Meeting Banner
Abstract #4387

Hierarchical non-negative matrix factorization using multi-parametric MRI to assess tumor heterogeneity within gliomas.

Nicolas Sauwen 1,2 , Diana Sima 1,2 , Sofie Van Cauter 3 , Jelle Veraart 4,5 , Alexander Leemans 6 , Frederik Maes 1,2 , Uwe Himmelreich 7 , and Sabine Van Huffel 1,2

1 Department of Electrical Engineering (ESAT), KU Leuven, Leuven, Belgium, 2 iMinds Medical IT, Leuven, Belgium, 3 Department of Radiology, University Hospitals of Leuven, Leuven, Belgium, 4 iMinds Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium, 5 Center for Biomedical Imaging, Department of Radiology, New York University Langone Medical Center, New York, NY, United States, 6 Image Sciences Institute, University Medical Center Utrecht, Utrecht University, Utrecht, Netherlands, 7 Biomedical MRI/MoSAIC, Department of Imaging and Pathology, KU Leuven, Leuven, Belgium

Tissue characterization within gliomas is challenging due to the co-existence of several intra-tumoral tissue types and the high spatial heterogeneity in high-grade gliomas. An accurate and reproducible method for brain tumor characterization and the detection of relevant tumor substructures could be of great added value for tumor diagnosis, treatment planning and follow-up. In this study, a hierarchical non-negative matrix factorization (hNMF) technique is applied to multi-parametric MRI data of 24 glioma patients. hNMF can be applied on a patient-by-patient basis, it does not require large training datasets and it provides a more refined voxelwise tissue characterization compared to binary classification.

This abstract and the presentation materials are available to members only; a login is required.

Join Here