Multi-task Learning Graph Neural Networks for Cancer Prognosis Prediction with Genomic Data
Part Of
Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
Start Page
1
End Page
5
ISBN
979-835037149-9
Date Issued
2024-07-15
Author(s)
DOI
10.1109/EMBC53108.2024.10782177
Abstract
Providing robust prognosis predictions for cancers with limited data samples remains a challenge for precision oncology. In this study, we propose a novel approach that combines multi-task learning (MTL) and graph neural networks (GNNs) to address this issue. By representing gene-gene interactions as a graph network, our approach leverages multi-task learning to effectively capture the relationships of genes relevant to the oncogenesis and progression of breast, lung, and colon cancer. We demonstrate that our approach improves the cancer prognosis prediction for cancers with fewer samples, such as colon adenocarcinoma, by leveraging the shared gene-gene interactions across different cancer types, obtaining increases in the area under the precision-recall curve (AUPRC) of 24%. Our work contributes to the field of smart healthcare by demonstrating the potential of MTL and GNNs for enhancing cancer prognosis prediction, even with limited data samples.
Event(s)
46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2024
SDGs
Publisher
Institute of Electrical and Electronics Engineers Inc.
Type
conference paper