Federated Transformer-GNN for Privacy-Preserving Brain Tumor Localization with Modality-Level Explainability
This work addresses privacy-preserving collaboration for healthcare institutions dealing with brain tumor analysis, but it is incremental as it extends existing methods within a known federated learning platform.
The paper tackles the problem of training deep learning models for brain tumor localization without sharing sensitive patient data by proposing a federated learning framework, which matches centralized performance by leveraging distributed data from multiple institutions. It also provides explainability analysis showing that deeper network layers increase attention to specific MRI modalities, validated with statistical significance.
Deep learning models for brain tumor analysis require large and diverse datasets that are often siloed across healthcare institutions due to privacy regulations. We present a federated learning framework for brain tumor localization that enables multi-institutional collaboration without sharing sensitive patient data. Our method extends a hybrid Transformer-Graph Neural Network architecture derived from prior decoder-free supervoxel GNNs and is deployed within CAFEIN\textsuperscript{\textregistered}, CERN's federated learning platform designed for healthcare environments. We provide an explainability analysis through Transformer attention mechanisms that reveals which MRI modalities drive the model predictions. Experiments on the BraTS dataset demonstrate a key finding: while isolated training on individual client data triggers early stopping well before reaching full training capacity, federated learning enables continued model improvement by leveraging distributed data, ultimately matching centralized performance. This result provides strong justification for federated learning when dealing with complex tasks and high-dimensional input data, as aggregating knowledge from multiple institutions significantly benefits the learning process. Our explainability analysis, validated through rigorous statistical testing on the full test set (paired t-tests with Bonferroni correction), reveals that deeper network layers significantly increase attention to T2 and FLAIR modalities ($p<0.001$, Cohen's $d$=1.50), aligning with clinical practice.