This project will turn promising methods based on artificial intelligence into products that will improve survival and quality of life for cancer patients. Cancer diagnostics and pathology are essential to provide the best treatment for the patient, but are facing three major challenges: 1) there is a lack of pathologists; 2) intra- and interobserver variation among pathologists are significant, potentially causing suboptimal treatment; 3) tumour heterogeneity implies that single samples from tumours often is insufficient to properly assess relevant tumour characteristics, causing inaccurate predictions. The DoMore! Lighthouse Project, funded by the Norwegian Research Council in 2016-2021, addresses these challenges by implementing methods for in silico pathology, i.e. automated analyses of pathology samples scanned with high-resolution scanners. A neural network-based method for automated prediction of patient outcome in colorectal cancer has been validated in a cohort of 1122 patients, providing large and treatment-relevant separation of patients with good and poor prognosis. Similar systems are under development for lung and prostate cancer. Additional inventions include a mitotic cells counter, with high counts indicating worse prognosis. Furthermore, we have initiated a project on automatic histological classification of bowel polyps, relevant to countries with bowel cancer screening programs and limited pathology resources. The challenges intended to be met by this project is to further support a planned prospective clinical trial, to implement a standalone module for the colorectal cancer classifier, to develop image standardisation modules for scanner types and tissue samples, to develop a prototype decision support system integrating information from the new markers with existing information and to ensure appropriate dissemination of results to primary users, such as pathologists and oncologists, to build confidence in the methods.
Project leader: Håvard E. Danielsen
Category: Helse Sør-Øst RHF
Institution: OSLO UNIVERSITETSSYKEHUS HF