Optical coherence tomography errors in glaucoma

Abstract

Purpose: To demonstrate the types and prevalence of errors associated with the use of optical coherence tomography in a cross section of glaucoma patients. The association of different types of artifacts and certain coexisting ocular diagnoses was also evaluated.

Patients and methods: The optical coherence tomography data of glaucoma patients in a 3-month period was evaluated for evidence of errors in peripapillary retinal nerve fiber layer (RNFL) map and the macular thickness map. Signal strengths, centering errors, individual scan errors, and association with ocular conditions were noted. Logistic regression was used to assess the significance of continuous and categorical variables in predicting the presence of artifacts.

Results: Macular scan artifacts were present in 16.8% and RNFL scan artifacts in 15.7% of 89 eyes of 89 patients studied. RNFL off-center scan was the most common error (34.8%). For macular thickness, 100% of the scans were artifact free for signal strength 8 or higher. However, for signal strength of < or =4, 64.3% had artifacts. For RNFL thickness, a signal strength > or =6, 96% of the scans had no artifacts. However, for signal strength < or =4, 86% had artifacts. Macular scan artifacts were present more commonly in patients with dry eye and RNFL centering errors were more frequent in eyes with cataracts.

Conclusions: A majority of glaucoma patients have artifact free scans. The final printouts of macular and RNFL thickness can identify most errors except for individual macular scan errors.