Reproducibility of coronary Fourier domain optical coherence tomography: quantitative analysis of in vivo stented coronary arteries using three different software packages.

Aims: Fourier domain optical coherence tomography (FD OCT) enables imaging of long coronary artery segments within few seconds, employing high data acquisitions, speed and fast automated catheter pullback. However, the reproducibility of these high-speed pullbacks in the clinical situation is unknown. We tested the reproducibility of in vivo, intracoronary FD OCT and assessed the influence of different computer-assisted algorithms on quantitative analysis. Methods and results: In patients undergoing elective coronary stenting, two repeated FD OCT pullbacks (20 mm/sec), were acquired. Lumen area (LA) and stent area (SA) were measured at 1 mm longitudinal intervals (n=18 pullbacks, n=326 frames). Inter-study variability in terms of absolute difference of mean LA, mean SA and minimum LA was very low (-0.06 +/- 0.28 mm(2), 0.05 +/- 0.29 mm(2) and 0.11 +/- 0.33 mm(2) in software 1) Sources of variability were incomplete visualisation of the vessel circumference, ambiguous luminal borders and drift of internal catheter calibration (Z-offset). Inter-software variability for LA and SA was low (R-2=0.98 similar to 1.00, p<0.01, respectively). Conclusions: FD OCT shows excellent reproducibility for consecutive pullbacks and represents a reliable tool for the in vivo assessment of stented coronaries. Computer-assisted quantitative analysis of FD OCT may be a valuable tool for future studies.