Fast Marching and Runge-Kutta Based Method for Centreline Extraction of Right Coronary Artery in Human Patients

Area of Science:

• Cardiovascular Imaging and Interventions
• Medical Image Analysis
• Computational Anatomy

Background:

• Coronary luminal stenosis assessment via CT angiography (CTA) necessitates accurate centerline extraction.
• Current automated centerline extraction methods lack real-time performance and high accuracy.

Purpose of the Study:

• To develop a hybrid algorithm combining fast marching and Runge-Kutta methods for coronary artery centerline extraction from CTA.
• To evaluate the accuracy of the developed method against Van's method using ground truth centerlines.
• To assess the accuracy of the derived coronary lumen area compared to intravascular ultrasound (IVUS) as the gold standard.

Main Methods:

• A hybrid approach integrating fast marching and Runge-Kutta methods was developed for centerline extraction.
• Method accuracy was evaluated using overlap measures (OV, OF, OT) against a ground truth centerline.
• Coronary lumen area accuracy was compared against IVUS-derived measurements.

Main Results:

• The proposed hybrid method demonstrated superior computational efficiency compared to Van's method.
• The method achieved significantly better overlap measures than Van's method (OV, OF, OT).
• The hybrid approach provided more accurate coronary lumen area measurements than Van's method when compared to IVUS.

Conclusions:

• The hybrid fast marching and Runge-Kutta method offers a computationally efficient and accurate solution for coronary artery centerline extraction from CTA.
• This technique provides precise lumen area measurements, outperforming existing methods.
• The developed approach holds significant potential for real-time clinical application in coronary stenosis assessment.