Purchase this article with an account.
Peter C Nicholas, Priyatham S Mettu, Scott W Cousins; Feasibility and acceptance of retinal imaging in a medical clinic by non-expert imagers. Invest. Ophthalmol. Vis. Sci. 2016;57(12):1661.
Download citation file:
© 2017 Association for Research in Vision and Ophthalmology.
To assess feasibility and acceptance of retinal imaging with non-mydriatic Color Fundus Photographs (CFP) and OCT imaging performed by non-expert personnel in a medical clinic setting. We hypothesized that non-experts could acquire diagnostic quality retinal images in a non-ophthalmology medical clinic, and that this process would be accepted by patients.
We recruited adults at risk for retinal disease from an endocrinology clinic (inclusion criteria: diagnosis of diabetes), and a geriatrics clinic (inclusion criteria: age>70y). After informed consent, participants underwent non-mydriatic retinal imaging on the iFusion device (Optovue, Fremont, CA). Imagers without expertise in ophthalmology were trained and certified to use the iFusion. After remote imaging, subjects completed a questionnaire about the experience and were scheduled for a reference standard dilated examination by a retinal specialist with imaging (CFP and OCT). Remote images were graded dichotomously by masked readers for interpretability and presence of referable retinal pathology, with each eye graded independently.
For the first 50 subjects (100 eyes) enrolled in the study, 68% of remote CFP and 97% of OCT were interpretable. 42 subjects (84 eyes) completed reference standard examination thus far, and of these, 31 eyes (37%) had referable retinal pathology. Sensitivity (SE) and specificity (SP) to detect referable retinal pathology for remote CFP was SE=90%, SP=57%; for OCT, SE=81%, SP=0.83% when uninterpretable images were considered equivalent to referable pathology. Only 33% of eyes with uninterpretable CFP had referable retinal pathology on reference standard exam. Questionnaires indicated that 57% favored remote imaging over in-person examination, 35% expressed no preference, and only 8% expressed a preference for in-person examination.
These results indicate that non-expert imagers can successfully acquire non-mydriatic OCT images (97% interpretable) using the iFusion system. Acquisition of interpretable CFP in these high-pathology populations is more challenging (32% uninterpretable). The high SE and lower SP for remote CFP was driven by uninterpretable CFP which forced referral. Survey results indicate a high degree of acceptance and preference for remote imaging versus full examination. Rigorous comparative effectiveness research will be required to evaluate emerging remote diagnosis systems.
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