[Evaluation of four new discriminant functions for HRT II in glaucoma diagnosis].

PURPOSE To assess the diagnostic usefulness of four linear discriminant functions for Heidelberg retina tomograph (HRT) II obtained in our population. METHODS 450 eyes of 450 patients were studied. Eyes were classified as those of healthy subjects (101), those with ocular hypertension (126), those suspected to have glaucoma (121) or those with glaucoma (103). Intraocular pressure, optic nerve head appearance in stereophotographs, and standard automated perimetry results were assessed. ANOVAs among the groups were calculated for all global parameters and the 4 formulas. Receiver operating characteristic (ROC) curves were plotted for the 4 multivariate functions designed in our hospital and the areas under the ROC curves were compared. Sensitivity at 80% and 90% fixed specificities were also calculated. RESULTS All functions discriminated well between healthy and glaucoma suspects, and between healthy and glaucomatous eyes. At 90% fixed specificity, sensitivities ranged from 54% to 60% in differentiating between healthy and glaucoma suspects, and from 65% to 68% in discriminating between control and glaucoma subjects. No differences were found between the areas under the ROC curves of these functions. CONCLUSIONS The evaluated linear discriminant functions increased the diagnostic ability of HRT II isolated parameters in detecting glaucoma. These functions performed better than the HRT-provided discriminant functions.