Mishra, Manisha MD*; Khurana, Poonam MD†; Meharwal, Zile S. MCh‡; Trehan, Naresh MD‡
Acute aortic dissection (AAD) is a cardiovascular emergency that requires prompt diagnosis and treatment. The mortality rate if untreated is approximately 25% during the first 24 hours, 70% during the first 2 weeks, and 90% after 2 weeks.1,2 The most common complications are rupture of the dissection into the pericardium with progressive cardiac tamponade, occlusion of the coronary or supraaortic vessels, and severe aortic regurgitation with heart failure. Aortography has been the traditional method of assessing suspected AAD. However, the procedure is invasive, time consuming, and has low sensitivity. This has prompted the development of other imaging techniques for this purpose, with reported sensitivity as low as 88% in the European multicenter study of Erbel et al.3
Transesophageal echocardiography (TEE) and magnetic resonance imaging are increasingly being used in the evaluation of AAD, with sensitivities and specificities of 95% to 100%. A recent study found that the sensitivity and specificity of multislice computed tomography (MSCT) compare well with those of magnetic resonance imaging and TEE.4
The dissection is termed “acute” when diagnosed within 14 days after the first symptoms appear; it is termed “chronic” when diagnosed later. Immediate appropriate treatment has improved the outcome of AAD: the overall in-hospital mortality rate is currently less than 30%. Patients with aortic dissection require continuous surveillance. Residual aortic disease deteriorates into life-threatening conditions that require surgery in 15% to 30% of patients after 10 years.5 Dilatation of the dissected region and progressive reduction of organ perfusion are the most frequent conditions.
When using MSCT to assess AAD, it is important to evaluate the entire aorta to determine the distal extent of the dissection and to detect abdominal ischemic disease that can increase the associated morbidity and mortality. Primary reconstructed transverse sections remain the mainstay of computed tomography angiographic interpretation. The presence, location, and extent of an intimal flap may be readily determined with MSCT, but the technique is limited in assessing coronary artery involvement and aortic regurgitation.6 Alternative visualization techniques, including multiplane and 3-dimensional reformation images, can substantially augment diagnosis and provide an efficient means of communicating critical anatomic relationships to referring clinicians.7
The recently introduced live 3-dimensional echocardiography is a promising new technique.8 This prospective study was undertaken to compare the accuracy of intraoperative live 3-dimensional epicardial echocardiography with that of multiplane TEE and MSCT imaging in the detection of thoracic aortic dissection, for the identification of intimal tear and for the involvement of coronary arteries and arch vessels.
To our knowledge, this has never been reported in the literature.
Of 24 patients studied from October 2003 to September 2004, 12 had DeBakey type I aortic dissection. All patients first underwent a multiplane TEE examination, followed by MSCT, preoperatively. Live 3-dimensional epicardial echocardiography was performed on all patients intraoperatively. Only patients who underwent all imaging procedures were included in the study. Of the 24 patients in whom all 3 examinations were performed, 12 had DeBakey type I aortic dissection, 7 had DeBakey type II, and 5 had ascending aorta dilation with severe aortic regurgitation. Patients with DeBakey type I aortic dissection comprised the study group. Ethical committee approval and informed consent were obtained for all patients. The mean age was 52.5 ± 12.6 years. There were 11 men and 1 woman. The demographic profile of the patients is shown in Table 1.
Multislice Computed Tomography
The MSCT studies were performed with a 16-slice Sensation Cardiac (Siemens) CT scanner. Noncontrast scans were done first to localize the aorta and to depict intimal calcification that might be obscured by contrast material. The MSCT scans were subsequently performed during dual phase power injection of 100 to125 mL nonionic contrast intravenously at the flow rate of 3.5 to 4.0 mL/s using a real-time bolus tracking technique. Craniocaudal transverse scans of the area from sternal notch to the groin were obtained using 0.75 collimators for data acquisition. A table speed of 15 mm per rotation and a gantry rotation period of 0.0.37 seconds were used. The voltage was 120 kV, with a current of 240 mA. Transverse sections were reconstructed at 1-mm intervals. Multiplanar reconstruction, maximum intensity projection, and 3-dimensional construction were performed with standard scanner software.
The classic feature of aortic dissection on computed tomographic imaging is a partition between the true and false channels, formed by the intimal flap. Secondary findings include internal displacement of intimal calcification or hyperattenuating intima, delayed enhancement of false lumen, widening of the aorta, and mediastinal, pleural, or pericardial hematoma.
