InfiltrNet: Dual-Branch CNN-Transformer Architecture for Brain Tumor Infiltration Risk Prediction

2026-05-04

AI Summary

This study addresses the challenge of accurately predicting glioma infiltration beyond the MRI-visible tumor boundary, a critical limitation in surgical and radiotherapy planning. The authors propose InfiltrNet, a novel dual-branch architecture that uniquely integrates convolutional neural networks (CNNs) and Swin Transformers, leveraging a cross-attention mechanism to fuse multimodal MRI data and generate a three-zone infiltration risk map. Key methodological innovations include a reproducible annotation strategy based on distance transforms, a boundary-aware loss function, intermediate-layer auxiliary supervision, and interpretability analysis via GradCAM++. Evaluated on the BraTS 2020 and 2025 datasets, the model significantly outperforms five baseline methods, with visualizations consistently highlighting clinically relevant peritumoral regions, thereby demonstrating both its predictive efficacy and interpretability.

Abstract

Gliomas are aggressive brain tumors that infiltrate surrounding tissue beyond the visible tumor margins observed on Magnetic Resonance Imaging (MRI). Predicting the spatial extent of this infiltration is essential for surgical planning and radiation therapy, yet existing deep learning approaches focus on segmenting the visible tumor rather than estimating infiltration risk in the surrounding tissue. This paper presents InfiltrNet, a novel dual-branch architecture that combines a convolutional neural network (CNN) encoder with a Swin Transformer encoder through cross-attention fusion modules to predict three-zone infiltration risk maps from multimodal MRI. A label generation strategy based on distance transforms is proposed to derive reproducible infiltration risk zones from standard Brain Tumor Segmentation (BraTS) annotations. InfiltrNet is trained with a combined Dice-CrossEntropy and boundary-aware loss augmented by auxiliary supervision heads at intermediate decoder levels. Extensive experiments on BraTS 2020 and BraTS 2025 demonstrate that InfiltrNet outperforms five established baselines. Explainability analysis using GradCAM++ and Occlusion sensitivity confirms that the model attends to clinically relevant peritumoral regions.

Problem

• brain tumor infiltration
• risk prediction
• multimodal MRI
• glioma
• peritumoral region

Innovation

• dual-branch architecture
• cross-attention fusion
• infiltration risk prediction
• distance transform labeling
• boundary-aware loss