‘Machine vision’ may be useful diagnosis aid

BRISBANE: A new international study shows an artificial intelligence machine outperformed doctors in identifying genetic disease from a person’s selfie photo, in a field where Queensland leads the way.

Queensland scientists are breaking down barriers in the area of “machine vision”, where algorithms are used to analyse images to determine an accurate diagnosis.

“At the University of Queensland there are different research groups using AI in medicine and biology. We are already tracking the growth of brain cells, automatically detecting different parts of the brain in scans, detecting skin cancer more accurately, scanning blood cells for disease and it is used in many pathology detecting systems,” UQ’s Dr Peter Stratton, computing neuroscientist and artificial intelligence researcher, told The Courier-Mail.

The success of the “selfie” study, sparks the debate that the “human touch” could be lost in a man-versus machine future.

Programmers ran through more than 17,000 images of patients with hundreds of distinct genetic syndromes with the AI. The machine was able to detect genetic syndromes such as foetal alcohol, Rett and Trisomy 18 with higher accuracy than clinicians.

“AI techniques tend to be adopted very quickly when they work well, however, it seems this research is still in the ‘proof of principle’ stage. This means that, while it worked well in this study, more research needs to be done before it can be used with real patients for diagnosing diseases in a doctor’s surgery or hospital,” Dr Stratton said.

The researcher believes Queensland is many years away from AI being the first point of contact for patients and that doctors’ jobs are safe.

“When people’s lives are on the line, like with medical diagnoses, we need to be sure that our AI will perform predictably under all circumstances and this is not something we can currently guarantee. In a few years we might see AI being used more commonly as tools by doctors and medical teams to help with their diagnoses,” he said.

McDowall mother-of-four Jade Jenkins, 28, said she welcomed the progression in medicine.

“I think it is amazing, as long as the findings from the machine are done in consultation with a doctor. My children’s health is a priority for me so the human element is important,” she said.