Microsatellite instability determines whether patients with solid tumors respond exceptionally well to immunotherapy. However, in clinical practice, not every patient is tested for MSI, because this requires additional genetic or immunohistochemical tests.
In this theme we will develop deep learning algorithms that can predict MSI or other mutational processes such as defects in DNA repair mechanisms, directly from histopathology slides. This approach has the potential to provide immuno- and targeted therapy to a much broader subset of patients with cancer. We will train and validate this on samples with Whole genome sequencing results and will be validated and applied to ongoing daily clinical MSI and genomic testing of cancer patients.
