Non-invasive biomarkers and advanced imaging for early detection of endometriosis: Diagnostic performance, real-world evidence, artificial intelligence, and implementation priorities

Endometriosis is a chronic, heterogeneous inflammatory disease for which delayed recognition and unequal access to specialist imaging remain important clinical problems. The transition from surgically centered diagnosis to an integrated, non-invasive strategy has increased interest in expert transvaginal ultrasonography, magnetic resonance imaging, circulating biomarkers, and saliva-based microRNA signatures. Objective: To critically appraise the diagnostic role of these modalities and describe how real-world data, causal inference and artificial intelligence can support safe implementation. Methods: