Exploratory Preoperative Risk Stratification of Uterine Mesenchymal Tumors in Postmenopausal Women with Complementary Transcriptomic Analysis

Background: Preoperative differentiation between uterine mesenchymal tumors and benign uterine fibroids remains challenging in postmenopausal women. This study aimed to develop an exploratory risk-stratification model for uterine mesenchymal tumors and to characterize complementary molecular features of uterine leiomyosarcoma using a public transcriptomic dataset. Methods: A retrospective case–control study was conducted in postmenopausal patients with uterine masses who underwent surgery between 2011 and 2021. A total of 23 uterine mesenchymal tumor cases and 92 frequency-matched fibroid controls were included. Clinical variables were analyzed using univariable and multivariable logistic regression to develop an exploratory risk-stratification model. Model performance was evaluated using ROC analysis, calibration analysis, decision curve analysis, and confusion matrix assessment. Exploratory transcriptomic analysis was performed using the GEO dataset GSE64763 to characterize molecular differences between uterine leiomyosarcoma and fibroid tissues. Results: Pelvic pressure, abnormal uterine bleeding, and tumor diameter were independently associated with uterine mesenchymal tumors. The model achieved an apparent AUC of 0.859 and an optimism-corrected AUC of 0.846 after 1000 bootstrap resamples. Exploratory transcriptomic analysis revealed distinct expression patterns and upregulation of proliferation-associated genes, including CCNB1, BUB1B, PRC1, TOP2A, and FOXM1, with enrichment of cell cycle–related pathways. Conclusions: An exploratory model based on routinely available preoperative variables demonstrated preliminary discriminatory ability within the development cohort. Independent external validation in representative prospective cohorts is required before any assessment of clinical utility.