Custom measurement of retinal nerve fiber layer thickness using STRATUS OCT in normal eyes

To evaluate variability of retinal nerve fiber layer (RNFL) thickness measurements in normal eyes and their correlation with optic disc diameter by using two different scan options of the ultimate commercial optical coherence tomography (OCT) unit (STRATUS OCT, Carl Zeiss Meditec, Inc., Dublin, CA).

In this observational case series and instrument validation study 30 eyes of 30 normal subjects were enrolled. Each eye underwent optic disc vertical diameter measurement by means of both stereoscopic photography and planimetry and OCT; RNFL thickness measurements were performed using OCT. Three repetitions of two series of scans were performed. Each eye was scanned at two different options (RNFL thickness 3.4 and Nerve Head Circle). For each option descriptive statistics, analysis of variance, intraclass correlation coefficients (ICCs), and coefficients of variation (COVs) were calculated. To verify the correlation between the two methods of optic disc diameter assessment and to study the influence of optic disc diameter on RNFL measurement using the two different OCT options, Pearson's correlation coefficients were calculated.

Optic disc diameter length ranged from 1.47 to 2.04 mm (mean 1.709 mm, SD +/- 0.147) with stereoscopic photographs, and from 1.47 to 2.02 mm (mean 1.703 mm, SD +/- 0.143) with OCT (Pearson correlation coefficient 0.999, p<0.001). Mean RNFL thickness was 89.29 mm (SD +/- 10.80 mm) using the RNFL thickness 3.4 scanning option and 89.88 mm (SD +/- 1.72 mm) using the Nerve Head Circle protocol (Pearson correlation coefficient 0.065, p=0.734). The intersubject variance is higher using the RNFL thickness 3.4 option than using the NHC protocol (sum of square: 1,014,760 vs. 25,741) (p<0.001); the intrasubject variance is very similar in the two groups (2,372 vs 2,360) (p=NS). The ICC is 99.89% when using the RNFL thickness 3.4 option, 95.62% with the NHC protocol (p=NS). COVs were 12.10% and 1.91% by using RNFL thickness 3.4 and Nerve Head Circle option, respectively. Pearson's correlation coefficient was 0.988 (p<0.001) when comparing optic disc diameter and RNFL thickness by using the RNFL thickness 3.4 option and -0.016 (p=0.932) when comparing optic disc diameter and RNFL thickness by using the Nerve Head Circle option.

These results suggest that both scan options give good RNFL thickness measurement reproducibility; the use of the Nerve Head Circle option leads to less interindividual variability and can minimize the effect of differences in optic disc diameter on RNFL thickness measurements in normal subjects.