Comparison of a smart phone retinal imaging headset system and binocular indirect ophthalmoscope in diagnosis of referral warranted retinopathy of prematurity in FEU-NRMF Medical Center / Mervyn H. Fernandez and Rachelle G. Anzures.

Includes appendices and bibliographical references.

Abstract: To determine the validity of a smart phone retinal imaging headset system for diagnosis of referral warranted retinophathy of prematurity in terms of sensitivity, specificity, and diagnostic accuracy using an indirect ophthalmoscope as gold standard. This is a descriptive cross-sectional study that compared smart phone retinal imaging headset system and binocular indirect ophthalmoscopy in the diagnosis of referral warranted retinopathy of prematurity in pediatric patients aged 1-day to 3-months old that were referred for retinopathy of prematurity screening were included in a selected tertiary hospital. Excluded in the study are those with systemic malformations, post retinal surgery or laser, ocular anomalies of the retina and choroid besides retinopathy of prematurity and ocular anomalies that obstructs the view of the posterior pole (ex. Cataract, corneal opacities and others). A total of 8 eyes were included in the study and undergone binocular indirect ophthalmoscopy by a specialist followed by a smart phone retinal imaging headset funduscopy system and images were stored. The images stored were then shown to a specialist to diagnose if there bis refferal warranted retinopathy of prematurity or not. The sensitivity, specificity and predictive values were then calculated. A total of 8 eyes from 4 infants were included in the pilot study results whose average age is around 26 days, deviating by 9.6. Majority of the infants are female (75%) while resulting average age of gestation is 32.3 weeks and mean post corrected age in weeks is 35.9. Moreover, the average birth weight of the infants was 1307.5 grams while their average current weight was 1567.5 grams. All infants are also positive in oxygen supplementation use. It shows that of the 8 eyes, none of them was positive for referral warranted retinopathy of prematurity (RW-ROP) according to binocular indirect ophthalmoscopy, while 6 also resulted to negative RW-ROP according to smart phone retinal imaging headset system. However, 2 eyes turned out to be ungradable due to poor image quality. Results showed 100% accuracy, as all results were negative RW-ROP for binocular indirect ophthalmoscopy (BIO) are also negative according to smart phone retinal imaging headset. Sensitivity was not determined as none of the eyes were positive for RW-ROP, while specificity is 100% (95% 54.1% to 100.0%). Likewise, since all eyes were predicted correctly by smart phone retinal imaging headset, both PPV and NPV are also 100%. In conclusion, although small number of patients limits the generalizability of this study, smart phone retinal imaging headset system showed high specificity for negative RW-ROP. This might be a useful tool for documentation, telemedicine consultation and screening for referral warranted retinopathy of prematurity.

Research - Department of Ophthalmology