Correlation and Agreement Between Cirrus HD-OCT "RNFL Thickness Map" and Scan Circle Retinal Nerve Fiber Layer Thickness Measurements

Purpose: To evaluate the correlation and agreement between optical coherence tomography (Cirrus HD-OCT) retinal nerve fiber layer (RNFL) thickness map and scan circle RNFL thickness measurements. Methods: ImageJ and custom Perl scripts were used to derive RNFL thickness measurements from RNFL thickness maps of optic disc scans of healthy and glaucomatous eyes. Average, quadrant, and clock-hour RNFL thickness of the map, and RNFL thickness of the areas inside/outside the scan circle were obtained. Correlation and agreement between RNFL thickness map and scan circle RNFL thickness measurements were evaluated using R-2 and Bland-Altman plots, respectively. Results: A total of 104 scans from 26 healthy eyes and 120 scans from 30 glaucomatous eyes were analyzed. RNFL thickness map and scan circle measurements were highly reproducible (eg, in healthy eyes, average RNFL thickness coefficients of variation were 2.14% and 2.52% for RNFL thickness map and scan circle, respectively) and highly correlated (0.55 <= R-2 <= 0.98). In general, the scan circle provided greater RNFL thickness than the RNFL thickness map in corresponding sectors and the differences tended to increase as RNFL thickness increased. The width of the 95% limits of agreement ranged between 5.28 and 36.80 mu m in healthy eyes, and between 11.69 and 42.89 mu m in glaucomatous eyes. Conclusions: Despite good correlation between RNFL thickness map and scan circle measurements, agreement was generally poor, suggesting that RNFL thickness assessment over the entire scan area may provide additional clinically relevant information to the conventional scan circle analysis. In the absence of available measurements from the entire peripapillary region, the RNFL thickness maps can be used to investigate localized RNFL thinning in areas not intercepted by the scan circle.