Advancing Women’s Health with Sovereign AI and Secure Storage on Gefion
At Hvidovre Hospital, researchers are using AI and large-scale data analysis to tackle some of the most under-researched conditions affecting women. Today, several research projects are powered by secure storage and high-performance AI compute on the Gefion supercomputer.
Women live significantly more years in poor health than men, often due to conditions linked to menstruation, pregnancy, and menopause, areas historically underrepresented in research. That is about to change.
One of the World’s Largest Women’s Health Datasets
The Department of Obstetrics and Gynaecology at Hvidovre Hospital is home to one of the largest and most unique women’s health datasets in the world. Over years of research, extensive cohorts have been established, combining medical records, questionnaires, biological samples, imaging data, and multi-omics datasets — amounting to nearly a petabyte of sensitive health data. With secure storage integrated with AI compute on Gefion, researchers can now fully unlock the potential of this unique dataset, without compromising data protection or sovereignty.
Watch the Case Story
In this video, the research team shares some of the opportunities — and responsibilities — of working with health data in a secure supercomputing environment where research quickly translates into real impact for patients.
AI Models for Faster Diagnosis
Among the first to benefit are women suffering from severe pain, infertility, and other complications caused by adenomyosis, a condition in which tissue grows into the muscular wall of the uterus. Today, diagnosing adenomyosis requires a specialist to manually review 3D scans of the uterus — a process that can take up to 30 minutes per patient.
The research team is now training AI models on Gefion to reduce that time to under a minute. The result: faster diagnosis, improved consistency, and the potential for earlier intervention.
Measurable Performance Gains
After migrating to Gefion, the team reports:
• Image classification models running up to 5× faster
• DNA variant calling workflows running up to 20× faster
By combining one of the world’s richest women’s health datasets with Gefion’s sovereign AI infrastructure, the team is setting a new standard for data-driven healthcare innovation.
About Technical University Hospital of Greater Copenhagen (TUH)
A partnership between DTU and hospitals in Region Hovedstaden, TUH connects health and technical sciences to foster interdisciplinary innovation.
About Department of obstetrics and gynaecology, Hvidovre Hospital
Denmark’s largest gynaecological and obstetric clinic and a unique, diverse patient population, the department leads pioneering research in women’s health.