Automated Strabismus Detecting Tool Shows Promise for Clinical Screening

An automated strabismus detecting tool may have the ability to rapidly screen for strabismus, according to a study published in Eye. The technology has the potential to be used by non clinicians for telemedical applications, the report suggests.

Researchers initiated the preliminary evaluation to determine the feasibility of using the automated strabismus detecting tool and included 26 individuals from 2 hospitals in the investigation. The team enrolled a cohort of control group participants (n=7) and individuals with ocular misalignments (n=19; mean age, 58.7 years), which included 8 individuals with esodeviations, 7 participants with exodeviations, and 10 patients with vertical deviations.

All study participants underwent full orthoptic assessment with alternate prism cover testing at 6 meters and the novel automated strabismus detecting tool on the same day. Examiners were masked to the results obtained using the novel screening tool, which utilized a virtual reality headset with integrated eye tracking worn over the patients’ habitual refractive correction.

In the first phase of the study, the automated strabismus detecting tool was developed by calculating the amplitude of horizontal refixation movements among control group participants. In phase 2, the device was evaluated among the participants with ocular misalignment.

The mean difference in horizontal deviation measurements using the standard alternate prism cover test and automated strabismus detecting tool was 2.1 (95% CI, -1.8-9.9) prism diopters. The 95% coefficient of repeatability was ±27.9 prism diopters, or ±17.3 prism diopters after removing 1 outlier. These data indicate both test outcomes were correlated (r, 0.62; P <.001).

The automated strabismus detecting tool’s ability to identify horizontal strabismus was high (area under the curve [AUC], 1.00; sensitivity, 100%; specificity, 100%). Introducing a small degree of noise into the data returned an AUC of 0.98, sensitivity of 87%, and specificity of 100%.

To test whether a shorter duration test may be feasible, the investigators calculated the accuracy during progressive seconds of the test and found that AUC values improved with each additional second of data. This finding shows that screening with the automated strabismus detecting tool for less than 60 seconds would yield less accurate results.

“[I]nstead of depriving the patient of their usual surroundings and distance cues, we allow the patient to view a faithful stereoscopic representation of the room they are in, using
augmented reality.” the study authors explain. In addition to the automated strabismus detecting tool’s accuracy and short screening duration, the investigators state that the device “is a rapid, tolerable screening test of strabismus that may in future be used by those unable to reach a traditional healthcare setting.”

Failure to perform inter- and intra-observer evaluations is an acknowledge limitation to the research.