A retrospective single-center investigation has developed and internally validated a specialized clinical formula to identify hepatic steatosis among adult patients diagnosed with primary biliary cholangitis.
The retrospective study cohort was assembled from adult patients with biopsy-confirmed primary biliary cholangitis evaluated between 2016 and 2025. In accordance with histological criteria, the investigators defined hepatic steatosis as the documented involvement of at least 5% of hepatocytes upon tissue examination. Among the 681 total participants enrolled in the cohort, histological steatosis was confirmed in 98 individuals, establishing an overall baseline prevalence of 14.4% within the study sample.
The investigators then deployed a genetic-algorithm formula-search procedure to systematically screen and evaluate potential combinations of routine predictor variables. This search process was evaluated through repeated nested five-fold cross-validation, benchmarked directly against least absolute shrinkage and selection operator regression, an explicitly prespecified clinical model, and six conventional steatosis indices.
The search procedure yielded a finalized, locked two-variable mathematical formula termed the primary biliary cholangitis steatosis index. The resulting equation is defined as five times serum triglycerides minus two times diabetes status, where recorded diabetes is coded as one and the absence of recorded diabetes is coded as zero.
Rigorous internal validation revealed that the index procedure achieved an out-of-fold area under the receiver operating characteristic curve of 0.703, with a 95% confidence interval ranging from 0.646 to 0.757.
When researchers assessed an exploratory out-of-fold probability cutoff of 0.131, the two-variable index demonstrated a sensitivity of 72.4% and a specificity of 59.3%. At this operating threshold, the positive predictive value was calculated at 23.1%, while the negative predictive value reached 92.8%. Statistical comparisons confirmed that the new index outperformed conventional tools, specifically HSI and TyG-BMI, following adjustment for multiplicity.
The authors concluded that the transparent, readily reproducible two-variable formula offers moderate internal discrimination and could help clinicians prioritize patients for further steatosis evaluation when imaging modalities are not readily available, though independent external validation remains necessary prior to clinical implementation.
Subject of Research: Evaluation of a clinical steatosis index in primary biliary cholangitis
Article Title: Development and internal validation of a steatosis index in primary biliary cholangitis: a retrospective single-center study
Article References: Li, L., Liu, H.-L., Danzeng, Q.-Z., Zhou, A.-L., Li, S.-S., & Yang, Y.-F. (2026). Development and internal validation of a steatosis index in primary biliary cholangitis: a retrospective single-center study. BMC Gastroenterology. https://doi.org/10.1186/s12876-026-05585-9
Image Credits: AI Generated
DOI: 10.1186/s12876-026-05585-9
Keywords: primary biliary cholangitis, hepatic steatosis, steatosis index, triglycerides, diabetes mellitus, liver biopsy, internal validation, genetic algorithm, area under the curve, Brier score, calibration slope, retrospective cohort study, cross-validation, LASSO regression, hepatic steatosis index, TyG index
News Source: Ophelia Keating. (November 19, 2026). Novel Two-Variable Steatosis Index Evaluated in Primary Biliary Cholangitis Cohort. Scienmag.



