M4: Multi-proxy multi-gate mixture of experts network for multiple instance learning in histopathology image analysis

• Published in: Medical image analysis Vol. 103; p. 103561
• Main Authors: Li, Junyu, Zhang, Ye, Shu, Wen, Feng, Xiaobing, Wang, Yingchun, Yan, Pengju, Li, Xiaolin, Sha, Chulin, He, Min
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.07.2025
• Subjects: Algorithms; Genetic mutation; Humans; Image Interpretation, Computer-Assisted - methods; Image Processing, Computer-Assisted - methods; Machine Learning; Multi-task learning; Multiple instance learning; Multiple-Instance Learning Algorithms; Mutation; Neoplasms - diagnostic imaging; Neoplasms - genetics; Neoplasms - pathology; Neural Networks, Computer; Whole slide image; Whole slide image; Multiple instance learning; Multi-task learning; Genetic mutation
• ISSN: 1361-8415; 1361-8423; 1361-8431; 1361-8423
• DOI: 10.1016/j.media.2025.103561

Multiple instance learning (MIL) has been successfully applied for whole slide images (WSIs) analysis in computational pathology, enabling a wide range of prediction tasks from tumor subtyping to inferring genetic mutations and multi-omics biomarkers. However, existing MIL methods predominantly focus on single-task learning, resulting in not only overall low efficiency but also the overlook of inter-task relatedness. To address these issues, we proposed an adapted architecture of Multi-gate Mixture-of-experts with Multi-proxy for Multiple instance learning (M4), and applied this framework for simultaneous prediction of multiple genetic mutations from WSIs. The proposed M4 model has two main innovations: (1) adopting a multi-gate mixture-of-experts strategy for multiple genetic mutation simultaneous prediction on a single WSI; (2) introducing a multi-proxy CNN construction on the expert and gate networks to effectively and efficiently capture patch-patch interactions from WSI. Our model achieved significant improvements across five tested TCGA datasets in comparison to current state-of-the-art single-task methods. The code is available at: https://github.com/Bigyehahaha/M4. [Display omitted] •Proposed M4 architecture to simultaneously predict multiple genetic mutations from WSIs.•Constructed an MMoE network with multi-proxy adaptation for MIL tasks.•Achieved notable improvements on five TCGA datasets over state-of-the-art methods.