Intra‐session repeatability of quantitative metrics using widefield optical coherence tomography angiography (OCTA) in elderly subjects

• Published in: Acta ophthalmologica (Oxford, England) Vol. 98; no. 5; pp. e570 - e578
• Main Authors: Hong, Jimmy, Tan, Bingyao, Quang, Nguyen Duc, Gupta, Preeti, Lin, Emily, Wong, Damon, Ang, Marcus, Lamoureux, Ecosse, Schmetterer, Leopold, Chua, Jacqueline
• Format: Journal Article
• Language: English
• Published: England Wiley Subscription Services, Inc 01.08.2020; John Wiley and Sons Inc
• Subjects: Angiography; Bayesian analysis; Datasets; Filters; macula; Medical imaging; optical coherence tomography angiography; Original; Perfusion; perfusion density; repeatability; Tomography; vessel density; repeatability; filters; optical coherence tomography angiography; perfusion density; macula; vessel density
• ISSN: 1755-375X; 1755-3768; 1755-3768
• DOI: 10.1111/aos.14327

Purpose To assess the repeatability of retinal vascular metrics using different postprocessing methods as obtained from the swept‐source optical coherence tomography angiography (SS‐OCTA). Methods Thirty‐two participants (63% males; mean [SD] age, 70 [7] years) underwent SS‐OCTA imaging (PLEX® Elite 9000, Carl Zeiss Meditec, Inc., Dublin, USA). Each participant underwent 2 repeated scans of 2 scan protocols: a macular‐centred 3 × 3‐mm2 and a widefield 12 × 12‐mm2 for a total of 4 acquisitions. Images of superficial vascular plexuses (SVP) and deep vascular plexuses (DVP) were processed using different filters to generate the perfusion density (PD) and vessel density (VD). Vessel enhancement filters ranged from vessel targeted (Hessian and Gabor filters), classical denoising (Gaussian filter), to a scale‐selective adaption (modified Bayesian residual transform [MBRT]). Intra‐session repeatability of the different filters and their correlation with the original data set were calculated with the intraclass correlation coefficient (ICC) and Pearson's r. Results Of the 32 eyes, 17 and 15 were right and left eyes, respectively. For 3 × 3‐mm2 scans, both MBRT and Gabor filters yielded very good repeatable PD and VD (both ICCs > 0.87) values. Gabor filter was the most correlated with the original data set for the OCTA metrics (r = 0.95–0.97). For 12 × 12‐mm2 scans, MBRT filter produced good‐to‐moderate ICC values for SVP (ICC>0.89) and DVP (ICC>0.73) metrics. Both the MBRT and Gabor filters were highly correlated with the original 12 × 12‐mm2 scan data set (r = 0.96–0.98). The ICCs for the agreement between 3 × 3‐mm2 and cropped 12 × 12‐mm2 were high only for the PD values at the SVP layer and were poor for the VD at SVP and DVP measurements (ICC < 0.50). Conclusion Our findings show that with the proper choice of postimaging processing methods, SS‐OCTA metrics can be obtained with high repeatability, which supports its use in various clinical settings.