Second-Step Risk Stratification Within the FIB-4 Gray Zone in MASLD: Derivation and Temporal Validation of a Four-Variable Clinical Model

Background & Aims: FIB-4 is recommended for first-line fibrosis risk assessment in MASLD, but many patients fall into the indeterminate gray zone (1.30–2.67), requiring further stratification. We tested whether a simple model using routine clinical variables improves second-step risk stratification after indeterminate FIB-4, especially when VCTE is unavailable. Methods Using NHANES data, we developed a weighted ridge logistic model in 2017–March 2020 (n = 2,709) and validated it in 2021–2023 (n = 1,543). Elevated liver stiffness was LSM ≥ 8.0 kPa. Predictors were age, BMI, ALT, and GGT. Performance was assessed by discrimination, calibration, prediction error, reclassification, decision-curve analysis, and prespecified gray-zone analyses. Results In temporal validation, the model had AUC 0.722 (95% CI 0.690–0.754) versus 0.581 (0.544–0.617) for FIB-4 (P < 0.001); survey-weighted AUCs were 0.735 and 0.562. Calibration was preserved (intercept 0.213; slope 1.014). Within the FIB-4 gray zone (n = 420; weighted LSM ≥ 8.0 kPa prevalence 23.7%), AUC was 0.697 versus 0.542 for FIB-4 (P < 0.001). Discrimination was higher than SAFE but similar to MAF-5. Thresholds 0.0923 and 0.2623 defined lower-, intermediate-, and higher-priority groups with event prevalences 8.2%, 23.8%, and 45.8%. At validation, sensitivity at the lower threshold was 90.8%; specificity at the upper threshold was 86.9%. At ~ 90% sensitivity, model-based prioritization reduced VCTE referral from 86.6% to 69.0%. Conclusion A simple four-variable model improved second-step risk stratification beyond FIB-4, particularly in the gray zone, with acceptable temporal calibration and clinically useful risk separation. It may help prioritize VCTE when access is constrained, but should complement, not replace, definitive fibrosis assessment, and requires external validation.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-20
DOI
https://doi.org/10.21203/rs.3.rs-11481076/v1
Primary Topic
Liver Disease Diagnosis and Treatment
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Second-Step Risk Stratification Within the FIB-4 Gray Zone in MASLD: Derivation and Temporal Validation of a Four-Variable Clinical Model

Xinyu Zhang, Yufeng Xing, Zhiyi Qiu
Research Square
Liver Disease Diagnosis and Treatment
preprint

Second-Step Risk Stratification Within the FIB-4 Gray Zone in MASLD: Derivation and Temporal Validation of a Four-Variable Clinical Model

Xinyu Zhang, Yufeng Xing, Zhiyi Qiu
preprint en

Abstract

Background & Aims: FIB-4 is recommended for first-line fibrosis risk assessment in MASLD, but many patients fall into the indeterminate gray zone (1.30–2.67), requiring further stratification. We tested whether a simple model using routine clinical variables improves second-step risk stratification after indeterminate FIB-4, especially when VCTE is unavailable. Methods Using NHANES data, we developed a weighted ridge logistic model in 2017–March 2020 (n = 2,709) and validated it in 2021–2023 (n = 1,543). Elevated liver stiffness was LSM ≥ 8.0 kPa. Predictors were age, BMI, ALT, and GGT. Performance was assessed by discrimination, calibration, prediction error, reclassification, decision-curve analysis, and prespecified gray-zone analyses. Results In temporal validation, the model had AUC 0.722 (95% CI 0.690–0.754) versus 0.581 (0.544–0.617) for FIB-4 (P < 0.001); survey-weighted AUCs were 0.735 and 0.562. Calibration was preserved (intercept 0.213; slope 1.014). Within the FIB-4 gray zone (n = 420; weighted LSM ≥ 8.0 kPa prevalence 23.7%), AUC was 0.697 versus 0.542 for FIB-4 (P < 0.001). Discrimination was higher than SAFE but similar to MAF-5. Thresholds 0.0923 and 0.2623 defined lower-, intermediate-, and higher-priority groups with event prevalences 8.2%, 23.8%, and 45.8%. At validation, sensitivity at the lower threshold was 90.8%; specificity at the upper threshold was 86.9%. At ~ 90% sensitivity, model-based prioritization reduced VCTE referral from 86.6% to 69.0%. Conclusion A simple four-variable model improved second-step risk stratification beyond FIB-4, particularly in the gray zone, with acceptable temporal calibration and clinically useful risk separation. It may help prioritize VCTE when access is constrained, but should complement, not replace, definitive fibrosis assessment, and requires external validation.

Research Square
Guangzhou University of Chinese Medicine (CN), Shenzhen Traditional Chinese Medicine Hospital (CN)
Liver Disease Diagnosis and Treatment
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon