Purpose:
To determine the agreement between scanning laser polarimetry (SLP; GDxECC, Carl Zeiss Meditec Inc.) and spectral domain optical coherence tomography (SD-OCT;Spectralis,Heidelberg Engineering GmbH) for assessing the retinal nerve fibre layer (RNFL).
Methods:
Eighty five eyes (31 eyes of 28 glaucomatous patients, 30 healthy eyes of 17 subjects and 24 eyes of 12 individuals suspected of having glaucoma) underwent a complete ophthalmological examination,visual field testing by standard automated perimetry (Humphrey Field Analyzer, Carl Zeiss Meditec) and RNFL thickness assessment by SLP and SD-OCT. SLP and SD-OCT were performed in a random order. Average, Superior and Inferior RNFL thickness measurements were included in the analysis. SLP RNFL measurements were plotted against SD-OCT RNFL measurements with the intercept forced through zero; the slope of the relationship was used to re-scale the SLP measurements. Agreement was assessed with Bland-Altman plots.
Results:
Least square linear regression showed similar slopes for Avarage (y=1.93x, Figure1), Superior (y=1.94x) and Inferior RNFL (y=1.93x). Avarage RNFL mean difference was 0.2 µm and the limits of agreement ± 20.4µm (Figure2). Superior and Inferior RNFL mean differences and limits of agreement were 0.9µm ± 35,6 µm and -0.3µm ± 29.4µm.
Conclusions:
Agreement between SLP and SD-OCT is better for average than for sector RNFL thickness. SLP over-estimates SD-OCT at low values while SD-OCT over-estimates SLP at high values. This may results from a difference in the offset for the two devices.
Keywords: nerve fiber layer