Purchase this article with an account.
U. Schiefer, L.M. Zangwill, P.A. Sample, W. Budde, J.B. Jonas, W. Lagrèze, J. Airaksinen, R. Vonthein, B. Selig, J. Paetzold; Evaluation of the Human Retinal Nerve Fibre Bundle (RNFB) Course by Fitting Trajectories to Aligned Digitised Fundus Images . Invest. Ophthalmol. Vis. Sci. 2005;46(13):615.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: (i) To extract and evaluate the RNFB course of human subjects, (ii) to fit RNFB trajectories, (iii) to analyse homogeneity. Methods: Fundus images of 46 eyes of 46 subjects (13 normals, 25 glaucomatous optic neuropathies, 2 myelatinated retinal nerve fibres, 2 optic disc swellings, 4 other pathologies) were digitised. The images were superimposed by translational shifting in order to centre the foveolar position and by rotational and zoom routines in order to align the centres of the optic discs. 1033 clearly visible RNFB courses were electronically traced with 9504 sampling points using a graphics software (CorelDraw 10.0, Corel Inc., Ottawa, CAN). Results of left eyes were mirrored along the vertical axis. The mean courses of RNFBs were fitted by natural cubic splines with 3 to 6 nodes heading for one of 36 sectors (10° each) of the optic disc. Homogeneity was analysed for each trajectory separately by counting the number of RNFBs crossing the related trajectory divided by the total number of RNFBs at a given location. This ratio is minimised by the fit. Results: On average, 21.4 RNFBs (range 2 to 97) could be traced on each image with a mean of 9.2 sampling points per RNFB (range 3 to 48). Superimposition of the electronically traced fibre segments resembles the typical RNFL course within 20° eccentricity. On average, 23% (range 7% to 42%) of the RNFBs crossed the related spline. Conclusions: There is a considerable uniformity of the RNFB course in humans, despite the inhomogeneity of the analysed patients. This consistency allows for mathematically modelling RNFB trajectories. This is an essential pre–requisite for assigning circumscribed glaucomatous visual field (VF) defects to the related RNFBs and to corresponding segments of the optic disc. The model should also substantiate local condensation of stimulus locations for increasing spatial resolution of VF examinations, as proposed for SCotoma–Oriented PErimetry (SCOPE).
This PDF is available to Subscribers Only