2020-NIHR-ACF-PES-Warwick-Platform Science and Bioinformatics 2
2020 NIHR Academic Clinical Fellowship in Priority Research Themes
HEE Local Office: West Midlands
Medical School: University of Warwick
Research Theme: Platform Science and Bioinformatics
Specialty Options: Histopathology
Plain English Summary
Aims and background of the research
Melanoma is the most lethal form of skin cancer and incidence rates in England have nearly doubled from 14 per 100,000 in 2000 to 26 in 2015 with the biggest rise being seen in the over 70 age group. As the population ages the disease burden can expect to increase. Management of the disease is dependent on accurate staging including lymph nodes.
The routine use of multiple sections and immunocytochemistry (A laboratory method that uses antibodies to check for certain antigens (markers) in a sample of cells) has improved performance. Current practice requires examining 20 sections at different levels from lymph nodes by a combination of stains, which takes up to 40 mins of pathologist reading time per lymph node.
This project aims to investigate the use of artificial intelligence to assist in the detection of metastatic disease in the biopsied tissue, to establish if this can improve pathologist performance and/or deliver savings on pathologist time and laboratory resource required to accurately stage metastatic spread of the disease.
Design and methods used
This project will be run at the UHCW NHS Trust Centre of Excellence for Digital Pathology. High throughput digital slide scanners routinely scan the entire histopathology workload of the trust allowing unprecedented access to digitised archive of cases extending back over 10 years.
The UHCW Pathology department has become the first one in the UK to 'go digital' for routine diagnostics after the publication of world's largest study for validation of digital pathology led by Snead. It has recently been shown that advanced deep learning methods, such as those developed in our lab, carry great potential to detect metastasis in breast cancer lymph node biopsies and that the algorithm's performance was comparable with an expert pathologist interpreting slides without time constraints.
The ACF appointed would be fully involved in this study with training in digital pathology and use of Artificial Intelligence techniques.
Patient and public involvement
We will be developing public patient involvement to develop methods of dissemination.
The ACF would be encouraged to present the data at relevant medical conferences. Write high impact publications and use data for fellowship applications.