Diagnostic ability of optical coherence tomography with a normative database to detect localized retinal nerve fiber layer defects.

PURPOSE: To evaluate the diagnostic ability of third-generation optical coherence tomography (Stratus OCT) with a normative database to detect localized retinal nerve fiber layer (RNFL) defects.

DESIGN: Cross-sectional study.

PARTICIPANTS: One eye of each of 55 subjects (43 normal-tension glaucoma and 12 primary open-angle glaucoma subjects) with localized, wedge-shaped RNFL defects by red-free RNFL photographs and corresponding visual field defects (mean deviation, -6.77+/-3.54; range, -1.25 to -9.20 dB).

METHODS: The segments of the borderline and outside normal limits in the RNFL thickness profile of OCT were considered to be significant RNFL defects by OCT and by comparison with their corresponding red-free RNFL photographs. The angular locations and widths of RNFL defects by OCT were compared with those of RNFL defects by red-free RNFL photographs.

MAIN OUTCOME MEASURES: Sensitivity and specificity of OCT with a normative database to detect the localized RNFL defects.

RESULTS: The mean age of the patients was 57.9+/-10.6 years (range, 25-78 years). The overall sensitivity and specificity of the OCT with a normative database for detecting localized RNFL defects were 85.9% and 97.4%, respectively. The OCT with a normative database showed good diagnostic agreement with red-free RNFL photographs (overall kappa value = 0.79). The angular locations and widths of RNFL defects by OCT were highly correlated with those of RNFL defects by red-free RNFL photographs (Pearson correlation coefficient [r] = 0.996 and 0.725, respectively).

CONCLUSIONS: The OCT with a normative database can detect localized RNFL defects with moderate sensitivity and high specificity, and shows good diagnostic agreement with red-free RNFL photographs.