Current issues of ACP Journal Club are published in Annals of Internal Medicine


MRI was more accurate than TEE in detecting aortic dissection

ACP J Club. 1992 May-June;116:85. doi:10.7326/ACPJC-1992-116-3-085

Related Content in the Archives
Noninvasive detection of thoracic aortic dissection

Source Citation

Nienaber CA, Spielmann RP, von Kodolitsch Y, et al. Diagnosis of thoracic aortic dissection. Magnetic resonance imaging versus transesophageal echocardiography. Circulation. 1992 Feb;85:434-47.



To evaluate magnetic resonance imaging (MRI) and transesophageal echocardiography (TEE) for the noninvasive diagnosis and classification of thoracic aortic dissection.


Assessment of independently interpreted MRI and TEE tests done on the same patients in random order and compared with angiography and follow-up findings.


Emergency department and intensive care and radiologic units in a university-affiliated hospital in Germany.


53 consecutive patients (mean age, 52 y; 35 men) with suspected dissection of the thoracic aorta. Most patients had hypertension; 4 patients had previous aortic surgery. Unstable angina, myocardial infarction, pulmonary embolism, and pneumothorax were excluded before the study.

Description of tests and diagnostic standard

2-dimensional TEE included color Doppler flow mapping and pulsed-wave Doppler echograms using a wide-angle, phased-array transducer. The esophageal probe was rotated, tilted, and gradually withdrawn to allow mapping of the descending and ascending aorta and arch. Examination required 5 to 14 minutes. Transverse MRI scans in 8- to 10-mm slices over the aorta of the supine patient included a 40-cm field. A whole-body, high-field-strength scanner was used. Examination required 15 to 65 minutes, which included cine-MRI in 43 patients.

The diagnostic standard was a composite of contrast angiography and intraoperative or necropsy findings or both (31 patients), or angiography and uneventful mean follow-up of 16 months (22 patients).

Main outcome measures

Sensitivity and specificity of the tests for detecting any dissection and for classifying dissections of type A (ascending aorta involved) and type B (ascending aorta not involved).

Main results

The diagnostic standard classified 20, 11, and 22 patients with type A, type B, or no dissection, respectively. MRI had sensitivity and specificity of 100% {and likelihood of a positive test (+LR) of infinity and of a negative test (-LR) of 0.0}* for either type of dissection. TEE had sensitivity, specificity, and likelihood ratio of a positive and negative test for any dissection of 100%, 68%, {3.1, and 0}*; for type A of 100%, 79%, {4.8, and 0}*; and for type B of 91%, 98%, {46, and 0.09}*.


Magnetic resonance imaging was completely accurate in detecting thoracic dissection in patients clinically suspected to have this condition. Transesophageal echocardiography was equally sensitive but less specific: 32% of those without dissection were incorrectly classified.

Source of funding: Not stated.

Address for article reprint: Dr. C.A. Nienaber, Department of Internal Medicine II, Division of Cardiology, University Hospital Eppendorf, D-2000 Hamburg 20, Martinistrasse 52, Germany.

*Numbers calculated from data in article.


Prompt diagnosis is crucial to the successful management of acute aortic dissection. Previous studies have shown that both MRI (1) and TEE (2, 3) have high sensitivity and specificity in making the diagnosis. This study confirms the accuracy of MRI and the high sensitivity of TEE, but finds a significantly lower specificity of TEE than has been previously described (2, 3).

These investigators examined a clinically relevant population of patients with suspected but not yet proven dissections. Multiple modalities were used to study each patient and interpreters were blinded to the final diagnosis.

The 100% sensitivity and specificity of MRI suggests that it may be the new "gold standard" for diagnosing dissections. Because MRI is noninvasive, it is all the more appealing. Unfortunately, urgent MRI is not available in many hospitals, and it may be undesirable for hemodynamically unstable patients. TEE is also noninvasive, can be done at the bedside, and is faster to do than angiography or MRI. The 100% sensitivity makes TEE an excellent screening test for dissection, but the low specificity suggests that a positive TEE should be confirmed by angiography or MRI. Surprisingly, a previous study (2) with a larger sample and similar methods found a specificity of 98% for TEE. Further evaluation of false positives is needed, and perhaps a stricter definition of a positive TEE (3) might improve specificity without compromising sensitivity. Institutional availability may ultimately dictate how each hospital uses these modalities.

Eric Isselbacher, MD
Joaguin Cigarroa, MDKim Eagle, MDHarvard Medical SchoolBoston, Massachusetts, USA


1. Kersting-Sommerhoff BA, Higgins CB, White RD, et al. Aortic dissection: sensitivity and specificity of MR imaging. Radiology. 1988;166:651-6.

2. Erbel R, Daniel W, Visser C, et al. Echocardiography in diagnosis of aortic dissection. Lancet. 1989;1:457-60.

3. Ballal RS, Nanda NC, Gatewood R, et al. Usefulness of transesophageal echocardiography in assessment of aortic dissection. Circulation. 1991; 84:1903-14.