BACKGROUND: Photographic screening for neovascular age-related macular degeneration (AMD) is not commonly employed because the prevalence of treatable disease is low and fluorescein angiography is considered necessary for the diagnosis of this form of AMD. However, there may be a role for colour retinal imaging in assisting with the diagnosis and triage of subjects with neovascular AMD. The purpose of this study was to evaluate the utility of colour fundus photographs for identifying subjects with potentially treatable neovascular AMD. METHODS: A total of 74 stereo pairs of Kodachrome colour slides of subjects with AMD were evaluated (i) nonstereoscopically, (ii) stereoscopically, and (iii) stereoscopically with visual acuity and visual symptom data. Two retina specialists read the images to identify active exudative lesions. RESULTS: The kappa statistic comparing the retinal specialists diagnosis of treatable neovascular AMD from color slides was excellent. The sensitivity and specificity of nonstereo images for the appropriate categorization of lesions was 0.95 and 0.90 respectively. The evaluation of stereo pairs was more sensitive, but less specific, 0.98, 0.83, as was the evaluation of stereo-pairs with clinical histories and visual acuities, 1.00, 0.77. CONCLUSIONS: The evaluation of colour images for subjects with suspected exudative macular degeneration can be diagnostic for neovascular AMD and may expedite the appropriate referral of patients for more timely angiography and treatment. Incorporating more clinical information for the image evaluators ((i) stereo image pairs and/or (ii) presenting symptomatology and visual acuity data) led to a decrease in the false-negative rate, but also decreased the screening specificity.
BACKGROUND: Photographic screening for neovascular age-related macular degeneration (AMD) is not commonly employed because the prevalence of treatable disease is low and fluorescein angiography is considered necessary for the diagnosis of this form of AMD. However, there may be a role for colour retinal imaging in assisting with the diagnosis and triage of subjects with neovascular AMD. The purpose of this study was to evaluate the utility of colour fundus photographs for identifying subjects with potentially treatable neovascular AMD. METHODS: A total of 74 stereo pairs of Kodachrome colour slides of subjects with AMD were evaluated (i) nonstereoscopically, (ii) stereoscopically, and (iii) stereoscopically with visual acuity and visual symptom data. Two retina specialists read the images to identify active exudative lesions. RESULTS: The kappa statistic comparing the retinal specialists diagnosis of treatable neovascular AMD from color slides was excellent. The sensitivity and specificity of nonstereo images for the appropriate categorization of lesions was 0.95 and 0.90 respectively. The evaluation of stereo pairs was more sensitive, but less specific, 0.98, 0.83, as was the evaluation of stereo-pairs with clinical histories and visual acuities, 1.00, 0.77. CONCLUSIONS: The evaluation of colour images for subjects with suspected exudative macular degeneration can be diagnostic for neovascular AMD and may expedite the appropriate referral of patients for more timely angiography and treatment. Incorporating more clinical information for the image evaluators ((i) stereo image pairs and/or (ii) presenting symptomatology and visual acuity data) led to a decrease in the false-negative rate, but also decreased the screening specificity.