KI researchers receive award for AI-based image analysis in breast cancer

The Athena Prize is awarded to Stratipath, which is made up of researchers, clinicians and developers from, among others, the Karolinska Institutet, Karolinska University Hospital and MedtechLabs, an interdisciplinary center for medical technology research. KI researchers who participated are Matthias RantalninAssociate Professor in the Department of Medical Epidemiology and Biostatistics Johann HartmannProfessor in the Department of Oncology and Pathology and a pathologist at Karolinska University Hospital.

The purpose of the research, through AI-based image diagnostics, was to partially increase the quality and speed of cancer diagnosis, and to offer a cost-effective alternative to the molecular analyzes that are increasingly used today in accurate diagnosis to determine treatment. The patient should receive it based on the severity of the tumor.

Our starting point was that the corresponding prognostic assessment could be made directly from the microscopic images because they contain a lot of information, says Johann Hartmann.

The research focused on breast cancer, where a pathologist’s evaluation plays a critical role in the patient’s treatment.

This solution for accurate diagnosis of breast cancer is based on the application of artificial intelligence (deep learning), where computer models are trained on very large data sets to capture patterns in tissue images that relate to the risk of disease recurrence. Models have the ability to see and measure these patterns in a way that the human eye can’t handle, says Matthias Rantalinen.

The results were published in the scientific journal Annals of Oncology in January of this year.

As an academic researcher, it is of course exciting to be able to contribute by taking research findings all the way to clinical implementation with the potential to improve cancer care, adds Matthias Rantalnin.

The award amount of SEK 100,000 will be used to stimulate and increase interaction between the academic research groups and the team at Stratipath.