Glaucoma Detection Using the OCT Normative Database

Abstract

Abstract: : Purpose: The current version of optical coherence tomography (OCT; StratusOCT, Carl Zeiss Meditec, Dublin, CA) includes a comparison to a normative database that provide a classification of the macular and peripapillary measurements into normal, borderline or abnormal. The purpose of this study was to define the best use of these data for discrimination of healthy from glaucomatous eyes as defined by clinical classification. Methods: One randomly selected eye from 44 glaucomatous and 21 healthy subjects was scanned by OCT using both peripapillary and macular fast scanning protocols. Glaucomatous eyes were defined as those with typical clinical findings and reproducible visual field abnormalities. From each circumpapillary scan, 17 nerve fiber layer (NFL) thickness measurements were considered: overall mean thickness, 4 quadrants and 12 clock hour segments. From each macular scan, 9 measurements were considered: 4 inner ring, 4 outer ring, and the central thickness. The analysis was conducted twice: first when defining borderline results as normal and then when defining the borderline as abnormal. Results:The presence of one abnormal NFL value alone was shown to have the highest positive predictive value (PPV) (0.98) specificity (0.95) and sensitivity (0.93) to detect glaucomatous changes. Defining the abnormality as 2 or more NFL values or by defining the borderline results as abnormal did not improve discriminating ability. Using macular thickness measurements the highest PPV, specificity and sensitivity was shown when defining the abnormality as two or more abnormal values when considering borderline results as abnormal (0.97, 0.95, 0.66, respectively). Incorporating the results of both peripapillary and macular scans did not improve the results than those observed by NFL alone. The highest negative predictive value (1.0) was shown when considering borderline results as abnormal with one abnormal NFL value or two abnormal values when combining peripapillary and macular results. Conclusions: OCT peripapillary and macular scans may be used to distinguish normal from glaucomatous eyes; the best predictor of glaucoma is the presence of one abnormal NFL value, as assessed by the OCT normative database. Support: NEI (RO1–EY13178, RO1–EY11289, P30–EY008098), Research to Prevent Blindness, Eye & Ear Foundation (Pittsburgh, PA)

Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • imaging/image analysis: clinical