Quantitative evaluation of vessel tracking techniques on coronary angiograms

Area of Science:

• Medical Imaging
• Cardiovascular Research
• Image Analysis

Background:

• Accurate coronary vascular tree analysis requires precise vessel centerline determination.
• Existing methods for automated vessel tracking need improvement in accuracy and precision for clinical applications.

Purpose of the Study:

• To develop and evaluate an automated vessel tracking technique for accurate centerline determination in coronary angiograms.
• To assess the performance of the technique in terms of accuracy, precision, and robustness in clinical images.

Main Methods:

• A modified sector-search approach was employed for automatic vessel tracking after manual point indication.
• Adaptive filtering techniques were used to enhance performance in high-background noise regions.
• Tracking accuracy was validated by comparing automated results with user-indicated centerlines.

Main Results:

• The automated technique successfully tracked 255 out of 256 ( >99%) coronary vessels, including those as small as 0.6 mm in diameter.
• The average distance between tracked and true centerlines was 0.8 pixels (range: 0.4–1.8 pixels).
• High accuracy was achieved with an average of 2-3 manually indicated points per vessel.

Conclusions:

• The developed automated vessel tracking technique provides a reliable method for accurate centerline determination.
• This technique is suitable for supporting both 2D and 3D analyses of the coronary vascular tree.
• The method demonstrates high accuracy and precision, making it valuable for clinical cardiovascular research.