Quantification of intratumoural heterogeneity via machine-learning

In oncology, lower intratumoural heterogeneity is closely linked with the increased therapy efficacy, and can be partially characterized via tumor biopsies. In the article, the authors show that intratumoural heterogeneity can be characterized spatially via phenotype-specific, multi-view learning classifiers trained with data from dynamic positron emission tomography (PET) and multiparametric magnetic resonance imaging (MRI).

Our members and colleagues Prateek Katiyar, Johannes Schwenck, Leonie Frauenfeld, Mathew R. Divine, Vaibhav Agrawal, Ursula Kohlhofer, Sergios Gatidis, Roland Kontermann, Alfred Königsrainer, Leticia Quintanilla Martinez-Fend, Christian la Fougère, Bernhard Schölkopf, Bernd J. Pichler & Jonathan A. Disselhorst wrote a great article in Nature Biomedical Engineering. Check it out: https://www.nature.com/articles/s41551-023-01047-9.epdf?sharing_token=QQ6s0qmLiSYQZMU3IHWkHdRgN0jAjWel9jnR3ZoTv0N2Zf3KSNC3HxgfS6pYyzHkWEf5X4QN2aYfKiiSZWS-1U5LjMAc8yDqlngR4hi5ZQr07RxSAmsMKb-CRjUYO7G3LKKlxir9ncPUZCd_GLQT6yZ2Iu29-OzpI_tOS2LlPZY%3D