The promise of machine learning applications in solid organ transplantation

Abstract

Solid-organ transplantation is a life-saving treatment for end-stage organ disease in highly selected patients. Alongside the tremendous progress in the last several decades, new challenges have emerged. The growing disparity between organ demand and supply requires optimal patient/donor selection and matching. Improvements in long-term graft and patient survival require data-driven diagnosis and management of post-transplant complications. The growing abundance of clinical, genetic, radiologic, and metabolic data in transplantation has led to increasing interest in applying machine-learning (ML) tools that can uncover hidden patterns in large datasets. ML algorithms have been applied in predictive modeling of waitlist mortality, donor-recipient matching, survival prediction, post-transplant complications diagnosis, and prediction, aiming to optimize immunosuppression and management. In this review, we provide insight into the various applications of ML in transplant medicine, why these were used to evaluate a specific clinical question, and the potential of ML to transform the care of transplant recipients. 36 articles were selected after a comprehensive search of the following databases: Ovid MEDLINE; Ovid MEDLINE Epub Ahead of Print and In-Process & Other Non-Indexed Citations; Ovid Embase; Cochrane Database of Systematic Reviews (Ovid); and Cochrane Central Register of Controlled Trials (Ovid). In summary, these studies showed that ML techniques hold great potential to improve the outcome of transplant recipients. Future work is required to improve the interpretability of these algorithms, ensure generalizability through larger-scale external validation, and establishment of infrastructure to permit clinical integration.

Fig. 1. Applications of ML in solid-organ transplantation.

a Artificial Neural Networks (ANNs) benefit from automatically learning from high-dimensional data and detecting complex nonlinear relationships between input variables and outcome of interest. ANNs report high accuracy in optimal identification of potential organ donors. b Convolutional Neural Network (CNNs) are neural network models that are popular for image classification tasks and help in efficient feature extraction through convolution operation and perform efficient segmentation of donor's liver through input data in the form of MRIs. c Random Survival Forest (RSF) approach is an Ensemble tree method resulting in better survival prediction and variable selection. Through RSF laboratory and hemodynamic variables affecting waitlist mortality can be identified through interpreting nonlinear relationships between the variables. d Multilayer perceptions are neural networks that identify complex nonlinear relationships in the data and can help in handling different data domains such as clinical and image features together to predict Hepatocellular Carcinoma (HCC) recurrence with high accuracy. e In Liver transplant recipients, Random Forest (RF) classifier is a tree-based classifier that generalizes classifications using decision trees and can efficiently identify important risk factors relevant to new-onset diabetes after transplantation (NODAT). f Gradient boosting machines employ sequential decision trees which reduce the error by training on the error residuals and can classify a subject into a candidate for risk of pneumonia, RBC transfusion etc. so that clinicians can efficiently filter patients requiring immediate support. g Important risk factors for Delayed Graft function (DGF) can be provided to ANNS, Support Vector Machine (SVMs) and tree-based models to identify patients at higher risk of DGF. ANNs can be applied on high-dimensional datasets, however, when complexity is low, SVMs and decision trees can provide more interpretable modeling.

Fig. 2. Flowchart of search strategy and selection of studies for inclusion.

Database search retrieved 155 papers for initial review. In total, 36 papers were included in the final review according to clinical significance and relevance to machine learning, transplantation, and donation.