Purchase this article with an account.
Alexander E Salmon, Rex Chen, Farid Atry, Mina Gaffney, Dana K Merriman, Daniel A Gil, Melissa Skala, Eric Buckland, Ramin Pashaie, Joseph Carroll; Assessment of a high-speed SD-OCT-A in a natural model of hypoperfusion: the hibernating 13-lined ground squirrel. Invest. Ophthalmol. Vis. Sci. 2020;61(7):199.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Hibernation offers a natural model of retinal hypoperfusion, a common manifestation of retinal or systemic disease in humans. Here we developed a high-speed spectral-domain OCT-A capable of detecting the wide range of blood flow velocities observed during torpor and euthermia in the 13-lined ground squirrel (13-LGS). We compared these images to those from a commercial OCT and a custom adaptive optics scanning light ophthalmoscope (AOSLO).
The custom OCT-A system includes a commercial spectrometer with a 1.5mm -6dB fall-off and a CMOS camera with a 130kHz maximum line-rate. An 850±80nm SLD was selected to obtain an axial resolution of 1.5µm to ensure adequate visualization of the ~3.5µm wide retinal capillaries. Three torpid and three winter-active 13-LGS were imaged with the custom OCT-A and a Bioptigen Envisu OCT (36kHz), then the torpid animals were warmed to a body temperature ≥ 26°C and re-imaged. Intraframe variance was used to generate vessel contrast in both devices. Vessel density was measured at the superficial, middle, and deep capillary plexuses (SCP, MCP, DCP, respectively) and compared between groups and devices. Videos of capillary blood flow were collected using an AOSLO to inform scan parameters for future studies.
In the winter-active animals, vessel density was not significantly different between the custom and commercial devices for any layer (p: 0.50, 0.50, 0.50; SCP, MCP, DCP, respectively; n-way ANOVA with Tukey post-hoc, used henceforth). Detectable vessel density was significantly lower in torpor before forced arousal (p: 3.09E-4); however, the density values were not significantly different between devices (p: 0.22) and there was not a significant interaction between thermic state and device (p: 0.20). Extremely slow blood flow (as low as ~0.2Hz with a 7% duty cycle in one capillary) was observed with AOSLO during torpor, which may account for the lack of detection with the relatively short scanning intervals used in both OCT systems.
The custom OCT-A, with substantially reduced scan time, yields similar density values to the commercial device. Further optimization of the custom OCT-A is needed to detect the extremely slow flow observed with AOSLO. The 13-LGS offers a unique opportunity to study a naturally reversible model of hypoperfusion and provides a valuable model for further OCT-A validation studies.
This is a 2020 ARVO Annual Meeting abstract.
This PDF is available to Subscribers Only