Forecasting Waitlist Trajectories for Patients With Metabolic Dysfunction–Associated Steatohepatitis Cirrhosis: A Neural Network Competing Risk Analysis

Background: The Model for End-Stage Liver Disease (MELD)-based liver allocation system has suboptimal concordance in predicting waitlist mortality for patients with Metabolic Dysfunction-Associated Steatohepatitis (MASH) cirrhosis and does not capture competing outcomes of death and transplantation on the liver transplant waitlist. Objective: A competing risk analysis using deep learning was conducted to forecast waitlist trajectories of MASH patients using data available at the time of waitlisting. Methods: A deep learning competing risk model was constructed using data from 17,551 waitlisted MASH cirrhosis patients in the Scientific Registry of Transplant Recipients (SRTR) based on the DeepHit model framework. Its performance was evaluated and compared to single-risk Cox Proportional Hazards (CoxPH) and Random Survival Forests (RSF) models in predicting death or transplant, using the Concordance index (C-index) and Brier scores. Additionally, a novel performance metric, the competing event coherence (CEC) score, was developed to evaluate model performance in the setting of competing risks. Features associated with death and transplant in the DeepHit model were identified using permutation importance. Models were externally validated. Results: In a competing risk scenario, DeepHit achieved the best CEC scores at one, three, six, and 12 months on the waitlist. The RSF model showed the highest C-indices for most individual events, except for death at 3 months and transplant at 1 month, and lower Brier scores except for transplant at 12 months. These results were consistent on external validation. MELD at listing along with its components as well as functional status, age, and blood type, were associated with both death and transplant on the waitlist. Conclusions: A deep learning competing risk analysis can be used to forecast the risks of both death and transplant in MASH patients on waitlist, helping to inform clinical decisions by identifying the most impactful covariates for each outcome.