融合两种肿瘤影像可更全面识别前列腺癌，但现有方法难兼顾两者精度。
When Two Tracers Disagree: An Investigation of Multimodal Fusion for Clinical PET/CT Segmentation
- 用双模态注意力网络融合PSMA与FDG影像，保留各自特征
- 融合后PSMA分割效果尚可（Dice 0.76），但FDG性能下降至0.57
- 单独使用单模态模型仍优于融合方案，提示需更好融合架构
PSMA与FDG PET/CT可互补显示前列腺癌的生物学信息。结合两者或能捕捉单一示踪剂遗漏的异质性肿瘤表型，但目前尚无公认的深度学习融合架构。本研究基于公开的DEEP-PSMA Challenge数据集，训练了针对不同示踪剂的3D nnU-Net基线模型，比较了（i）早期融合：单编码器单解码器（OEOD）或双解码器（OETD），以及（ii）中间融合：双编码器交叉注意力U-Net（DECA-UNet）。单模态基线表现优异（PSMA Dice = 0.93；FDG = 0.81）。融合策略结果参差：OEOD在较简单的非特异性任务中达联合Dice 0.90，而双模态融合模型在PSMA/FDG上分别达到0.69/0.64（OETD）和0.76/0.57（DECA-UNet）。尽管融合对PSMA分割有合理表现，但FDG性能普遍下降，且无一策略持续超越单模态基线。在当前设置下，模态特异性模型仍是更强基线；临床有效的多模态融合需能更好保持模态特异性表示的架构。代码已开源：https://github.com/JackJ3636/DEEP_PSMA_code
原文摘要 · Abstract (English)
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
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。