Diagnostic Accuracy of a Retinal Birefringence Scanning Device Compared With a Traditional Autorefraction Device

Photoscreening devices are commonly used to identify children at risk for eye disease, although they often demonstrate low specificity as a trade-off for high sensitivity. A newly introduced photoscreening device that uses a newer mechanism, retinal birefringence scanning, purports to detect amblyopia and strabismus directly, and may theoretically increase specificity.

To compare the diagnostic performance of 2 photoscreening devices, the Spot Vision Screener (Welch Allyn), hereafter referred to as a traditional autorefraction device, and the Blinq vision scanner (Rebion), hereafter referred to as a retinal birefringence scanner, for detecting amblyopia, strabismus, or visually significant refractive error (VSRE).

This was a prospective, randomized, masked, diagnostic accuracy cohort study. The study included consecutive children 18 years or younger referred as a new patient to the University of California, San Francisco, pediatric ophthalmology clinic between January 2024 and February 2025 for suspected vision problems.

Each participant underwent screening with the traditional autorefraction device and the retinal birefringence scanner in randomized testing order, followed by a comprehensive eye examination. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC) were calculated for each device relative to the clinical reference standard.

Sensitivity, specificity, and AUROC of both screening devices to detect amblyopia, strabismus, or VSRE.

Of 195 enrolled participants (median [IQR] age, 50 [25-66] months), 139 (mean [SD] age, 49.7 [35.6] months; 75 female [54.0%]) completed both photoscreening tests. The prevalence of amblyopia, strabismus, and VSRE was 15.1%, 32.4%, and 30.1%, respectively. The traditional autorefraction device demonstrated higher sensitivity than the retinal birefringence scanner for detecting amblyopia (29%; 95% CI, -2% to 50%; P = .03) and VSRE (41%; 95% CI, 17%-59%; P < .001), with no difference in sensitivity for detecting strabismus identified (-6%; 95% CI, -28% to 18%; P = .69). No differences in specificity were identified. Across all vision disorders, the AUROC for the retinal birefringence scanner was lower than that for the traditional autorefraction device, particularly for amblyopia (-0.13; 95% CI, -0.23 to -0.03; P = .01).

In this single-site diagnostic accuracy study, the traditional autorefraction device demonstrated greater overall diagnostic sensitivity than the retinal birefringence scanner for detecting amblyopia and refractive errors, with no difference in specificity for all conditions. These findings suggest that the traditional autorefraction device may be more reliable for comprehensive pediatric vision screening.