Abstract
Abstract: :
Purpose: i) To evaluate the agreement of software-corrected retinal nerve fibre layer thickness (RNFLT) thickness measurements made with the fixed corneal compensator (FCC) position of the scanning laser polarimeter (SLP) and RNFLT measurements made with a new variable corneal compensator (VCC). ii) To assess the sensitivity for glaucoma diagnosis and correlation with visual function of FCC-derived and VCC RNFLT measurements. Methods: Peripapillary (PP) images were taken of 56 eyes of 56 normal subjects and 55 glaucomatous eyes of 55 patients (average mean defect -7.2dB) with two settings of the VCC SLP: FCC (population mode setting) and VCC (set to neutralize the retardation pattern in the macula). Macula images were also taken with the FCC setting. RNFLT measurements were made in 16 segments in annuli 1.4 mm (macula) and 3.0 mm (PP) in diameter. FCC RNFLT measurements were corrected by subtracting macula from PP RNFLT measurements (FCCm). The difference between VCC segmental RNFLT measurements and FCC and FCCm measurements was calculated. The mean and standard deviation (SD) of these differences are presented as summary measures of agreement. The lower limit for normal RNFLT thickness was set at the (lower) 1st percentile in normal subjects. This was calculated for FCC, FCCm and VCC PP RNFLT. A normal or glaucomatous eye was defined as "outside normal limits" if the RNFLT was below the normal range in any segment. Correlation of RNFLT (measured RNFLT/age-corrected average RNFLT) with visual function (measured visual field sensitivity/age-corrected average sensitivity) was performed by linear regression analysis. Results: Agreement between methods: the mean difference and SD of differences was 20.7 µm and 24.7 µm for FCC-VCC and -37.2 µm and 16.5 µm for FCCm-VCC. Sensitivity for glaucoma diagnosis: with specificity fixed at 85%, sensitivity (95% confidence interval) was 47% (34-61%), 62% (49-75%) and 75% (63-86%) for FCC, FCCm and VCC, respectively. Correlation with visual function: R2 values were 0.03, 0.28 and 0.43 for FCC, FCCm and VCC, respectively. Conclusion: Use of the VCC is associated with greater sensitivity to detect glaucoma and greater correlation with visual function when compared with the FCC and FCCm. The FCCm correction improves agreement between FCC and VCC RNFLT measurements. FCCm improves sensitivity for diagnosis and correlation with visual function, but not to the level of the VCC.
Keywords: 432 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • 484 nerve fiber layer • 624 visual fields