Abstract
Purpose: :
To assess the reproducibility of retinal nerve fiber layer (RNFL) thickness reported by Spectralis and Stratus OCT in patients with ocular hypertension (OHT) or open-angle glaucoma (OAG), and to attempt to create a conversion formula between the two instruments.
Methods: :
26 eyes of 26 subjects with OHT or OAG but no other ocular pathology participated in this study. Each patient underwent three Spectralis and three Stratus scans by the same operator during a single visit. One eye was randomly selected from each patient for the data analysis. Reproducibility was evaluated with intra-class correlation coefficients (ICCs). Pearson correlation coefficients (R) were used to assess the degree of agreement between instruments. Spectralis values were plotted against Stratus values and a linear regression line was created for each OCT quadrant in an attempt to create a conversion formula.
Results: :
ICCs ranged from 0.95 (nasal and temporal quadrants) to 0.97 (inferior quadrant) for Stratus and ranged from 0.96 (nasal quadrant) to 0.99 (superior and inferior quadrants) for Spectralis. In every quadrant, the Spectralis ICC value was significantly higher than the Stratus ICC value (P < .05). The correlation coefficients between Stratus and Spectralis measurements ranged from R=0.65 (nasal quadrant) to R=0.92 (inferior quadrant). Correlation coefficients for the superior and temporal quadrants were 0.82 and 0.85, respectively. The nasal quadrant had a significantly lower correlation coefficient than all other quadrants (P < .05). When average RNFL thickness values were compared between the two instruments, they were within 5 υm less than 35% of the time with a range in discrepancies from -29 to +17 υm and a nonsignificant trend toward lower values reported by Spectralis (83.6 ± 14.9 υm) compared to Stratus (84.8 ± 18.3 υm). When Stratus estimated a nasal RNFL thickness of less than 50 υm, Spectralis estimated significantly lower values; conversely, when Stratus estimated a nasal RNFL thickness of greater than 80 υm, Spectralis estimated significantly higher values (P < .05).
Conclusions: :
In patients with OHT or OAG, RNFL measurements by Spectralis are more reproducible than those generated by Stratus which favors use of the former for the monitoring of changes in RNFL thickness. Correlations are strong between the two instruments but the measurements are not interchangeable as differences exceed 5 υm more often than not. No meaningful conversion formula could be derived.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • imaging/image analysis: clinical • clinical (human) or epidemiologic studies: systems/equipment/techniques