Learning to segment anatomy and lesions from disparately labeled sources in brain MRI

2025-03-24

AI Summary

Addressing the challenge of jointly segmenting healthy anatomical structures and pathological regions in brain MRI—particularly when lesions induce anatomical deformations and paired annotations for both structures are unavailable—this paper proposes the first anatomy-lesion dual-path decoupled segmentation framework. The method employs an image-adaptive mechanism to suppress lesion-induced interference on healthy tissue segmentation, and integrates multi-sequence MRI modeling, attention-based feature fusion, and image-level adaptive inference, coupled with meta-learning and co-training strategies to enable end-to-end optimization on heterogeneous datasets containing only single-class labels (anatomy or lesion). Evaluated on a public glioblastoma dataset, our approach outperforms state-of-the-art methods in both anatomical structure and lesion segmentation metrics, significantly reducing reliance on scarce jointly annotated training data.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Segmenting healthy tissue structures alongside lesions in brain Magnetic Resonance Images (MRI) remains a challenge for today's algorithms due to lesion-caused disruption of the anatomy and lack of jointly labeled training datasets, where both healthy tissues and lesions are labeled on the same images. In this paper, we propose a method that is robust to lesion-caused disruptions and can be trained from disparately labeled training sets, i.e., without requiring jointly labeled samples, to automatically segment both. In contrast to prior work, we decouple healthy tissue and lesion segmentation in two paths to leverage multi-sequence acquisitions and merge information with an attention mechanism. During inference, an image-specific adaptation reduces adverse influences of lesion regions on healthy tissue predictions. During training, the adaptation is taken into account through meta-learning and co-training is used to learn from disparately labeled training images. Our model shows an improved performance on several anatomical structures and lesions on a publicly available brain glioblastoma dataset compared to the state-of-the-art segmentation methods.

Problem

• Segmenting brain MRI with disparate labels
• Handling lesion disruptions in anatomy segmentation
• Training without jointly labeled datasets

Innovation

• Decouples healthy tissue and lesion segmentation paths
• Uses attention mechanism to merge multi-sequence information
• Employs meta-learning for image-specific adaptation during training