Priya Goel1, Matthias Karrasch2, Jan A. den Hollander3, James Macdowell Markert4, Louis Burt Nabors5, Narsimha Shastry Akella1
1Biomedical Engineering, The University of Alabama at Birmingham, Birmingham, AL, USA; 2Clinical Research and Development, MediGene AG, Munich, Germany; 3Medicine, The University of Alabama at Birmingham, Birmingham, AL, USA; 4Neurosurgery, The University of Alabama at Birmingham, Birmingham, AL; 5Neurology, The University of Alabama at Birmingham, Birmingham, AL, USA
Quantitative assessment of malignant brain tumor therapy response remains challenging due to limited in-vivo biomarkers that can differentiate infiltrative tumor from healthy brain tissue. Diffusion Tensor Imaging derived isotropy "p" and anisotropy "q" maps were plotted on a 2D feature space and partitioned to characterize constituent tissue using an adaptive fuzzy clustering algorithm. The resulting segmentation permits longitudinal assessment of early and subtle white matter changes around the tumor and quantitative analyses which is useful in therapy efficacy evaluation.