Phone: (+39) 0813995453

Diabetic Retinopathy Telemedicine Outcomes with Artificial Intelligence-Based Image Analysis, Reflex Dilation, and Image Overread

HumanInsight Diabetic Retinopathy Telemedicine Outcomes with Artificial Intelligence-Based Image Analysis, Reflex Dilation, and Image Overread

Am J Ophthalmol. 2022 Aug 12:S0002-9394(22)00310-5. doi: 10.1016/j.ajo.2022.08.008. Online ahead of print.


PURPOSE: To examine real-world telemedicine outcomes of diabetic retinopathy screening with artificial intelligence (AI)-based image analysis, reflex dilation, and secondary image overread in a primary care setting.

STUDY OF SCREENING TEST: Validity and reliability analysis METHODS: Single institution review of 1,052 consecutive adult patients who received diabetic retinopathy photoscreening in the primary care setting over an 18-month period. Nonmydriatic fundus photographs were acquired and analyzed by the IDx-DR AI-based system. When nonmydriatic images were ungradable, reflex dilation (1% tropicamide) and mydriatic photography were performed for repeat AI-based analysis. Manual overread was performed on all images. Patient demographics, clinical characteristics, and screening outcomes were recorded.

RESULTS: 91.7% (965/1052) of patients had AI-gradable fundus photographs. 55.1% (580/1052) had gradable nonmydriatic imaging and 93.2% (440/472) of those with ungradable nonmydriatic photographs had reflex dilation. 14.3% (138/965) of patients were AI-graded as "positive" (>mild NPDR) and 85.7% were "negative" (827/965), with 100% sensitivity (95%CI 90.8-100%), 89.2% specificity (95%CI 87.0-91.1%), 27.5% positive predictive value (95%CI 24.0-31.4%), and 100% negative predictive value (95%CI 99.6-100%) compared to manual overread assessment of >mild NPDR requiring further evaluation with a comprehensive dilated examination. Image gradeability was inversely related to patient age: 93.5% gradable (61.9% nonmydriatic) for patients aged <70 years versus 85.3% (31.0% nonmydriatic) for patients aged 70+ (p<0.001).

CONCLUSION: Incorporation of AI-based image analysis into real-world primary care diabetic retinopathy screening yielded no false negative results and offered excellent image gradeability within a protocol combining nonmydriatic fundus photography and pharmacologic dilation, as needed. Image gradeability was lower with increasing patient age.

PMID:35970206 | DOI:10.1016/j.ajo.2022.08.008

Powered by WPeMatico

P.IVA 08738511214
Privacy Policy
Cookie Policy

Sede Legale
Viale Campi Flegrei 55
80124 - Napoli

Sede Operativa
Via G.Porzio 4
Centro Direzionale G1
80143 - Napoli

AI 4394
© Copyright 2022 - Humaninsight Srls - All Rights Reserved
Privacy Policy | Cookie Policy
envelopephone-handsetmap-marker linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram