A high myopia-specific normative database of peripapillary retinal nerve fiber layer (pRNFL) thickness more accurately detects glaucoma among eyes with high myopia than the built-in database used in swept-source optical coherence tomography (SS-OCT) devices, according to research published in Ophthalmology.
Researchers conducted a cross-sectional, multicenter study to develop and validate the performance of their normative database for differentiating high myopia from highly myopic glaucoma (intraocular pressure [IOP] >21 mm Hg). The team included 459 eyes of 408 patients (high myopia, n=323 ; highly myopic glaucoma, n=136 eyes with HMG) in the internal validation cohort, and 322 eyes of 197 participants (high myopia, n=131; highly myopic glaucoma, n=191) in the external validation cohort.
Study participants underwent comprehensive ophthalmic examinations and SS-OCT assessment. The investigators examined thresholding strategies of pRNFL-specified values derived from the normative database, including global and quadratic pRNFLT below the lowest 5th or the lowest 1st percentile of the normative database, and evaluated the accuracy, sensitivity, and specificity of the high myopia-specific normative database for detecting highly myopic glaucoma.
When setting the threshold as the 5th percentile of the global pRNFLT using the high myopia-specific normative database, the accuracy of detecting highly myopic glaucoma among eyes with high myopia was 0.93 (95% CI, 0.90-0.95) in the internal validation cohort and 0.85 (95% CI, 0.81-0.89) in external validation cohort.
When setting the threshold as the 1st percentile of the global pRNFLT using the high myopia-specific normative database, the accuracy for detecting highly myopic glaucoma among eyes with high myopia was 0.85 (95% CI, 0.81-0.88) in the internal validation cohort and 0.70 (95% CI, 0.65-0.75) in external validation cohort. The diagnostic ability to detect highly myopic glaucoma was significantly better using the 5th percentile of pRNFLT compared with the 1st percentiles of pRNFLT (P<.001), the report shows.
Using the 5th percentile of the global pRNFLT, the investigators observed high sensitivities (internal: 0.75; 95% CI, 0.67-0.82; external: 0.75; 95% CI, 0.68-0.81) and specificities (internal: 1.00; 95% CI, 0.99-1.00; external: 1.00; 95% CI, 0.97-1.00). Compared with the built-in statistical database of the SS-OCT device, the team determined that the high myopia-specific normative database had a significantly higher sensitivity and specificity for all corresponding pRNFL thickness values below the 5th or 1st percentile (P<.001 for all).
“To meet the need for screening and diagnosis of glaucoma, we developed a normative database based on SS-OCT images and validated its performance in both internal and external data sets,” according to the study authors. “Our normative database of the [pRNFL thickness] demonstrated a relatively high diagnostic performance in detecting glaucomatous RNFL abnormalities in [highly myopic] eyes.”
Study limitations include an ethnically homogenous cohort, which may limit the globalization of the study’s findings.
Disclosure: Some study authors declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.
Song Y, Li F, Chong RS, et al. High myopia normative database of peripapillary retinal nerve fiber layer thickness to detect myopic glaucoma in Chinese population. Published online July 25, 2023. Ophthalmology. doi:10.1016/j.ophtha.2023.07.022