Purchase this article with an account.
I. Kovacs, Z. Szepessy, J. Feher; Assessment of Retinal Alterations With Automatic Retinal Image Anyalsis in Early AMD Patients. Invest. Ophthalmol. Vis. Sci. 2007;48(13):5099.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the correlation between retinal alterations and visual functions in early AMD patients using automatic image analysis with pixel-to-pixel resolution.
Retinal alterations were determined on digital fundus photographs of 186 eyes with a newly developed software which transforms data of the RGB image into ‘intensity’, ‘hue’ and ‘saturation’ values, and analyse the data in each pixel. The sum area of drusen and pigmentary abnormalities was analysed in two areas centered on the fovea (diameters of 1000 and 6000 µm) and was correlated with ETDRS visual acuity and foveal sensitivity. Grading was also performed according to the definitions of the International Classification and Grading System for AMD. To measure drusen size, circles with diameters of 63, 125, and 250 µm were drawn using a special software and the sum area of retinal alterations was determined.
: Results of automatic digital analysis correlated with measurement values according to ICGS (p<0.05). Visual acuity and foveal sensitivity correlated with retinal alterations of the 6000 µm area (p<0.05), but not the 1000 µm area (p>0.05). The presence of clinically visible retinal alterations in the foveola did not correlate with ETDRS visual acuity or foveal sensitivity (p>0.05).
Total area of retinal alterations in early AMD can be accurately measured with automated image analysis. Our findings suggest that visual impairment in early AMD depends on ophthalmoscopically undetectable dysfunction of the RPE/photoreceptor complex rather than the presence of drusen in the fovea. Visible retinal alterations of the 6000 µm diameter area presumably represent the clinically undetectable functional impairment of the fovea better, than that of the 1000 µm diameter area. The assesment of changes in fundus alterations and corresponding visual functions in AMD as a measurement of treatment efficacy can be improved using this software with a capability of pixel-to-pixel resolution in image analysis.
This PDF is available to Subscribers Only