Verknüpfte ProjekteNAKO-709

ZusammenfassungMorphologic changes of the large thoracic and abdominal blood vessels commonly occur in pulmonary and cardiovascular disease. Detection of these changes by medical imaging can provide valuable prognostic or diagnostic information. Large epidemiological MR studies, such as the NAKO Magnetic Resonance Imaging (MRI) study, offer the unique possibility to study these vascular changes in a large population by non contrast-enhanced MR angiography (NE-MRA). The overwhelming amounts of data acquired in these studies however demand an automated approach for quantitative analysis. The aim of this proposed study is two-fold. 1) An algorithm will be implemented for automated segmentation of large vessels in whole body MR data based on convolutional neural networks. 2) The significance of changes in vascular configuration and morphology in the context of cardiovascular and pulmonary disease will be studied. We hypothesize that cardiovascular risk factors and existent cardiac or pulmonary diseases are associated with concurrent phenotypic changes of large vessels.

Schlüsselwörter-

EinrichtungenUniversitätsklinikum Tübingen, University of Freiburg, DKFZ Heidelberg, University of Wisonsin - Madison