发表机构
Seoul National University; OUTTA; Chung-Ang University; Samsung Medical Center, Sungkyunkwan University School of Medicine; Samsung Changwon Hospital; NVIDIA AI Technology Center（首尔国立大学; OUTTA; Chung-Ang 大学; 三星医疗中心，全北大学医学院; 三星昌原医院; NVIDIA AI 技术中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对3D MRI无创预测PNI的挑战，提出MMA-Former，采用粗-细变压器结构及窗口特定混合注意力机制，能并行多尺度提取特征，实现空间自适应特征提取，在回顾性数据集上AUC达0.752，优于其他架构。
AI 中文摘要
神经周围侵犯（PNI）是胆管癌的关键预后因素。从3D MRI进行无创预测具有挑战性，需要能有效捕捉细粒度细节和全局上下文的模型。我们提出了多窗口混合注意力头变压器（MMA-Former），这是一种新颖的端到端3D架构，具有用于并行多尺度特征提取的粗-细变压器（CFT）结构。我们通过集成一种新颖的窗口特定混合注意力（WS-MoH）机制来改进此结构。与标准多头自注意力（MSA）不同，WS-MoH为每个3D窗口生成一个表示，并将整个窗口动态路由到专门的或通用的注意力头。这实现了针对每个窗口的局部上下文进行空间自适应特征提取，在不增加参数的情况下增强了专业化并减少了冗余。在168例T1加权MRI扫描的回顾性数据集中进行评估，MMA-Former的AUC为0.752，优于其他3D架构，包括最佳的CNN（AUC为0.708）和变压器基线（AUC为0.681）。
英文摘要
Perineural invasion (PNI) is a critical prognostic factor in cholangiocarcinoma. Non-invasive prediction from 3D MRI is challenging, demanding models that efficiently capture both fine-grained details and global context. We propose the Multi-window Mixture-of-Head Attention Transformer (MMA-Former), a novel end-to-end 3D architecture featuring a Coarse-Fine Transformer (CFT) structure for parallel multi-scale feature extraction. We advance this structure by integrating a novel Window-Specific Mixture-of-Head attention (WS-MoH) mechanism. Unlike standard Multi-Head Self Attention (MSA), WS-MoH generates a representation for each 3D window and dynamically routes the entire window to specialized or common attention heads. This enables spatially adaptive feature extraction tailored to the local context of each window, enhancing specialization and reducing redundancy without increasing parameters. Evaluated on a retrospective dataset of 168 T1-weighted MRI scans, MMA-Former achieved an AUC of 0.752, outperforming other 3D architectures, including the best CNN (AUC of 0.708) and Transformer baselines (AUC of 0.681).
CommentsPublished in the 2026 IEEE 23rd International Symposium on Biomedical Imaging (ISBI 2026); accepted for oral presentation
Journal ref2026 IEEE 23rd International Symposium on Biomedical Imaging (ISBI), 2026
DOI：10.1109/ISBI61048.2026.11515401
