Abstract
Purpose: :
To compare the ability of GDxVCC (scanning laser polarimetry), Stratus OCT (optical coherence tomography), HRT II (confocal scanning laser ophthalmoscope), RTA (laser optical cross–sectioning) and Discam (stereo planimetry) to discriminate between healthy eyes and eyes with glaucomatous visual field (VF) loss.
Methods: :
46 patients with glaucomatous VF loss and 34 healthy subjects were recruited. Recruitment was abandoned at these numbers due to software/hardware revisions of two devices. Individuals underwent imaging with GDx VCC, HRT II, Stratus OCT (RNFL 3.4 and Fast Optic Disc protocols), RTA (Glaucoma Full protocol) and Discam (stereo planimetry). Subject selection criteria were based on VF (glaucomatous =≥2 contiguous points p < 0.01 or greater, or ≥3 contiguous points p < 0.05 loss or greater, or 10–dB difference across nasal horizontal midline at ≥2 adjacent points in total deviation plot), and intraocular pressure (normal < 21mmHg, glaucoma ≥ 21mmHg), but not optic nerve head (ONH) or nerve fibre layer appearance. Area under the receiver operating characteristic (ROC) curves and sensitivities at fixed specificities (95%) were calculated for each parameter of each device. Areas under the ROC for the ‘best’ parameter were compared.
Results: :
Mean (range) mean deviation of the visual field test result was –4.24 (–12.0 to 1.15) and –1.10 (–3.3 to 1.0) for patients and healthy subjects respectively. Areas under the ROC (sensitivity at 95% specificity) for the best parameter for each device follows: Stratus OCT RNFL 3.4 inferior thickness 0.941 (84.8), GDx NFI 0.907 (73.9), Stratus OCT Horizontal Integ Rim Width (rim area) 0.899 (56.5), HRT II vertical cup disc ratio 0.853 (63.0), RTA rim area 0.840 (39.0), Discam vertical cup disc ratio 0.787 (40.9), RTA Peri–Foveal Minimum Thickness (µm) 0.665 (24.4). No statistically significant differences were found between Stratus OCT RNFL inf thickness, GDx NFI, Stratus OCT rim area. A statistically significant difference (p < 0.05) was found between Stratus OCT RNFL inf thickness and HRT II vertical cup disc ratio, RTA rim area, RTA Peri–Foveal Minimum Thickness and Discam vertical cup disc ratio.
Conclusions: :
The intended study power was not achieved because of device revision. However, some differences between best parameters were observed even at this sample size. RNFL parameters performed better than ONH parameters. Performances reported agree with previous reports by other groups.11 Medeiros et al. Arch Ophthalmol. 2004 Jun;122(6):827–37
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • optic disc • nerve fiber layer