Improving cancer care through digital insight and AI tools
The Institute for Cancer Genetics and Informatics (ICGI) is a research and innovation department at Oslo University Hospital, located at the Norwegian Radium Hospital. For more than 20 years, ICGI has brought together cancer- and AI-researchers, pathologists, laboratory engineers, software developers and technical experts to develop methods and digital tools that support cancer diagnostics, prognostication and clinical decision-making.
ICGI is recognised for our work within the field of deep learning and artificial intelligence in health care, digital pathology and advanced image analysis. Our AI-based tools are developed in close connection with clinical needs in hospitals, with the aim of supporting more precise diagnostics, better risk assessment and more efficient use of specialist resources.
Find current projects, research activities and news at Oslo University Hospital's website.
Introducing ICGI
Watch this video from 2021 for a short introduction to the Institute.
Research at the ICGI
The number of cancer patients is steadily increasing, and successful cancer treatment requires a correct diagnosis, as well as biomarkers that predict prognosis and response to treatment. Prostate cancer, for example, is a disease with more than 1,4 million new cases each year, where about half of these patients would have been much better off if their cancer had not been diagnosed and treated, whereas the other half completely depend on proper treatment to survive. We have little, and no exact, knowledge about what distinguishes indolent cancer from a lethal one, and this is the overall research question at ICGI.
The Institute combines the science of genetics and informatics to obtain new knowledge about cancer and to develop a new methodology for cancer diagnostics by focusing on the differences in DNA- and gene expression in tumour cells from indolent and lethal cancers. Whereas most genomic studies are performed on DNA or RNA isolated and extracted from the tumour tissue, due to the heterogenic nature of tumours, we prefer to perform our analysis directly in intact cells and tissues, which is where the informatics comes into play. We use artificial intelligence (AI) to analyse DNA, its organisation and expression through high-resolution image analysis of the cells and tissue of the tumour.
Although we belong to an organisation with a line structure and as such consist of three sections, our research is organised around our projects in a matrix structure. These projects are grouped according to their goal, which is either clinical use, mechanistic understanding, or new technology to support the first two groups.
Clinical decision support using Histotyping
This video explains our deep neural network algorithm for determining recurrence and survival prognosis in colorectal cancer.
Collaboration and cancer knowledge
ICGI’s national and international collaborations strengthen Oslo University Hospital’s role in cancer research and innovation. With broad expertise and strong digital infrastructure, the institute is an attractive partner for research groups and institutions developing, testing and implementing new approaches in cancer diagnostics and care.
A flat and project-oriented way of working allows colleagues with different expertise to collaborate closely across disciplines.
ICGI is also part of the European collaborations EUnetCCC and EU-CIP. In the latter, the institute contributes experience from Kreftlex.no, the quality-assured online cancer resource managed at ICGI. Kreftlex.no provides information about cancer, diagnostics, treatment and follow-up for cancer patients and their loved ones.