AI pilot launched for endometriosis

Endometriosis can hide in plain sight on a scan, which is one reason people with the disease are often told for years that their pain is “normal” before anyone finds the cause. The disease happens when tissue similar to the uterine lining grows outside the uterus and triggers inflammation, scarring, pain, and sometimes infertility. (who.int) Doctors now use imaging more than they used to because looking for endometriosis through surgery first can slow diagnosis. The American College of Obstetricians and Gynecologists said in March 2026 that imaging tests and clinical findings can support a presumptive diagnosis instead of waiting for surgical proof. (acog.org 1) (acog.org 2) Magnetic resonance imaging is one of those imaging tools. Think of it as a high-detail map of the pelvis that can show deep disease and scar tissue in places an ordinary exam cannot reach. (radiologyinfo.org) (pubmed.ncbi.nlm.nih.gov) The catch is that these scans are hard to read when the signs are subtle. A recent review in the American Journal of Roentgenology said routine ultrasound and routine magnetic resonance imaging perform worse than dedicated endometriosis imaging protocols, which helps explain why cases get missed. (ajronline.org) That is the opening Matricis.ai is trying to use. The Paris company said on April 10, 2026 that it started its first United States clinical pilot of EndomAI with SimonMed Imaging, one of the country’s largest outpatient imaging providers. (manilatimes.net) (markets.businessinsider.com) EndomAI is not a robot doctor making the call by itself. Matricis.ai describes it as software that helps radiologists interpret pelvic magnetic resonance imaging scans by flagging patterns linked to endometriosis and aiming to improve sensitivity and specificity. (matricis.ai) The company says the software was trained on what it calls one of the world’s largest annotated pelvic magnetic resonance imaging datasets, built from multicenter data across three continents. In plain English, that means experts labeled many past scans so the system could learn what disease patterns look like. (matricis.ai) SimonMed is the kind of partner that turns a lab claim into a real-world test. The imaging company says it has more than 170 clinics nationwide, so a pilot there can show whether the tool fits into everyday radiology workflows instead of only working under ideal study conditions. (simonmed.com) Matricis.ai also says it plans Food and Drug Administration clearance in the fourth quarter of 2026, which means this pilot is an evidence-gathering step rather than a full commercial green light. That matters because clinical software has to show it helps in practice, not just in a demo. (matricis.ai) The fertility angle is easy to miss unless you know how common the disease is in infertility care. The World Health Organization says endometriosis affects an estimated 10% of reproductive-age women worldwide, and a World Health Organization bulletin notes prevalence estimates of 5% to 50% among infertile women, which is why earlier detection can change counseling long before pregnancy starts. (who.int 1) (who.int 2)