Objective: To compare the accuracy of infrared (IR)-reflex assessment using a prototype imaging device to standard non-mydriatic red-reflex screening with direct ophthalmoscope (DO) in the diagnosis of neonatal and childhood cataract.
Methods: The comparison of the techniques was made in two distinct cohorts: in the first, newborns underwent IR and red-reflex testing by a medical student, with results compared to a reference red-reflex examination by an experienced midwife. In the second, an enriched cohort of children attending a specialist paediatric ophthalmology clinic had IR and red-reflex testing by a medical student to reference examination by a paediatric ophthalmologist. The medical students were considered inexperienced screeners due to their limited exposure to ophthalmology. The sensitivity and specificity of the IR and red-reflex assessments in respect to reference examination were calculated. Diagnostic accuracy was compared in Caucasian and non-Caucasian eyes.
Results: IR and red-reflex imaging were possible in all 180 neonatal eyes examined. A total of 5% of newborn eyes were found to have embryological remnants in the anterior segment of the eye with IR-reflex imaging which were not detected on reference red-reflex examination. IR-reflex assessment had significantly better sensitivity (100 vs 71%, p < 0.05) and specificity (100 vs 63%, p < 0.01) than red-reflex assessment in the diagnosis of childhood cataract. Red-reflex specificity was particularly poor in non-Caucasian eyes compared to Caucasian eyes (32 vs 72%, p < 0.05).
Conclusion: This pilot study indicates that IR-reflex imaging has the potential to improve the diagnostic accuracy of eye screening for cataract by inexperienced healthcare staff, particularly in non-Caucasian children.