Author + information
- Received September 10, 2001
- Revision received December 26, 2001
- Accepted January 10, 2002
- Published online April 3, 2002.
- Ehud Schwammenthal, MD*,§,
- Yvonne Schwammenthal, MD†,*,§ (, )
- David Tanne, MD†,§,
- Alexander Tenenbaum, MD*,§,
- Alex Garniek, MD‡,§,
- Michael Motro, MD, FACC*,§,
- Babeth Rabinowitz, MD, FACC*,§,
- Michael Eldar, MD, FACC*,§ and
- Micha S. Feinberg, MD*,§
- ↵*Reprint requests and correspondence:
Dr. Ehud Schwammenthal, Heart Institute, Sheba Medical Center, Tel Hashomer, Israel
Objectives The goal of the present study was to examine whether suprasternal harmonic imaging (SHI) (i.e., harmonic imaging from the suprasternal windows) can visualize protruding arch atheromas (PAAs) and reliably predict the presence or absence of significant lesions.
Background Protruding arch atheromas are a major source of cerebral and peripheral embolism and probably the most frequent cause of stroke during cardiac catheterization and open-heart surgery. Preprocedural screening by transesophageal echocardiography (TEE) would be desirable but is limited by the nature of the examination.
Methods Of 354 patients who underwent a TEE study in our laboratory during the study period, 106 were referred for detection of a source of embolism. Findings were classified based on the French Aortic Plaque study criteria as: 1) no or minimal atherosclerotic changes; 2) PAAs <4 mm; 3) PAAs ≥4 mm or presence of a mobile component.
Results Adequate transcutaneous image quality could be achieved in 89 patients (84%). Protruding arch atheromas were present in 42 patients (47%) and absent in 47 (53%). Positive and negative predictive values for large PAAs on TEE were 91% and 98%, respectively. In one case, SHI detected a complex PAA inaccessible for TEE due to interposition of the left bronchus as demonstrated by dual helical computed tomography. Inter-observer agreement for SHI was 91%.
Conclusions Suprasternal harmonic imaging reliably predicted or excluded the presence of PAAs in a sizable, consecutive group of patients referred to TEE for detection of a source of embolism. It represents an excellent screening test and provides complimentary views of regions, which may be blind spots for TEE.
Aortic arch atheromas are a major source of cerebral and peripheral embolism (1–6)and probably the most frequent cause of stroke during cardiac catheterization and open-heart surgery (7–15). Stroke that complicates surgery involving cardiopulmonary bypass occurs at least six times more frequently in patients with protruding arch atheromas (PAAs) (11,15), has an in-hospital mortality rate of almost 40%, and leaves many survivors severely disabled (15). Preprocedural screening by transesophageal echocardiography (TEE) would be desirable (16–23)but is limited by the nature of the examination, involving the need to place an endoscopic device in the esophagus. The recently introduced technique of harmonic imaging combines good penetration with preserved resolution and an excellent signal-to-noise ratio (24). We therefore hypothesized that harmonic imaging can be used to directly visualize aortic arch atheromas from the jugulum or the supraclavicular fossae and reliably predict the presence or absence of significant lesions.
During the study period, 354 patients underwent a TEE in our laboratory; 106 of these patients were referred for detection of a source of embolism. Seventeen patients (16%) were excluded because of insufficient transthoracic image quality. The remaining 89 patients (84%) represent the study group. There were 29 women and 60 men ranging in age from 19 years to 84 years (61 ± 16 years). Sixty-three patients (71%) were referred after a transitory ischemic attack, and 15 (17%) after a cerebrovascular accident. Two patients (3%) were referred after peripheral embolization, and nine (10%) before planned cardiac surgery.
Suprasternal harmonic imaging (SHI)
Examinations were performed with a System Five ultrasound system (GE Vingmed, Horten, Norway) (1.6 to 1.7 MHz probe), a Sequoia C 256 Acuson (Mountain View, California) (1.75 MHz) or a Hewlett Packard Sonos 5500 (Andover, Massachusetts) (1.8 MHz). Individual patients, in whom a fundamental frequency of 1.8 MHz did not allow sufficient penetration, were restudied with a system that provided lower frequency transmission. The transducer was placed at the jugulum or the supraclavicular fossae and angulated to obtain a long-axis view of the complete aortic arch (Fig. 1). Views of the ascending aorta and proximal arch were obtained by anterior tilting of the transducer; views of the distal arch and descending thoracic aorta were obtained by posterior tilting of the transducer. Special care was taken to identify the origin of the three major branches of the aorta. The aortic arch was systematically scanned using acoustic zoom intermittently to facilitate detail recognition and adjusting gain to optimize signal-to-noise ratio (Fig. 2 to 5). ⇓⇓⇓Based on the classification of the French Study of Aortic Plaques in Stroke group (4,5), findings of the examination were classified as follows: 1) absence of significant atherosclerotic lesions (normal arch or only intimal thickening); 2) presence of aortic arch atheromas protruding not more than 3.9 mm into the lumen; 3) presence of aortic arch atheromas protruding 4 mm or more into the lumen or presence of a mobile component regardless of atheroma size.
