J Glaucoma
J GlaucomaMay 2025Journal Article

Performance of Optic Disc Optical Coherence Tomography Normative Database in a Large, Diverse, Real-World Cohort.

Optic Nerve & DiscOCT & Imaging

Summary

OCT normative databases should accurately reflect diverse populations to avoid misclassification by RNFL thickness color codes. Larger data sets should be leveraged to encompass the full spectrum of healthy optic nerve anatomy.

Abstract

PRCIS

Current optical coherence tomography normative sample data may not represent the diversity of human optic nerve anatomy needed to accurately classify all individuals with true glaucomatous optic neuropathy.

PURPOSE

To compare optic nerve head (ONH) measurements between published values from an optical coherence tomography (OCT) normative database and a larger, more diverse cohort of healthy individuals.

PATIENTS AND METHODS

ONH parameters from healthy participants of the Michigan Screening and Intervention for Glaucoma and Eye Health through Telemedicine (MI-SIGHT) program and the Topcon Maestro-1 normative cohort were compared. χ 2 tests compared MI-SIGHT retinal nerve fiber layer (RNFL) quadrant color-code labels with the expected distribution and multinomial logistic regression identified factors associated with label classifications.

RESULTS

In all, 1084 MI-SIGHT and 399 Topcon eyes were evaluated. The MI-SIGHT cohort was older (54 vs. 46 y), with more individuals identifying as black (61% vs. 20%), fewer as Hispanic (4% vs. 18%), and spherical equivalents closer to plano (-0.6 vs. -1.2 diopters) compared to the Topcon cohort (all P <0.001). Black/African American MI-SIGHT participants had larger cup-to-disc ratios and cup volumes, while white MI-SIGHT participants had smaller ONH values, except for rim area and rim volume, compared to Topcon participants (all P <0.001). The MI-SIGHT cohort's RNFL color codes did not follow the expected distribution ( P <0.05); more MI-SIGHT RNFL quadrant measurements were assigned as white (10.6% and 6.3% MI-SIGHT vs. 5% Topcon) and red codes (2.2% and 1.8% MI-SIGHT vs. <1% Topcon) than expected in the superior and inferior quadrants, respectively.

CONCLUSIONS

OCT normative databases should accurately reflect diverse populations to avoid misclassification by RNFL thickness color codes. Larger data sets should be leveraged to encompass the full spectrum of healthy optic nerve anatomy.

Keywords

biasbig datadiversityglaucomanormative dataoptical coherence tomography (OCT)screening

In the Knowledge Library

Discussion

Comments and discussion will appear here in a future update.