Predicting short-term mortality in severe cirrhosis: An interpretable machine learning model integrating routine clinical indicators

Shun Zhang, Rui Liu, Zhengjie Li, Tao Pan and Xudong Wen

PLOS ONE, 2026, vol. 21, issue 3, 1-20

Abstract: Background: The critical need for precise risk stratification in severe liver cirrhosis is underscored by its substantial 30-day mortality rates, demanding reliable tools to guide clinical interventions. Objective: To establish a machine learning-driven prognostic model for short-term mortality prediction in decompensated cirrhosis through comprehensive analysis of critical care data. Methods: This retrospective cohort study analyzed 1,044 carefully curated cases from the MIMIC-IV database, randomly divided into training (n = 740) and validation (n = 304) sets. We developed a machine learning model incorporating multidimensional clinical parameters, with rigorous evaluation and internal validation. Short-term survival was analyzed via bootstrap-validated Cox proportional hazards regression. Prognostic heterogeneity across international normalized ratio (INR)-based strata was examined. Results: The final prediction model incorporated eight significant predictors: age (OR 1.051, 95% CI 1.033–1.070), INR (OR 1.423, 95%CI 1.231–1.644), creatinine (OR 1.171, 95%CI 1.071–1.208), platelets (OR 0.995, 95%CI 0.993–0.997), white blood cell (OR 1.116, 95%CI 1.078–1.155), total bilirubin (OR 1.027, 95%CI 1.002–1.052), peptic ulcer (OR 0.336, 95%CI 0.134–0.845), and Aspartate Aminotransferase/Alanine Aminotransferase (AST/ALT) (OR 1.508, 95%CI 1.294–1.757). The model demonstrated excellent discrimination with an AUC of 0.846 in the training cohort. Cox regression analysis confirmed these findings and identified additional associations with aspartate aminotransferase and red blood cell levels. Furthermore, the indicators within the model provide accurate predictions for the clinical outcomes of patients suffering from severe cirrhosis. Subgroup analysis revealed significant mortality variations across different INR ranges (P

Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0328952 (text/html)
https://journals.plos.org/plosone/article/file?id= ... 28952&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pone00:0328952

DOI: 10.1371/journal.pone.0328952

Access Statistics for this article

More articles in PLOS ONE from Public Library of Science
Bibliographic data for series maintained by plosone ().