Endometrial cancer (EC) is a common malignancy in postmenopausal women where early detection is important for prognosis. Traditional circulating tumour markers such as CA125 and HE4 have been used in gynaecological oncology but have limitations when applied alone for early-stage disease. Peripheral blood immunological indices related to lymphocytes have emerging interest as potential diagnostic adjuncts, but their incremental value when combined with established tumour markers and metabolic risk factors in postmenopausal women was unclear in the source study.
This was a single-centre case-control study that enrolled two matched groups of postmenopausal women: 72 with early-stage endometrial cancer (FIGO I–II; EC group) and 72 with benign uterine lesions (BL group). Baseline clinical data were collected, including metabolic comorbidities such as diabetes, hypertension and dyslipidaemia. Laboratory measurements included peripheral blood immunological markers—total lymphocyte count (LYM) and lymphocyte-to-monocyte ratio (LMR)—and circulating tumour markers CA125 and HE4. The diagnostic performance of individual markers and combinations was assessed by receiver operating characteristic (ROC) curve analysis. Multivariate logistic regression models were used to compare model fit, including comparison with Akaike information criterion (AIC) and statistical testing between models.
The abstract reports group sizes, the set of variables analysed (LYM, LMR, CA125, HE4, metabolic factors), and the main statistical approaches (ROC curves and multivariate logistic regression). The source abstract does not provide details such as laboratory assay methods, marker cut-offs, exact regression coefficients, internal or external validation procedures, or the handling of potential confounders beyond the variables listed.
Clinical characteristics and laboratory comparisons between groups showed several statistically significant differences.
The EC group had higher rates of metabolic comorbidities: diabetes, hypertension and dyslipidaemia (all reported p < .05).
Immunological and tumour marker differences: patients with early EC had lower LYM and LMR (both p < .01) compared with the BL group, and higher CA125 and HE4 levels (both p < .001).
Diagnostic performance of single markers and combined models was reported using area under the ROC curve (AUC) and sensitivity for individual markers where given:
Among single markers, HE4 had the highest AUC at 0.725, though its sensitivity was limited (45.8%).
The combined use of CA125 + HE4 produced an AUC of 0.759.
Adding peripheral lymphocyte indicators (LYM and LMR) to CA125 and HE4 increased the AUC to 0.839.
Incorporation of metabolic factors (history of diabetes, hypertension, dyslipidaemia) into the model further increased the AUC to 0.865. This full model had an Akaike information criterion (AIC) value of 144 and was statistically superior to the other tested models (reported p < 0.01).
The abstract highlights that incremental additions—first lymphocyte-related parameters, then metabolic comorbidities—improved diagnostic discrimination beyond tumour markers alone.
The study compared diagnostic discrimination across several configurations:
Single-marker performance: HE4 outperformed other single markers by AUC (0.725) but had limited sensitivity (45.8%). The abstract does not report the AUC or sensitivity for CA125 alone or for LYM/LMR as single predictors.
Two-marker panel: CA125 + HE4 yielded improved AUC (0.759) versus HE4 alone.
Immune marker augmentation: adding LYM and LMR to the CA125 + HE4 panel produced a marked improvement in discrimination (AUC 0.839), indicating that simple peripheral blood lymphocyte indices added independent diagnostic information in this sample.
Full model including metabolic factors: the model combining CA125, HE4, LYM, LMR, and metabolic comorbidities achieved the highest reported AUC (0.865) and the best model fit by AIC (AIC = 144). Statistical comparison showed this model was superior to other tested combinations (p < 0.01).
The abstract does not report confidence intervals for AUCs, sensitivity and specificity for the combined models, net reclassification indices, calibration metrics, or performance in subgroups.
In this single-centre case-control cohort of postmenopausal women, the authors report that combining peripheral blood lymphocyte indicators (LYM, LMR) with established tumour markers (CA125, HE4) and metabolic risk factors improved diagnostic performance for early-stage endometrial cancer. The fully adjusted model achieved an AUC of 0.865 and had superior goodness-of-fit (AIC = 144) compared with models lacking lymphocyte or metabolic variables. The authors propose that this multi-parameter, non-invasive approach may provide a more reliable early diagnostic strategy in postmenopausal women.
The study evaluated whether adding simple immune cell counts and clinical metabolic information to routine cancer blood tests could better identify early endometrial cancer in women after menopause. Compared with benign uterine conditions, women with early endometrial cancer had lower lymphocyte counts and lymphocyte-to-monocyte ratios, higher CA125 and HE4, and more metabolic disorders such as diabetes and hypertension. Combining these measures improved the ability to distinguish early cancer from benign disease, with the best-performing model including tumour markers, lymphocyte indicators and metabolic factors.
The abstract provides the primary findings and statistical comparisons but does not report several methodological details commonly important for interpreting diagnostic studies: specific assay methods and cut-offs for CA125 and HE4, exact regression coefficients or thresholds for the combined models, confidence intervals for reported AUCs, measures of sensitivity/specificity for combined models, and whether any internal or external validation was performed. These details were not reported in the available source abstract.