Intrathoracic Mesenchymal Tumors on Computed Tomography: Imaging Spectrum and Predictors of Malignancy

To characterize computed tomography (CT) features of primary intrathoracic mesenchymal tumors and identify imaging findings that differentiate benign from malignant lesions.

This retrospective study included 81 histopathologically confirmed intrathoracic mesenchymal tumors that were evaluated using pretreatment contrast-enhanced CT between January 2010 and January 2025. The tumor location, morphology, attenuation, invasive behavior, and associated findings were assessed and compared between the benign and malignant groups.

Tumors most commonly arose in the mediastinum (32%) and the lungs (32%). Benign tumors (n = 39) were significantly smaller and exhibited a round or oval shape with smooth or lobulated margins (77%). Malignant tumors (n = 42) were larger (mean, 11.2 vs. 6.7 cm), more frequently demonstrated irregular or ill-defined margins, and commonly showed heterogeneous hypoenhancement with extensive low-attenuation areas. On multivariable analysis, tumor necrosis, irregular or spiculated margins, and increasing tumor size were independent predictors of malignancy, whereas calcification showed an inverse association and favored benign pathology.

Intrathoracic mesenchymal tumors demonstrate distinctive CT features that aid in differentiation. Malignant tumors characteristically present as large lesions with necrosis and irregular margins, whereas benign tumors are typically well circumscribed and may demonstrate calcification. Recognition of these patterns supports the early identification of mesenchymal etiology, guides biopsy planning, and informs multidisciplinary management.