Abstract

AIMS:

To evaluate the diagnostic accuracy of cardiovascular magnetic resonance (CMR) imaging from a risk-stratification and therapeutic-management perspective in patients with suspected cardiac tumours.

METHODS AND RESULTS:

Cardiovascular magnetic resonance exams of 41 consecutive patients (aged 61 ± 14 years, 21 men) referred for evaluation of a suspected cardiac mass were reviewed for tumour morphology and signal characteristics in various unenhanced and contrast-enhanced sequences. Cardiovascular magnetic resonance-derived diagnosis and treatment were compared with clinical outcome and histology in patients undergoing surgery or autopsy (n = 20). In 18 of 41 patients, CMR excluded masses or reclassified them as normal variants; all were treated conservatively. In 23 of 41 patients, CMR diagnosed a neoplasm (14 'benign', 8 'malignant', and 1 'equivocal'); 18 of these patients were operated on, 2 managed conservatively, and 3 by palliation. During follow-up of 705 (inter-quartile range 303-1472) days, 13 patients died. No tumour-related deaths occurred in conservatively managed patients. Patients with a CMR-based diagnosis and treatment of benign tumour had a similar survival as patients without detectable tumour. Compared with histology, CMR correctly classified masses as 'benign or malignant' in 95% of the cases. Tumour perfusion, invasiveness, localization, and pericardial fluid were valuable to distinguish between malignant and benign tumours. Soft tissue contrast and signal intensity patterns in various sequences were valuable for excluding neoplastic lesions and helped to obtain tissue characterization at the histological level in selected tumour cases, respectively.

CONCLUSION:

Comprehensive CMR provides a confident risk-stratification and clinical-management tool in patients with suspected tumours. Patients where CMR excludes tumours can be managed conservatively.