当错位成为监督：有监督合成CT生成中的结构化标签噪声
When Misalignment Becomes Supervision: Structured Label Noise in Supervised Synthetic CT Generation
浏览论文内容
中文总结 AI 辅助
本研究揭示配准残差作为结构化标签噪声影响有监督合成CT生成，提出基于Segment Anything编码器的感知损失以提升分割性能和结构连贯性，并主张补充解剖导向评估。
中文摘要 AI 辅助
有监督合成CT（sCT）生成通常作为体素级回归任务，与配准后的参考CT图像进行训练和评估。在实践中，MRI-CT和CBCT-CT图像对通过配准流程对齐，但会残留错位。这些残差并非独立的强度噪声，而是空间上连贯的几何差异，表现为结构化标签噪声。我们研究了这种由配准引入的偏差如何影响有监督的MRI-to-CT和CBCT-to-CT合成，涉及1,784例配对患者，覆盖五个解剖区域。体素级分数强烈依赖于构建训练目标所用的配准与评估所用配准之间的一致性：当两种约定匹配时模型得分最佳，这表明网络部分学习了配准流程的几何约定，且标准指标对此给予奖励。在解剖上更一致的配准上训练可减少预测变异性并改善分布外鲁棒性，而仅使用CT的对照实验表明，仅配准本身产生的指标误差就处于顶级挑战提交结果的范围内。为缓解体素级监督的局限性，我们引入了一种在预训练Segment Anything编码器特征空间中计算的感知损失。与仅MAE和基于VGG的目标相比，它改善了下游分割，并生成更锐利、结构更连贯的sCT。在不完美对齐下，感知指标与体素级指标存在分歧，而在评估几何可靠时两者一致。这些结果将配准引入的偏差确定为有监督sCT生成中的核心混杂因素，并主张以解剖导向的评估标准补充体素级一致性。
英文摘要
Supervised synthetic CT (sCT) generation is commonly trained and evaluated as voxel-wise regression against registered reference CT images. In practice, MRI-CT and CBCT-CT pairs are aligned through registration procedures that leave residual misalignments. These residuals are not independent intensity noise but spatially coherent geometric discrepancies that act as structured label noise. We investigate how this registration-induced bias affects supervised MRI-to-CT and CBCT-to-CT synthesis on 1,784 paired patients covering five anatomical regions. Voxel-wise scores strongly depend on the consistency between the registration used to build the training targets and the one used for evaluation: models score best when both conventions match, showing that networks partly learn the geometric convention of the registration pipeline and that standard metrics reward it. Training on more anatomically consistent registrations reduces prediction variability and improves out-of-distribution robustness, and CT-only controls show that registration alone produces metric errors in the range of top challenge submissions. To mitigate the limits of voxel-wise supervision, we introduce a perceptual loss computed in the feature space of a pretrained Segment Anything encoder. Compared with MAE-only and VGG-based objectives, it improves downstream segmentation and yields sharper, more structurally coherent sCT. Perceptual and voxel-wise metrics disagree under imperfect alignment and agree when the evaluation geometry is reliable. These results identify registration-induced bias as a central confounder in supervised sCT generation and argue for complementing voxel-wise agreement with anatomy-oriented evaluation criteria.
发表机构
- Univ. Rennes（雷恩大学）
- CLCC Eugene Marquis（欧仁·马奎斯癌症中心）
- INSERM（法国国家健康与医学研究院）
- LTSI - UMR 1099（信号与图像处理实验室 - UMR 1099）
机构由 AI 辅助整理，请以论文原文为准。
