Multimodal Masked Autoencoder Pre-training for 3D MRI-Based Brain Tumor Analysis with Missing Modalities

TL;DR

This paper introduces BM-MAE, a pre-training method for multimodal MRI data that enables models to handle missing modalities effectively, improving downstream task performance and modality reconstruction in brain tumor analysis.

Contribution

The paper presents a novel masked autoencoder pre-training strategy that adapts to any subset of modalities without architectural changes, addressing missing modality issues in medical imaging.

Findings

Outperforms or matches baselines with separate pre-training for each modality subset.

Surpasses training from scratch on multiple downstream tasks.

Efficiently reconstructs missing modalities, aiding clinical applications.

Abstract

Multimodal magnetic resonance imaging (MRI) constitutes the first line of investigation for clinicians in the care of brain tumors, providing crucial insights for surgery planning, treatment monitoring, and biomarker identification. Pre-training on large datasets have been shown to help models learn transferable representations and adapt with minimal labeled data. This behavior is especially valuable in medical imaging, where annotations are often scarce. However, applying this paradigm to multimodal medical data introduces a challenge: most existing approaches assume that all imaging modalities are available during both pre-training and fine-tuning. In practice, missing modalities often occur due to acquisition issues, specialist unavailability, or specific experimental designs on small in-house datasets. Consequently, a common approach involves training a separate model for each desired modality combination, making the process both resource-intensive and impractical for clinical use. Therefore, we introduce BM-MAE, a masked image modeling pre-training strategy tailored for multimodal MRI data. The same pre-trained model seamlessly adapts to any combination of available modalities, extracting rich representations that capture both intra- and inter-modal information. This allows fine-tuning on any subset of modalities without requiring architectural changes, while still benefiting from a model pre-trained on the full set of modalities. Extensive experiments show that the proposed pre-training strategy outperforms or remains competitive with baselines that require separate pre-training for each modality subset, while substantially surpassing training from scratch on several downstream tasks. Additionally, it can quickly and efficiently reconstruct missing modalities, highlighting its practical value. Code and trained models are available at: https://github.com/Lucas-rbnt/BM-MAE