Purchase this article with an account.
Sophie Kubach, Jennifer Yen Luu, Scott Lee, Jesse J Jung, Philip J Rosenfeld, Giovanni Gregori; Importance of Sampling Density in the Analysis of Optical Coherence Tomography Angiography. Invest. Ophthalmol. Vis. Sci. 2016;57(12):4613.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
In OCTA imaging, blood flow is assessed from repeated measurements of the light scattered from the same location over time. The analysis of OCTA images strongly depends on the ability to detect the slowest blood flow and resolve the finest blood vessels. These two factors must be considered when designing the scan patterns and trade-off between scan repetition and sampling density is often required to keep the total scan time below an acceptable duration for patient comfort.
Three Angio scan patterns of different sampling density all covering a 6mm x 6 mm field of view with 2 B-scan repeated at every position are used to assess the image quality and clinical relevance of the Angio scans. Study included two patients with Diabetic Retinopathy (DR), two patients with Branch Retinal Vein Occlusion (BRVO) and two others with Choroidal NeoVascularization (CNV). All were imaged on a 100 KHz Swept-Source OCT prototype instrument (ZEISS, Dublin, CA). The definition of the scan patterns are listed in the table below.Highest resolution scan (Baseline)420 A-scans x 420 B-scansSampling resolution: 14 umNumber of samples per beam: ~ 1.5Medium resolution scan333 A-scans x 333 B-scansSampling resolution: 18 umNumber of samples per beam: ~ 1Low resolution scan250 A-scans x 250 B-scansSampling resolution: 24 umNumber of samples per beam: ~ 0.8
In DR patients, detection of areas of ischemia was found to be fairly insensitive to the scan definition, however capillary dropout and enlargement of the foveal avascular zone were more clearly identified in the baseline (highest resolution) angiography image. In BRVO patients, remodeling of the retinal vasculature was more noticeable in the baseline image, and capillary dropout visible in the baseline image was much less visible in the lower resolution angiography image. In CNV patients, the baseline image revealed more details of the new vascular network with sub-branches that are not resolved in the lower resolution angiography images.
Although detection of areas of ischemia in DR patients was found to be fairly insensitive to the scan definition, sampling density was found to be of importance in the identification of anomalies of the superficial and deep capillary network such as capillary dropout in DR; capillary tortuosity and remodeling in BRVO; and analyzing the abnormal branching network and vascular detail in CNV.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only