Skip to content
Main menu hidden.
Published: 2021-10-06

Can AI Fight Cancer?

NEWS The Faculty of Medicine at Umeå University organized an AI day to open up for collaboration between different disciplines, and the big question was discussed: can AI help us fight cancer and other diseases?

Text: Petra Wester

Every third person in Sweden will receive a cancer diagnosis during their lifetime. Researchers know that the increase in cancer will also increase significantly, where the largest increase is expected to occur in countries that do not have the same assets and resources as the Western world when it comes to healthcare. Increasingly, AI is being used in medical research to more effectively detect and identify cancer.

Great Interest
More than a hundred participants were invited to the Faculty of Medicine's invitation to discuss the possibilities for the use of Artificial Intelligence in healthcare. Prominent researchers lectured and had panel discussions to promote the possibility of collaboration between different disciplines. One of today's Key-note speakers; Nigel Mongan, University of Nottingham, highlighted the sharp increase in cancer facing the world, especially in Asia and Africa, and the problem of helping patients in less resourceful countries, as well as the shortage of pathologists in these areas. Nigel Mongan therefore highlighted AI as an opportunity to diagnose and treat cancer in a cost-effective way, with benefits that extend globally.

With the help of large databases of images of cancer, an AI can, with the help of machine learning and deep neural networks, read and identify cancer, as well as facilitate pathologists' work in sorting, scaling and grading the cancer. In addition to detecting cancer, AI can also make treatment more accurate and beneficial to the patient. Cytostatics, as a treatment, are usually used to a greater extent than is necessary ¬– something that can give the patient negative consequences. With the help of AI, the researchers believe that a more tailored and effective treatment can be offered to the patient.

Difficulties with AI
Difficulties associated with AI were also highlighted. To create rewarding algorithms with good results, there are many parameters that must be included in the calculations. The panel therefore discussed the availability or rather the lack of data - to continue the development, larger amounts of data are needed than we have today. Lars Lindsköld, coordinator of the AI ​​Council in health care, highlighted the problem of AI and confidentiality, if we teach a machine to keep secrets, how do we know that the answer we get is correct? In addition, he also discussed the place of Europe and the Nordic countries in the digital future and AI, where he believed that success requires a high degree of transparency and the importance of security and trust in AI.
Collaboration between the disciplines
Erik Elmroth, professor of computer science and head of department, sat in one of the panel discussions. He is convinced that the best possible results are achieved by collaborating more across the various disciplines. One of the main problems, however, is that there are no conditions for structuring and managing the work within the organization. By creating cross-cutting projects where computer scientists collaborate with sociologists and medical scientists, we can achieve better results, faster. Several of today's lecturers highlighted the same problem and said that more interdisciplinary work is of great importance in the fight against cancer and other serious diseases.