Building analytical models for predicting de novo malignancy in pancreas transplant patients: A machine learning approach

Abstract

This research aimed to predict de novo malignancy in patients who underwent pancreas transplants using a machine learning approach. We constructed various predictive models based on data from the Organ Procurement and Transplantation Network (OPTN) to classify malignant patterns in transplant patients. The models were trained using medical records spanning from 1987 to 2020 to identify essential prognostic components associated with malignancy development. Various datasets from the United Network for Organ Sharing Standard Transplant Analysis and Research (UNOS-STAR) database were used to evaluate each model's performance using the areas under the receiver operating characteristic (ROC) and precision-recall (PR) curves. Our findings demonstrated the effectiveness of machine learning in predicting de novo malignancy in pancreas transplant patients. Specifically, we identified the recipient’s B2 and DR1 antigens and the donor’s DDAVP hormone as significant factors associated with malignancy. Our study highlights the importance of integrating medical records from various sources to enhance the accuracy of predictive models for organ transplantation.