Predicting post–liver transplant outcomes in patients with acute-on-chronic liver failure using Expert-Augmented Machine Learning - Journal article abstract

Aug 5, 2026

Upload by Aria

Liver transplantation treats acute-on-chronic liver failure (ACLF), yet post-transplant mortality remains high and existing prediction models for ACLF are limited. This study builds an Expert-Augmented Machine Learning (EAML) model to predict post–liver transplant outcomes. ACLF patients receiving transplant were identified from the University of California Health Data Warehouse, RuleFit extracted decision rules, and human expert ratings were integrated to form final EAML models. Performance for 1-year and 90-day death and comparisons with baseline methods are reported, and discrepancies highlight biomarker ranking differences.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages28

Reading Time1 h 11 min

AccessFree to download

Abstract

Study approach

Model performance and comparisons

Expert-vs-ML discrepancy and implications

UCSF UC San Francisco Previously Published Works Title Predicting post–liver transplant outcomes in patients with acute-on-chronic liver failure using Expert-Augmented Machine Learning Permalink [https://escholarship.org/uc/item/10q1z73c](https://escholarship.org/uc/item/10q1z73c) Journal American Journal of Transplantation, 23(12) ISSN 1600-6135 Authors Ge, Jin Digitale, Jean C Fenton, Cynthia et al. Publication Date 2023-12-01 DOI 10.1016/j.ajt.2023.08.022 Peer reviewed [eScholarship.org](eScholarship.org) Powered by the California Digital Library University of California Author Manuscr ipt Author Manuscr ipt Author Manuscr ipt Author Manuscript | | HHS Public Access
Author manuscript
Am J Transplant. Author manuscript; available in PMC 2024 September 01. | | --- | --- | Published in final edited form as: Am J Transplant. 2023 December ; 23(12): 1908–1921. doi:10.1016/j.ajt.2023.08.022 . Predicting post–liver transplant outcomes in patients with acute-on-chronic liver failure using Expert-Augmented Machine Learning Jin Ge 1,* , Jean C. Digitale2 , Cynthia Fenton3 , Charles E. McCulloch2 , Jennifer C. Lai 1 , Mark J. Pletcher2 , Efstathios D. Gennatas2 1 Division of Gastroenterology and Hepatology, Department of Medicine, University of California– San Francisco, San Francisco, California, USA 2 Department of Epidemiology and Biostatistics, University of California–San Francisco, San Francisco, California, USA 3 Division of Hospital Medicine, Department of Medicine, University of California–San Francisco, San Francisco, California, USA Abstract Liver transplantation (LT) is a treatment for acute-on-chronic liver failure (ACLF), but high post-LT mortality has been reported. Existing post-LT models in ACLF have been limited. We developed an Expert-Augmented Machine Learning (EAML) model to predict post-LT outcomes. We identified ACLF patients who underwent LT in the University of California Health Data Warehouse. We applied the RuleFit machine learning (ML) algorithm to extract rules from *Corresponding author. Jin Ge, 513 Parnassus Avenue, S-357 San Francisco, CA 94143, USA., [jin.ge@ucsf.edu](jin.ge@ucsf.edu) (J. Ge) . Author contributions Authorship was determined using ICMJE recommendations. J. G.: Study concept and design; data extraction; analysis and interpretation of data; drafting of manuscript; critical revision of the manuscript for important intellectual content; statistical analysis; obtained funding; study supervision J.C.D.: Analysis and interpretation of data; drafting of manuscript; critical revision of the manuscript for important intellectual content C. F.: Data extraction; critical revision of the manuscript for important intellectual content C.E.M.: Interpretation of data; critical revision of the manuscript for important intellectual content J.C.L.: Study concept and design; interpretation of data; critical revision of the manuscript for important intellectual content; study supervision M.J.P.: Interpretation of data; critical revision of the manuscript for important intellectual content E.D.G.: Study concept and design; analysis and interpretation of data; critical revision of the manuscript for important intellectual content Declaration of competing interests The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jin Ge reports financial support was provided by National Center for Advancing Translational Sciences. Jin Ge reports financial support was provided by American Association for the Study of Liver Diseases. Jin Ge reports financial support was provided by American Society of Transplantation. Jin Ge reports financial support was provided by UCSF Liver Center. Charles E. McCulloch reports financial support was provided by National Center for Advancing Translational Sciences. Mark J. Pletcher reports financial support was provided b

FAQs

What clinical problem does this study address?

It targets predicting outcomes after liver transplantation in patients with acute-on-chronic liver failure, where post-transplant mortality is high and existing models are limited.

How is the Expert-Augmented Machine Learning (EAML) model constructed?

The study uses the RuleFit algorithm to extract rules from ACLF data, then incorporates human expert ratings to generate final EAML models.

What data source and patient cohort are used?

ACLf patients who underwent liver transplantation were identified from the University of California Health Data Warehouse.

Aug 5, 2026

Upload by Aria