The use of artificial intelligence for interpretation of x-ray images
| Reference number | |
| Coordinator | Danderyds sjukhus AB |
| Funding from Vinnova | SEK 532 500 |
| Project duration | May 2019 - December 2020 |
| Status | Completed |
| Venture | Medtech4Health innovators |
Purpose and goal
The purpose of the project was to develop a product that is integrated with the hospital´s PACS system and that meets all eligibility requirements for a class of IIb CE labeling of medical devices.
Expected results and effects
During the project we deployed and evaluated the interpretation of radiographs with use of artificial intelligence. During the evaluation phase, we have come to the conclusion that what is most appreciated by clinics is a detailed classification of injuries. It was perceived that a pure detection of damage, i.e. fracture or not, brought little added value while a classification of fractures by 50 different subgroups brought a new dimension of knowledge that would otherwise have been used mainly for research purposes. During the project, we have therefore expanded the categories.
Planned approach and implementation
We have implemented the system at Danderyds Sjukhus AB. The evaluation has been based on a small group of radiologists who work in their clinical everyday life with primary examination of orthopedic radiographs.