Purchase this article with an account.
M. V. Boland, N. G. Strouthidis, D. F. Garway-Heath, H. A. Quigley; A Method for Comparing Quantitative Measures of Glaucoma Progression Using Longitudinal Data. Invest. Ophthalmol. Vis. Sci. 2010;51(13):4914.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To facilitate comparison of putative measures of disease progression in glaucoma, we propose a scale-independent means of defining the signal-to-noise ratio (SNR).
The proposed method assumes that, over the time frame of most studies, changes in measures of glaucoma are approximately linear. The first step is to estimate the real change in the parameter using linear regression. The SNR is calculated as shown in the equation below where f(t) is the predicted value at time t, fmean is the average of all points in the regression function and y(t) is the value of the measured parameter. The signal component (numerator) is therefore a measure of the change in the regression function over time and the noise component (denominator) measures the deviation of the actual measurements from the predicted values. To test this method, we used data collected from subjects followed with repeated visual field testing and optic nerve imaging. This cohort included 193 subjects with ocular hypertension and 19 matched controls. We calculated the SNR for the HFA mean deviation, the HRT rim area, and for our combined structure function index (SFI).
The SNR values for field mean deviation, rim area, and SFI were 1.53, 1.76 and 1.63. None of the values was statistically different from the others. The subjects found to be progressing by the various measures are not the same, however. Of 41 subjects who progressed by any measure, only 4 were progressive by all three and 30 by only one.
The fact that mean deviation, rim area, and the SFI have similar SNR values suggests that they will be similarly able to detect progression, at least in this group of ocular hypertensives with at most early disease. A scale independent measure of SNR like this one will be helpful in comparing proposed measures of glaucoma progression. Those parameters with larger SNR values will be able to detect progression more quickly.
This PDF is available to Subscribers Only