发表机构
University of Oxford; Nuffield Department of Medicine; Department of Oncology; Big Data Institute; Nuffield Department of Population Health（牛津大学; 纳菲尔德医学系; 肿瘤学系; 大数据研究所; 纳菲尔德人口健康系）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对前列腺癌PSMA与FDG PET/CT的多模态融合分割，训练示踪剂特异性3D nnU-Net基线并对比多种融合策略，发现融合未始终优于单示踪剂模型，需更优架构保留示踪剂特异性表征
AI 中文摘要
PSMA与FDG PET/CT可可视化前列腺癌中互补的生物学信息，结合两种示踪剂或能捕获单一示踪剂可能遗漏的异质性肿瘤表型，但目前尚无关于融合这些模态的有效深度学习架构的共识。我们评估了用于自动全身PET/CT病灶分割以估算总肿瘤负荷的多模态图像融合策略，使用公开的DEEP-PSMA挑战赛数据集，我们训练了示踪剂特异性的3D nnU-Net基线模型，并比较了以下方法：（i）早期融合，采用单编码器单解码器（OEOD）或双解码器（OETD）；（ii）中间融合，采用双编码器交叉注意力U-Net（DECA-UNet）。示踪剂特异性基线模型表现出色（PSMA Dice=0.93；FDG=0.81）。融合结果喜忧参半：OEOD在更简单的非示踪剂特异性任务上的组合Dice为0.90，而示踪剂特异性融合模型的PSMA/FDG表现为0.69/0.64（OETD）和0.76/0.57（DECA-UNet）。尽管融合通常能提供合理的PSMA分割结果，但FDG性能出现下降，且没有任何策略能始终优于单示踪剂基线模型。在本评估设定下，示踪剂特异性模型仍是更优的基线；多模态融合的临床实用增益可能需要能更好保留示踪剂特异性表征的架构。我们的代码可在此处获取：this https URL
英文摘要
PSMA and FDG PET/CT visualise complementary biological information in prostate cancer. Combining both tracers could capture heterogeneous tumour phenotypes that may be missed by either alone, yet there is no consensus on effective deep learning architectures for fusing these modalities. We evaluated multimodal image-fusion strategies for automatic whole-body PET/CT lesion segmentation to estimate total tumour burden. Using the public DEEP-PSMA Challenge dataset, we trained tracer-specific 3D nnU-Net baselines and compared (i) early fusion with a single encoder and one decoder (OEOD) or two decoders (OETD), and (ii) intermediate fusion via a dual-encoder cross-attention U-Net (DECA-UNet). Tracer-specific baselines performed strongly (PSMA Dice = 0.93; FDG = 0.81). Fusion yielded mixed results: OEOD produced a combined Dice of 0.90 (on an easier, non-tracer-specific task), whilst the tracer-specific fusion models reached PSMA/FDG = 0.69/0.64 (OETD) and 0.76/0.57 (DECA-UNet). Whilst fusion often provided reasonable PSMA segmentation, FDG performance degraded and no strategy consistently exceeded the single-tracer baselines. Under the evaluated setting, tracer-specific models remain the stronger baseline; clinically useful gains from multimodal fusion will likely require architectures that better preserve tracer specific representations. Our code is available at: https://github.com/JackJ3636/DEEP_PSMA_code
Comments10 pages (8 pages main content and 2 pages of references), 2 figures, 2 tables, accepted to MICCAI 2026 Cancer Prevention, Detection, and IntervenTion (CaPTion) Workshop
