<![CDATA[Introduction: Endometriosis is a chronic condition affecting about one in ten women of reproductive age, resulting in reduced quality of life. Non-invasive diagnostic methods based on single serum biomarkers or isolated clinical parameters have shown limited reliability. Consequently, integrating serum biomarkers with clinical variables has been proposed as a more effective diagnostic approach. This study aimed to assess the diagnostic accuracy of combined serum biomarkers and clinical variables for endometriosis diagnosis. Methods: A systematic review and meta-analysis were carried out in line with PRISMA recommendations. Literature synthesize were conducted across PubMed, Cochrane, ScienceDirect, Epistemonikos, Scopus, ProQuest, and EBSCO to February 2025. Eligible studies included diagnostic investigations that compared combined serum biomarker and clinical-variable models with serum biomarker or clinical-variable only reference standards. Study quality was evaluated using the QUADAS-2 tool. Pooled effect analysis was generated as odds ratios (ORs) using 95% confidence intervals (CI). Results: Seventeen studies met eligibility criteria. Combined diagnostic models demonstrated significantly higher sensitivity (OR 2.64; 95% CI 1.54–4.53) but non-significant specificity (OR 1.03; 95% CI 0.43–2.49) compared with single diagnostic approaches. However, the area under the curve (AUC) analysis significantly improved in combined models (OR 1.46; 95% CI 1.10–1.93). Moderate to high heterogeneity with publication bias was observed across studies. Conclusion: Combined serum biomarkers and clinical variables provide better diagnostic performance than single tests and support the use of non-invasive multimodal strategies prior to surgical confirmation. Further prospective validation is required to ensure consistency and clinical applicability. ]]>
