Abstract
Abstract: :
Purpose: Genetic studies in age–related macular degeneration (AMD) often require retinal photography of relatives which can only be acquired in the community. We have previously reported on the diagnostic accuracy of the Nidek 200D hand held camera in comparison to the TOPCON TRC50EX. Our data demonstrated a high degree of agreement except for the diagnoses of small lesions such as microaneurysms and small drusen. The aim of this study is to determine if red–free images allow improved diagnostic accuracy in the detection of drusen in comparison with standard colour images taken with the NM200D hand held camera. Methods: We selected images from 60 consecutive relatives of patients who had indicated a positive family history for AMD. Images were taken in the community using the NM200D hand held fundus camera through a dilated pupil in digital format. The images, originally saved in Nidek format (.hfc extension), were converted to TIFF format using the Nidek software. The images were then saved to a CD. Using Adobe Photoshop the images were then converted to red–free format using the channel mixer adjustment and applying 100% of the green channel. No other enhancements were made to the images. These were then saved to a separate CD. The colour and red free images from each patient were anonymised, saved in a coded fashion and in random order and were graded on separate days by a retinal specialist to ensure that grading was carried out in a masked fashion. Results: Images were graded for presence and size of drusen. In the <63um category drusen were identified in 22.22% of colour images and 45.12% of red–free images. In the >63um&<125um category drusen were identified in 31.11% of colour images and 32.93% of red–free images. In the >125um category drusen were identified in 46.67% of colour and 21.95% of red–free images. Overall the rate of detection of drusen was 45% greater with the red–free images. Conclusions: The results confirm that smaller drusen <63um are more easily detected in the red–free image. This technique will be helpful in grading images from relative of AMD patients who often have subtle findings.
Keywords: imaging/image analysis: clinical • drusen • age-related macular degeneration