Endometriosis study reveals distinct molecular response patterns that could guide clinical trials

Endometriosis is typically treated as a single condition, yet patients often experience very different outcomes with the same therapy.

A new preprint from Paris-based biotech company endogene.bio, Beyond one-size-fits-all: single-cell transcriptomic signatures predict drug efficacy and reveal responder subgroups in endometriosis, suggests that this variability reflects a fundamental biological reality: endometriosis is molecularly heterogeneous across patients, lesion sites, and cell types.

This molecular heterogeneity has important implications for how endometriosis treatments are studied and developed. Treating the disease as a single entity may help explain why many clinical trials yield mixed results or fail, while matching patients to therapies based on biology could improve trial success and efficiency.

To address this heterogeneity, the researchers combined single-cell RNA sequencing – which measures gene activity in individual cells – with a computational drug-response framework. The approach focuses on the specific cell populations most likely to drive lesion persistence, rather than averaging signals across an entire tissue.

For single-cell datasets spanning eutopic endometrium and endometriotic lesions, the analysis identifies disease-associated molecular mechanisms concentrated in stromal, endothelial, and stem-like cell populations, which consistently showed disease-associated transcriptional changes linked to lesion maintenance. The model then predicts which therapeutic mechanisms may be most likely to counter these disease states and identifies responder subgroups, suggesting that a drug effective for one subset of patients may be less relevant for another.

Dr Cristina Fernández Molina, co-founder and Head of Science at endogene.bio, commented: “Endometriosis is currently classified based on symptoms and anatomical or surgical features, but these approaches don’t capture the biological programs active inside the tissue. Our work shows that what looks like unpredictable variability between patients actually reflects recurring cellular states that shape how the disease behaves and responds to treatment. This creates an opportunity to move toward more biologically informed patient stratification in research and clinical trials.”

For decades, endometriosis has been approached much like breast cancer once was – treated as a single, uniform disease rather than a spectrum of biologically distinct subtypes. In oncology, recognizing HER2-positive, hormone-receptor-positive, and triple-negative breast cancers as fundamentally different diseases transformed both patient outcomes and clinical trial design. The emerging molecular picture of endometriosis points to a similar need for precision.

The findings add to a growing body of evidence that endometriosis is not one disorder but many, each driven by distinct cellular programs and therapeutic vulnerabilities. This shift suggests that clinical trials enrolling biologically defined subgroups may be more informative, less ambiguous when results appear mixed, and less likely to fail late due to hidden molecular heterogeneity, a major contributor to high development costs in women’s health.

A notable implication discussed by the authors is that drug response–associated signatures identified in lesion stromal cells are also detectable in eutopic endometrial stromal cells, which are naturally shed during menstruation. Because lesion tissue is typically only accessible during surgery, this raises the prospect of using menstrual blood sampling as a less invasive route for patient stratification and longitudinal monitoring.

Access to lesion tissue over time remains one of the central practical challenges in endometriosis research. Lesions are usually identified and sampled during surgery, making repeated sampling difficult and limiting researchers’ ability to track disease biology or treatment response. Approaches that leverage accessible tissues – such as endogene.bio’s collection and analysis of menstrual blood–derived cell populations – could help overcome this long-standing barrier.

Together with an earlier preprint published in July, which demonstrated the potential of menstrual-blood-derived methylation signatures as a non-invasive diagnostic tool for endometriosis, the new findings point to a broader shift in how the disease could be understood, studied, and managed. Taken as a whole, endogene.bio’s work suggests that both diagnosis and treatment response may be accessible through molecular signals captured without surgery.

For the medical field, this represents a move toward the same biology-driven framework that has reshaped oncology, defining diseases by their underlying cellular programs rather than by symptoms and anatomical features alone.

For pharmaceutical companies, this approach has the potential to reduce clinical trial failure rates, turn molecular heterogeneity from a liability into a design feature, and lower development costs by aligning therapies with the patients most likely to benefit.

For patients, the implications are profound: fewer invasive procedures, clearer explanations for why treatments work or fail, and a more direct path toward therapies chosen because they match the biology driving their disease.

Source: endogene.bio  https://www.endogene.bio/fr

Reference: Beyond one-size-fits-all: single-cell transcriptomic signatures predict drug efficacy and reveal responder subgroups in endometriosis | bioRxiv

Image: Canva

Staff Writer