Effects of treatment on brain tissue classification with serial MRI-based ISODATA cluster analysis in an experimental subarachnoid hemorrhage model

Introduction: Multiparametric MRI of different aspects of cerebral ischemic lesions may improve diagnosis of patients with cerebrovascular injury [1]. Nevertheless, spatial and temporal variations in the extent of various pathophysiological processes, complicates straightforward characterization of brain tissue status. In recent years, unsupervised algorithms have been developed to voxel-wise classify ischemic lesion areas in distinct categories based on multiparametric MRI data [1]. However, most of these analysis strategies have not incorporated dynamics of lesion changes into tissue classification. We have previously introduced a lesion classification approach that includes the temporal evolution of diffusion and T2 changes in an experimental stroke model [2]. In the current study we extend this approach in an experimental subarachnoid hemorrhage (SAH) model, where we evaluate lesion characteristics in a treatment and control group based on the progression of diffusion, T2 and perfusion parameters.