| |
DOG Congress Home
Invitation
Organization, Deadlines
Overview of the Congress
Scientific Programme
Opening Ceremony
Ceremony 150 Years of DOG
Thursday, 20.September
Friday, 21.September
Saturday, 22.September
Sunday, 23.September
Poster Sessions
Symposia
Courses
Satellite Programme
Information
Social Programme
Sponsors, Exhibitors
DOG Homepage
|
|
AbstractSO.15.04 Combined analysis of frequency doubling perimetry (FDT) and optic disc measurements (HRT) for glaucoma screening Horn F., Jertila M., Mardin C., Michelson G., Jünemann A., Adler W., Lämmer R. Erlangen Glaucoma Research SFB 539, University Erlangen Objective: Glaucoma detection in open angle glaucomas (OAG) using a structural and a functional method in two subject groups. Methods: To study the diagnostic value of combined analysis, measurements were performed in two independent samples. A training sample included participants of the Erlangen glaucoma registry: 81 preperimetric OAG, 100 perimetric OAG, and 85 control subjects. An independent second sample was recruited from the Erlangen university eye hospital (39 preperimetric, 65 perimetric OAG and 116 healthy subjects). All subjects had conventional perimetry, evaluation of fundus photographs, FDT screening and scanning of the optic disc (HRT) with disc size correction analogue to Moorfield Regression Analysis (MRA). FDT evaluation used a published method with casewise calculation of a FDT-score including all missed localized probability levels. Age: 35-66 years. Diagnostic value of procedures has been judged from the area under the ROC-curve. In addition to single parameters validity, the value of a discriminant function has been calculated for all individuals with the score delivered by FDT and global MRA. The same formula as gained in the training-sample has been applied in the test-sample and was compared to machine-learning algorithms including other parameter of the devices. Results: In the training data set the areas under the ROC-curves for FDT, MRA and the combined analysis were 84, 81 and 88% compared to 83, 73 and 86% in the independent test set. In perimetric patients, area under ROC is larger for the FDT-score than for MRA. In the preperimetric group, however, ROC of the MRA is partly above the FDT-score. A combination of both methods with a simple formula is superior to single tests in both samples. Sectoral information increase sensitivity. Conclusions: The present study indicates that combination of different techniques, which include structural and functional methods, can improve validity of the procedures.
|
|