All TEE examinations were performed using a multiplane probe (5 to 7 MHz). The ascending aorta was visualized in a long-axis view and also in a short-axis view above the coronary sinuses. After imaging the descending aorta, the probe was gradually pulled back until the curve of the distal arch was visible. The probe was then rotated clockwise to study the proximal arch, and if possible, the distal portion of the ascending aorta using multiple imaging planes.
Classification of aortic arch findings by SHI and TEE were compared using contingency tables (chi-square test). If the studies were grossly discrepant (more than one grade), dual helical computed tomography (DHCT) was performed in an attempt to clarify the cause of the discrepancy (25). Positive and negative predictive values of SHI were calculated for the diagnosis of atheromas of any size as well as only large (≥4 mm) or mobile ones. The SHI studies were read blinded to the findings on TEE. Inter-observer agreement was determined as the ratio of concordant classifications to all comparisons performed. A p value of <0.05 was considered significant.
Forty-seven patients (53%) had a normal aortic arch or only intimal thickening by TEE; 42 patients (47%) had PAAs. In 21 patients (24%), PAAs protruded <4 mm into the lumen, in 21 patients (24%), 4 mm or more. Classification of the aortic arch findings by SHI and TEE/DHCT agreed well (p < 0.0001, chi-square = 122) (Fig. 6). In four of 45 patients classified as having no PAA by SHI, an atheroma of 1 to 2 mm was detected by TEE. In two patients classified as having PAAs of <4 mm by SHI, only intimal thickening (at the most) was found on TEE. In one case, a PAA ≥4 mm was underestimated by SHI. A major discrepancy was present in two cases, and consequently, these patients underwent DHCT. In the first patient, a moderately sized plaque was found by both ultrasound techniques, but SHI suggested the presence of an additional large PAA (in retrospect a misinterpreted artifact) not seen on TEE and DHCT. In the second patient, a freely mobile component compatible with a thrombus superimposed on a plaque was detected by SHI (Fig. 7) but could not be visualized by TEE. Dual helical computed tomography demonstrated that the protruding component of the complex atheroma was obscured by interposition of the left bronchus between aortic arch and esophagus (Fig. 8). In one of four cases, a mobile component on a correctly identified large PAA was missed by SHI. All 12 cases of hypoechoic components and three of four cases of ulcerative components were correctly appreciated by SHI.
Positive and negative predictive values of SHI for the diagnosis of PAA were 95% and 92%, respectively, and for diagnosis of only large or mobile PAAs, 91% and 98%, respectively. Sensitivity and specificity for diagnosing large PAAs were 95% and 97%, respectively. When inadequate studies were counted as transthoracic misses of the TEE findings, sensitivity and specificity were 77% and 83%, respectively (negative predictive value was still 85%). The two independent observers agreed in 81 of the 89 cases (91%, p < 0.0001, chi-square = 123); there were no major discrepancies (Fig. 6).
The present study demonstrates that PAAs can be well visualized transcutaneously using SHI (i.e., harmonic imaging from suprasternal windows). This technique reliably predicted the presence or absence of PAAs in a sizable, consecutive group of patients referred to TEE for detection of a source of embolism. It represents not only an excellent screening test but provides complementary views of regions, which may be blind spots for TEE (Fig. 7). Both the anatomical orientation and the localization of detected atheromas with respect to the origin of the major aortic branches are more readily achieved with SHI than TEE, because SHI is facilitated by the long-axis view of the complete aortic arch (Fig. 1).
Weinberger et al. (26)first demonstrated the feasibility of the transcutaneous approach for imaging atherosclerotic plaques in the aortic arch of stroke patients compared with TEE, as well as for correlating plaque morphology to cerebrovascular symptoms (27). However, because of the small sample size, no attempts were made to assess the diagnostic accuracy of the method (26). Although the high-frequency probe the authors used may provide excellent resolution, it does so at the expense of ultrasound penetration. It may therefore be difficult to obtain a complete study of the proximal thoracic aorta in adult patients, especially in the elderly. In contrast, harmonic imaging provides good penetration, preserved resolution, and reduced signal-to-noise ratio, particularly in the near field (24), which is important given the relatively small distance between jugulum and aortic arch.
