For the nearly 200 million people worldwide living with endometriosis, the diagnostic odyssey — averaging four to eleven years — reflects a fundamental problem: the disease is classified by what surgeons see, not by what patients experience. A large-scale symptom-phenotyping study now offers a data-driven framework that could reshape how clinicians identify, stratify, and ultimately treat these conditions.
Drawing on data from over 22,000 individuals in the NIH-sponsored All of Us Research Program collected between 2018 and 2023, researchers applied latent class analysis to 22 distinct symptoms, yielding four reproducible phenotypic clusters. These included a high-pain group, a gastrointestinal-dominant group, a psychological-neurological cluster, and a relatively lower-burden class. Critically, when concomitant adenomyosis was present, patients were disproportionately concentrated in the high-symptom classes, with no identifiable minimal-symptom subgroup emerging — suggesting adenomyosis may amplify disease burden rather than represent a parallel but milder condition. High-burden phenotypes showed measurably worse quality-of-life scores across physical, emotional, and social domains.
This work is analytically significant for several reasons. Symptom-based phenotyping using latent class analysis is not new — it has been applied in conditions like irritable bowel syndrome and fibromyalgia — but applying it at this scale to endometriosis represents a meaningful methodological advance. The findings reinforce the growing consensus that endometriosis is a systemic condition with neurological, gastrointestinal, and psychiatric dimensions, not merely a pelvic one. Practically, identifying a psychoneurological phenotype carries treatment implications: this group may respond differently to central sensitization-targeted therapies than to hormonal suppression alone. The primary limitation is the cross-sectional design, which prevents causal inference, and reliance on both electronic health records and self-report introduces ascertainment variability. Nonetheless, the cohort size and phenotypic resolution make this one of the more methodologically robust endometriosis characterization studies to date — incremental in method, potentially paradigm-shifting in clinical application.