Quantitative multi-modality imaging analysis of a fully bioresorbable stent: a head-to-head comparison between QCA, IVUS and OCT

The bioresorbable vascular stent (BVS) is totally translucent and radiolucent, leading to challenges when using conventional invasive imaging modalities. Agreement between quantitative coronary angiography (QCA), intravascular ultrasound (IVUS) and optical coherence tomography (OCT) in the BVS is unknown. Forty five patients enrolled in the ABSORB cohort B1 study underwent coronary angiography, IVUS and OCT immediately post BVS implantation, and at 6 months. OCT estimated stent length accurately compared to nominal length (95% CI of the difference: -0.19; 0.37 and -0.15; 0.47 mm(2) for baseline and 6 months, respectively), whereas QCA incurred consistent underestimation of the same magnitude at both time points (Pearson correlation = 0.806). IVUS yielded low accuracy (95% CI of the difference: 0.77; 3.74 and -1.15; 3.27 mm(2) for baseline and 6 months, respectively), with several outliers and random variability test-retest. Minimal lumen area (MLA) decreased substantially between baseline and 6 months on QCA and OCT and only minimally on IVUS (95% CI: 0.11; 0.42). Agreement between the different imaging modalities is poor: worst agreement Videodensitometry-IVUS post-implantation (ICCa 0.289); best agreement IVUS-OCT at baseline (ICCa 0.767). All pairs deviated significantly from linearity (P < 0.01). Passing-Bablok non-parametric orthogonal regression showed constant and proportional bias between IVUS and OCT. OCT is the most accurate technique for measuring stent length, whilst QCA incurs systematic underestimation (foreshortening) and solid state IVUS incurs random error. Volumetric calculations using solid state IVUS are therefore not reliable. There is poor agreement for MLA estimation between all the imaging modalities studied, including IVUS-OCT, hence their values are not interchangeable.