Stroke prevention is of paramount importance in an era characterized by an increasing number of cardiovascular procedures performed in an aging population. Suprasternal harmonic imaging may facilitate preprocedural screening of patients at risk for stroke during cardiac interventions, particularly cardiac surgery involving cardiopulmonary bypass or port-access using an intra-aortic endoclamp. In patients in whom a significant PAA can be clearly demonstrated by SHI, or no evidence of atheroma can be detected despite good imaging quality, TEE would probably not alter the management strategy. In cases in which image quality is inadequate to reliably rule out atheromas, or in which there are inconclusive findings or evidence of plaque with insufficient image quality to evaluate plaque characteristics, TEE should be performed. In addition, transcutaneous screening of unselected asymptomatic patients within a primary prevention program could provide data about the natural risk of asymptomatic patients with PAAs unrelated to cardiac interventions.
When imaging is performed from the posteriorly located esophagus, the left bronchus (at its origin from the trachea) may obscure a portion of varying size of the aortic arch. The size of this blind segment depends on the diameter of the bronchus, its take-off angle and the curvature of the aortic arch. The segment is readily accessible from the anteriorly positioned suprasternal window (Fig. 8). In the present study a mobile thrombus superimposed on an aortic arch atheroma was clearly demonstrated by SHI (Fig. 7), but missed by TEE, in a patient without prior embolic event who was scheduled for cardiac catheterization and open-heart surgery. Intervention was deferred, coumadine therapy was instituted, and the thrombus gradually and uneventfully dissolved. Every TEE examination should be complemented by careful imaging from the suprasternal window to close the blind gap and perform a truly comprehensive ultrasound examination of the proximal thoracic aorta.
Achieving a high degree of versatility and accuracy in SHI requires a learning process. Comparison of the ultrasound studies with findings on DHCT in the sagittal plane in patients with known PAAs (not included in the study) was extremely useful (25). Adjustment of gain and contrast settings is mandatory in order to avoid missing “soft,” hypoechoic, noncalcified plaques (Figs. 3 and 4), morphologic features, which are associated with high embolic risk and may represent lipid-rich atheromas with or without thrombus formation (19,20).
Adequate image quality for evaluating the aortic arch transcutaneously cannot be obtained in 16% of the patients using current state-of-the-art equipment, which seems acceptable for screening purposes. Frequently, a positive diagnosis can be made even when imaging quality is less than perfect, because in patients who are difficult to scan, image quality is usually at its worst in areas proximal to the inominate artery (where significant plaques are rare) (28)and distal to the left subclavian artery (irrelevant for stroke) but is relatively preserved in the arch. Because the descending thoracic aorta cannot usually be visualized completely, TEE is recommended in patients referred for detection of a source of peripheral (and not cerebral) embolism, if suprasternal imaging does not detect significant findings.
The authors thank Lili Ezer, RCDS, and Ziva Dror, RCDS, for their dedication to this project, and Meirav Moreno for her expert secretarial assistance.
☆ Supported by a grant from the Israel Science Foundation.
- dual helical computed tomography
- protruding aortic arch atheroma
- suprasternal harmonic imaging
- transesophageal echocardiography
- Received September 10, 2001.
- Revision received December 26, 2001.
- Accepted January 10, 2002.
- American College of Cardiology Foundation
- Tunick P.A.,
- Kronzon I.
- Dávila-Román V.G.,
- Barzilai B.,
- Wareing T.H.,
- Murphy S.F.,
- Schechtman K.B.,
- Kouchoukos N.T.
- Tunick P.A.,
- Rosenzweig B.P.,
- Katz E.S.,
- Freedberg R.S.,
- Perez J.L.,
- Kronzon I.
- Keeley E.C.,
- Grines C.L.
- Katz E.S.,
- Tunick P.A.,
- Rusinek H.,
- et al.
- Wareing T.H.,
- Dávila-Román B.G.,
- Barzilai B.,
- et al.
- Pop G.,
- Sutherland G.R.,
- Koudstaal P.J.,
- Sit T.W.,
- de Jong G.,
- Roelandt J.R.T.C.
- Montgomery D.H.,
- Ververis J.J.,
- McGorisk G.,
- Frohwein S.,
- Martin R.P.
- Tunick P.A.,
- Kronzon I.
- the FAPS Investigators,
- Cohen A.,
- Tzourio C.,
- Bertrand B.,
- et al.
- Vaduganathan P.,
- Newton A.,
- Nagueh S.F.,
- et al.
- Weinberger J.,
- Azhar S.,
- Danisi F.,
- et al.
- Katibzadeh M.,
- Mitusch R.,
- Stierle U.,
- Gromoll B.,
- Sheikhzade A.