Risk Prediction of Post-Transplant Lymphoproliferative Disease in Heart Transplant Patients.

Abstract

Post-transplant lymphoproliferative disease (PTLD) represents the second most frequent malignancy in cardiac allograft recipients and constitutes up to 10% of de novo cancers. This study attempts to determine risk factors in adult heart transplant patients to develop a risk prediction model using the Scientific Registry of Transplant Recipients database and the lasso regression model. Data on 55,150 adult heart transplant patients (1987-2021) were extracted. The χ2/Wilcoxon tests were performed to identify significant variables (p < 0.05). The dataset was divided into two. One set was used for model development/validation, and the other for simulating external validation. Lasso logistic regression models were developed to predict disease at 1, 3, and 5 years post-transplant. Cyclosporine, positive donor Epstein-Barr Virus (EBV) IgG, induction with OKT3, and the donor's human leukocyte antigen (HLA) B38 antigen had a higher risk at 3 and 5 years post-transplant. African American recipients have a lower risk of developing PTLD as compared with other ethnic groups. This is the first report of a lasso regression model with good discriminatory power (c statistic of >0.7) in testing and validation cohorts. Future studies need to explore advanced modeling technologies and artificial intelligence systems capable of capturing patient diversity.