Purchase this article with an account.
Abhishek Rege, Yusi Liu, Yici Jing, Jonathan Howarth, M. Jason Brooke, Osamah Saeedi; Non-invasive imaging of retinal blood flow with high spatio-temporal resolution. Invest. Ophthalmol. Vis. Sci. 2017;58(8):4851.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Noninvasive monitoring of vascular blood flow and tissue perfusion in the retina may offer new insights on ophthalmologic, neurologic and cardiovascular pathophysiology. Here, we demonstrate the ability of a novel speckle-based retinal imaging investigational device—the XyCAM ID™—to noninvasively capture retinal blood flow velocities (BFV) and its fluctuations with high spatio-temporal resolution.
The XyCAM ID was mounted on a slit lamp base and used to acquire speckle images from the right eye of two healthy individuals (authors MJB, YL) through self-experimentation. Imaging was performed twice, five days apart. On each day, images were acquired through a dilated pupil when the subject was rested (baseline), and after the subject had performed vigorous exercise as confirmed by elevation of the heart rate to 120 ± 10 beats/minute. Each imaging session comprised of serial acquisition of speckle images at a frame rate of 120 Hz for a period of 5 seconds, of a region of interest (ROI) centered on the optic disc illuminated with red laser. A plethysmogram was simultaneously recorded for reference. The acquired speckle images were processed in sliding batches of 11 frames each to obtain maps of retinal BFV. BFV values were averaged over all pixels in the ROI, and within pixels comprised in the central half-diameter of six manually identified vessel segments in each imaged eye, for assessment of temporal characteristics.
Fig. 1 shows an exemplary sequence of BFV maps and temporal trend in BFV fluctuations obtained from a human retina. The high spatio-temporal resolution permits probing of individual vessels (Figs. 2A-C) which exhibit pulsatility with a similar frequency as other vessels in the retina, but dissimilar amplitudes. The frequency of pulsations in the XyCAM based flow measurements varied commensurately with heart rate. In both pre- and post-exercise conditions, the correlation between measured retinal BFV and associated plethysmogram was high (Fig. 2D), but varied with image acquisition time (greater than 0.49 ± 0.17; p<0.001 for 4 second acquisition).
Preliminary results from our investigation validate the ability of the XyCAM ID to resolve regional and vascular blood flow fluctuations over multiple cardiac cycles with high fidelity. Further development is necessary to address relative inter-frame motion that can permit automated assessments over longer acquisition periods.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.
This PDF is available to Subscribers Only