Comparison between different approaches to evaluate fibrous cap thickness in sequential OCT studies.

BACKGROUND In this in vivo human study we tested the reproducibility for optical coherence tomography (OCT) assessment of lumen area (LA) and plaque components measurements, such as lipid arc extension and fibrous cap thickness (FCt). METHODS We tested the variability of LA, lipid arc and FCt assessments in two repeated OCT pullbacks from the same diseased coronary segment matched using fiduciary anatomical landmarks. In particular, for the reliability of minimal FCt measurement we compared four different approaches based on continuous (longitudinal) or segmental (spot) individuation of smaller thickness: 1) comparison of single minimal FCt individuated alongside all plaque extension in the two pullbacks (Longitudinal (L)-spot minimal FCt value); 2) comparison of the mean FCt values of the plaque in the two pullbacks (L-plot mean FCt value); 3) comparison between the single minimal FCt value obtained in the first pullback and the single FCt obtained in the matched CS of second pullback (L- spot CS matched FCt value); 4) comparison of measurements obtained by visual selection of CS with minimal FCt s in the two pullbacks (Single-spot minimal FCt value). RESULTS From the paired analyses of 20 non culprit lesions (accounting for a total of 387 matched CS), we found a suboptimal in-segment correlation for LA (Intra-Class Coefficient [ICC] 0.731), but a good in-segment correlation for lipid arc (ICC 0.963). Regarding FCt measurement, a high reproducibility was obtained applying continuous assessment; in particular, the best correlation was observed with L-spot minimal FCt value and the L-plot mean FCt (ICC 0.893 and 0.952, respectively) with small inter-pullback differences (confidence intervals less than 0.04 mm). CONCLUSIONS In this methodological study we observed a good reproducibility for quantitative plaque measurements with OCT confirming its reliability for serial assessment. In particular, longitudinal measurement in multiple adjacent frames seems to be the more accurate and reproducible approach for sequential FCt assessment.