IVUS Longitudinal and Axial Registration for Atherosclerosis Progression Evaluation

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computational Anatomy

Background:

• Intravascular ultrasound (IVUS) provides detailed cross-sectional vessel imaging.
• Accurate registration of longitudinal IVUS pullbacks is crucial for tracking disease progression and treatment efficacy.
• Existing methods often require complex feature extraction, limiting real-time application.

Purpose of the Study:

• To present a novel, automated two-stage registration framework for temporal IVUS pullback alignment.
• To enable accurate assessment of pathophysiological changes and treatment response using sequential IVUS data.
• To develop a computationally efficient method suitable for real-time clinical applications.

Main Methods:

• A two-stage registration framework combining Dynamic Time Warping (DTW) for temporal alignment and an intensity-based method using Harmony Search for axial registration.
• The intensity-based method maximizes Mutual Information for precise pullback matching.
• The framework is fully automated, requiring only two global image-based measurements.

Main Results:

• Synthetic data: Achieved alignment error of 0.1853 frames, rotation error of 0.93°, and translation error of 0.0161 mm.
• Real clinical data: Longitudinal registration error of 4.3±3.9 frames; Axial registration distance error of 0.56±0.323 mm and rotational error of 12.4°±10.5°.
• The method demonstrates competitive or superior performance in axial registration compared to state-of-the-art, with computationally lighter steps.

Conclusions:

• The proposed automated two-stage registration framework effectively aligns longitudinal and axial IVUS pullbacks.
• The method's computational efficiency makes it suitable for real-time clinical decision support.
• This approach can aid clinicians in diagnosing and managing vascular diseases based on sequential IVUS examinations.