Patient-Agnostic Synthetic Pretraining for Efficient Patient-Specific Intraoperative 2D/3D Registration

2026-07-25

AI Summary

This work addresses the high computational cost and low deployment efficiency of conventional patient-specific 2D/3D registration, which typically requires training a new model from scratch for each patient. The authors propose an efficient registration framework that first performs patient-agnostic pretraining on digitally reconstructed radiographs (DRRs) synthesized from multiple CT scans, employing a segmentation-free domain randomization strategy that encompasses variations in intensity, projection physics, field of view, occlusion, and fluoroscopic artifacts without requiring anatomical annotations. Subsequently, by integrating spherical similarity learning with differentiable Levenberg-Marquardt optimization, the method enables rapid adaptation to new patients using only a few target CT projections. Experiments demonstrate that the approach significantly reduces patient-specific training requirements across multiple anatomical datasets while maintaining high-accuracy intraoperative registration performance at minimal adaptation cost.

Abstract

Intraoperative 2D/3D registration aligns preoperative CT volumes with intraoperative X-ray or fluoroscopic images and is essential for image-guided interventions. Recent learning-based and differentiable registration methods have shown promising accuracy, especially in patient-specific settings where abundant digitally reconstructed radiographs (DRRs) can be synthesized from the target CT. However, training a separate patient-specific model from scratch for every new patient is computationally inefficient and limits practical deployment. In this work, we propose an efficient patient-specific 2D/3D registration framework based on patient-agnostic synthetic pretraining and spherical similarity learning. The model is first pretrained on synthetic DRRs generated from multiple CT volumes to learn transferable pose-sensitive representations, and is then adapted to a new patient using only a limited number of synthetic projections from the target CT. To improve synthetic-to-real robustness without requiring anatomical labels, we introduce a segmentation-free domain randomization strategy that perturbs image intensity, projection physics, field-of-view, occlusion, and fluoroscopic artifacts. The adapted model provides an initial pose estimate, which is further refined using spherical similarity learning and differentiable Levenberg-Marquardt optimization. Experiments on multiple anatomical datasets evaluate whether patient-agnostic synthetic pretraining can improve the efficiency of patient-specific registration, with particular focus on the trade-off between adaptation cost and registration accuracy. The results demonstrate that patient-agnostic synthetic pretraining can significantly reduce patient-specific training requirements while preserving accurate intraoperative 2D/3D registration.

Problem

• intraoperative 2D/3D registration
• patient-specific
• computational inefficiency
• model training
• image-guided interventions

Innovation

• patient-agnostic pretraining
• synthetic-to-real domain randomization
• spherical similarity learning
• differentiable 2D/3D registration
• efficient patient-specific adaptation