Artificial Intelligence for HISTOpathological images towards personalised immunotherapy
Exploiting the body’s ability for an immune response against tumour cells is now a well-established strategy to treat cancers. However, identifying which patients would benefit most from immunotherapy is hard. Within this project, together with Ellogon AI, a start-up of AI researcher at the University of Amsterdam, artificial intelligence algorithms will be build to investigate the relationship between histopathological images, genomics and treatment response in patients treated with immunotherapy, to guide assist in stratifying patients for immunotherapy .
Immunotherapies are revolutionising the management of cancer patients by producing durable responses with minimal toxicities in a subset of patients. This suggests that a sensitive companion diagnostic test is required to select these patients. Recently a cancer-immunogram was proposed (Blank et al. 2016, doi:10.1126/science.aaf2834) for describing the different cancer-immune interactions in individual patients, and to guide treatment choice. In clinical practice however, quantifying these cancer-immune interactions requires additional and expensive genetic tests and a analysis of histopathological images by pathologists which is subjective and suffers from a significant intra-rater variability.
Having access to histopathological images of tumors of thousands of cancer patients treated with immunotherapy at The Netherlands Cancer Institute-Antoni van Leeuwenhoek and from Nationwide consortia capturing clinical, histopathology and genomics data we can objectively evaluate the contribution and correlation of histopathology imaging, genomic markers and clinical outcome. This creates an exceptional opportunity to deploy the latest advances in artificial intelligence (AI) to quantify cancer-immune interactions and correlate them to treatment outcome .
At the end of the project the goal is to have built AI-algorithms within digital pathology for biomarker discovery by integrating clinical, pathological and genetic factors with histopathology imaging analysis to make better treatment decisions, and develop a protype system ready for clinical validation in an operational environment.
