Neurodegenerative and cerebrovascular diseases, and mental disorders, are the leading cause of physical and mental disability in the population of developed countries and thus have a huge and increasing socio-economic impact. Recent researches show that these diseases and disorders also affect the physical structure of human brain. Currently, magnetic resonance (MR) tomographic imaging is by far the most sensitive technique for visualizing normal and pathological brain structures, but is also being increasingly used for their quantification. This Thesis advances state-of-the-art in the field of automated brain MR image segmentation in several major aspects: (a) a novel method is proposed for estimation of a mixture model of normal brain structures that is robust to unbalanced samples with outliers; (b) a novel model of otherwise complex whole-brain multi-sequence MR intensities is proposed that employs spatial stratification of the complex model into several simplified models and their independent, robust estimation; (c) a method for segmenting normal and pathological brain structures based on locally-adaptive MR intensity model of normal-appearing brain structures is proposed. A quantitative and comparative evaluation of the proposed and several state-of-the-art methods for segmenting normal-appearing structures and white-matter lesions in MR images of multiple sclerosis patients, having different disease severity characterized by total lesion load, revealed that herein developed methodological contributions substantially improve the performance of segmentation of both normal appearing and pathological brain structures. Based on the observed accuracy, reliability and efficiency the proposed methods seem as good tools for the extraction of neuroimaging biomarkers.
F.35 Other
COBISS.SI-ID: 11167316The latest guidelines for treatment of patients with multiple sclerosis (MS) involve the identification and assessment of newly-appearing lesion and the increase in brain atrophy based on magnetic resonance (MR) imaging. Alternative to standard radiologic assessment of these often barely visible morphologic changes in the brain is the use of automated MR image analysis. Compared to the radiologic assessment, which is descriptive and qualitative, the automated image analysis yields quantitative measurements of important structures directly from the MR image. These measurements are very accurate and reliable, given that the quality of MR image is also high. The MR image acquisition should follow the current guidelines for MS, such that the MR study consists of 3D T1-weighted and 3D FLAIR image with isotropic 1 mm sampling, using the same scanner and the same protocols for a particular patient. The field of view should capture the whole brain and the skull. With MR images acquired in this way, the image-based measurements are highly accurate and reliable. We have developed a software-as-a-service platforms that enables a direct transfer of MR images to a remote server, where the images are processed and a report containing measurements of important brain structures (e.g. lesion volume, brain atrophy) is output as a result. The software is currently extensively validated on clinical MR images of patients with MS from two major clinical centres in Slovenia.
F.21 Development of new health/diagnostic methods/procedures