Diagnostic evaluation reveals high diagnostic accuracy for ultrasound attenuation algorithms in uncomplicated fatty liver, indicating potential for standardized disease monitoring.
Key Points
To determine the relationship between ultrasound backscatter intensity attenuation and diffuse liver disease to assess its utility for quantitative objective characterization of fatty liver.
Enrolled 24 healthy volunteers and 28 patients with elevated liver enzymes who underwent liver needle biopsy.
Applied an automated far-field slope (FFS) algorithm to noncompensated B-scan images captured on a phased-array ultrasound system to calculate backscatter amplitude loss across beam depth.
Experienced ultrasound radiologists read and graded repeat scans blindly, and histology results were reviewed without knowledge of imaging findings.
Across all patient groups, FFS analysis yielded 67% sensitivity, 77% specificity, 77% PPV, 67% NPV, and 71% accuracy, compared to radiologist mean scores of 82% sensitivity, 66% specificity, and 72% accuracy.
In patients with uncomplicated fatty infiltration, FFS achieved 100% sensitivity, 80% specificity, 89% PPV, 100% NPV, and 92% accuracy, outperforming the best radiologist score (100% sensitivity, 60% specificity, 85% accuracy).
False-negative FFS readings occurred in five patients with fatty liver who had coexistent moderate to severe hepatic inflammation.