Eliminating Registration Bias in Synthetic CT Generation: A Physics-Based Simulation Framework

TL;DR

This paper introduces a physics-based simulation framework for generating synthetic CT images that eliminates registration bias, improving geometric fidelity and clinical relevance over traditional intensity-based metrics.

Contribution

The authors propose a novel physics-based CBCT simulation method that produces geometrically aligned training data, reducing registration bias in synthetic CT generation.

Findings

Synthetic data improved geometric alignment metrics (NMI=0.31 vs 0.22).

Clinical observers preferred synthetic outputs in 87% of cases.

Normalized Mutual Information predicted observer preference reliably.

Abstract

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.