Predicting the outcome for patients in a heart transplantation queue using deep learning
Insights
Deep learning models predict heart transplant waiting list outcomes. This approach improves patient status prediction, offering valuable insights for managing organ transplant waiting lists.
Area of Science:
- Medical Informatics
- Cardiology
- Machine Learning
Background:
- Heart transplantation offers extended survival for end-stage heart disease but is limited by donor organ scarcity.
- Patients face prolonged waiting periods, averaging 200 days, with significant individual variability.
- Predicting patient outcomes during the waiting period is crucial for resource allocation and patient management.
Purpose of the Study:
- To develop and evaluate deep learning models for predicting patient outcomes on the heart transplant waiting list.
- To assess model performance at multiple time points (180, 365, and 730 days).
- To identify key predictors influencing patient status during the waiting period.
Main Methods:
- Utilized a two-layer neural network architecture implemented with the Keras framework.
- Trained the model on adult patient data (>17 years) from the United Network for Organ Sharing (UNOS) registry (January 2000 - December 2011).
- Employed a backward elimination procedure to determine the 10 most significant predictive parameters.
Main Results:
- Achieved F1 macro scores of 0.674, 0.680, and 0.680 at 180, 365, and 730 days, respectively.
- Demonstrated a significant improvement over a baseline model with a score of 0.271.
- Identified the top 10 parameters most influential in predicting patient outcomes.
Conclusions:
- Deep learning models can effectively predict patient outcomes on heart transplant waiting lists.
- The identified significant parameters offer insights into factors affecting patient survival and transplantation status.
- This predictive capability can aid in optimizing waiting list management and patient care strategies.
Abstract:
Heart transplantations have made it possible to extend the median survival time to 12 years for patients with end-stage heart diseases. This operation is unfortunately limited by the availability of donor organs and patients have to wait on average about 200 days in a waiting list before being operated. This waiting time varies considerably across the patients. In this paper, we studied the outcome for patients entering a transplantation waiting list using deep learning techniques. We implemented a model in the form of two-layer neural networks and we predicted the outcome as still waiting, transplanted or dead in the waiting list, at three different time points: 180 days, 365 days, and 730 days. As data source, we used the United Network for Organ Sharing (UNOS) registry, where we extracted adult patients (>17 years) from January 2000 to December 2011. We trained our model using the Keras framework, and we report F1 macro scores of respectively 0.674, 0.680, and 0.680 compared to a baseline of 0.271. We also applied a backward elimination procedure, using our neural network, to extract the 10 most significant parameters predicting the patient status for the three different time points.
