0.39)	0.46 (0.45)	0.87 (1.07)	0.56 (0.49)
Direct bilirubin (mg/dL)	0.15 (0.05)	0.12 (0.04)	0.14 (0.08)	0.20 (0.17)	0.15 (0.08)
Alkaline phosphatase (U/L)	68.2 (27.7)	65.0 (14.0)	71.6 (20.7)	74.3 (26.1)	69.2 (20.7)
Aspartate aminotransferase (U/L)	21.0 (4.1)	24.5 (7.0)	19.7 (6.9)	18.7 (4.2)	21.2 (6.0)
Alanine transaminase (U/L)	18.0 (9.7)	27.2 (19.3)	17.0 (9.8)	14.7 (4.0)	19.7 (12.8)
Total protein (g/dL)	7.5 (0.5)	7.5 (0.3)	7.5 (0.6)	7.1 (0.1)	7.4 (0.4)
Globulin (g/dL)	2.8 (0.5)	2.7 (0.2)	3.0 (0.4)	2.9 (0.3)	2.9 (0.4)

Variables are summarized as mean (standard deviation) or counts (percent).

^aParticipants self-reported race based on the following categories: American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or other Pacific Islander, White, or Other.

IPMI SWE safety

None of the LFT upper 95% confidence limits crossed the corresponding non-inferiority margins at any time point, indicating no evidence of liver injury from IPMI SWE (Figure 2). Furthermore, all participant-level LFT changes at 1, 2, and 7 days were less than the non-inferiority margins. The mean [95% CI] increase in liver function tests from baseline were: AST, −0.9 U/L [−2.3, 0.6], P = .24; ALT, 0.3 U/L [−1.0, 1.7], P = .63; and ALP, 1.7 U/L [−1.0, 4.5], P = .21 at day 1; AST, −1.0 [−2.4, 0.3], P = .12; ALT, −0.6 [−2.5, 1.3], P = .5; and ALP, 1.3 [−0.6, 3.2], P = .17 at day 2; and AST, −0.5 [−2.5, 1.5], P = .6; ALT, −0.3 [−2.6, 2.1], P = .81; and ALP, 2.2 [−0.6, 5.0], P = .12 at day 7 (Figure 3). Adjusted 95% confidence intervals using a linear mixed model for days 1, 2, and 7 were [−2.6, 0.9], [−2.7, 0.8], and [−2.2, 1.2] for AST, [−1.7, 2.3], [−2.7, 1.3], and [−2.3, 1.7] for ALT, and [−0.9, 4.4], [−1.5, 3.9], and [−0.4, 4.9] for ALP.

Figure 2.

Mean increase in liver function tests after imaging. A thin vertical line shows the mean difference for each LFT between post-imaging and pre-imaging. The 95% CI (thick horizontal bar) and range of values (thin horizontal line) are shown for each time point. None of the 95% CI overlaps the non-inferiority margins (vertical dotted lines), indicating that increased push mechanical index shear wave elastography did not cause liver injury. The non-inferiority margins were selected as the standard deviation for each biomarker. LFT = liver function testing, CI = confidence interval, ALP = alkaline phosphatase, ALT = alanine transaminase, AST = aspartate aminotransferase.

Figure 3.

Change in liver function tests by body habitus and day of testing. Each data point represents the change in a liver function test from pre-imaging baseline to (A) 1-day post-imaging, (B) 2-day post-imaging, and (C) 7-day post-imaging. Horizontal lines reflect the range of values. Positive values correspond to increasing injury. BMI groups are normal (1), overweight (2), class 1 obesity (3), and class 2 obesity (4). BMI = body mass index, ALP = alkaline phosphatase, ALT = alanine transaminase, AST = aspartate aminotransferase.

No serious adverse events attributable to IPMI SWE occurred during the course of the study. Two participants reported adverse events. Both participants reported mild transient abdominal tingling or pulsing sensations starting 1-2 days after study visit 1, which self-resolved prior to the study follow-up phone visit. Adjudication by study physicians determined that neither incident was likely to be related to study participation given the temporal association. Neither event met institutional review board reporting guidelines.

