Related Experiment Video
Updated: Apr 10, 2026
Author Spotlight: A Non-Invasive Tool to Assess and Differentiate Fat Patterns in Liver Using 3D Dixon MRI
Published on: October 20, 2023
A Two-Step Approach Combining Steatosis Biomarkers and Multiparametric Quantitative US Enhances Noninvasive Diagnosis
Jie Zeng1, Kaimin Cai2, Jian Zheng3
1Department of Medical Ultrasound, Guangdong Provincial Key Laboratory of Diabetology & Guangzhou Municipal Key Laboratory of Mechanistic and Translational Obesity Research, Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China.
Objective:
To develop optimized models integrating multiple quantitative US parameters and steatosis biomarkers for discriminating steatosis in patients with suspected metabolic dysfunction-associated steatotic liver disease (MASLD).
Methods:
This prospective cross-sectional multicenter study consecutively enrolled participants with suspected MASLD from five tertiary hospitals in China (July 2023-April 2024). All participants underwent multiparametric US (attenuation coefficient [AC], liver texture index [LTI], and hepatorenal index [HRI]) and MRI proton density fat fraction (PDFF). Steatosis discrimination models were constructed using US parameters alone or combined with clinical indicators. A two-step approach involving steatosis biomarkers followed by a US model was proposed. Area under the receiver operating characteristic curve (AUC) analysis was used to evaluate the diagnostic performance.
Results:
Among the 234 participants (mean age, 38.12 ± 9.56 y; 110 male), 173 (73.93%) had steatosis. AC (odds ratio [OR], 8.14 [95% CI: 2.51, 26.39]; p < 0.001), LTI (OR, 4.96 [95% CI: 1.92, 12.83]; p < 0.001), and fasting blood glucose (FBG) (OR, 2.81 [95% CI: 1.10, 7.21]; p = 0.03) were independently associated with steatosis. The AUCs for discriminating steatosis in the model (AC with LTI) and model (AC with LTI and FBG) were 0.94 and 0.95, respectively (p = 0.15). The two-step approach, the initial visceral adiposity index (VAI) or the triglyceride × glucose (TyG) index followed by the model (AC with LTI), improved the diagnostic accuracy from 0.85 (95% CI: 0.80, 0.90) to 0.88 (95% CI: 0.84, 0.92) (p < 0.001).
Conclusions:
The two-step approach, initial steatosis biomarkers followed by a multiparametric US model, might be a practical solution for discriminating steatosis in clinical practice.
