Purchase this article with an account.
Ashkan Abbey, Cagri G Besirli, Paul P Lee, David C Musch, Chris Andrews, Antonio Capone, Kimberly A Drenser, David Keith Wallace, Michael Thomas Trese; A Computerized Telemedicine System for Screening of Retinopathy of Prematurity (ROP) and Validation of ROPtool. Invest. Ophthalmol. Vis. Sci. 2014;55(13):5927.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To determine if tortuosity assessment by a computer program (ROPtool) that traces retinal blood vessels and measures their tortuosity compares favorably with remote clinical grading of plus disease via telemedicine by 3 retinopathy of prematurity (ROP) experts (MT, AC, KD).
335 high-quality RetCam images (84 with plus disease and 251controls) of premature infants were obtained by neonatal intensive care unit nurses, and were sent electronically using the FocusROP software program to 3 ROP experts who graded vessel tortuosity (plus disease). These same images were analyzed using ROPtool, which assigns a numerical value to the level of tortuosity present in each quadrant. The maximum recorded value was then compared to the grading of the ROP experts.
Using 4.906 as a minimum value for declaring Referral-Merited ROP (presence of plus disease) creates a test with 96% sensitivity and 52% specificity relative to the gold standard of ROP expert reading. The receiver operating characteristic (ROC) curve revealed an area under the curve (AUC) of 0.918. 23 images (6.9%) were not gradable by ROPtool but could be read by the doctors.
ROPtool was found to be highly sensitive for the detection of plus disease as compared to expert physician graders. Our discussion will demonstrate how a paradigm combining regular digital fundus imaging, remote interpretation, and use of ROPtool would ensure appropriate follow-up, minimize medical error, and reduce cost of the delivery of quality care to infants at risk for developing ROP.
This PDF is available to Subscribers Only