Prognostic Value of Preoperative Inflammatory, Hematological, and Metabolic Blood Indices in Endometrioid-Type Endometrial Cancer: A Retrospective Cohort Study

Abstract

Background and Objectives: Peripheral blood indices have been proposed as prognostic biomarkers in endometrial cancer, but most reports include mixed histological subtypes, do not control for benign uterine pathology that independently alters hematological profiles, and evaluate these indices in isolation. We evaluated all three categories head-to-head in a histologically homogeneous endometrioid cohort and formally tested whether they add prognostic information beyond established clinicopathological variables. Materials and Methods: Of 195 patients operated for endometrioid-type endometrial cancer between 2018 and 2024, 142 were included; adenomyosis or leiomyoma on final histopathology, active infection and hematological disease were exclusion criteria. Fourteen preoperative blood-derived indices spanning the three categories were analyzed as continuous variables. A pre-specified clinical model (age category, grade, FIGO stage) was compared with models adding the biomarkers using likelihood-ratio tests, Harrell’s concordance index, time-dependent areas under the curve, bootstrap internal validation, and influence diagnostics. Results: Median follow-up was 52.5 months; 34 patients (23.9%) died and 5-year overall survival was 74.1%. None of the conventional inflammatory ratios were associated with survival (SII p = 0.893; all others p ≥ 0.26). In the multivariable model, histologic grade and NRBC% remained associated with survival (NRBC% HR 3.30 per 1%, 95% CI 1.09–9.97, p = 0.034). Adding NRBC% to the clinical model improved fit (likelihood-ratio p = 0.039) but not discrimination (ΔC 0.006, 95% CI −0.041 to 0.053, p = 0.814), and no index retained significance after Benjamini–Hochberg adjustment. The NRBC% association disappeared on dichotomization at the detection limit and could not be estimated once the five patients with values above 0.30% were excluded. Conclusions: In a predominantly early-stage endometrioid endometrial cancer cohort, preoperative blood-derived indices added no measurable discriminatory information beyond established pathological variables. The NRBC% association rests on very few observations at the upper end of its distribution and is hypothesis-generating only.

The authors' abstract, as published at the source. Medicina, 2026 · DOI ↗