MULTIPLANE TRANSESOPHAGEAL ECHOCARDIOGRAPHY
Two-dimensional and Doppler color flow echocardiograms were obtained using Hewlett-Packard Sonos 5500 imaging systems (Andover, MA) and a multiplane dual frequency phased-array 3.7/5 MHz transesophageal transducer. The protocol for imaging in all patients included the standard images from 1) a transesophageal four-chamber, 3-chamber, and longitudinal view; and 2) transgastric mitral valve, midpapillary, apical, and long-axis views to assess the function of valves and contractility of both ventricles. The proximal and mid–ascending aorta was seen with TEE through the midesophageal window. The midesophageal ascending aorta short-axis view was developed by locating the ascending aorta in the center of the image and adjusting the multiplane angle until the vessel appears circular, usually between 0° and 60°. The multiplane angle was rotated forward between 100° and 150° to develop a midesophageal ascending aorta long axis view, in which anterior and posterior walls appear parallel to each other. The diameter of the ascending aorta at the sinotubular junction and at specified distances from the sinotubular junction or aortic valve annulus was measured from long-axis and short-axis images.
The descending aorta was visualized by rotating the transducer to the left from the midesophageal four-chamber view until circular, short-axis image of the vessel was located in the center of the near field of the display producing the descending aortic short-axis view. The image depth was decreased to 6 to 8 cm to increase the size of the aorta in display. The multiplane angle was rotated forward from 0° to between 90° and 110° to yield a circular, oblique, and finally descending aortic long-axis view, in which the walls of the descending aorta appeared as 2 parallel lines. The probe was then introduced into the stomach and gradually withdrawn, while recording all images from 0° to 180°, until posterior to the aorta at the level of the distal aortic arch. The aortic arch was imaged with the multiplane angle at 0° by withdrawing the probe, while maintaining an image of the descending thoracic aorta until the upper esophageal window was reached. This technique allowed the development of the upper esophageal aortic long-axis view. The proximal arch is to the left of the display and the distal arch is to the right. The multiplane angle was rotated forward to 90° to develop the upper esophageal aortic short-axis view.
Withdrawing the transducer further allowed imaging of the proximal left subclavian artery and left common carotid artery. The right brachiocephalic artery is more difficult to image because of the interposition of the air-filled trachea. As the transducer was withdrawn, it was turned to the left to follow the left subclavian artery, distally. In the upper esophageal aortic arch view, the origin of the great vessels is often identified at the superior aspect of the arch to the right of the display.9 The morphology of the entire aorta was evaluated, providing information regarding the site of intimal tear at entry, direction of the propagation of this tear, presence of hematoma in the false lumen, communication between true and false lumens, size of the true lumen and false lumens, and involvement of the coronary arteries and neck vessels.
The differentiation of true and false lumens is important and is based on the following features: 1) the true lumen is usually small and the false lumen significantly larger; 2) the false lumen usually has some degree of thrombus formation; 3) the blood flow in the false lumen is generally slower than that in the true lumen as demonstrated by spontaneous echocontrast (“swirling”) in the false lumen; and 4) the velocity, direction, and timing of the color flow Doppler signal within the lumen and at the intimal tear help distinguish the false and true lumens. With large proximal entry, flow into the nearby segments of the false lumen may have the same direction and timing as the true lumen flow and may reverse in diastole. With smaller or more distal tears, false lumen flow is less similar to true lumen flow and may be directed in the opposite direction. Peak velocity may also occur later in the cardiac cycle, representing delay of flow into the false lumen. Multiple communications between the true and false lumen were identified on careful scanning of the entire thoracic aorta.
Doppler color flow imaging was performed using enhanced and variance mapping, with an optimal Doppler signal obtained by angling the transducer tip to interrogate blood flow in multiple planes in all views. Patients were starved for 1 hour if stable.
Intraoperatively, the probe was introduced immediately after induction of general anesthesia and endotracheal intubation, and the examination validated the findings noted on the preoperative TEE and MSCT. Initial images were obtained before the skin incision.
LIVE THREE-DIMENSIONAL ECHOCARDIOGRAPHY
All live 3-dimensional epicardial echocardiography studies were performed after sternotomy, using a Philips Sonos 7500 ultrasound system and a broadband 2 to 4 MHz, X4 MATRIX transducer, fully sampled phased array transducer with both elevation and lateral dimensional control for real time 3-dimensional volume image acquisition and rendering using 4 cardiac beats in the electrocardiographically triggered mode. These volumes can be time aligned to render a full-volume 3-dimensional image, along with respiratory gating by turning off the ventilator. Brightness and contrast function keys were used to optimize 3-dimensional image quality, and the images were cropped using X, Y, and Z cutting planes to obtain the 3-dimensional perspective of the dissection flap in the aorta. Acquisition of 3-dimensional full volume images from real-time 3-dimensional images and cropping in the 3 axes took less than 5 minutes.
