Eliminating Registration Bias in Synthetic CT Generation: A Physics-Based Simulation Framework
This addresses the issue of corrupted evaluation metrics in medical imaging for clinicians, though it is incremental as it builds on existing synthetic CT generation methods.
The paper tackled the problem of registration bias in synthetic CT generation from CBCT by proposing a physics-based simulation framework to create geometrically aligned training pairs, resulting in models that achieved superior geometric alignment (e.g., Normalized Mutual Information: 0.31 vs 0.22) and were preferred by clinical observers in 87% of cases.
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.