Automatic Lumen Segmentation in Intravascular Optical Coherence Tomography Images Using Level Set.

Area of Science:

• Medical Imaging
• Cardiovascular Technology
• Image Processing

Background:

• Accurate lumen segmentation in intravascular optical coherence tomography (IVOT) is crucial for diagnosing and treating coronary artery disease.
• Challenges include irregular lumen shapes due to plaque and bifurcations, guide wire shadows, and blood artifacts.
• Existing methods struggle with these complex imaging artifacts.

Purpose of the Study:

• To develop a novel automatic level set-based segmentation algorithm for IVOT images.
• To enhance robustness against irregular lumen shapes, guide wire shadows, and blood artifacts.
• To improve the accuracy and reliability of lumen segmentation in IVOT analysis.

Main Methods:

• A novel automatic level set-based segmentation algorithm is proposed.
• A narrow image smoothing filter is employed to reduce artifacts and prevent level set leakage.
• A divide-and-conquer strategy is utilized to effectively handle guide wire shadows.

Main Results:

• The proposed method significantly reduces the impact of irregular lumens, guide wire shadows, and blood artifacts.
• Experimental evaluation on 880 images from 5 patients demonstrated high accuracy.
• An average Dice Similarity Coefficient (DSC) of 98.1% ± 1.1% was achieved, indicating robust performance.

Conclusions:

• The developed automatic level set algorithm offers a competent solution for lumen segmentation in challenging IVOT images.
• The method demonstrates robustness and accuracy, outperforming existing approaches in handling common artifacts.
• This advancement facilitates more reliable diagnosis and treatment planning for coronary artery disease.