Purchase this article with an account.
E. J. Sabet-Peyman, B. Bertoni, J. Dougall, D. Eliott, M. Humayun, R. Pappuru, S. R. Sadda; Assessing Accuracy of Intraocular Measurements Made by Camera Systems Using Prosthetic Retinal Arrays. Invest. Ophthalmol. Vis. Sci. 2010;51(13):2284.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To assess the accuracy of retinal measurements provided by commonly used camera systems by using the ArgusTM I and II retinal implant arrays with standardized intraocular dimensions.
225 total measurements were obtained from five subjects with retinal prostheses, two with ArgusTM II implants (60 electrode array) and three with ArgusTM I (16 electrode array) implants positioned on the retinal surface, who had undergone standard fundus photography and angiography using camera systems (Zeiss FF450 with Escalon 6MP CCD for all cases and additionally a Topcon 50IX with 12MP CCD for one case). These camera systems had been certified for clinical trial use by multiple reading centers through the use of a model eye to calibrate measurements. The diameters of the circular electrodes in the respective arrays were measured and compared with their known values to calculate absolute and percent error. Reproducibility between visits and camera systems was assessed using images from different dates for both Argus I and II arrays.
For a particular visit for the Argus II eyes, the mean percentage error ranged from -11.9 to +6.4%, with a mean absolute percentage error for all visits of 6.6% (+5.2% +/- SD of 1.1% for subject1 and -8.0% +/- SD of 3.4% for subject 2) . Thus, measurements were overestimated in one Argus II subject, and underestimated in the other. The maximum difference in percentage error between two visits for a single subject was 2.2% (Argus II), suggesting good reproducibility between visits. Furthermore, one Argus I subject, imaged using both the Zeiss and Topcon systems, demonstrated a mean percentage error of only 2% between camera systems.
Although there are potential confounding factors that require further study, our data suggest that there may be variability in the accuracy of camera systems in measuring the size of structures on the retina. These may be due to differences between real and model eyes which may need to be considered when evaluating the results of quantitative fundus image data from clinical studies.
This PDF is available to Subscribers Only