Efficacy of IPMI vs. Conventional SWE

The average shear wave speed across participants was 1.34 ± 0.21 m/s (range 1.03 m/s—1.97 m/s) for conventional SWE and 1.36 ± 0.20 m/s (range 1.05 m/s—1.94 m/s) for IPMI SWE with a mean difference of 0.02 m/s [95% CI: −0.04 m/s, 0.07 m/s]. Mean IQRM was 0.066 (range 0-0.122) for conventional SWE and 0.047 (range 0.012-0.090) for IPMI SWE (Figure 4). The 29.2% [95% CI: 7.7%, 50.9%] relative reduction in IQRM was statistically significant (P = .01). Mean IQRM was lower for IPMI SWE than conventional SWE across all BMI groups: normal BMI, 0.051 vs 0.068; overweight, 0.047 vs 0.065; class 1 obesity, 0.046 vs 0.062; and class 2 obesity, 0.039 vs 0.074. Of note, 1 conventional SWE examination had only 1/20 diagnostic images, yielding an IQRM of 0 (Figure 5). The IQRM was paradoxically higher for the IPMI SWE acquisition because more measurements were diagnostic.

Figure 4.

Measurement variability between conventional and IPMI elastography. Measurement variability was 29.2% lower for IPMI examinations compared to conventional push pulse shear wave elastography (P = .01). Lower measurement variability indicates improved examination quality. IPMI = increased push mechanical index, IQRM = interquartile range to median ratio.

Figure 5.

Representative conventional and increased push mechanical index elastography images from varying body habitus. (A) 32-year-old female, BMI 21.3 kg/m², with conventional elastography images and (B) increased push mechanical index elastography images showing homogenous fill. All conventional and increased push mechanical index images, and both examinations were diagnostic. (C) 59-year-old male, BMI 29.0 kg/m², with conventional elastography images and (D) increased push mechanical index elastography images showing homogenous fill. All conventional and increased push mechanical index images, and both examinations were diagnostic. (E) 39-year-old female, BMI 30.5 kg/m², with conventional elastography images showing small areas of signal loss. 2/12 images were non-diagnostic, though the examination was diagnostic. (F) In the same participant, a representative increased push mechanical index image shows a more homogenous elastogram, indicating improved quality. All 10 measurements and the examination were diagnostic. (G) 65-year-old male, BMI 35.3 kg/m², with a conventional elastography image showing signal degradation due to signal attenuation. A reliable shear wave speed estimate could not be obtained. Only 1/20 acquisitions resulted in a satisfactory shear wave speed estimate. The examination was non-diagnostic. (H) In the same participant, a representative increased push mechanical index image shows improved image quality with a satisfactory measurement. The 10/15 measurements and the aggregate examination were diagnostic. BMI = body mass index.

The mean number of conventional SWE attempts, 11.41 (range 10-20), was higher than IPMI SWE, 10.73 (range 10-20), though the difference of 0.68 [IQR: 0.0.75] was marginally statistically significant, P = .058 (Figure 6). Among all cases, IPMI required fewer measurements in 6/22 (27.2%) cases and more measurements in 1/22 (4.5%) cases. Among the 10 participants with BMI ≥ 30.0 kg/m², IPMI required fewer measurements in 4 (40.0%) cases, and none required more measurements.

Figure 6.

Probability of examination completion versus measurement number. The cumulative percent of successful examinations is plotted versus the number of measurement attempts. The higher values associated with IPMI elastography indicate that more examinations are completed than conventional elastography for the same number of measurement attempts. The difference approaches statistical significance (P = .058). IPMI = increased push mechanical index.

Imaging hardware remained operational for the duration of the study. Specifically, 3 transducer elements were non-functional at the completion of the study, less than the a priori threshold of 10.

Discussion

The primary aim of our study was to evaluate for adverse bioeffects from IPMI SWE. We found no significant change in serologic biomarkers of liver damage 1-7 days after IPMI SWE. In fact, all participant-level measures of liver health fell within the pre-defined safety margin. These results indicate that IPMI SWE did not cause liver damage. Importantly, safe utilization was observed in both obese participants, who are most likely to benefit from this technology, and normal body habitus participants, who may be at relatively increased risk of injury without thick subcutaneous fatty tissue to attenuate the increased US energy.

The secondary aim assessed for improvement in image quality using IPMI SWE. IQRM, the most widely used clinical measure of SWE examination quality, decreased by 0.0193, corresponding to a relative reduction of 29.2%. For reference, an IQRM of 0.15 is the typical clinical threshold to distinguish diagnostic and non-diagnostic examinations. Therefore, the observed IQRM improvement is sufficiently large to be clinically relevant. Furthermore, the mean number of imaging attempts required for IPMI SWE was 0.68 less than for conventional SWE (P = .058). While not meeting the threshold for statistical significance in this study, a reduction in measurement requirements would enable greater workflow efficiency when multiplied across the large numbers of routinely obtained examinations. Additionally, 40% of obese participants saw a reduction in required measurement attempts with IPMI. These results suggest that IPMI SWE leads to reduced measurement variability, improved examination quality, and may increase image acquisition efficiency.

