Background : Uterine leiomyosarcoma (ULMS) is a rare, aggressive malignancy with high recurrence and poor survival, necessitating prognostic biomarkers and therapeutic targets. Cyclooxygenase-2 (COX-2), implicated in tumorigenesis and angiogenesis, remains understudied in ULMS. This systematic review and meta-analysis assessed COX-2’s prognostic role in ULMS. Methods : Following PRISMA guidelines, six studies (n=185) were retrieved from PubMed, EMBASE, and Scopus (2001–2024). Inclusion criteria required ULMS cohorts with COX-2 expression data and survival outcomes. Risk of bias was assessed using QUADAS-2, and evidence certainty via GRADE. A random-effects meta-analysis calculated pooled effect estimates with 95% confidence intervals (CIs), while heterogeneity was evaluated using I² statistics. Result : COX-2 expression correlated moderately with epithelial components (pooled effect: 0.51, 95% CI: 0.26–0.77) and weakly with mesenchymal components (0.26, 95% CI: 0.06–0.45). High heterogeneity (I² = 89.5% and 82.2%) reflected differences in study design, tumor subtypes, and COX-2 measurement thresholds. QUADAS-2 indicated a low risk of bias, and GRADE confirmed high evidence certainty. Stronger epithelial correlations were observed in Asian cohorts, highlighting geographic variability. Conclusions: COX-2 plays a more significant role in epithelial-driven ULMS carcinogenesis. Despite heterogeneity, robust methodologies support these findings. Future studies should standardize COX-2 assessment, expand cohort sizes, and integrate multi-omics approaches to refine prognostic and therapeutic applications.