advertisement
Abstract #53889 Published in IGR 15-2
Glaucoma classification based on visual pathway analysis using diffusion tensor imaging
El-Rafei A; Engelhorn T; Wärntges S; Dörfler A; Hornegger J; Michelson GMagnetic Resonance Imaging 2013; 31: 1081-1091
Most of the existing methods for diagnosing glaucoma analyze the eye with a main focus on the retina, despite the transsynaptic nature of the fiber degeneration caused by glaucoma. Thus, they ignore a significant part of the visual system represented by the visual pathway in the brain. The advances in neuroimaging, especially diffusion tensor imaging (DTI), enable the identification and characterization of white matter fibers. In this work, we propose a system based on DTI analysis of the visual pathway fibers in the optic radiation for detecting and discriminating different glaucoma entities. The optic radiation is identified semi-automatically. DTI provides information about the fiber orientation as well as a set of derived parameters describing the degree of diffusion anisotropy and diffusivity. Features for each DTI derived measure are extracted from a specified region of interest on the optic radiation. The features are grouped into three sets: Histogram, co-occurrence matrices, and Laws features. For feature selection, the features are ranked using a support vector machine classifier. The highest ranked features are used for classification. A support vector machine classifier is used for classification in a 10-fold cross validation setup. The system is applied to three age-matched subjects' categories containing 27 healthy, 39 primary open angle glaucoma (POAG), and 18 normal tension glaucoma (NTG) subjects. The discrimination accuracy between healthy and glaucoma (POAG and NTG) subjects is 94.1% with an area under the ROC of 0.97. Classification accuracy of 92.4% is obtained for the normal and the POAG groups while it increased to 100% in case of healthy and NTG groups. In addition, the system could differentiate between glaucoma types (POAG and NTG) with an accuracy of 98.3%. A complementary analysis was performed to estimate the selection bias in the obtained accuracy. The bias ranged from 10% to 20% depending on the group pair under consideration. The classification results indicate the high performance of the system compared to retina-based glaucoma detection systems. The proposed approach utilizes visual pathway analysis rather than the conventional eye analysis which presents a new trend in glaucoma detection. Analyzing the entire visual system could provide significant information that can improve the glaucoma examination flow and treatment.
Pattern Recognition Lab, Department of Computer Science, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany; Erlangen Graduate School in Advanced Optical Technologies (SAOT), Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany. Electronic address: ahmed.el-rafei@informatik.uni-erlangen.de.
Full articleClassification:
2.16 Chiasma and retrochiasmal central nervous system (Part of: 2 Anatomical structures in glaucoma)
6.30 Other (Part of: 6 Clinical examination methods)
