Background and Objectives: Uterine FIGO grade 3 endometrioid carcinoma (EC) is an uncommon but aggressive subtype of endometrial cancer with limited biomarker data to guide prognosis and management. This study aimed to evaluate the prognostic significance of programmed death-1 (PD-1) and programmed death-ligand 1 (PD-L1) expression in tumor tissue (TT) and tumor microenvironment (TME). Materials and Methods: We retrospectively analyzed tumor samples from 53 patients with FIGO grade 3 EC. Immunohistochemistry was performed to assess PD-1 and PD-L1 expression in TT and TME. Clinicopathological data including age, stage, lymph node invasion (LNI), lymphovascular space invasion (LVSI), depth of myometrial invasion (MI), adjuvant therapy, and survival outcomes were collected. Survival analyses were conducted using Kaplan–Meier and Cox proportional hazards models. Results: PD-1 expression was identified in 34% of TT and 41.5% of TME, while PD-L1 was expressed in 22.6% of TT and 34% of TME. Except for PD-1 in TME, positive expression of these immune checkpoint molecules correlated with significantly shorter survival (log-rank p < 0.05) outcomes. In univariate analysis, PD-1 and PD-L1 expression in TT, deep MI, LNI and LVSI were associated with adverse outcomes. Multivariate analysis confirmed PD-1 and PD-L1 positivity in TT as independent prognostic factors (PD-1: HR 3.2, 95% CI 1.4–7.0; PD-L1: HR 3.3, 95% CI 1.4–7.8). Patients with concurrent PD-1 and PD-L1 expression in TT showed the poorest overall survival, suggesting a cumulative negative effect. Conclusions: PD-1 and PD-L1 expression in tumor tissue are independent predictors of poor prognosis in FIGO grade 3 EC. These findings support their role as clinically relevant biomarkers and potential therapeutic targets. Incorporating checkpoint evaluation into routine pathological assessment could improve prognostic accuracy and guide treatment strategies, particularly in high-risk patients who might benefit from immunotherapy approaches.
Bagir et al. (Fri,) studied this question.