Comparison of glaucoma diagnosis by ophthalmologist, optometrist, and telemedicine with OCT

Abstract

Purpose : Telemedicine is a novel, emerging medium addressing the challenges of accessing in-person glaucoma care by utilizing objective tools such as optical coherence tomography (OCT) for remote diagnosis. This study assessed the agreement of the diagnosis of glaucoma performed by an in-person ophthalmologist (MD), in-person optometrist (OD) and 2 telemedicine ophthalmologists (TMD1 and TMD2) with that of OCT alone.

Methods : This was a cross-sectional study of patients presenting to an academic tertiary referral glaucoma clinic who were independently examined for glaucoma with a dilated exam and optic nerve testing (visual field, optical coherence tomography, and photos) by both an MD and OD. Optic nerve testing was then reviewed by TMD1 and TMD2 with blinding to the MD and OD diagnoses. The OCT software labelled each scan as within normal limits, borderline or outside normal limits. “Borderline” was considered as both outside normal limits (OCT1) and within normal limits (OCT2). Agreement between each method of diagnosis (MD, OD, and TMD1 and TMD2) of normal vs. glaucoma (open angle glaucoma, normal tension glaucoma, other type of glaucoma, ocular hypertension, glaucoma suspect) compared to OCT1 and OCT2 for each eye was calculated using Cohen’s unweighted Kappa statistic and 95% confidence intervals. The predictive ability of OCT alone compared to TM2 was assessed by the area under the curve (AUC) of the receiver operating characteristics curve.

Results : 100 study participants (200 eyes) with a mean age of 66.7 years (IQR 59-72) were 40% male and 62% white. There was moderate agreement between OCT1 and telemedicine (TMD1: K=0.45 [95% CI 0.32-0.57], TMD2: K=0.46 [0.34-0.58]) in determining a diagnosis of normal from glaucoma. There was only fair agreement between OCT1 and in-person review by an MD (K=0.32 [0.19-0.45]) and OD (K=0.35 [0.22-0.48]). These results were similar for OCT2 agreement with telemedicine (TMD1, K=0.39 [0.28-0.51], TMD2, K=0.40 [0.28-0.51]), MD (K=0.34 [0.22-0.46]), and OD (K=0.29 [0.17-0.41]). When assessing the ability of OCT1 to diagnose glaucoma as compared to TMD2, the AUC was 0.73 (95% CI: 0.67-0.79) (Figure 1).

Conclusions : Although OCT is an important tool, it is not sufficient for the diagnosis of glaucoma in a telemedicine model as other parameters such as visual fields and optic nerve photos provide key information.

This is a 2021 ARVO Annual Meeting abstract.