Ultrasound trained with AI for the treatment of endometriosis

Researchers at FAU hope to improve the diagnosis and treatment of the “chameleon” of gynecological disorders

Endometriosis is a gynecological disorder that is largely unknown, even though it is estimated that it affects one in ten to fifteen women of reproductive age in Germany. A multi-disciplinary team of researchers at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) and Universitätsklinikum Erlangen are investigating the options for the treatment and diagnosis of the gynecological condition with three million euros of funding from the Bavarian State Ministry of Public Health and Care Services.

Endometriosis is a chronic hormone-dependent condition, often referred to as the “chameleon” disease due to wide range of phenotypical symptoms it presents. It is usually noticeable because of the severe pain it causes in the lower abdomen during menstruation. In endometriosis, tissue similar to the endometrium, or the lining of the uterus, grows in other places in the body outside the uterus, mostly in the lesser pelvis, and can lead to problems with bowel movements and when urinating. The condition is also associated with infertility, anxiety disorders and depression. The number of unreported cases of the condition is high due to its wide variety of symptoms.

Creating a 3D patient model

The aim of the new study is to improve non-invasive diagnosis and increase the support for effective treatment of patients with endometriosis. The novelty of the research lies in the use of AI methods to transfer various imaging methods such as ultrasound and MRI into a holistic 3D patient model, which is supplemented by further surgical and post-operative data. Ultimately, the aim is to evaluate whether this improves the situation not only for patients, but also for hospitals.

Five female researchers at FAU and Universitätsklinikum Erlangen and two partners from Universität Würzburg and TU München are collaborating on the EndoKI (Endometriose und künstliche Intelligenz) project in order to carry out interdisciplinary research “by women, for women” on the condition. The project, which is due to run for three years, is being funded as part of the digital and innovative health and care projects (BayDiGuP) and women’s health and artificial intelligence.

Up to eight years on average until diagnosis

For Prof. Dr. Franziska Mathis-Ullrich, Professor of Surgical Robotics at the Department of Artificial Intelligence in Biomedical Engineering, it is important that “we increase the awareness of women but also society for this unknown and often undetected condition, which affects so many.” The seven female researchers come from the fields of AI, medical engineering, computer science, gynecology and gender studies. “We want to collect the data of at least 300 patients and find out how a diagnosis can be made earlier and in a more standardized way using imaging, thus preventing pain from becoming chronic,” says Mathis-Ullrich. This is because it takes an average of eight years to reach a diagnosis for endometriosis due to the wide variety of symptoms it presents. The condition is often only detected during an endoscopic examination. It is treated by prescribing pain relief, hormone treatments or with surgery.

“The purpose of the collected data is also to reduce the number of operations required,” says Mathis-Ullrich. Operations in conjunction with endometriosis alone cost almost 100 million euros per year. There is a potential for savings of 20 million euros per year. “For patients, it would be ideal if endoscopy is not only used for diagnosis, but also to completely remove the diseased tissue at the same time,” explains Mathis-Ulrich. Currently, several procedures are sometimes required. The aim of EndoKI is to facilitate faster and more precise diagnosis. “More precise pre-operative detection of endometriosis lesions could prevent repeated surgery. This would benefit patients, since each operation is associated with a certain level of risk,” says the researcher.

Database to be created

In the long term, a pseudonomized database is to be created, in which MRI datasets and histopathological information will be available, among others, to train AI models and the data will also serve as the basis for further research. To improve the understanding of diagnostic and treatment processes, a qualitative and ethnographic sub-study is planned during which gynecologists, patients and researchers will be interviewed to find out their viewpoints and needs. These findings are then to be used as a basis for recommendations in guidelines for the diagnosis and treatment of endometriosis, ideally also for UN organizations such as the WHO.

“We also want to increase physicians’ awareness of the condition with our study.” In general, the aim of the project is not for the findings to just remain within the scientific community, but rather to create more awareness for women’s health among the general public as well. With this in mind, a conference on the topic of endometriosis has been planned at FAU for 2028 for physicians, researchers and those affected by the condition. “I have rarely seen a project involving female researchers from such a wide range of disciplines get off to such a highly-motivated start. It is something close to our hearts as researchers,” says Mathis-Ullrich.

In addition to researchers from FAU and the University Hospital Erlangen, scientists from the University of Würzburg and the Technical University of Munich are also responsible for managing the project:

Prof. Franziska Mathis-Ullrich (Coordination, FAU)
Prof. Jana Hutter (UKER)
PD Dr. Med. Stefanie Burghaus (UKER)
Prof. Katharina Breininger (Julius Maximilian University of Würzburg)
Prof. Maria Rentetzi (FAU)
Prof. Julia Schnabel (Technical University of Munich)
Dr. Sina Martin (FAU)

Further information:

Prof. Dr. Franziska Mathis-Ullrich
Professorship for Medical Robotics
franziska.mathis-ullrich@fau.de

PD Dr. Stefanie Burghaus
Endometriosis Center at Universitätsklinikum Erlangen
stefanie.burghaus@uk-erlangen.de