Non-confocal split-detection adaptive optics scanning light ophthalmoscope with small pupil for vascular imaging

Abstract

Purpose : To evaluate a novel adaptive optics scanning light ophthalmoscope (AOSLO) designed for retinal microvasculature imaging by balancing image resolution, field of view (FOV), isoplanatic patch and depth of focus through the reduction of pupil size.

Methods : The peripapillary vasculature of 8 normal subjects (age 27±2.6 years) was imaged using nearly identical custom AOSLOs with FOVs inversely proportional to their pupil diameters (4.3 & 7.75 mm). Stacks of non-confocal split-detection image sequences spanning the entire depth of the vascularized retina were used to calculate perfusion maps (standard deviation from 100-frame sequences). These maps were then combined through maximum-intensity projection and compared against the average of 3x3 mm registered perfusion maps from an optical coherence tomographer (OCT) by OptoVue.

Results : Images from the small-pupil AOSLO showed retinal vessels with comparable contrast and signal-to-noise ratio to those of the large-pupil AOSLO. The reduced pupil allowed for increased FOVs with the same resonant scanner (Lagrange invariant) and depth of focus, thus reducing the number of image sequences required to achieve the same retinal coverage by over a factor of four. The maximum-intensity projections allow resolving all retinal capillaries within the field of view and depth of focus, unlike the commercial OCT angiography instrument tested (see figure).

Conclusions : This first demonstration of a non-confocal split-detection AOSLO with a small pupil produced comparable images to those from a similar AOSLO with 80% larger pupil. This suggests that the multiple scattering image contrast that creates the split-detection images is not strongly dependent on pupil diameter. As anticipated, the reduced pupil increased the isoplanatic patch, allowing for good image sharpness over larger FOVs. The increased retinal coverage and extended depth of focus drastically reduced imaging time, bringing this technology closer to clinical utility for personalizing treatment of retinal vasculopathies and as a research tool to evaluate vascular changes in conditions such as glaucoma.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.

 

Motion contrast-derived retinal perfusion maps of a healthy subject acquired with an OCT (top and middle left), and two AOSLOs with 7.5 (high-res) and 4.30 (low-res) mm pupil diameters, respectively. Arrows in low-res AOSLO panel and image intensity plot below it indicate capillaries.