Purchase this article with an account.
Ivania Pereira, Stephan Holzer, Hemma Resch, Barbara Kiss, Georg Fischer, Clemens Vass; Intersubject correlation between circumpapillary distribution of retinal nerve fiber layer and retinal vessels: comparison between optical coherence tomography and scanning laser polarimetry. Invest. Ophthalmol. Vis. Sci. 2013;54(15):4800.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To assess circumpapillary retinal vessel thickness (cRVT) profiles and correlate them with retinal nerve fiber layer (RNFL) thickness measured by high resolution optical coherence tomography (HR-OCT) and scanning laser polarimetry (SLP).
A sample of 106 healthy volunteers underwent complete ophthalmologic examination, including HR-OCT scanning (Cirrus® Carl Zeiss Meditec Inc.) and SLP (GDx-VCC® and GDx-ECC ® Carl Zeiss Meditec Inc.). A proprietary software was developed in MATLAB® (Version R2009b, The Mathworks Inc.) for individual manual assessment of the optic disc (OD) border and cRVT, using the SLO image from Cirrus HR-OCT centered in the OD. For each measurable vessel, we established a line connecting its center point at the OD border with the OD center. The angle between this line and a horizontal line passing through OD center was determined. Individual cRVT measurements were convoluted with a Gaussian window, generating a cRVT profile dependent on the number of sectors measured in each device (256 sectors for HR-OCT or 64 sectors for both GDxVCC and GDxECC). Intersubject variability was determined using MATLAB.
The median values for intersubject correlation between cRVT and RNFL thickness and the numbers of significant sectors as measured with Cirrus HR-OCT, GDx-ECC and GDx-VCC were 0.255 (186/256 sectors statistically significant), 0.192 (32/64) and 0.195 (33/64). 90% percentiles of the intersubject correlation were 0.400, 0.355 and 0.336, respectively.
Both SLP methods (GDx-VCC and GDx-ECC) yield comparable results regarding the intersubject correlation of the cRVT profile and the RNFL with about 50% of the sectors showing significant correlation. However, regarding the HR-OCT, our data showed a tendency towards a better correlation which was statistically significant in 70% of the sectors. Based on our results, up to 25% of the interindividual variation of the RNFL thickness might be explained by the variation of the cRVT.
This PDF is available to Subscribers Only