Diagnostic performance of computed tomography angiography compared with transesophageal echocardiography for the detection and the analysis of aortic atheroma


  • Conflicts of interest: Ariel Cohen received research grants from Sanofi-Aventis, CPAM/consulting fees from Sanofi-Aventis and ECOACH/lecture fees from Bayer, Bohringer-Ingelheim, Daiichi Sankyo and Sanofi-Aventis.

Correspondence: Nadia Benyounes*, Foundation Rothschild Hospital, 25 Rue Manin, 75019 Paris, France.

E-mail: nbenyounes@fo-rothschild.fr

Aortic arch atheroma is a major source of embolism [1, 2]. Transesophageal echocardiography (TEE) is the gold standard for the detection of aortic atheroma. However, we have witnessed in these recent years an increasing number of requests for computed tomography angiography (CT angiography) to assess the ascending aorta and the aortic arch. This study compared CT angiography and TEE performances for the detection of aortic arch atheroma.

Among the 192 patients admitted to our hospital in 2008–2009 for acute ischemic stroke and in whom transthoracic echocardiography failed to identify a source of embolism, we have retrospectively selected the 47 patients who underwent both TEE and CT angiography. A scoring was established and resulted in the following (Table 1):

  • No atheroma: 0
  • Plaque thickness < 4 mm: 1
  • Plaque thickness ≥ 4 mm (protrusive aortic atheroma): 2
  • Ulcerated plaque: 3
  • At risk atheroma: 2 or 3.
Table 1. Results in the 47 patients of the study
Atheroma scoringn = 47
CT Reader 1 n (%)CT Reader 2 n (%)CT Consensus n (%)


n (%)

018 (38·3)9 (19·2)11 (23·4)16 (34·0)
116 (34·0)24 (51·1)23 (49·0)12 (25·5)
26 (12·8)7 (14·9)5 (10·6)9 (19·2)
37 (14·9)7 (14·9)8 (17·0)10 (21·3)
At-risk atheroma (2 or 3)13 (27·7)14 (29·8)13 (27·7)19 (40·4)

Interobserver reproducibility for CT angiography (two readers) was found to be 57·45% for each score, with a kappa index of 0·4020. For at-risk plaques, interobserver reproducibility was 89·36% (kappa index 0·7403).

The agreement between CT angiography and TEE for each score was 61·70%, with a Kappa index of 0·4816. For at-risk plaques, the agreement between the two techniques was of 74·47% (kappa index 0·4416).

Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of CT angiography compared with TEE for the detection of at-risk plaques were calculated.

The sensitivity was 52·60% (95% CI: 33·50–79·70%) and the specificity was 92·60% (95% CI: 75·70–99·10%).

The PPV was 84·60% (95% CI: 54·60–98·10%) and the NPV was 75·80% (95% CI: 57·70–88·90%).

In conclusion, CT angiography is very specific but lacks sensitivity for the detection of at-risk aortic plaques.