AbstractBackgroundWe aimed to analyze the differences in the peripheral blood cells and tumor biomarkers between the patients with endometriosis and healthy people, and establish a more efficient combined diagnostic model.MethodsWe retrospectively analyzed the differences in the peripheral blood cells and tumor biomarkers between the patients with endometriosis and healthy people. Binary logistic regression analysis was used to establish a combined diagnostic model. We plotted the receiver operator characteristic (ROC) curve to analyze the diagnostic efficiency of different diagnostic indexes.ResultsCompared with patients in the control group, patients in the endometriosis group had significantly lower eosinophil% (p = 0.045), neutrophil (p = 0.001), lymphocyte (p < 0.001), red blood cells (RBCs) (p < 0.001), and hemoglobin (HGB) (p < 0.001), and had significantly higher monocyte% (p = 0.008), monocyte‐to‐lymphocyte ratio (MLR) (p = 0.001), platelet‐to‐lymphocyte ratio (PLR) (p < 0.001), carbohydrate antigen (CA)‐199 (p < 0.001), CA125 (p < 0.001), human epididymis protein (HE)‐4 (p < 0.001), and the risk of ovarian malignancy algorithm (ROMA) (p < 0.001). The combined diagnostic model of HGB, CA199, CA125, and HE4 was established by binary logistic regression analysis. The ROC curve showed that the combined diagnostic model reached a sensitivity of 85.4%, a specificity of 78.83%, and an area under the curve of 0.900, which was significantly higher than that of the individual index in endometriosis diagnosis.ConclusionThe combined diagnostic model of HGB, CA199, CA125, and HE4 may provide a new approach for the early non‐invasive diagnosis of endometriosis.