Purchase this article with an account.
Jonathan Denniss, Allison M McKendrick, Andrew Turpin; Individualised Structure-Function Mapping: Influence of Variability in Clinical Measurements of Anatomy on Mapping Resolution. Invest. Ophthalmol. Vis. Sci. 2014;55(13):966.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We have published a model that maps the visual field to 1° optic nerve head (ONH) sectors according to individual anatomy (Denniss et al, IOVS 2012: 6981-6990). Errors in clinical measurement of input anatomical parameters will cause variability in mapping. We aimed to quantify this variability in order to determine minimum feasible ONH sector sizes.
Variability in fovea-ONH distance and angle measurements made from retinal nerve fibre layer optical coherence tomography was estimated from 10 repeat scans of 10 healthy participants. Errors in estimating axial length from refractive error were determined from the 95% prediction interval of a linear fit to published population data. The model generated structure-function maps for 24-2 visual field locations using simulated anatomical parameters under conditions where axial length was either measured accurately or estimated from refractive error (axial length 23 to 26mm or refractive error -6.00 to +4.00D, fovea-ONH distance 4.0 to 5.0mm and angle -15 to 3°). For each unique parameter set (n=210), 200 maps were generated, each using parameters sampled from the measurement/estimation error distributions. Mapped 1° ONH sectors at each visual field location from each parameter set were normalised to difference from their mean, and variability (90% ranges) in normalised mapped sectors was calculated.
Standard deviations of differences from their mean of repeat measurements of fovea-ONH distance and angle were 61µm and 0.97° respectively. Neither measure varied systematically with its mean (Spearmans’s rho=0.26, p=0.47 for distance, rho=-0.31, p=0.39 for angle). Across the visual field, variability (90% ranges) in normalised mapped sectors ranged from 3 to 18° when axial length was accurate, and from 6 to 32° when axial length was estimated from refractive error. Variability in mapping across the visual field was commensurate with the influence of axial length and ONH position on the model as reported previously.
For general clinical use with 24-2 visual fields, our data suggest that 30° ONH sectors capture the likely effects of measurement error in anatomical parameters. For research purposes minimum ONH sector size varies across the visual field and should reflect the 90% ranges reported.
This PDF is available to Subscribers Only