Radiomics Feature Analysis Using Native T1 Mapping for Discriminating Between Cardiac Tumors and Thrombi

This study evaluates the diagnostic performance of radiomics features extracted from native T1 mapping in differentiating cardiac tumors from thrombi using cardiac magnetic resonance imaging (MRI), compared to late gadolinium enhancement (LGE) imaging. The retrospective study included 41 patients with 22 cardiac tumors and 21 thrombi who underwent cardiac MRI. Fifty-six radiomics features were extracted from native T1 images using AVIEW software (Coreline Soft). Feature selection and rad score extraction were performed using the least absolute shrinkage and selection operator method. The rad score showed superior diagnostic performance, with an area under the receiver operating characteristic curve (AUC) of 0.98, compared to mean T1 (AUC 0.86) and LGE ratio (AUC 0.82). The rad score achieved 95.4% sensitivity, 95.2% specificity, and 95.2% accuracy. Combining the rad score with mean T1 significantly improved diagnostic performance. This method offers a valuable alternative for patients who cannot tolerate contrast agents.