用物理仿真生成配准对，解决合成CT中的配准偏差问题
Eliminating Registration Bias in Synthetic CT Generation: A Physics-Based Simulation Framework
- 通过物理模拟生成几何对齐的训练数据，避免真实扫描配准误差
- 模型在几何一致性上显著优于传统方法（NMI 0.31 vs 0.22）
- 临床医生更偏好几何保真度高的结果，而非强度匹配的伪真值
从CBCT生成合成CT通常依赖已配准的训练样本，但独立获取的扫描间难以实现完美配准，导致训练模型继承配准偏差，并污染标准评估指标。这可能使更高基准性能反映的是对配准伪影的复制，而非解剖真实性。本文提出基于物理的CBCT仿真框架，通过构造几何对齐的训练对，结合以输入CBCT为参考的几何对齐度量进行评估。在两个独立骨盆数据集上，使用合成数据训练的模型在几何对齐性上表现更优（归一化互信息：0.31 对比 0.22），尽管其常规强度指标较低。强度指标在形变配准数据中与临床评估呈现反向相关，而归一化互信息在不同配准方法下始终能预测观察者偏好（rho = 0.31, p < 0.001）。临床观察者在87%情况下更偏好合成训练生成的结果，表明几何保真度才是临床需求的关键。
原文摘要 · Abstract (English)
Supervised synthetic CT generation from CBCT requires registered training pairs, yet perfect registration between separately acquired scans remains unattainable. This registration bias propagates into trained models and corrupts standard evaluation metrics. This may suggest that superior benchmark performance indicates better reproduction of registration artifacts rather than anatomical fidelity. We propose physics-based CBCT simulation to provide geometrically aligned training pairs by construction, combined with evaluation using geometric alignment metrics against input CBCT rather than biased ground truth. On two independent pelvic datasets, models trained on synthetic data achieved superior geometric alignment (Normalized Mutual Information: 0.31 vs 0.22) despite lower conventional intensity scores. Intensity metrics showed inverted correlations with clinical assessment for deformably registered data, while Normalized Mutual Information consistently predicted observer preference across registration methodologies (rho = 0.31, p < 0.001). Clinical observers preferred synthetic-trained outputs in 87% of cases, demonstrating that geometric fidelity, not intensity agreement with biased ground truth, aligns with clinical requirements.
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。