Can multi-phase contrast-enhanced CT be used to differentiate between intra-abdominal and retroperitoneal fat-poor liposarcoma and leiomyosarcoma?

BACKGROUND: This study aimed to assess the diagnostic performance of contrast-enhanced CT in differentiating intra-abdominal and retroperitoneal fat-poor liposarcoma (FPLPS) from leiomyosarcoma (LMS). METHODS: A retrospective analysis was conducted on patients with pathologically confirmed intra-abdominal or retroperitoneal non-uterine LMS and FPLPS. Twelve CT imaging features were assessed, and quantitative parameters, including arterial enhancement fraction (AEF) and extracellular volume fraction (ECV), were calculated. Multivariate logistic regression identified independent predictors, and a combined nomogram model was constructed. Model performance was assessed using ROC analysis, calibration, and decision curve analysis. RESULTS: A total of 56 intra-abdominal or retroperitoneal sarcomas, including non-uterine LMS (n = 27) and FPLPS (n = 29). AEF (OR = 1.07, P = 0.006), mean tumor diameter (OR = 1.29, P = 0.040), and patchy enhancement (OR = 7.58, P = 0.021) were independent predictors for distinguishing FPLPS from LMS. The dynamic nomogram ( https://liexiantu1.shinyapps.io/dynom/ .) demonstrated excellent diagnostic accuracy (AUC = 0.885) with satisfactory calibration, and clinical utility. CONCLUSIONS: Quantitative (AEF) and qualitative (tumor size and enhancement pattern) of CT features reliably distinguish FPLPS from LMS in intra-abdominal and retroperitoneal locations. The nomogram offers a practical, non-invasive tool to support preoperative diagnosis.