Harnessing AI to revolutionise endometriosis diagnosis

An interdisciplinary research team from Adelaide University are using ground-breaking AI technologies to make endometriosis diagnosis faster, more cost-effective, and less invasive with IMAGENDO^®. 

Endometriosis affects hundreds of millions of people globally and nearly a million Australians. The condition, which occurs when tissues similar to the lining of the uterus grow outside the womb, can cause pelvic inflammation, headaches, fatigue, and even infertility. The debilitating pain can affect one’s ability to work, relationships, fitness, education and mental health. And yet, despite its undeniable toll, pathways to diagnosis and treatment remain slow, expensive and invasive.

Currently, the wait time for diagnosis averages more than six years from symptom onset. Painful symptoms are often overlooked or normalized, and the most common diagnostic tool, a keyhole surgery, is costly and intrusive. Fortunately, Adelaide University researchers are working to improve care for endometriosis with a new form of non-invasive imaging called IMAGENDO^®.

Two Adelaide University research institutes – the Robinson Research Institute and the Australian Institute for Machine Learning – are harnessing the emergent power of artificial intelligence (AI) to revolutionise endometriosis diagnosis. By combining ultrasound, magnetic resonance imaging (MRI), and AI, the researchers aim to make IMAGENDO^® a cost-effective, accessible, and accurate new way to diagnose the condition without surgery.

“Endometriosis symptoms can have a devastating impact on patients’ lives, but an earlier diagnosis enabled by the IMAGENDO^® technology will lead to prompt treatment and better quality of life by avoiding unnecessary hospitalisations and repetitive surgery,” she says.

“[Earlier diagnosis] provides an opportunity to manage and treat pain appropriately to prevent complications and to preserve fertility via egg freezing and fertility planning.” 

Interpreting images from IMAGENDO® takes a niche skillset, so Professor Hull and her team are collaborating with Adelaide University’s Australian Institute for Machine Learning to overcome this potential workforce challenge. 

“Endometriosis scanning is a specialist skill and not all sonographers, radiologists and gynaecologists will have the opportunity to receive training to perform specialist endometriosis scans or MRIs,” says Professor Hull.

“Fortunately, we have a solution to this potential obstacle.”

Using a diagnostic dataset of ultrasounds provided by endometriosis expert Professor George Condous, and MRIs from Benson Radiology and South Australian Medical Imaging as a training sample, the researchers built a program that uses AI to read specialist scans. After learning from a large number of ultrasounds and MRIs, the program’s algorithm is able to recognise certain endometriosis imaging markers. 

Initial tests show the software’s diagnostic accuracy is similar to that of a specialist endometriosis doctor, and the IMAGENDO^® team expect it to improve even further as the machine learning training continues. 

On the horizon

The IMAGENDO^® team is aiming to make the algorithm available for clinical use in the next two to five years after thorough testing and approval.

Now with further funding from an NHMRC Ideas Grant to explore adolescent endometriosis imaging and an AEA Ignite Grant to expand their data set, IMAGENDO^® is increasingly recognised globally for its use of AI in analysing data from ultrasound and MRI scans to diagnose endometriosis. The program is expanding across Australia and internationally with collaborations in North America, South America, Southeast Asia, the United Kingdom, and Europe.

The breakthrough is part of a rapidly evolving global movement in which AI is being used in conjunction with medical imaging technologies to transform patient outcomes. With these innovations accelerating, IMAGENDO^® is just the beginning of a new healthcare frontier.