Abstract

Background/Aim: To compare the ability of confocal scanning laser tomography (CSLT), scanning laser polarimetry (SLP) and optical coherence tomography (OCT) in recognizing localized retinal nerve fiber layer (RNFL) defects.

Methods: 51 eyes from 43 patients with glaucoma were identified by two observers as having RNFL defects visible on optic disc photographs. 51 eyes of 32 normal subjects were used as controls. Three masked observers evaluated CSLT, SLP and OCT images to determine subjectively the presence of localized RNFL defects.

Results: Interobserver agreement was highest with OCT, followed by SLP and CSLT (mean kappa: 0.83, 0.69 and 0.64 respectively). RNFL defects were identified in 58.8% of CSLT, 66.7% of SLP and 54.9% of OCT (P = 0.02 between SLP and OCT) by at least two observers. In the controls 94.1% of CSLT, 84.3% of SLP and 94.1% of OCT scans respectively were normal (P=0.02 between CSLT and SLP, and SLP and OCT).

Conclusion: Approximately 20-40% of localized RNFL defects identified by colour optic disc photographs are not detected by CSLT, SPL or OCT. SLP showed a higher number of false positive results than the other techniques, but also had a higher proportion of correctly identified RNFL defects in the glaucoma population.