Liver Transplantation In Highest Acuity Patients (MELD 40 Or More): Identifying Factors Predicting Survival
*Srinath Chinnakotla, *Michael Evans, *David Vock, Arthur Matas
University of Minnesota, Minneapolis, MN
Objective Although survival benefits for highest acuity patients is indisputable, transplanting this subgroup is a medical and economic-challenge. We sought to identify and develop a prognostic model for 1-year survival outcomes in recipients with a MELD score of 40. Methods Review of 5,309 liver-transplants with a MELD score of 40 were identified between 2001-2016 in SRTR-database. Forty-four recipient, donor, and perioperative-factors were identified as potential predictors of transplant outcomes. The relationships between these factors and 1-year patient-survival were examined using random-survival-forests (RSF) methodology. Variable importance measures were used to identify the factors with the strongest influence on survival, and partial dependence plots were used to display the dependence of survival on the target variable while adjusting for all other predictors. Results Predictors of 1-year patient-survival are shown in Figure 1, ranked in order of variable importance. Strongest predictors of futility include patient on life-support, prior-liver-transplant, donor-age, cold-ischemia-time, recipient-age, and DCD-donor. These predictors and their relationships to 1-year-survival are shown in partial-dependence plots in Figure 1. Conclusions Despite high medical acuity, liver transplant-outcomes continue to improve for high-acuity-patients(Figure1B). Using ensemble learning methods, recipient and donor-characteristics available at the time of transplant can be used to provide predicted-survival probabilities for future transplant cases(Figure2).
Back to 2020 Abstracts