Abstract 4754: Determining the Optimal Cross-sectional Analysis Interval for OCT Assessment of Coronary Stenting
Introduction: Intracoronary Optical Coherence Tomography (OCT) provides high resolution images and allows detailed stent strut level analysis (SLA). SLA was proposed recently as a surrogate endpoint for DES safety. We aim to define the “optimal” sampling interval for SLA: combining minimal data variability with optimal post-processing time of analysis.
Methods: A total of 14.716 cross-sections from 192 patients and 4 different stents type (Endeavor, Taxus, Cypher and Liberte (BMS)) were studied. A total of 456 stent segments were analyzed. Data sampling every five frames (0.3 mm) for cross-sectional area (0.06mm) for SLA were considered the reference method. Data sampling at increments 5 frames for area measurement and 1 frame for percentage of strut coverage were compared with the reference interval to define variability. Variability was calculated as absolute difference from the reference method at different analysis interval.
Results: A low variability (<10%) was observed for minimal lumen and stent area measurements for all tested interval analyses. At the maximum tested interval (40 frames, 2.4 mm) the variability (%, SD) for minimal lumen and stent area was 9.6±16.2, 5.8±6, respectively. Detection of any uncovered struts showed good agreement (Kappa>0.8) up to every-7-frame SLA sampling (0.42 mm intervals). However, the variability for percentage of uncovered struts was high even at the lowest (every 2 frames) analysis interval (11.7±25) and reached unacceptable 51.9±132 at 15 frames (~1 mm) analysis interval.
Conclusion: Lumen and stent area measurements by OCT showed low variability in all tested analysis intervals. However, percentage of uncovered struts has high variability even at high data sampling. The ideal cross-section data sampling for OCT stent analysis is highly dependent on the variable of interest.