Predicting Graft Loss after Living Donor Kidney Transplantation: Informing Optimal Donor Selection for a Potential Kidney Transplant Recipient

Team: Precision Care Medicine: Rose

Program: Biomedical Engineering

As of late 2019, there were nearly 800,000 people with end-stage kidney disease in the United States. Living donor kidney transplantation (LDKT) is the treatment of choice for such individuals, as it is associated with significantly improved long-term survival and quality of life compared to other renal replacement strategies. For any potential LDKT recipient, it is important to predict the risk of transplanted organ (graft) failure so as to inform selection of the best available donor kidney for said recipient. This is a challenging task, as graft failure results from many complex interactions between donor and recipient characteristics. Our team developed several predictive models of graft loss using survival analysis techniques and random forest and XGBoost algorithms. In all cases, we compared performance to an earlier model for graft loss risk created by our clinical mentors. We ultimately aim to develop a novel risk index which will facilitate the matching of potential living kidney donors to recipients so as to maximize transplant survival.

Allan Massie
Dorry Segev
Raimond L Winslow
Joseph L Greenstein

Team Members

  • Christian Gonzalez
  • Vedant Jain
  • Jinghan Lin
  • Ayon Mitra
  • Amy van Ee
  • Emma Weeding

Project Links