Skip to primary content

MAGIC: AI for Breast Cancer Diagnostics

Women in Denmark between the ages of 50 and 69 are offered a mammogram every two years; an X-ray examination of the chest, where the purpose of the screening is to detect breast cancer in the early stages. The earlier treatment is initiated, the better the prognosis for survival.

Alle

PROJECT PERIOD

Start: January 2020
End: December 2023

Annually, approx. 70,000 screenings for breast cancer are conducted in the Region of Southern Denmark, where four X-rays are taken at each screening. Every single X-ray is assessed by two independent radiologists to ensure the highest quality. It is resource-intensive, and therefore there is a need for a clinical tool that can reduce the need for radiologists and at the same time maintain and/or increase quality.

AIM

The overall aim of the project was to develop and test an AI tool for diagnosing breast cancer to ensure earlier and more accurate diagnostics, which could lead to faster initiation of treatment – ​​and thus a better chance of survival and increased quality of life.

The two overall activities in the project were:

  • Automation of data extraction
    Data extraction for AI projects is essential, but time-consuming to do manually. Therefore the process must be time-optimised through automation of data extraction. This will benefit future AI validation projects as well as support future data infrastructure for clinical AI use.
  • Validation of AI tool for diagnostic imaging
    The project will test and validate AI to support clinical decisions when diagnosing breast cancer. AI has the potential to increase the quality of X-ray assessment and optimise available medical resources. The project is being tested at all imaging departments in the Region of Southern Denmark and includes both women from the regular screening programme (screening mammography) and women referred by their own doctor (clinical mammography).

Watch a video introducing the project.

PARTNERS

Ole Graumann, Head of Research, and Benjamin Schnack Rasmussen, Postdoc, from the research unit UNIFY at the Department of Radiology at Odense University Hospital instituted the project.

The project was a cooperation with the radiology departments at all hospitals in the Region of Southern Denmark.

Since the project involved artificial intelligence, it was also anchored at the Centre for Clinical Artificial Intelligence (CAI-X).

Thea Damkjær Syse

Thea Damkjær Syse

Project manager - On maternity leave

Odense University Hospital, Dept. of Clinical Development - Innovation, Research & HTA


(+45) 2323 7449
Benjamin S. Rasmussen

Benjamin S. Rasmussen

Associate Professor, MD (Radiologist), Head of Clinical Research at CAI-X

Odense University Hospital, Department of Radiology and Centre for Clinical AI (CAI-X)


(+45) 2434 1749 CAI-X
APPFWU02V