Research News AI-driven tissue analysis offers new insights into endometriosis

25 March 2025

Researchers at Radboudumc are now using advanced AI-driven tissue analysis to gain new insights into the cellular interactions within endometriotic lesions, aiming to identify potential therapeutic targets. By mapping the cellular composition of these lesions, they hope to uncover key drivers of disease progression and develop more effective treatment options. They published their findings in Human Reproduction journal.

Endometriosis is a chronic condition affecting 6–10% of women of reproductive age, often leading to infertility and a reduced quality of life. It is characterized by the growth of endometrial-like tissue outside the uterus, triggering inflammation, pain, and fibrotic tissue changes that contribute to its symptoms. Despite its prevalence, the underlying causes remain poorly understood, and current treatments only manage symptoms rather than addressing the root of the disease.

A multidisciplinary team from the Department of Medical BioSciences (Roland Brock, Scott Korman, and Mark Gorris) and the Department of Obstetrics and Gynaecology (Annemiek Nap and Guus Vissers) has adapted advanced cancer tissue analysis techniques to study endometriosis. Using multiplex immunohistochemistry combined with AI-driven analysis, developed at Radboudumc, researchers can precisely identify and count cells within the complex tissue structures of endometriotic lesions.

In an initial study, the team successfully determined the fraction of dividing immune cells within different tissue substructures of endometriosis. The next step is a detailed mapping of immune cell types present in the lesions, which could reveal key drivers of lesion growth and fibrosis.

Figure: Artificial intelligence-based segmentation of an endometriosis lesion into different tissue substructures.

What's next?

Beyond mapping, the research team is exploring targeted therapies that could transform the treatment of endometriosis. One approach involves selectively eliminating specific cell types that contribute to lesion growth and fibrosis, allowing researchers to test their hypotheses about disease mechanisms and identify key drivers of progression. Another promising strategy is the use of therapeutic mRNAs—genetic instructions that enable cells to produce proteins capable of reversing inflammation and fibrotic changes in the affected tissues. By delivering these mRNAs directly to the lesions, researchers aim to restore normal cellular function and potentially halt disease progression.

Early experiments in these areas are already underway, with researchers testing both selective cell-targeting methods and mRNA-based interventions. If successful, these strategies could provide a foundation for developing precision treatments for endometriosis, moving beyond symptom management to address the root causes of the disease.

This project is part of the Radboudumc research programs: 3D Biology and Disease Mechanisms and Sex and Gender Sensitive Health and Reproduction.

About the publication

Scott E Korman, Guus Vissers, Mark A J Gorris, Kiek Verrijp, Wouter P R Verdurmen, Michiel Simons, Sebastien Taurin, Mai Sater, Annemiek W Nap, Roland Brock, Artificial intelligence-based tissue segmentation and cell identification in multiplex-stained histological endometriosis sections, Human Reproduction, Volume 40, Issue 3, March 2025, Pages 450–460, https://doi.org/10.1093/humrep/deae267

Related news items