SU-E-I-90: Fast and Robust Algorithm Towards Vessel Lumen and Stent Strut Detection in Optical Coherence Tomography

Area of Science:

• Medical Imaging
• Cardiovascular Technology
• Biomedical Engineering

Background:

• Optical Coherence Tomography (OCT) provides high-resolution intravascular imaging.
• Accurate segmentation is crucial for quantitative analysis of vascular changes like neointimal hyperplasia (NIH).
• Existing methods may lack automation and robustness for complex intravascular OCT data.

Purpose of the Study:

• To develop and validate a novel automatic segmentation technique for intravascular OCT images.
• To enable precise lumen area extraction and stent strut detection.
• To facilitate quantitative morphological analysis of in-stent neointimal hyperplasia (NIH).

Main Methods:

• A hybrid approach combining Fuzzy C-Means (FCM) clustering and Wavelet Transform (WT) was employed.
• Frequency-domain OCT scans of the human femoral artery were analyzed.
• The method focused on inner luminal contour extraction and stent strut localization.

Main Results:

• The automatic segmentation achieved high accuracy in extracting lumen contours and stent struts.
• An average overlap value of 0.917 ± 0.065 was observed compared to manual segmentation.
• The algorithm successfully identified an average of 6.7 ± 0.5 struts per OCT image.

Conclusions:

• A fast and robust automatic segmentation algorithm using FCM and WT was successfully developed.
• The technique provides accurate lumen border extraction and stent strut detection in intravascular OCT.
• This automated approach is suitable for quantitative morphological analysis of in-stent neointimal hyperplasia.