1016 CARDIAC MAGNETIC RESONANCE IMAGING IN DIFFERENTIAL DIAGNOSIS OF CARDIAC MASSES

Abstract

Abstract Background Differential diagnosis of cardiac masses represents a challenging issue with important implications for therapeutic management and patient's prognosis. Cardiac Magnetic Resonance (CMR) is a non-invasive imaging technique used to characterize morphologic and functional features of masses. Integration of these information can lead to an accurate diagnosis. Purpose To evaluate the diagnostic role of CMR in defining the nature of cardiac masses. Methods One hundred-fourteen patients with cardiac masses evaluated with CMR were enrolled. All masses had histological certainty. CMR sequences allowed a qualitative morphologic description as well as tissue characterization. Evaluation of masses morphology included localization, size and borders assessment, detection of potential multiple lesions and pericardial effusion. Tissue characterization resulted from an estimation of contrast enhancement - early gadolinium enhancement (EGE) and late gadolinium enhancement (LGE) sequences - and tissue homogeneity in T1 and T2 weighted acquisitions. The descriptive analysis was carried out by comparing benign vs malignant lesions as well as dividing patients into 4 subgroups: primitive benign tumours, primitive malignant tumours, metastatic tumours and pseudotumours. Results The descriptive analysis of morphologic features showed that diameter > 50 mm, invasion of surrounding planes, irregular margins and presence of pericardial effusion were able to predict malignancy (p < 0.001). As for tissue characteristics, heterogeneous signal intensity - independently from T1 and T2 weighted acquisitions - and EGE were more common in malignant lesions (p <0.001). When analysing the four subgroups, CMR features did not discriminate between primitive malignant masses and metastasis. Conversely, hyperintensity signal in T2 w-TSE, EGE and heterogeneity after gadolinium enhancement were more common in benign primitive lesions compared to pseudotumours. Furthermore, using classification and regression tree (CART) analysis, we developed an algorithm to differentiate masses: invasion of surrounding planes was a common characteristic of malignancy and identifies itself malignant tumours. In the absence of invasive features, gadolinium enhancement was evaluated: the lack of contrast uptake was able to increase the probability of a pseudotumour. Last step of decision algorithm included ejection fraction assessment to discriminate between benign tumours: a reduced ejection fraction increased the probability of pseudotumour diagnosis and reduced the probability of primary benign tumour diagnosis. Conclusions Cardiac magnetic resonance is a very powerful diagnostic tool for differential diagnosis of cardiac masses as it correctly addresses malignancy. Furthermore, an accurate evaluation of the several CMR features may discriminate primary benign masses and pseudotumours.