Stent detection with very thick tissue coverage in intravascular OCT

Area of Science:

• Cardiovascular Imaging
• Medical Artificial Intelligence
• Interventional Cardiology

Background:

• Coronary stenting (PCI) is crucial for coronary artery disease but complications arise from improper deployment.
• Intravascular optical coherence tomography (OCT) offers high-resolution imaging for assessing stent placement and tissue coverage.
• Manual analysis of OCT images is time-consuming and labor-intensive, limiting clinical efficiency.

Purpose of the Study:

• To develop and validate a deep learning method for automated stent analysis using intravascular OCT.
• To enable accurate analysis of stents with both thin and very thick tissue coverage.
• To create an algorithm for precise stent area measurement in cases of single and multiple stent implantation.

Main Methods:

• A deep learning model was trained on 25,203 images from 56 OCT pullbacks across 41 patients.
• The method was evaluated for analyzing stents with thin (≤0.3mm) and thick (>0.3mm) tissue coverage.
• An algorithm was developed to calculate stent area in vessels with single and multiple stents.

Main Results:

• High precision (0.932±0.009) and sensitivity (0.939±0.007) were achieved for stents with thin tissue coverage.
• Good precision (0.856±0.019) and sensitivity (0.874±0.011) were obtained for stents with thick tissue coverage.
• Strong correlations were found between automated and manual stent area measurements (r=0.954 for single, r=0.918 for multiple stents).

Conclusions:

• The developed deep learning method accurately detects stent struts even with significant tissue coverage.
• The algorithm effectively analyzes stent area in complex cases with multiple stents.
• This automated approach can significantly aid in evaluating stent implantation and post-procedure tissue coverage.