The study, led by the School of Computer Science and Technology, was published as a research paper in Scientific Reports, an online journal from the publishers of renowned science journal, Nature.

Titled ‘Deep learning for early detection of pathological changes in X-ray bone microstructures: case of osteoarthritis’, the paper reported an 11% increase in accuracy of early diagnosis of osteoarthritis (OA) – the world’s most common musculoskeletal condition – by using Artificial Intelligence (AI) technology.

OA is a major cause of disability in adults and new methods of early diagnosis are urgently needed in order to improve patients’ treatment outcomes.

Dr Vitaly Schetinin, Senior Lecturer in Computational Intelligence, led the study and says the research is essential for helping both the NHS and those who suffer with OA. He said:

“Our team had an ambition to develop a new AI-based technology capable of using X-Ray to detect signs of OA, which cannot yet be recognised by a standard radiology test.

“With the study showing a diagnostic accuracy improvement of 11%, we have demonstrated that the AI technology will help healthcare practitioners – it is essential this technology becomes more accessible as it can save the NHS a significant amount on expensive bone density scanners.”

Supported by the EU Regional Development Fund, the study was conducted in collaboration with Fusion Radiology (a UK-based NHS contractor), Wien Technical University, Stavropol Medical University and University of Exeter. .

Visiting University of Bedfordshire lecturer, Dr Livija Jakaite, was the study’s lead researcher. She said:

“It was really exciting to see how tiny pathological variations of bone microstructures in patients’ X-ray images could be identified by Artificial Intelligence in order to assist radiologists with detection of the pathology at the early stage.”

Since the research paper was published, the EU Regional Development Fund has continued to support the team, who have recently completed a consultancy project for a GP on using AI for Medical Decision Making.

After being published in Nature journal in early 2021, the paper has attracted a great deal of readership and media attention – ranking in the top 4% of all Nature articles published around the same time.

To read the research paper, ‘Deep learning for early detection of pathological changes in X-ray bone microstructures: case of osteoarthritis’, visit the Nature website.

For information about courses and research opportunities with the University of Bedfordshire’s School of Computer Science and Technology, visit: www.beds.ac.uk/howtoapply/departments/computing/