Battling cancer with AI
Cancer arises from genetically driven errors in cell division and growth processes. Simultaneously, different processes can be affected, resulting in functionally diverse tumor cells. Such heterogeneous tumors consist of cells with distinct genetic profiles. To effectively treat them as a whole, a single therapeutic approach is insufficient. Instead, the characteristics of all different tumor cells must be considered to combat the entire tumor.
Prof. Frederick Klauschen and his team of physicians and scientists at the Pathological Institute of the LMU employ imaging techniques and -omics analyses to characterize tumors in terms of structure, proteomics (the entirety of contained proteins), genomics (DNA and RNA), and immune behavior. They develop AI algorithms that make sense of the immense quantity and complexity of these biological features to help understand how cancer cells function.
Furthermore, AI algorithms capable of automatically evaluating histopathological slides have significant clinical application potential. If they are allowed to be used as either assistive tools or routine tools in the future, it could address the severe shortage of radiological expertise.
