Purpose: To evaluate (1) the feasibility of non-ophthalmologist healthcare workers (HCWs) to obtain images of sufficient quality for retinopathy of prematurity (ROP) screening using a FDA-approved portable, non-contact, narrow-field fundus camera (i.e., Pictor™), and (2) the accuracy of grading these images to identify infants who developed treatment-warranted (type 1) ROP.
Design: Prospective cohort study.
Subjects: Infants undergoing routine ROP screening examinations (i.e. birth weight ≤1500 grams and/or gestational age ≤30 weeks or selected infants with a birth weight of 1500-2000g or gestational age >30 weeks and an unstable clinical course).
Methods: We prospectively recruited infants undergoing ROP screening examinations at a community hospital. On the same day an ophthalmologist examined them, a trained HCW imaged their retinas using the non-contact camera. Two masked ROP experts graded these images remotely. We calculated both the percentage of gradable images (i.e. having at least 3 quadrants with sufficient image quality), as well as the accuracy of identifying infants who developed type 1 ROP.
Main outcome measures: Percentage of gradable images and the sensitivity and specificity of each grader for identifying infants with type 1 ROP by grading for the presence of pre-plus or plus disease.
Results: Ninety-nine infants were included. Overall, 92.4% and 94.2% of all infant imaging sessions were considered gradable by graders 1 and 2, respectively. Amongst gradable images, the sensitivity of both graders for identifying type 1 ROP by grading for the presence of pre-plus or plus disease was 100% (95% confidence interval (CI): 95-100%) and the specificity 91% (95% CI: 83-95%) for grader 1 and 93% (95% CI: 86-96%) for grader 2.
Conclusions: It was highly feasible for trained HCWs to obtain digital retinal images of sufficient quality for ROP screening using a non-contact fundus camera. By grading for the presence of pre-plus or plus disease, graders identified infants who developed type 1 ROP with high sensitivity and specificity. The use of portable, non-contact retinal cameras by trained HCWs could increase our workforce in ROP screening and identify infants needing an indirect ophthalmoscopy examination by an ophthalmologist.