The Accuracy of the Retcam as a Wide-Field
Digital Retinal Imaging Device Relative to Binocular
Indirect Ophthalmoscopy for the Diagnosis of
Retinopathy of Prematurity:
A Systematic Review of Diagnostic Test Accuracy

Shannon Brown BAppSc MOrth

Discipline of Orthoptics, School of Allied Health, Human Services and Sport, La Trobe University, Melbourne, Australia

Objective: To assess the diagnostic test accuracy of the RetCam, a wide-field digital retinal imaging device, for the detection of referral-warranted or treatment-requiring retinopathy of prematurity (ROP), using binocular indirect ophthalmoscopy (BIO) as the reference standard for screening in premature infants. ROP is a sight-threatening condition of the immature retina requiring diagnosis by a trained ophthalmologist. An alternative method for performing a comprehensive retinal examination of a neonate for the diagnosis and staging of ROP is necessary to meet the increasing demand for  screening, improve access and reduce costs. 

Method: Database searches of CINAHL, PubMed, Embase, SCOPUS and CENTRAL were conducted on the 20th May 2021 for articles published from 2010 onwards. Studies where infants were tested with both the RetCam and BIO were included. The QUADAS-2 quality assessment checklist was utilised to assess any bias or applicability concerns for each study. Study data was synthesised in narrative and tabular form, a meta-analysis was not performed due to limited scope and resources. 

Results: Ten studies were included in the final review for the primary objective with 1,663 total study participants, and four for the secondary objective of  detecting any stage of ROP. The sensitivity ranged from 80-100% and specificity from 35-100% for referral-warranted or treatment-requiring ROP. For any  stage ROP, the sensitivity and specificity ranged from 58.6-86% and 97.8-100%, respectively. Only two studies had a low risk of bias and the remaining studies  had unclear or problematic methodological concerns.

Conclusion: The accuracy demonstrated by the included studies was not generalisable as they either had small sample sizes or were under the influence of bias, however the RetCam can be useful as an adjunct to BIO. Further research is required with larger cohorts in well controlled, prospective studies to establish the RetCam as a feasible substitute for BIO.