Understanding Machine Learning Applications in Lung Transplantation: A Narrative Review
Bieke Vercauteren1,2, Balin Özsoy1,3,4, Jasper Gielen2
1Department of Chronic Diseases and Metabolism, Laboratory of Respiratory Diseases and Thoracic Surgery (BREATHE), KU Leuven, Leuven, Belgium.
Summary
Machine learning (ML) can improve lung transplantation (LTx) outcomes by analyzing complex data for organ allocation and predicting patient results. Overcoming data challenges is key for integrating ML into clinical practice.
Area of Science:
- Transplantation research
- Medical informatics
- Artificial intelligence in medicine
Background:
- Lung transplantation (LTx) is vital for end-stage lung disease but faces challenges like donor scarcity and graft failure.
- Machine learning (ML) offers potential solutions by analyzing complex data to uncover insights and enhance outcomes.
Purpose of the Study:
- To review machine learning studies in lung transplantation.
- To explain ML methodologies and their application in LTx.
- To highlight the potential and challenges of ML in improving LTx outcomes.
Main Methods:
- Review of existing literature on ML applications in LTx.
- Explanation of various ML techniques including support vector machines, deep learning, random forests, and transfer learning.
- Discussion of ML's role in analyzing multi-omics data and imaging.
Main Results:
- ML shows promise in organ allocation and predicting LTx outcomes.
- Specific ML techniques are effective for risk stratification and improving interpretability.
- Transfer learning aids model development in limited data scenarios.
- ML is increasingly used for multi-omics and imaging diagnostics.
Conclusions:
- Barriers to clinical adoption include small datasets, data inconsistency, poor interpretability, and lack of validation.
- Future progress depends on multicenter collaborations, transparent methods, and clinical workflow integration.
- ML should augment, not replace, clinical judgment to improve LTx outcomes.
