Morphology-Based Non-Rigid Registration of Coronary Computed Tomography and Intravascular Images Through Virtual...

Area of Science:

• Cardiovascular Imaging
• Medical Image Analysis
• Biomedical Engineering

Background:

• Coronary computed tomography angiography (CCTA) offers 3D insights into coronary artery disease but lacks vessel wall detail.
• Intravascular imaging provides high-resolution cross-sections but struggles with 3D spatial context.
• Co-registering CCTA and intravascular images is crucial for research but is hindered by distortions and manual effort.

Purpose of the Study:

• To develop a morphology-based framework for accurate rigid and non-rigid co-registration of intravascular images to CCTA.
• To optimize the virtual catheter path within CCTA to match intravascular morphology.
• To reduce manual effort and enable large-scale, multi-modal cardiovascular imaging studies.

Main Methods:

• A morphology-based framework was developed for matching intravascular images to CCTA.
• The method identifies an optimal virtual catheter path in CCTA space to replicate intravascular morphology.
• Validation was performed on a multi-center cohort (40 patients) using bifurcation landmarks for registration accuracy.

Main Results:

• The proposed registration framework significantly outperformed existing methods in aligning coronary artery bifurcations.
• The approach effectively handles non-rigid distortions inherent in intravascular imaging data.
• Demonstrated accurate longitudinal and rotational registration of multi-modal vascular data.

Conclusions:

• The morphology-based framework provides a robust solution for co-registering CCTA and intravascular images.
• This method significantly reduces manual labor in multi-modal cardiovascular imaging research.
• Enables advancements in machine learning applications for vascular image co-registration and analysis.