发表机构
University of Calgary（卡尔加里大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出多阶段动态提示nnU-Net，通过多阶段动态提示调制编码器特征，在BraTS GOAT数据集上较基线nnU-Net提升了脑肿瘤分割的Dice分数，增强了模型泛化性。
AI 中文摘要
从磁共振成像（MRI）中准确分割脑肿瘤对诊断、治疗规划、手术引导及疾病监测至关重要。然而，开发能在不同肿瘤特征、成像协议、采集站点和患者人群间实现泛化的自动分割模型仍具挑战性，肿瘤形态与成像分布的差异会大幅降低训练域外的性能。因此，提升基于深度学习的分割模型的鲁棒性与泛化性成为医学图像分析的关键目标。为提高分割鲁棒性，本文提出Multi-Stage Dynamic Prompt nnU-Net，它是nnU-Net的提示条件扩展，在最深层编码器阶段插入三个独立的动态提示模块，每个模块包含可学习的10个256维提示向量库，利用全局池化的编码器特征生成图像特异性提示表示，这些表示被投影为逐特征缩放（γ）与偏移（β）参数，通过逐特征线性调制（FiLM）调制编码器特征图，实现多语义层级的自适应特征条件化。在BraTS GOAT验证数据集上的评估显示，所提Multi-Stage Dynamic Prompt nnU-Net在多数评估指标与肿瘤亚区上优于基线nnU-Net，其平均病灶级Dice分数为：增强肿瘤（ET）76.16%、肿瘤核心（TC）80.04%、全肿瘤（WT）86.42%，而基线模型对应分数为74.38%、78.14%、84.01%。结果表明，多阶段动态提示条件化可提升脑肿瘤分割的准确性与边界描绘效果。
英文摘要
Accurate brain tumor segmentation from magnetic resonance imaging (MRI) is essential for diagnosis, treatment planning, surgical guidance, and disease monitoring. However, developing automated segmentation models that generalize across diverse tumor characteristics, imaging protocols, acquisition sites, and patient populations remains challenging. Variations in tumor morphology and imaging distributions can substantially degrade performance outside the training domain. Consequently, improving the robustness and generalization of deep learning-based segmentation models has become a key objective in medical image analysis. To improve segmentation robustness, we propose Multi-Stage Dynamic Prompt nnU-Net, a prompt-conditioned extension of nnU-Net. Three independent dynamic prompt modules are inserted into the deepest encoder stages. Each module contains a learnable bank of ten 256-dimensional prompt vectors and uses globally pooled encoder features to generate image-specific prompt representations. These representations are projected into feature-wise scaling $(γ)$ and shifting $(β)$ parameters that modulate encoder feature maps through Feature-wise Linear Modulation (FiLM), enabling adaptive feature conditioning at multiple semantic levels. Evaluation on the BraTS GOAT validation dataset demonstrated that the proposed Multi-Stage Dynamic Prompt nnU-Net outperformed the baseline nnU Net across the majority of evaluated metrics and tumor subregions. The proposed model achieved average lesion-wise Dice scores of 76.16% (ET), 80.04% (TC), and 86.42% (WT), compared with 74.38%, 78.14% and 84.01% for the baseline model. The results demonstrate that multi-stage dynamic prompt conditioning improves segmentation accuracy and boundary delineation for brain tumor segmentation.
CommentsAccepted to BraTS MICCAI 2026 Task 3 (Generalizability Across Tumors)
