Key result
CT-based radiomics model shows preliminary ability to differentiate cardiac masses with AUCs up to ~93%.
Why the study?
Cardiac masses present diagnostic challenges on imaging, and differentiating between tumors and non-neoplastic lesions like thrombi is essential for appropriate management.
Can a CT-based radiomics model accurately differentiate between types of cardiac masses (thrombus vs. myxoma vs. other)?
Observational (n=89)
No
Can a CT-based radiomics model accurately differentiate between types of cardiac masses (thrombus vs. myxoma vs. other)?
Effect estimate: AUC 77.1% to 93.3%
A CT-based radiomics model shows proof-of-concept feasibility for noninvasively differentiating between cardiac myxomas, thrombi, and other masses.
May support radiomics differentiation of cardiac masses on CT; hypothesis-generating and requires prospective validation.
INTRODUCTION: Cardiac masses can present a diagnostic challenge on cardiac imaging. Differentiating between tumors such as myxomas and non-neoplastic lesions like thrombi is essential for appropriate management and treatment planning. OBJECTIVES: The aim of this study was to explore the feasibility of radiomics-based "virtual biopsy" in distinguishing between different types of cardiac masses on computed tomography (CT). PATIENTS AND METHODS: 89 patients were included in the analysis. 42 (47.2%) were diagnosed with myxoma, 37 (41.6%) were diagnosed with thrombus and 10 (11.2%) had other diagnoses [fibroelastoma in 6 (6.7%) cases, sarcoma in 2 (2.2%) cases, angioma and metastasis in one (1.1%) case each]. The ROI selection, and calculation of the radiomic features was done with syngo.via Frontier Radiomics. 3D ROI of the cardiac masses for CT examination has been manually delineated. A multivariate LASSO-based model was developed for both binary (thrombus vs. tumor) and multiclass (thrombus vs. myxoma vs. other) classification. RESULTS: The exploratory multiclass radiomics model demonstrated preliminary discriminatory performance, with validation AUCs ranging from 77.1% to 93.3%; however, class-specific estimates, particularly for the heterogeneous "other" category, should be interpreted with caution due to the small validation cohort. These findings suggest that radiomics features can capture subtle differences between distinct cardiac mass types, enabling noninvasive tissue characterization. CONCLUSION: This proof-of-concept study provides the first evidence that a multiclass CT-based radiomics model can differentiate between myxomas, thrombi, and other cardiac masses. Radiomics-based "virtual biopsy" may potentially complement conventional imaging; however, these exploratory findings require validation in larger, multicentre cohorts before clinical application.
No takes yet. Share an insight, caveat, or question.
Trochimiuk et al. (2026) conducted an observational in Abnormal cardiac masses (n=89). CT-based radiomics model (virtual biopsy) was evaluated on Discriminatory performance for multiclass classification (thrombus vs. myxoma vs. other) (AUC 77.1% to 93.3%). A multiclass CT-based radiomics model demonstrated preliminary discriminatory performance in differentiating between distinct cardiac mass types, with validation AUCs ranging from 77.1% to 93.3%.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: