Automated analysis of normal and glaucomatous optic nerve head topography images

PURPOSE. To classify images of optic nerve head (ONH) topography obtained b y scanning laser ophthalmoscopy as normal or glaucomatous without prior man ual outlining of die optic disc. METHODS. The shape of the ONH was modeled by a smooth two-dimensional surfa ce with a shape described by 10 free parameters. Parameters were adjusted b y least-squares fitting to give the best fit of the model to the image. The se parameters, plus others derived from the image using the model as a basi s, were used to discriminate between normal and abnormal images. The method was tested by applying it to ONH topography images, obtained with the Heid elberg Retina Tomography, from 100 normal volunteers and 100 patients with glaucomatous visual field damage. RESULTS. Many of the parameters derived from the fits differed significantl y between normal and glaucomatous ONH images. They included the degree of s urface curvature of the disc region surrounding the cup, the steepness of t he cup walls, the goodness-of-fit of the model to the image in the cup regi on, and measures of cup width and cup depth. The statistics of the paramete rs were analyzed and were used to construct a classifier that gave the prob ability, P(G), that each image came from the glaucoma population. Images we re classified as abnormal if P(G) > 0.5. The probabilities assigned to each image were in most cases close to 0 (normal) or 1 (abnormal). Eighty-seven percent of the sample was confidently classified with P(G) < 0.3 or P(G) > 0.7. Within this group, the overall classification accuracy was 92%. The o verall accuracy of the method (the mean of sensitivity and specificity, whi ch were similar) in the whole sample was 89%. CONCLUSIONS. ONH images can be classified objectively and dependably by an automated procedure that does not require prior manual outlining of disc bo undaries.