Platelet-to-Lymphocyte Ratio as a Predictor of Lymphovascular Space Invasion in Endometrioid Endometrial Cancer:...

Background and Objectives: The relationship between preoperative inflammatory markers and lymphovascular space invasion (LVSI) in endometrioid-type endometrial cancer (EC) remains incompletely defined and warrants evaluation using robust statistical methods. This study aimed to evaluate the independent association of preoperative inflammatory markers, analyzed strictly as continuous variables, with the presence of LVSI, and to develop a refined predictive nomogram adjusted for established clinical confounders. Materials and Methods: Data from 156 patients who underwent standard staging surgery for endometrioid-type EC were retrospectively analysed. To preserve statistical power and avoid structural artifacts from data forcing, preoperative glucose-to-lymphocyte ratio (GLR), platelet-to-lymphocyte ratio (PLR), and neutrophil-to-lymphocyte ratio (NLR) were modeled on their original continuous scale. Multivariable logistic regression analysis was performed to identify independent risk factors for LVSI, adjusting for patient age and maximum tumor diameter. Internal validation was conducted using bootstrap resampling (1000 iterations). Results: In the multivariable logistic regression model, continuous PLR emerged as a significant independent risk factor for the presence of LVSI (adjusted OR: 1.013 per 1-unit increase, 95% CI: 1.001-1.024; p = 0.033). Among clinical parameters, maximum tumor diameter demonstrated the strongest independent association with LVSI (adjusted OR: 1.595 per 1 cm increase, 95% CI: 1.211-2.099; p = 0.001). Continuous NLR (p = 0.513) and GLR (p = 0.545) did not retain statistical significance due to overlapping explanatory variance and shared hematological components. The optimized 3-variable nomogram (PLR, tumor size, and age) demonstrated an apparent C-index of 0.816 (95% bootstrap CI: 0.719-0.920) and a robust optimism-corrected C-index of 0.794. The bootstrap-corrected calibration slope was 0.909, and Decision Curve Analysis (DCA) demonstrated a positive net clinical benefit across clinically relevant threshold probabilities. Conclusions: Preoperative PLR, evaluated as a continuous parameter, provides a statistically stable framework for preoperative risk stratification in endometrioid EC. When integrated with tumor size and age, the proposed nomogram demonstrates promising discriminative performance and potential clinical utility pending external validation for predicting LVSI. However, given the limited number of LVSI-positive events (n = 17), these findings should be regarded as exploratory and hypothesis-generating and require external validation before clinical use.