The following epicardial views were used as recommended for 2-dimensional transesophageal echocardiography9,10:
1. Aortic root view: This view is obtained by placing the probe on the aortic root to permit evaluation of the aortic valve and proximal ascending aorta as well as the LVOT.
2. Short-axis and long-axis equivalent views of the ascending aorta: By moving the epicardial probe in a cephalad direction from the aortic root position, a short-axis view of the ascending aorta can be obtained. Turning the probe 90° permits visualization of the ascending aorta in the long axis.
3. Proximal aortic arch equivalent view: Moving the probe slightly more cephalad from the ascending aorta position provides a long-axis view of the proximal aortic arch and great vessels, which corresponds to the TEE aortic arch long-axis view. This view may be particularly advantageous in patients with an acute intraoperative aortic dissection to assess the exact extension of the dissection and to identify a true or false lumen.
All images of each study were evaluated in a blinded fashion by an expert in interpreting results of that imaging technique. Imaging results in terms of detection of aortic dissection, site of intimal tear, and involvement of coronary arteries and arch vessels were confirmed during operative exploration by the operating surgeon as the reference standard.
Sensitivity, specificity, and accuracy of each imaging technique in the detection of aortic dissection, site of intimal tear, and involvement of supraaortic vessel were calculated and compared by mean of statistical analysis with the McNemar test for paired data. Findings were considered statistically significant when the P value was less than 0.05. 95% confidence intervals were calculated for the sensitivity, specificity, and accuracy of each imaging method.
For all imaging techniques and in all patients, the diagnosis was made based on the visualization of an intimal flap in at least 1 part of the thoracic aorta. Partial thrombosis of the lumen was demonstrated in 5 patients. None of the patients had complete thrombosis of any segment of the aorta. There were no findings indicating the presence of an intramural hematoma or complete thrombosis of the false lumen without detectable intimal flap separating the 2 perfused lumina in any segment of thoracic aorta.
All patients were symptomatic. In 7 patients, the flap arose from the aortic root; in 4 patients, the coronary arteries were also involved, whereas the arch vessels were involved in 9 patients. One patient had undergone coronary artery bypass grafting and one had aortic valve replacement 3 years prior.
Results of live 3-dimensional epicardial echocardiography, multiplane TEE, and MSCT are presented in Table 2. MSCT had an accuracy of 100% in detection of aortic dissection in all segments of the thoracic aorta (ascending, aortic arch and descending aorta). Even though the definitive diagnosis of aortic dissection was made preoperatively by MSCT and TEE, live 3-dimensional epicardial echocardiography enhanced the imaging details to the surgical team. The intimal flap was not visualized as a linear echo but as sheet of tissue by live 3-dimensional epicardial echocardiography, depicting the splitting of the aortic wall by the dissection process.
Differences between live 3-dimensional epicardial echocardiography, multiplane TEE and MSCT in the detection of aortic dissection were not statistically significant. For the detection of site of intimal tear, the sensitivity of 3-dimensional epicardial echocardiography and multiplane TEE was 92%.The accuracy of 3-dimensional epicardial echocardiography was slightly higher than multiplane TEE (88% versus 83%). In 1 patient, the site of intimal tear was in the proximal aortic arch and could not be detected by multiplane TEE, but was detected intraoperatively by 3-dimensional epicardial echocardiography.
Of the arch vessels, the brachiocephalic artery was the most commonly involved (75%), followed by the left subclavian artery (50%), and the left carotid artery (25%) was involved the least. In 3 patients, dissection was demonstrated to be extending into all 3 supraaortic branches. There were no patients in whom the supraaortic vessels were perfused solely by the false lumen, and no patients had any clinical signs of arch vessel involvement. Thirty-day mortality was 8.3% (Table 3), and 1 patient suffered a stroke.
Multislice computed tomography and live 3-dimensional epicardial echocardiography had significantly better accuracy in the assessment of arch vessel involvement (P < 0.05) compared with multiplane TEE. However, live 3-dimensional epicardial echocardiography and TEE were superior in detection of the site of intimal tear (P < 0.05).
