New AI Tool Spots Endometriosis Signs in 18ms

Australian researchers at Adelaide University have developed EndoFusion, an AI framework that accurately identifies two major indicators of advanced endometriosis in pelvic scans, producing results in just 18 milliseconds, according to News-Medical's reporting on the study, published in the journal Artificial Intelligence in Medicine.

The tool addresses a genuine, well-documented limitation in how endometriosis is currently detected. "Current scanning methods each have their own strengths when it comes to detecting two common markers that indicate the likelihood of endometriosis and patients will often only have access to one of them," said study author Associate Professor Jodie Avery, Research Co Lead of Chronic Reproductive Conditions in the Endometriosis Research Group at Adelaide University's Robinson Research Institute. MRI and ultrasound imaging are each better at detecting different signs of the condition, meaning patients scanned with only one method may miss indicators the other would have caught.

EndoFusion is designed specifically to combine data from both imaging methods rather than relying on either one alone. "Our AI tool can help address these shortcomings by combining data from both imaging tools, giving the framework the knowledge it needs to detect both signs of endometriosis through a single scan more effectively and efficiently," Avery said. The tool was trained using four datasets containing more than 9,000 female pelvic MRI scans and more than 800 transvaginal ultrasound sliding scans, according to Medical Xpress's reporting on the study's methodology.

EndoFusion at a Glance

The study evaluated how well EndoFusion could distinguish between positive and negative cases of endometriosis, finding the tool provided a correct diagnosis 83% of the time, more accurate than all competing models tested, according to News-Medical's reporting. That combination of speed and accuracy is genuinely significant given the current diagnostic reality: definitive endometriosis diagnosis has historically required invasive surgery to visually identify lesions, a process carrying real surgical risk, high financial cost, and significant delay, according to Archynewsy's reporting on the study, with patients often waiting years for a confirmed diagnosis.

EndoFusion wasn't developed by a single lab in isolation. The research team partnered with multiple institutions, including Flinders University, Benson Radiology, Omni Ultrasound and Gynaecological Care, the University of Surrey, McMaster University Medical Centre, and the Mohamed bin Zayed University of Artificial Intelligence, according to Archynewsy's reporting, giving the study genuine international, multi-institutional backing.

Study co-author Yuan Zhang was direct about the tool's current stage and where the research goes next. "We envisage that clinicians will be able to use these tools to help make a determination about the presence of endometriosis from a single scan," Zhang said, according to Medical Xpress's reporting. Researchers plan to expand the training dataset to include additional endometriosis markers, and Zhang pointed to genuinely broader potential applications beyond this specific condition: the underlying multi-modal AI approach "could also potentially provide insights for research into other diseases that require the use of multimodal imaging, such as gynecological disorders, prostate and breast cancer, and fetal abnormalities."

This research is worth understanding for any business in medical diagnostics, women's health technology, or AI-powered imaging specifically, since EndoFusion's core approach, combining data from two existing, already-standard imaging methods rather than requiring new hardware, represents a genuinely lower-cost path to improved diagnostic accuracy than developing entirely new scanning technology.

For businesses in healthcare AI more broadly, this connects directly to the broader wave of AI-driven medical diagnostic breakthroughs we've tracked closely this year, including University of Rochester researchers using AI to map the brain's hidden waste-clearing system, reinforcing a genuine pattern of AI extracting meaningfully more diagnostic value from imaging data hospitals already routinely collect.

What is EndoFusion, and what does it detect?
EndoFusion is an AI framework developed by Adelaide University researchers that detects two major indicators of advanced endometriosis by combining data from MRI and ultrasound pelvic scans, producing results in 18 milliseconds.

How accurate is EndoFusion at diagnosing endometriosis?
The tool correctly distinguished between positive and negative endometriosis cases 83% of the time in testing, outperforming all competing diagnostic models the researchers evaluated.

Is EndoFusion available for clinical use yet?
No. The framework remains in early stages of development, with researchers planning to expand the training dataset to include additional diagnostic markers before broader clinical deployment.

Adelaide University researchers developed EndoFusion, an AI tool that detects two major indicators of advanced endometriosis by combining MRI and ultrasound scan data, producing results in just 18 milliseconds with 83% diagnostic accuracy. The tool addresses a genuine gap in current diagnostic methods, since MRI and ultrasound each detect different endometriosis markers, and patients often only receive one type of scan, contributing to a diagnostic process that has historically required invasive surgery and years-long waits. Researchers, working with international partners including McMaster University and the Mohamed bin Zayed University of Artificial Intelligence, say the tool remains in early development but could eventually extend to diagnosing other conditions requiring multimodal imaging.