AI-based technology for wound analysis

Hard-to-heal wounds cause a lot of suffering and cost Swedish healthcare around SEK 4 billion every year. Thanks to new technology based on Artificial Intelligence, the size of such wounds can now be determined with great precision, which contributes to faster healing and less suffering for the patient.
Professor Folke Sjöberg examins a patient's wound Photo credit Emma Busk Winquist

Research resulted in more accurate assessments

AI technology in this area has its background in the research carried out at the Burn Centre at Linköping University Hospital in collaboration with Linköping University and others. This research initially focused on assessing the healing capacity of wounds caused by burns. A research project launched in 2015 was to document burn wounds using various technologies, with the aim of assuring an objective assessment of the healing capacity of such wounds. This research resulted in an imaging technology that can determine this with an accuracy of 95%, which is far better than a trained physician’s eye can achieve.

Better quality of life for large patient groups

In parallel with this, the researchers found an additional effect that has had a strong impact on one of the largest patient groups in Sweden, i.e. the around 50,000 people living with hard-to-heal wounds, among them diabetics and patients with multimorbidity. The important side effect was that, thanks to AI technology, this method could be developed to the extent that it was possible to measure wound size with square millimetre exactness, and thereby facilitate the healing process.

Hard-to-heal wounds has been a neglected research area in Sweden, in spite of the physical and mental suffering they cause to so many people. Such wounds may cause long-term pain, they smell and have a negative impact on the patient’s quality of life.
Folke Sjöberg Photo credit Emma Busk Winqui

“Finding a special method that can be applied to this patient group is a huge breakthrough that not only has an impact on their quality of life but also reduces annual health care costs by several million SEK. Wounds heal faster, and hospital stays are shortened,” says Folke Sjöberg, professor in burns care at Linköping University and medical director of the Burn Centre at the University Hospital.

Three doctoral theses and 20 other publications have led to these results, financed by Vinnova and others.

AI enables the use of this method in everyday healthcare

“In our trials, we have worked with advanced optical systems. The problem with the methods used was that they were complex and hard to apply in everyday care. But we discovered early on that we could use normal photography. The concern then was that interpreting the images would be more difficult. That’s why we have been working with Artificial Intelligence,” says Folke Sjöberg.

When using photography to document hard-to-heal wounds, a reference card is placed next to the wound so that the computer will know which scale to use. A phone is then used to take a photo. The image is then analysed by an AI algorithm in the phone. The algorithm is based on thousands of images of wounds, and several hundred million different variables. The computer only takes seconds to analyse the information in the image, and delivers an assessment of the wound size that is much more precise than the human eye could ever achieve. Few technologies have practical uses in wound care. This wound analysis method is therefore at the cutting edge internationally, and has now been nominated for the Swedish MedTech4Health Innovation Award, which highlights innovative medical technology research. The method has also been launched commercially via the Linköping-based company DermaCut AB. Putting it into commercial use can lead to new businesses and more jobs.
Photo credit Emma Busk Winquist

New advances for effective wound treatment

The number of burns is decreasing, in Sweden as well as in other parts of the world. Therefore, half of the Burn Centre’s patients are now people with hard-to-heal wounds unrelated to burns. According to Burn Centre staff, the AI method has provided them with an objective measuring tool, which makes their work with hard-to-heal wounds safer and more successful. Both parties having the same information also facilitates patient communication.

Folke Sjöberg and his research colleagues are confident that the AI method will contribute to more advances making wound care even more effective, such as being able to tell at an early stage whether a wound is infected, or in the detection of dead tissue.

Collaboration at Linköping University