Clinical and Ultrasound Characteristics of High‐Grade Endometrial Stromal Sarcoma

OBJECTIVES: To investigate the clinical features, ultrasound manifestations and diagnostic value of high-grade endometrial stromal sarcoma (HGESS), analyze their correlation with pathological findings, and provide evidence for preoperative diagnosis, differential diagnosis and prognostic evaluation. METHODS: A total of 36 patients with surgico-pathologically confirmed HGESS (2023-2026) were retrospectively enrolled. Control groups included 28 low-grade endometrial stromal sarcoma (LGESS) and 30 uterine leiomyosarcoma (ULMS). This retrospective singlecenter study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting checklist. Clinical, ultrasound, and pathological data were collected. Multivariate logistic regression with bootstrap internal validation, Receiver-operating characteristic (ROC) curves, Kappa test, Spearman's correlation, Kaplan-Meier survival analysis, and Log-rank test were performed. We acknowledged the potential overfitting risk for regression analysis given the limited sample size. RESULTS: The predominant symptoms of HGESS were irregular vaginal bleeding and lower abdominal distension and pain, with markedly shorter disease course than LGESS (p < 0.001). Serum LDH, CA125, and HE4 were significantly elevated in HGESS versus control groups (all p < 0.001). Typical ultrasound findings included ill-defined borders, heterogeneous hypoechogenicity, frequent internal necrosis and cystic degeneration, predominant Grade III blood flow and decreased resistance index (RI) (all p < 0.001). Multivariate logistic regression identified ill-defined borders, heterogeneous internal echo, Grade III blood flow and elevated LDH as independent risk factors for HGESS. Bootstrap internal validation supported model stability. The combined ultrasound-tumor-marker model yielded the area under the curve (AUC) = 0.918, sensitivity = 100%, specificity = 74.1%. Caution should be exercised interpreting this high diagnostic performance given singlecenter retrospective design and sample-size limitation. Lesion maximum diameter (r = 0.621, p < 0.001) and blood-flow abundance (r = 0.584, p < 0.001) positively correlated with the degree of histological malignancy. Shorter progression-free survival (PFS) was observed for patients with lesion > 5 cm (p < 0.001), RI < 0.40 (p = 0.003), and lymph-node metastasis (p < 0.001) in Log-rank tests. CONCLUSIONS: HGESS is highly malignant with rapid progression and characteristic ultrasound signs. Combined ultrasound and tumor-markers improve preoperative differential diagnosis. Ultrasound parameters provide prognostic clues for HGESS.