Background/Aims Prediction of short-term mortality in patients with acute-on-chronic liver failure (ACLF) admitted to the intensive care unit (ICU) may enhance effective management.
Methods To develop, explain, and validate a predictive machine learning (ML) model for short-term mortality in patients with ACLF with two or more organ failures (OFs). Utilizing a large ICU cohort with detailed clinical information, we identified ACLF patients with two or more OFs according to the EASL-CLIF and NACSELD definitions. ML model was developed for each definition to predict 30-day mortality. The Shapley value was estimated to explain the models. Validation and calibration of these models were performed.
Results Of 5,994 patients with cirrhosis admitted to ICU, 1,511 met NACSELD criteria, and 1,692 met EASL-CLIF grade II or higher criteria. The CatBoost ACLF (CBA) model had the greatest accuracy in the NACSELD cohort (area under curve [AUC] of 0.87), while the Random Forest ACLF (RFA) model performed best in the EASL-CLIF cohort (AUC of 0.83). Both models showed robust calibration. The models were explained by SHAP score analysis, yielding a rank list, and the top twelve predictors were selected. Both simplified models demonstrated similar performance (CBA model: AUC 0.89, RFA model: AUC 0.81) and significantly outperformed contemporary scoring systems, including CLIF-C ACLF and MELD 3.0. The models were validated in both internal and external cohorts. A simple-to-use online tool was created to predict mortality rates.
Conclusions We presented explainable, well-validated, and calibrated predictive models for ACLF patients with two or more OFs, which outperformed existing predictive scores.
Citations
Citations to this article as recorded by
Unbiased clustering of acute-on-chronic liver failure patients using machine learning in a real-world ICU cohort Mengyi Zhang, Fanpu Ji, Jian Zu, Yingli He, Tao Chen, Yi Liu, Hirsh D. Trivedi, Ju Dong Yang, Vinay Sundaram, Yuan Wang, Xiaodan Sun, Zhujun Cao, Chun-Ying Wu, Yee Hui Yeo, Rajiv Jalan Nature Communications.2026;[Epub] CrossRef
Background/Aims The role of reactivation of human cytomegalovirus (HCMV) in determining outcomes of cirrhotic patients with acute decompensation (AD) is unknown. We aimed to investigate HCMV incidence and potential correlation with hepatic outcomes in AD patients.
Methods Two prospective multicentre cohorts with AD patients were investigated. Patients in cohort 1 were recruited from 4 centres, while patients in cohort 2 were randomly selected from a second multicentre cohort. HCMV reactivation was established with quantitative real-time polymerase chain reaction assay in seropositive patients.
Results HCMV reactivation was found in 35 patients from cohort 1 (n=722) and 14 from cohort 2 (n=291), with an incidence of 4.8% in both cohorts. Bacterial infection and liver failure were independently correlated with HCMV reactivation. HCMV reactivation was an independent predictor of 90-day mortality. Among bacterial infection populations in these two cohorts, patients with HCMV reactivation had worse prognosis compared to those without. Incidence of acute-on-chronic liver failure (ACLF) was higher in patients with HCMV reactivation compared to those without and was also independently correlated with development of ACLF. In a total of 49 HCMV reactivation cases, 8 patients were treated with ganciclovir, in whom a significantly lower 90-day mortality compared with those not treated was observed. All 3 patients who underwent liver transplantation with reactivation of HCMV died.
Conclusions In AD patients, HCMV reactivation was common, especially in those with bacterial infection or liver failure, and they were more prone to having ACLF and 90‑day mortality. The data propose the need for active surveillance for HCMV infection in AD patients.
First
Prev