The mean investigation time was 5.0 ± 3 minutes for live 3-dimensional epicardial echocardiography, 5 ± 4 minutes for multiplane TEE, and 18 ± 8 minutes for MSCT.
Early detection of acute aortic dissection of the thoracic aorta is important, because the mortality in these patients can be as high as 1% to 2% per hour during the first 2 days if not treated.5 Surgical planning for aortic dissection depends on several factors. Most surgeons require information regarding the extent of the dissection, the location of primary tears, concomitant coronary artery disease, morphology of the aortic valve, presence of associated aortic regurgitation, left ventricular function, and other valvular pathology, and the presence and size of pericardial and pleural effusions.
Transesophageal echocardiography, by virtue of the close proximity of the esophagus to the aorta, can image the entire thoracic aorta thoroughly. The principal benefits of perioperative TEE are identification of safe cannulation sites, determination of aortic valve repair strategies, identification of the need for coronary reimplantation with root replacement, and confirmation of adequate cardiopulmonary bypass. In addition, during aortic reconstruction TEE is useful for assessing hemodynamic status (eg, after cross-clamping), to document entry and exit sites, to confirm decompression of the false lumen, and to determine whether valve surgery is needed.11 Time to diagnosis and intervention is critical when aortic dissection is suspected. The morphology of the entire aorta, information regarding the site of intimal tear at entry, direction of the propagation of this tear, presence of hematoma in the false lumen, communication between true and false lumens, size of the true and false lumens, and involvement of the coronary arteries and neck vessels can all be assessed confidently by TEE.3,12
Involvement of the coronary arteries by acute dissection has been estimated to occur in 10% to 20% of cases. Adequate views of the ostia and proximal vessels can be obtained in 88% of left main arteries and 50% of right coronary arteries. Erbel and the European Cooperative Study Group for Echocardiography published a multicenter study showing the diagnostic accuracy in 164 consecutive patients with suspected aortic dissection. The sensitivity and specificity were 99% and 98%, respectively, for TEE compared with 83% and 100%, respectively, for computed tomographic scanning, and 88% and 94%, respectively, for aortic angiography.3 In this study, the accuracy of TEE was 96% in detecting aortic dissection, 83% in detecting the site of intimal tear, and 54% in assessing the involvement of arch vessels.
There are also pitfalls to this diagnostic modality. It is difficult to optimally measure the size of the aorta in a patient with a tortuous, ectatic aorta, and to accurately determine the true level of the imaging plane. Patients with an ectatic or aneurysmaly dilated aorta generally have some degree of mural thrombus. In these patients it may be difficult to separate a thrombosed dissection from a longstanding mural thrombus.12
For MSCT, sensitivities of 83% to 94% and specificities of 87% to 100% have been reported in large, prospective studies for the evaluation of aortic dissection with conventional incremental of dynamic CT.3,13,14 Benefits of MSCT include the considerably shorter examination time and the potential for better evaluation of vascular structures, because more images will be obtained during peak levels of enhancement owing to better tracking of the contrast material bolus.15 MSCT also gives high-quality 2- and 3-dimensional reconstructions.16 The images can be produced in several colors and provide a more realistic 3-dimensional views.7 In a prospective study of aortic dissection, MSCT results have been excellent, with an accuracy of 100% in the detection of thoracic aortic dissection and an accuracy of 88% in the assessment of vessel involvement and 54% in detection of site of intimal tear.17
There are few reports of the use of live 3-dimensional echocardiography in the diagnosis of aortic dissection. However Htay et al.8 reported the use of live 3-dimensional transthoracic echocardiography in 10 patients with aortic dissection. In their experience, live 3-dimensional transthoracic echocardiography increased the confidence level of the diagnosis even though it had been diagnosed earlier by other methods. In our experience, live 3-dimensional epicardial echocardiography had 100% sensitivity for detecting aortic dissection, 92% sensitivity for detecting the site of intimal tear where it was beyond the arch of aorta because that area could not be visualized by an epicardial imaging technique, and 92% sensitivity for assessing arch vessel involvement. Imaging the structures in the near field is difficult unless a good standoff is achieved. The mean duration for a comprehensive examination was 5.0 ± 3 minutes, which is consistent with other studies.10
In summary, our initial experience with live 3-dimensional epicardial echocardiography demonstrates its accuracy and usefulness in the diagnosis of thoracic aortic dissection.
The authors thank Mr. Sudhir Shekhawat for providing assistance in statistical analysis and Mr. Sanjeev Kumar for secretarial assistance.
© 2005 Lippincott Williams & Wilkins, Inc.