Double-Decomposition Motion Tracking of Intraoperative 3D Structures via Cross-Spatio-Temporal Semantics Alignment

Area of Science:

• Medical Imaging
• Surgical Navigation
• Computational Anatomy

Background:

• 3D motion tracking in X-ray image-guided operations is crucial but challenged by spatio-temporal misalignment and respiratory phase differences between 3D priors and 2D X-ray images.
• Existing methods struggle to accurately estimate 3D dynamic organ structures due to these inherent limitations.

Purpose of the Study:

• To propose a novel Double-Decomposition Tracking (DD-Track) framework to overcome challenges in 3D motion tracking from X-ray images.
• To accurately estimate 3D dynamic organ structures by addressing spatio-temporal misalignment and respiratory phase differences.

Main Methods:

• DD-Track employs a dual-alignment strategy: temporal alignment compensates for respiratory phase differences using a dual-excitation mechanism, and spatial alignment integrates 2D motion features for accurate 3D prior warping.
• Motion tracking is decomposed into common trajectory and organ-specific deformation to better represent multi-organ motion and avoid excessive organ stretching.

Main Results:

• DD-Track demonstrated superior performance over state-of-the-art methods in comprehensive quantitative and qualitative experiments on simulated and clinical multi-organ datasets.
• The framework showed generalization capabilities for tracking intra-organ lesions on simulated data.

Conclusions:

• The proposed DD-Track framework effectively addresses spatio-temporal misalignment and respiratory phase differences for accurate 3D motion tracking in X-ray image-guided operations.
• DD-Track offers a robust solution for multi-organ motion estimation and shows potential for tracking intra-organ lesions, enhancing surgical guidance.