Purchase this article with an account.
Gerald Seidel, Gerold Aschinger, Christoph Singer, Anton Haas, Martin Weger, Sereina Annik Herzog, Gerhard Garhofer, Leopold Schmetterer; Blood flow velocity evaluation with conventional optical coherence tomography. Invest. Ophthalmol. Vis. Sci. 2016;57(12):5917.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
While conventional spectral domain optical coherence tomography (OCT) is widely conceived as a purely morphological imaging technique, indirect information on blood flow can be obtained by evaluating the orientation of erythrocytes in the lumen of retinal vessels. On OCT this translates into a sandglass shaped intraluminal pattern. The faster the blood flows the more pronounced the contrast of this pattern is. We evaluated the potential of grading retinal blood flow by assessment of this intraluminal reflectance pattern of retinal vessels on OCT.
We took high-resolution OCT cross-sections of the inferior and superior temporal vascular arcade. By assessing the resulting intraluminal backscattering profile 3 independent graders estimated the blood flow velocity. Within 30 minutes after the OCT all eyes underwent a video fluorescein angiography (FLA) and the fluorescein traveling velocity was determined to serve as an estimate for the true blood flow velocity, which was then categorized into “stasis” (< 1500µm/sec), “reduced flow” (<3000µm/sec), and “normal flow” (>3000µm/sec).
Thirty eyes from 30 patients (9 central retinal artery occlusions, 8 branch retinal artery occlusions, and 13 controls) were included, resulting in 60 measured vessels. Two vessels were excluded due to poor OCT imaging quality (both were controls).In the remaining 58 vessels the inter-rater agreement between the categories on FLA and the human graders as measured by Cohen’s kappa (κ) was substantial in one grader (κ 0.77, 95% CI 0.6 to 0.93), and almost perfect in the other two graders (κ 0.64, 95% CI 0.44 to 0.83 and κ 0.87, 95% CI 0.74 to 0.99). This was statistically significant in all cases (p<0.001). A blood flow velocity under 1500µm/sec (15 vessels, 45 ratings) was correctly identified in all but 4 ratings (once as “normal flow”, trice as “reduced flow”). A normal flow (36 vessels, 108 ratings) was only once graded as stasis. The sensitivity to detect a healthy vessel ranged from 89.7% to 100% depending on the grader.
The retinal blood flow can be estimated by evaluating the intraluminal backscattering pattern on OCT. While it is more difficult to identify only slightly reduced flow, the discrimination between very little flow and relatively normal flow is robust.
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