Variation in cross-correlation as a discriminator for microvessel imaging using clinical intravascular optical coherence tomography systems

Cross-correlation of Intravascular Optical Coherence Tomography (IV-OCT) images is affected by image distortion due to the non-uniform rotational velocity of the imaging catheter. It results in non-representative cross-correlation maps such that for a static scan, the coefficients fluctuate from high to low correlation values. The variation in cross-correlation at flow locations is muted, in comparison to stationary regions. In the present study, the variation of correlation values and its standard deviation (SD) is used to suppress the distortion related noise effects and to extract flow maps from static scan images. The standard deviation of the cross-correlation variation can distinguish flow locations from the surrounding tissue region. The advantage of this technique is its ability to identify slow flow, even Brownian flow, in the presence of motion artifacts. The SD mask used for generating flow maps, is optimized using tissue mimicking phantoms. Finally, the ability of this technique to suppress noise and capture flow maps is demonstrated by imaging microflow through excised porcine coronary artery wall and mucosa membrane imaging.