Using Artificial Intelligence and Deep Machine Learning to Improve Treatment and Diagnosis of Prostate Cancer
Prof Clare Verrill is a pathologist at the Nuffield Dept of Surgical Sciences, and holds an honorary contract with the Oxford University Hospitals NHS Foundation Trust as a consultant pathologist, specialising in urological pathology. Clare’s primary interests are testis and prostate. Within the Verrill Pathology Group research is focused on digital pathology and image analysis. Clare holds a variety of local and national roles including: Thames Valley Supraregional Lead for germ cell tumour pathology (since 2011); Chief Investigator of Oxford Radcliffe Biobank (since 2015); RCPath Thames Valley Regional College Advisor (Since June 2017) and Co- lead for Testicular Genomic Clinical Interpretation Partnership (GeCIP) for 100,000 Genomes Project (Genomics England) (since 2015). She is also the National Cancer Research Institute (NCRI) Cellular Molecular Pathology Initiative (CM-Path) Technology and Informatics Workstream Lead.
By applying Artificial Intelligence (AI) in the cellular pathology setting, Clare and her team hope to deliver benefits for NHS patients in terms of efficiency, accuracy and quality of pathology assessment. Initially, the team hope to develop digital systems that impact clinical care pathways through improving existing workflow efficiency (i.e. potentially reducing turn-around times and making cost savings). In the long term the team hope to develop novel diagnostic strategies relying on machine learning algorithms reliant on tissue morphology features not obvious to a human observer.
One example of Clare’s work is the collaboration with the Finnish Institute of Molecular Medicine (FIMM), which has world leading expertise in digital pathology and image analysis. An image analysis algorithm has been developed, using deep machine learning, which can assess and count tumour infiltrating lymphocytes in testicular germ cell tumours on H&E sections. This helps to avoid the problems of inter-observer variability and subjectivity with pathologist assessment.
Within Oxford Clare collaborates with the Institute of Biomedical Engineering (Prof Jens Rittscher), the Department of Oncology (Prof Andrew Protheroe), the Big Data Institute (Nuffield Department of Medicine) (Dr David Wedge).
National collaborators are Prof Johan de Bono and Dr Clare Turnbull at the Institute of Cancer Research, Dr Matthew J Murray (University of Cambridge), and VisioPharm.