Inflammatory markers and the systemic immune-inflammation index in endometrial carcinoma and hyperplasia: a retrospective cohort study

Background Endometrial carcinoma (EC) presents commonly with abnormal uterine bleeding, and definitive diagnosis requires invasive endometrial biopsy. Haematological inflammatory markers—including the neutrophil-to-lymphocyte ratio (NLR), platelet distribution width (PDW), and the systemic immune-inflammation index (SII)—have been proposed as adjuncts in the differential diagnosis of endometrial pathologies. We evaluated whether these markers can independently discriminate EC from endometrial hyperplasia with atypia/endometrial intraepithelial neoplasia (EHA/EIN), endometrial hyperplasia without atypia (EH), and normal endometrium, after adjustment for age and other confounders.Methods Data from 238 patients who underwent endometrial biopsy for abnormal uterine bleeding between 2015 and 2020 were retrospectively analysed. Patients were categorised as EC (n = 74), EHA/EIN (n = 28), EH without atypia (n = 66), and controls with normal endometrium (n = 70). Haematological parameters, NLR, platelet-to-lymphocyte ratio (PLR), and SII were derived from preoperative complete blood count results. Univariable comparisons used the Kruskal–Wallis test with Dunn’s post-hoc pairwise comparisons. Hierarchical multivariable logistic regression models were fitted, adjusting sequentially for age, menopausal status, and (where available) diabetes mellitus and hypertension. Receiver operating characteristic (ROC) analysis was performed for markers showing significant univariable discrimination, and PDW was analysed in inverted form (1–PDW direction) to reflect its inverse association with malignancy.Results NLR and SII were significantly elevated, and PDW was significantly reduced, in EC patients compared with controls (all p < 0.05). After adjustment for age and menopausal status, NLR (adjusted OR [aOR] 2.06, 95% CI 1.04–4.09; p = 0.037), SII per 100 units (aOR 1.49, 95% CI 1.09–2.04; p = 0.014) and PDW (aOR 0.93, 95% CI 0.89–0.96; p < 0.001) remained independently associated with EC versus controls. By contrast, after adjustment for age, NLR and SII no longer independently discriminated EC from EH (NLR aOR 1.27, p = 0.245; SII aOR 1.14, p = 0.165), suggesting that age accounts for a substantial proportion of the apparent univariable differences. In ROC analysis with PDW values inverted, AUCs were 0.79 (95% CI 0.72–0.87) for PDW, 0.68 (95% CI 0.59–0.76) for NLR, and 0.62 (95% CI 0.52–0.70) for SII against the control group. No marker—adjusted or unadjusted—significantly distinguished EC from EHA/EIN.Conclusions NLR, SII, and (low) PDW are independently associated with EC compared with normal endometrium, even after age and menopausal status adjustment, but their performance is modest and they do not independently discriminate EC from premalignant endometrial lesions. Histopathological evaluation remains indispensable. These markers may inform risk stratification but are not sufficient as standalone tests.