An artificial intelligence-generated model predicts 90-day survival in alcohol-associated hepatitis: A global cohort study
Winston Dunn, Yanming Li, Ashwani K. Singal, Douglas A. Simonetto, Luis A. Díaz, Francisco Idalsoaga, Gustavo Ayares, Jorge Arnold, María Ayala-Valverde, Diego Perez, Jaime Gomez, Rodrigo Escarate, Eduardo Fuentes-López, Carolina Ramirez-Cadiz, Dalia Morales-Arraez, Wei Zhang, Steve Qian, Joseph C. Ahn, Seth Buryska, Heer Mehta, Nicholas Dunn, Muhammad Waleed, Horia Stefanescu, Andreea Bumbu, Adelina Horhat, Bashar Attar, Rohit Agrawal, Joaquín Cabezas, Victor Echavaría, Berta Cuyàs, Maria Poca, German Soriano, Shiv K. Sarin, Rakhi Maiwall, Prasun K. Jalal, Fátima Higuera-de-la-Tijera, Anand V. Kulkarni, P Nagaraja Rao, Patricia Guerra-Salazar, Lubomir Skladaný, Natália Kubánek, Veronica Prado, Ana Clemente-Sanchez, Diego Rincon, Tehseen Haider, Kristina R Chacko, Gustavo A. Romero, Florencia D. Pollarsky, Juan C. Restrepo, Luis G. Toro, Pamela Yaquich, Manuel Mendizabal, Maria L. Garrido, Sebastián Marciano, Melisa Dirchwolf, Victor Vargas, César Jiménez, David Hudson, Guadalupe García-Tsao, Guillermo Ortiz, Juan G. Abraldes, Patrick S. Kamath, Marco Arrese, Vijay H. Shah, Ramon Bataller, Juan P. Arab- Hepatology
Background & Aims:
Alcohol-associated hepatitis (AH) poses significant short-term mortality. Existing prognostic models lack precision for 90-day mortality. Utilizing artificial intelligence (AI) in a global cohort, we sought to derive and validate an enhanced prognostic model.
Approach and Results:
The Global AlcHep initiative, a retrospective study across 23 centers in 12 countries, enrolled AH patients per NIAAA criteria. Centers were partitioned into derivation (11 centers, 860 patients) and validation cohorts (12 centers, 859 patients). Focusing on 30 and 90-day post-admission mortality, three AI algorithms (Random Forest, Gradient Boosting Machines, and eXtreme Gradient Boosting) informed an ensemble model, subsequently refined via Bayesian updating, integrating the derivation cohort’s average 90-day mortality with each center’s approximate mortality rate to produce post-test probabilities. The ALCoholic Hepatitis Artificial INtelligence (ALCHAIN) Ensemble score integrated age, gender, cirrhosis, and 9 laboratory values, with center-specific mortality rates. Mortality was 18.7% (30-day) and 27.9% (90-day) in the derivation cohort, versus 21.7% and 32.5% in the validation cohort. Validation cohort 30 and 90-day AUCs were 0.811 (0.779 – 0.844) and 0.799 (0.769 – 0.830), significantly surpassing legacy models like Maddrey’s Discriminant Function, MELD variations, ABIC, Glasgow, and modified Glasgow Scores (
Conclusions:
Harnessing AI within a global consortium, we pioneered a scoring system excelling over traditional models for 30 and 90-day AH mortality predictions. Beneficial for clinical trials, steroid therapy, and transplant indications, it’s accessible at: https://aihepatology.shinyapps.io/ALCHAIN/.