AI key to easing back pain
An innovative measurement tool developed by an international collaboration led by James Cook University researchers could lead to significant improvements in treating lower back pain, thanks to artificial intelligence.
JCU Electrical and Electronic Engineering Senior Lecturer Dr Mostafa Rahimi Azghadi and PhD Candidate Alzayat Saleh led a team to design and develop a machine-learning computer algorithm that can automatically detect the size of a Transverse abdominis (TrA) muscle from an ultrasound image, similar to what a human operator does but with the advantages of better accuracy and consistency.
A dysfunctional TrA, also known as a “corset” muscle, can play a key role in diagnosing lower back pain.
“Our measurement tool can be used to evaluate abdominal muscle dimensions to aid patient rehabilitation,” Dr Azghadi said.
“It can, therefore, make a tangible difference in improving treatment outcomes. Our tool can also make a difference by assisting healthcare professionals to perform measurements quicker and more accurately.”
Dr Azghadi believes the method developed by his team and Canadian collaborators will decrease the variability in measurements that may happen if the same measurement is performed twice by the same human operator, or if the measurements are done by two or more different operators.
“This may help health professionals make more accurate diagnoses and evaluations of rehabilitation progress. In addition, image measurement may be expedited by the tool, providing benefit in terms of saving clinical time and cost and improving patient outcomes,” Mr Saleh said.
The team is hoping to trial their newly developed treatment method on other ultrasound images of different body organs and are interested in integrating the tool in clinical practice.
Based on self-reported data from the Australian Bureau of Statistics’ 2017–18 National Health Survey, about four million Australians, or 16 per cent of the population, have back problems.
Media enquiries: firstname.lastname@example.org