Evaluating a Smartphone-Based Optic Nerve Imaging Device as a Glaucoma Screening Tool in an Outpatient Clinic Setting

Abstract

Purpose : Ideally, population based eye screening could be performed by a layperson with minimal training at low cost with image interpretation occurring at telemedicine grading centers. This would be particularly useful in underserved communities in order to identify high risk individuals and best optimize delivery of ophthalmic care. This study aims to evaluate a smartphone-based ophthalmoscope, D-EYE, as a potential screening device for glaucoma.

Methods : Patients presenting to university based primary care and glaucoma clinics were recruited for this study. Clinical charts were reviewed to determine glaucoma status for each eye of every recruited patient. Without pupillary dilation, bilateral optic nerve videos were captured in 30-second-long clips with the D-EYE portable smartphone attachment for fundus photography. Videos were later independently graded by four ophthalmology residents and an ophthalmology intern in a blinded fashion. Grading criteria included the ability to visualize the optic nerve (yes or no), subjective image quality (5 point Likert scale [1 = minimal visibility of optic nerve head, 5 = optimal visibility of optic nerve head]), cup to disc ratio (0.01 to 0.99), and likelihood of referring patients for additional glaucoma care (5 point Likert scale [1 = very unlikely to refer, 5 = very likely to refer]).

Results : Optic nerve video capture was attempted on a total of 236 eyes (210 normal, 26 glaucomatous). A total of 186 of the 236 videos (78.8%) were of sufficient quality to visualize the nerve (163 normal, 23 glaucomatous). The mean image quality score was 3.11 out of 5. The average cup to disc ratio of normal and glaucomatous eyes was 0.299 and 0.420, respectively (intraclass correlation coefficient [ICC] = 0.436). The likelihood of referral of normal and glaucomatous eyes was 1.671 and 2.565, respectively (ICC = 0.509).

Conclusions : This pilot study used a smartphone-based photography system in undilated eyes which minimizes cost and facilitates ease of adoption. There was a moderate degree of inter-grader agreement for videos and, overall, glaucomatous eyes were more likely to be referred for further workup compared to healthy eyes. Agreement between raters and accurate likelihood of referral for glaucomatous eyes may be increased with dilated imaging and more experienced graders.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.