Automated 3D coronary sinus catheter detection using a scanning-beam digital x-ray system

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Interventional Cardiology

Background:

• Scanning-beam digital x-ray (SBDX) enables 3D catheter tracking but requires accurate detection for motion compensation.
• Respiratory motion significantly impacts the precision of 3D catheter localization during electrophysiology procedures.

Purpose of the Study:

• To develop and validate an automated coronary sinus (CS) catheter detection algorithm for SBDX systems.
• To improve respiratory motion-compensated 3D catheter tracking accuracy.

Main Methods:

• Developed a 3D shape-constrained search algorithm utilizing SBDX localization and prior catheter shape knowledge.
• Implemented a cost function minimization to identify the most probable CS catheter candidate.
• Tested the algorithm offline on phantom data simulating various challenging detection scenarios and motion.

Main Results:

• Achieved a 98.1% true positive rate and 100% true negative rate for CS catheter detection across 2755 frames.
• Demonstrated significant reduction in Root Mean Square (RMS) error for ablation catheter tracking (1.41 mm with compensation vs. 10.35 mm without).
• Average processing time was 12.7 ms/frame, indicating computational efficiency.

Conclusions:

• A computationally efficient automated 3D CS catheter detection method for SBDX has been successfully developed.
• This algorithm facilitates accurate motion-compensated 3D catheter tracking.
• The technique aids in the registration of 3D cardiac models to tracked catheters, enhancing interventional procedures.