DINO-3DRA：利用二维基础模型语义实现三维颅内动脉瘤分割

中文摘要 AI 辅助

三维旋转血管造影（3DRA）中的动脉瘤精确分割面临三类挑战：极端类别不平衡、动脉瘤与血管形态相似、缺乏大规模三维预训练。二维视觉基础模型从17亿张图像中编码了密集结构先验，但直接按切片迁移会破坏解剖连续性并导致优化不稳定。本文提出DINO-3DRA，一种双路径框架，通过将冻结的DINOv3特征经Room-Lite空间混合与校准残差融合注入3D U-Net主干，实现高效跨维度语义迁移。在多中心3DRA数据上，DINO-3DRA以仅572万可训练参数实现了最优动脉瘤分割性能：Dice系数达0.758，95%豪斯多夫距离（HD95）为2.75毫米，较nnU-Net提升13%。 ablation研究证实，性能提升源于结构化跨维度迁移而非仅损失函数设计，桥接的基础特征改善了动脉瘤与母血管间的解剖连续性。在未对CADA和SHINY-ICARUS数据集微调的情况下，DINO-3DRA消除了基线架构中所有灾难性失败案例，展现出对异构成像协议的强泛化能力。

英文摘要

Accurate aneurysm segmentation in 3D rotational angiography (3DRA) is hindered by extreme class imbalance, morphological similarity to vessels, and absent large-scale 3D pretraining. 2D vision foundation models encode dense structural priors from 1.7 billion images, yet naïve slice-wise transfer fragments anatomical continuity and destabilises optimisation. We propose DINO-3DRA, a dual-path framework achieving effective cross-dimensional semantic transfer by injecting frozen DINOv3 features into a 3D U-Net backbone via Room-Lite spatial mixing and calibrated residual fusion. On multi-centre 3DRA data, DINO-3DRA achieves state-of-the-art aneurysm segmentation (Dice: 0.758; HD95: 2.75 mm; +13% over nnU-Net) with only 5.72M trainable parameters. Ablation studies confirm that gains arise from structured cross-dimensional transfer rather than loss design alone, with bridged foundation features improving anatomical continuity between aneurysms and parent vessels. Without fine-tuning on CADA and SHINY-ICARUS, DINO-3DRA eliminates all catastrophic failure cases observed in baseline architectures, demonstrating robust generalisation across heterogeneous imaging protocols.

发表机构

• University of Manchester（曼彻斯特大学）
• Christabel Pankhurst Institute, University of Manchester（曼彻斯特大学克里斯塔贝尔·潘克赫斯特研究所）
• NIHR Manchester Biomedical Research Centre（NIHR曼彻斯特生物医学研究中心）
• Manchester Academic Health Sciences Centre（曼彻斯特学术健康科学中心）

机构由 AI 辅助整理，请以论文原文为准。

补充信息

备注

Accepted by MICCAI 2026