Verknüpfte ProjekteNAKO-722

ZusammenfassungEvery day, vast amounts of medical imaging data are acquired to confirm or rule out a suspected medical condition. Usually this is done in a subjective manner by the radiologist in charge. While the accuracy of this approach is limited due to time constraints and inter-reader variability - even more importantly - the majority of information embedded in the data remains unused because no efficient and reliable analysis strategies are available to extract it in an efficient and robust way. Thus, potentially valuable diagnostic and prognostic information to improve prevention and prognostication remains locked in the image. With recent advances in artificial intelligence a new possibility for high-throughput, quantitative image analysis has become available with the potential to change radiologic image interpretation in a fundamental way. Decoding and phenotyping of unused data from medical imaging has the potential to improve clinical decision making and precision medicine in an unprecedented way and introduce a paradigm shift in radiology by translating the field from a primarily subjective to a more objective and from primarily diagnostic to a more prognostic specialty. The proposed study will develop and validate quantitative artificial intelligence-based measures of organ/tissue phenotypes and aging from medical imaging data and will assess whether these measures improve prevention and prognostication of cardiometabolic disease and cancer.

Schlüsselwörter-

EinrichtungenRadiologie Freiburg, Universitätsklinikum Freiburg, Universitätsklinikum Freiburg, Klinik für Diagnostische und Interventionelle Radiologie, University of Freiburg, Uniklinik Freiburg - Klinik für Radiologie