Hepatic Steatosis Quantification with Ultrasound

PROJECT SUMMARY
Quantification of liver steatosis has weighty implications in management of Nonalcoholic Fatty Liver Disease (a
condition affecting 75-100 million Americans), diabetes mellitus, and cardiovascular disease. Serum
biomarkers, computed tomography, and existing ultrasound methods have low sensitivity or specificity for
steatosis staging. Proton Density Fat Fraction (PDFF) measured by MRI has high accuracy, but is limited by
accessibility and cost. Here we propose a novel ultrasound technology, Spectrum Normalization Attenuation
Imaging (SNAI), to quantify ultrasound attenuation coefficient for accurate liver steatosis staging. SNAI does
not require a calibration phantom, and is robust to rib shadowing as well as phase aberration and reverberation
clutter from the body wall. Accuracy of SNAI for steatosis staging is demonstrated by a promising correlation
coefficient of 0.91 with MRI-PDFF in 50 patients. In this project, we will prototype, optimize, and evaluate SNAI
on a low-cost, pocket-sized, wireless ultrasound probe, which can be conveniently used for screening and
follow-up at the point-of-care setting such as the office of a family doctor or hepatologist.
Specific Aim 1: Technical Development. We will use phantom and patient studies to advance and optimize
SNAI on the the wireless ultrasound probe. Acquisition parameters and post processing algorithms will be
optimized for both fundamental and harmonic SNAI imaging. We will suppress reverberation clutter and correct
for sound speed mismatch in liver for more accurate steatosis quantification.
Specific Aim 2: Patient study. We will use the SNAI prototypes optimized in Aim 1 to study 250 steatosis
patients with clinically indicated MRI-PDFF to investigate the efficacy of SNAI for steatosis quantification.
Correlation analysis will be performed to assess the association of SNAI with MRI-PDFF. Steatosis will also be
categorized as S0, S1, and S2/S3 according to PDFF. Receiver operating characteristic analyses will be
performed to evaluate performance of SNAI for detecting ≥S1 and ≥S2. The agreement between SNAI and
PDFF classification will be evaluated using the Kappa statistic. Fibroscan CAP will be used for benchmarking.
Specific Aim 3: Reproducibility study. Two sonographers and two hepatology residents will repeatedly scan
a subset of subjects (50 patients) studied in Aim 2. Intraclass correlation coefficients will be used to evaluate
the inter-operator agreement for SNAI measurements. The within patient variance component from the model
will provide an estimate of the inter-operator variance, which represents a lower bound for the minimum
detectable difference for longitudinal follow-ups.
Successful completion of this project will result in a safe, cost-effective, and easily accessible ultrasound
solution for accurate quantification of liver steatosis for diagnosis and frequent follow-up of this very large
patient population at point-of-care settings such as the office of a hepatologist or family doctor.

Grant Number: 5R01DK127978-04
NIH Institute/Center: NIH
Principal Investigator: Shigao Chen