Our findings are consistent with a prior study by Nightingale et al.,²³ which highlighted that acoustic radiation force impulse-based imaging methods suffer from depth penetration limitations and high rates of technical failure in obese patients. Their findings suggested that the number of technically successful measurements increased, and the noise decreased, when using increased acoustic pressures. This aligns well with our study, where IPMI SWE demonstrated improved measurement success and reduced variability.

Nightingale et al. comprehensively discussed the safety margins associated with current FDA guidelines for MI. They argued that the guidelines are based on conservative assumptions that may not accurately reflect valid safety thresholds, particularly in tissues not known to contain gas bodies. Our study’s findings support this argument, as we observed no adverse bioeffects from using IPMI SWE, even at increased MI levels. The safety and efficacy data suggest a favorable risk-benefit balance for IPMI SWE for liver imaging.

Regarding study limitations: (1) While there are no obvious unrepresented factors that would place patients at increased risk of liver damage, this small single-center proof-of-concept study may not generalize to all the diverse patient characteristics represented in the larger population. Specifically, patients with BMI > 40 were excluded, as they were thought to be at less risk of liver injury compared to thinner patients because of the attenuating effects of the thick body wall. Moreover, as this group potentially could be at higher risk of having both conventional and IPMI SWE fail, we targeted the population that would be most likely to benefit from the intervention in this proof-of-concept trial. Given this, future trials that include participants with BMI > 40 are required to determine the safety and efficacy of IPMI SWE in this group. (2) LFT measurement may lack the sensitivity to detect minimal liver injury. However, our LFT-based hepatic injury detection criterion was far more stringent than in typical interventional trials. The levels of injury that this criterion would miss are unlikely to have clinical consequences. (3) While our study assessed measurement precision, we could not assess measurement accuracy given the absence of liver biopsy histology in this volunteer research subject group. Future studies to enable clinical translation should focus on demonstration of safety in patients with liver fibrosis, assessment of safety across more diverse patient populations and imaging centers, characterization of imaging accuracy with a ground truth reference standard, and assessment for non-hepatoxic adverse events, such as in-situ thrombosis in an animal model. If IPMI SWE is ultimately FDA-cleared, the technique should undergo further evaluation for potential improvements in diagnostic accuracy.

In conclusion, our study demonstrates IPMI SWE results in improved image quality and reduced measurement variability, especially in obese patients, without compromising patient safety. Thus, this IPMI SWE should be developed to improve noninvasive testing and monitoring of MASLD patients for better long-term clinical outcomes. More generally, all diagnostic US imaging modes are adversely affected by obesity, and our study shows a potential pathway to mitigate this common factor.

Author contributions

Theodore T. Pierce (Conceptualization, Data curation, Formal analysis, Methodology, Writing—review & editing), Kim Naja (Data curation, Writing—original draft, Writing—review & editing), Scott J. Schoen Jr (Conceptualization, Data curation, Formal analysis, Methodology, Writing—review & editing), Rimon Tadross (Conceptualization, Methodology, Software, Writing—review & editing), Michael H. Wang (Conceptualization, Methodology, Resources, Software, Writing—review & editing), Arinc Ozturk (Conceptualization, Data curation, Methodology, Writing—review & editing) , Kathleen R. Pope (Data curation, Writing—review & editing), David T. Hunt (Data curation, Methodology, Writing—review & editing) Lauren A. Ling (Data curation, Writing—review & editing), Sunethra K. Dayavansha (Data curation, Writing review & editing), Mary Peters (Data curation, Writing—review & editing), Ann B. Iafrate (Data curation, Writing—review & editing), Nathaniel Mercaldo (Formal analysis, Writing—review & editing), Michael J. Washburn (Conceptualization, Funding acquisition, Methodology, Resources, Writing—review & editing), Viksit Kumar (Conceptualization, Methodology, Writing—review & editing), Kurt Sandstrom (Conceptualization, Methodology, Software, Writing—review & editing), TaeYun Kim (Conceptualization, Methodology, Software, Writing—review & editing), and Anthony E. Samir (Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing—review & editing)

Supplementary material

Supplementary material is available at Radiology Advances online.

Funding

This project was supported by GE HealthCare.

Conflicts of interest

GE HealthCare markets ultrasound devices and shear wave elastography technology. GE HealthCare employees did not analyze or control the data used in this research. Salary support for non-industry co-authors was administered through Massachusetts General Hospital. A patent related to the work has been submitted by GE HealthCare and Massachusetts General Hospital. Individual author conflicts are reported separately within supplemental materials.

Data availability

Data generated or analyzed during the study are available from the corresponding author upon request. Data sharing is subject to local institutional approval and applicable regulatory requirements.