MSR:Hybrid Field Modeling for CT-MRI Rigid-Deformable Registration of the Cervical Spine with an Annotated Dataset

2026-04-30

AI Summary

Cervical spine CT–MRI registration faces significant challenges due to anatomical complexity, high inter-subject variability, and a scarcity of high-quality, multimodal annotated data, particularly lacking an effective hybrid rigid–deformable modeling paradigm. This work proposes the first MSR framework, which integrates rigid alignment of individual vertebrae with a deformable module that synergistically combines Mamba for global modeling and Swin Transformer for local modeling. An adaptive gating mechanism is introduced to enable collaborative optimization of rigid and non-rigid deformation fields. Additionally, the study releases R-D-Reg, the first fully annotated cervical CT–MRI dataset. The proposed method substantially improves registration accuracy and anatomical consistency, with both code and dataset made publicly available to advance the field.

Abstract

Accurate CT-MRI registration of the cervical spine is essential for preoperative planning because this region is anatomically complex,highly variable,and vulnerable to injury of the vertebral arteries and spinal cord. However,cervical CT-MRI registration remains underexplored,particularly for rigid-deformable hybrid modeling,and the lack of high-quality annotated multimodal data further limits progress. To address these challenges, we construct and release a comprehensively annotated CT-MRI dataset, R-D-Reg, and propose MSR, a rigid-deformable hybrid registration framework for complex joint structures. Specifically, MSR includes a rigid registration module for independent local rigid alignment of individual vertebrae and a deformable registration module with an MSL block that combines Mamba-based global modeling and Swin Transformer-based local modeling through adaptive gating. The rigid and deformable deformation fields are then fused to generate a hybrid field that better preserves local anatomical consistency. The code and dataset are publicly available at https://github.com/ssc1230609-spec/MSR-registration.

Problem

• CT-MRI registration
• cervical spine
• rigid-deformable hybrid modeling
• annotated dataset
• multimodal image registration

Innovation

• rigid-deformable registration
• Mamba
• Swin Transformer
• hybrid field modeling
• cervical spine