Microaneurysm Imaging using Multiple En Face Optical Coherence Tomography Angiography Image Averaging: Morphology and Visualization

2019 
Abstract Purpose In diabetic retinopathy (DR), optical coherence tomography angiography (OCTA) could not image all fluorescein angiography (FA)-detected microaneurysms. We investigated if multiple image averaging could enhance its microaneurysm detection capability in DR. Design Prospective- and cross-sectional observational study. Participants Consecutive 31 patients (n = 62 eyes) with DR. Methods All eyes underwent FA, and 3 x 3 mm fovea-centered OCTA images were obtained using two devices (RTVue XR Avanti: Optovue Inc. and OCT HS-100: Canon Inc.). OCTA imaging (HS-100) was performed ten consecutive times. Microaneurysm detection capability was compared among five OCTA images (single image, 3x, 5x, and 10x averaged images, and single scan image with RTVue XR Avanti). Main Outcome Measures Microaneurysm detection capability, and the correlation between microaneurysm clinical characteristics or morphology, and the extent of image averaging required for OCTA detection Results A total of 415 microaneurysms could be analyzed in 31 eyes from 25 cases. Microaneurysms detected on single image, 3x, 5x, 10x averaged OCTA images were 144 (34.7%), 227 (54.7%), 285 (68.7%), 306 (73.7%), respectively. Microaneurysm detection capability was significantly increased with increased image averaging. Microaneurysm detection on OCTA was not correlated with retinal thickness, FA leakiness and indocyanine green angiogram detection, nor the number of averaged images, whereas there was significant correlation between microaneurysm morphology and microaneurysm visibility by image averaging process for four morphologies, particular the focal bulge types (P Conclusions In DR, multiple image averaging is useful for increasing OCTA microaneurysm detection capability, especially for focal bulge type microaneurysms.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    24
    References
    4
    Citations
    NaN
    KQI
    []