Retinopathy of Prematurity Management uing ingle-Image v Multiple-Image Telemedicine Examination Original Reearch Article

Purpoe To compare performance of ingle-image v multiple-image telemedicine examination for retinopathy of prematurity (ROP) diagnoi. Deign Propective comparative tudy. Method A total of 248 eye from 67 conecutive infant underwent wide-angle retinal imaging by a trained neonatal nure at 31 to 33 week and/or 35 to 37 week potmentrual age (PMA) at a ingle academic intitution. Data were uploaded to a web-baed telemedicine ytem and interpreted by three maked retinal pecialit. Diagnoe were provided baed on ingle image, and ubequently on multiple image, from both eye of each infant. Finding were compared to a reference tandard of indirect ophthalmocopy by a pediatric ophthalmologit. Primary outcome meaure were recommended follow-up interval, preence of plu dieae, preence of type-2 or wore ROP, and preence of viible peripheral ROP. Reult Among the three grader, mean enitivity/pecificity for detection of infant requiring follow-up in le than one week were 0.85/0.93 by ingle-image examination and 0.91/0.88 by multiple-image examination at 35 to 37 week PMA. Mean enitivity/pecificity for detection of infant with type-2 or wore ROP were 0.82/0.95 by ingle-image examination and 1.00/0.91 by multiple-image examination at 35 to 37 week PMA. Mean enitivity/pecificity for detection of plu dieae were 1.00/0.86 by ingle-image examination and 1.00/0.87 by multiple-image examination at 35 to 37 week PMA. There were no tatitically-ignificant intragrader difference between accuracy of ingle-image and multiple-image telemedicine examination for detection of plu dieae. Concluion ingle-image and multiple-image telemedicine examination perform comparably for determination of recommended follow-up interval and detection of plu dieae. Thi may have implication for development of creening protocol, particularly in area with limited acce to ophthalmic care.