Purchase this article with an account.
P. J. Banerjee, G. Vafidis, K. Scanlon, V. Vora; Extended Nasal Field Imaging and its Effect on Grading When Screening for Diabetic Retinopathy: A Comparison With Screening Guidelines. Invest. Ophthalmol. Vis. Sci. 2007;48(13):1410.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The National Screening Programme for Diabetic Retinopathy recently issued guidelines to capture 2 fundus images for diabetic retinopathy screening: 1 fovea centred and 1 disc centred image. A field with the disc orientated at the temporal edge of the image increases the net gain in area viewed by approximately 10 fold. Previous research into optimal fields for fundus imaging shows conflicting results.
A prospective analysis was performed on104 patients selected at random attending our department over 2 months. A standard non-mydriatic fundus camera was used to capture three 45-degree fundus images:1 fovea centred, 1 disc centred and 1 disc at the temporal side of the image (nasal field). Images of 86 patients were deemed suitable for inclusion in the study (ungradable maculae images were excluded). Eyes were randomised into equal groups of left and right. Two accredited diabetic retinopathy graders, masked to the study, were asked to grade a total of 344 images presented as 172 pairs, according to the National Screening Framework criteria (i.e Ro, - R3 for retinopathy and Mo ,MNR or M1 for maculopathy). The pairs consisted of a macula centred image in addition to either a disc centred image or nasal field. The images were viewed on a 1600X1200 pixel accepted standard screen.
Considering Grader 1: The results showed excellent reliability with a Kappa value of 0.87. No retinopathy was found in 32.5% of disc centred images compared with 22% of nasal field images. Although of low statistical significance (Χ2 = 2.37 (p<0.2)) there is a clear trend.13% of nasal images were graded as referable retinopathy compared with 7% found using disc centred views.All discrepancies in retinopathy grading involved a more serious grade of retinopathy made in association with the nasal field compared with the disc centred view.Considering Grader 2: The results showed excellent reliability with a Kappa value of 0.83. Again a trend detecting more retinopathy was found (28% nasal field compared with 22% disc centred field). There was no significant difference between percentages of referable retinopathy graded.The concordance between the graders was fair (Kappa 0.61). There was no significant difference in the numbers of ungradable images.
The use of a predominantly nasal field image instead of a disc centred second image when screening for diabetic retinopathy may increase its detection and referral to secondary care. This pilot study shows interesting trends following which a larger study may confirm statistical significance.
This PDF is available to Subscribers Only