A multiscale dynamic programming procedure for boundary detection in ultrasonic artery images

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Ultrasound Technology

Background:

• Conventional ultrasonic measurements of artery walls rely on manual tracing, which is prone to interobserver variability and inefficiency.
• Accurate assessment of carotid and femoral artery walls is crucial for diagnosing and monitoring vascular diseases.

Purpose of the Study:

• To develop and evaluate a novel automated method for ultrasonic measurement of human artery walls.
• To reduce interobserver variability and analysis time associated with manual tracing techniques.

Main Methods:

• Application of a multiscale dynamic programming (DP) algorithm for estimating vessel wall positions.
• Utilizing a cost function incorporating image features and geometrical characteristics, with weights adjusted via expert tracings.
• Integrating operator interventions within a global optimality framework.

Main Results:

• Significant decrease in interobserver variability compared to manual methods.
• Substantial reduction in overall analysis time for ultrasonic artery wall measurements.
• Demonstrated robustness and improved performance in a clinical environment.

Conclusions:

• The automated DP-based method effectively replaces manual tracing for ultrasonic artery wall measurements.
• This approach leads to improved accuracy, reduced variability, and increased efficiency in clinical practice.
• The integration of human knowledge enhances the algorithm's robustness and reliability.