A Leiden research group is investigating whether artificial intelligence (AI) can help notice subtle changes in joints on MRI images of patients with rheumatoid arthritis. The researchers will receive a €680,000 grant from the Netherlands Organisation for Scientific Research (NWO) for this purpose.
Rheumatoid arthritis is a common disease characterised by inflammation in the joints of the hands and feet. Although drugs are available to prevent extensive damage to joints, rheumatoid arthritis is a chronic disease for many patients. Early treatment at a preclinical stage can help prevent rheumatoid arthritis from developing into a chronic disease.
Automated analysis of MRI images
Researchers from the LUMC are therefore conducting research into early detection of subtle changes in joints of patients with rheumatoid arthritis. The research is part of the AIMIRA project, in which AI experts, radiologists and rheumatologists from LUMC are joining forces. Together, they are investigating the use of AI to analyse MRI images in an automated way.
The researchers are deploying AI to detect abnormalities on these images. They then convert these abnormalities into a number. This allows early effects of rheumatoid arthritis treatment to be shown more accurately.
"Until now, MRI images of joints are assessed visually and a grade is given for the severity of the abnormality. This method was originally developed for obvious inflammatory changes, but at a preclinical stage we expect these changes to be very subtle. Visual scoring is not sensitive enough for this, which means that the effect of a drug may go completely undetected," said Berend Stoel, computer scientist at the Clinical and Experimental Imaging Lab.
Other AI applications in healthcare
The LUMC is not the only organisation using AI for medical purposes. For example, the British company Ultomics has already developed an AI system to analyse ultrasound scans. The company claims that this can improve the accuracy of diagnosing heart disease from 80 per cent to over 90 per cent.
DeepMind, part of Google’s parent company Alphabet, previously developed an AI system capable of diagnosing fifty different eye conditions. The system does this with the same accuracy as specialist ophthalmologists. The system also recommends the most appropriate treatment for patients and identifies which patients should be treated as a priority.
In August 2019, Philips invested, through its investment arm, in DEARhealth, a health tech spin-off from the University of California, Los Angeles (UCLA) in the US. DEARhealth offers a platform that provides AI-based recommendations for care pathways in chronic care. The Dutch company Aidence is working on Veye Chest, an AI application that supports radiologists in the treatment of lung cancer.
Fujitsu has been providing an AI solution since 2019 that helps process unstructured medical notes faster into patients' medical records.
Author: Wouter Hoeffnagel
