S2674 Use of the LiverRisk Score for Prediction of Moderate to Advanced Liver Fibrosis in United States Adults

Introduction: Resmetirom is the first approved treatment in the United States (US) for noncirrhotic metabolic dysfunction-associated steatohepatitis (MASH; formerly known as nonalcoholic steatohepatitis [NASH]), conditionally approved for use in adults with moderate to advanced liver fibrosis consistent with stages F2 to F3. Patient selection for treatment with resmetirom may be based on evidence of fibrosis from noninvasive tests (NITs). Liver stiffness measurement (LSM) of 8-20 kPa by vibration-controlled transient elastography (VCTE) (with exclusion of cirrhosis) is considered a potential signal of fibrosis stages F2-F3. The LiverRisk Score (LRS) is a NIT based on blood-based biomarkers and demographics, developed for prediction in the general population of liver fibrosis and future liver-related outcomes. This study evaluated in US adults the performance of the LRS for prediction of LSM ≥8 kPa and ≥20 kPa compared to the Fibrosis-4 (FIB-4) Index. Methods: An observational analysis was conducted of the National Health and Nutrition Examination Survey (NHANES) 2017-2020 cycle, including participants who had complete information to calculate the LRS (i.e., age, sex, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, fasting glucose [imputed from non-fasting when missing], total cholesterol, and platelet count). Performance of the LRS vs FIB-4 for prediction of LSM ≥8 kPa and ≥20 kPa was evaluated as area under the curve (AUC), estimated by survey-weighted logistic regression of LSM ≥8 kPa/≥20 kPa vs the LRS or FIB-4. Results: The unweighted study population included N = 7,005 participants. The LRS demonstrated significantly (P < 0.05) superior discrimination of LSM ≥8 kPa vs FIB-4 in all adults, with AUC (95% confidence interval [CI]) of 0.73 (0.71-0.75) for LRS and 0.63 (0.61-0.65) for FIB-4. Discrimination remained numerically superior for LSM ≥20 kPa, with AUC (95% CI) of 0.79 (0.74-0.84) for LRS and 0.73 (0.67-0.79) for FIB-4. Improved accuracy was driven by lower false positive rates (at fixed sensitivity) for LSM ≥8 kPa, and higher true positive rates (at fixed specificity) for LSM ≥20 kPa. Conclusion: In US adults, the LRS demonstrated superior accuracy vs FIB-4 for prediction of LSM ≥8 kPa and ≥20 kPa, which could facilitate screening for noncirrhotic MASH with fibrosis stages F2-F3.