ZusammenfassungIn this proposal, we aim to develop standardizing and normalizing methods for MRI-based radiomic features of the whole body in the present large dataset provided by the German National Cohort, with the objective to develop a MRI radiomics atlas. This will provide a standard range of MRI radiomic features of organs, tissue group and imaging sequences. In addition, we will put special focus on standardizing and normalizing methods for MRI radiomics of the renal parenchyma in order to establish a methodology for MRI based tissue analysis with regard to kidney function (performed in a second step). Here, we will develop and evaluate machine learning algorithms for automatic segmentation of the kidneys, automatic feature extraction for standardizing and normalizing renal parenchyma and detection of features with prognostic value with regard e.g. to age or kidney function.

Schlüsselwörter-

EinrichtungenLMU, LMU München, Maastricht University, Klinikum der Universität München