BACKGROUND Aortic dissection requires prompt and reliable diagnosis to reduce the high mortality. The purpose of this study was to assess the reliability of both ECG-triggered magnetic resonance imaging (MRI) and transesophageal two-dimensional echocardiography combined with color-coded Doppler flow imaging (TEE) for the diagnosis of thoracic aortic dissection and associated epiphenomena. RESULTS Fifty-three consecutive patients with clinically suspected aortic dissection were subjected to a dual noninvasive imaging protocol in random order; imaging results were compared and validated against the independent morphological "gold standard" of intraoperative findings (n = 27), necropsy (n = 7), and/or contrast angiography (n = 53). No serious side effects were encountered with either imaging method. In contrast to a precursory screening transthoracic echogram, the sensitivities of both MRI and TEE were 100% for detecting a dissection of the thoracic aorta irrespective of its location. The specificity of TEE, however, was lower than the specificity of MRI for a dissection (TEE, 68.2% versus MRI, 100%; p less than 0.005), which resulted mainly from false-positive TEE findings confined to the ascending segment of the aorta (TEE, 78.8% versus MRI, 100%; p less than 0.01). In addition, MRI proved to be more sensitive than TEE in detecting the formation of thrombus in the false lumen of both the aortic arch (p less than 0.01) and the descending segment of the aorta (p less than 0.05). There were no discrepancies between the two imaging techniques in detecting the site of entry to a dissection, aortic regurgitation, or pericardial effusion. CONCLUSIONS Both MRI and TEE are atraumatic, safe, and highly sensitive methods to identify and classify acute and subacute dissections of the entire thoracic aorta. TEE, however, is associated with lower specificity for lesions in the ascending aorta. These results may still favor TEE as a semi-invasive diagnostic procedure after a precursory screening transthoracic echogram in suspected aortic dissection, but they establish MRI as an excellent method to avoid false-positive findings. Anatomic mapping by MRI may emerge as the most comprehensive approach and morphological standard to guide surgical interventions.