Eur. Radiol. 8, 1099±1105 (1998) Ó Springer-Verlag 1998

European Radiology

Non-invasive vascular imaging

Review article Non-invasive imaging of the great vessels of the chest G. Marchal, J. Bogaert Department of Radiology, University Hospitals Leuven, Herestraat 49, B-3000 Leuven, Belgium Received 8 April 1998; Accepted 16 April 1998

Abstract. Several excellent imaging modalities are available for studying the great vessels of the chest noninvasively. Besides computed tomography (CT), magnetic resonance imaging (MRI) and echocardiography (in particular the transesophageal approach) can accurately depict abnormalities of the thoracic vasculature, and are a valuable substitute for contrast angiography in most circumstances. The aim of this paper is to provide a comprehensive review of the current contribution of CT and MRI to the diagnosis of great vessel pathology of the chest. Key words: Aorta ± Pulmonary arteries ± Computed tomography ± Magnetic resonance

Computed tomography With the advent of computed tomography (CT) in the mid-1970s, the human body could be studied tomographically in the axial or transverse planes with a high spatial resolution and a contrast resolution far superior to that of conventional radiographic techniques. Blood vessels could be well visualized after intravenous administration of iodinated contrast medium. However, rapid dilution and extravasation of the contrast material into the extracellular space, as well as long scan times, limited the study of long vessel segments (e. g., the thoracoabdominal aorta in a patient with an aortic dissection). Further technical innovations, such as faster CT scanners and the introduction of the spiral or helical CT mode, opened new avenues for studying the great vessels of the chest [1]. This technique ist called computed tomography angiography (CTA). With scan times of 1 s (or less) using the helical CT mode, the vessel of Correspondence to: G. Marchal Mini Categorical Course ECR '99

interest can be studied during peak enhancement following a bolus injection of contrast material (e. g., 100 ml of 300 mg I/ml non-ionic contrast agent at an injection rate of 2 ml/s for the pulmonary arteries starting 15 s after injection). To insure consistent vascular opacification throughout the acquisition, the bolus duration is made equivalent to the scan duration and the bolus should arrive at the target volume coincident with the initiation of the scan. To determine the time delay to achieve peak vascular enhancement after the start of the injection, a test injection of intravenous contrast material is advisable. Some manufacturers provide trigger systems starting the helical scan acquisition once the contrast concentration reaches a predetermined level. CTA is able to demonstrate the presence of vessels as small as 1±2 mm in diameter. Detection of stenoses, however, is dependent on the direction of the vessel, vessels oriented in the axial plane being less well visualized than vessels oriented in other planes. This is related to the anisotropic resolution of the pixels, which have a size of 0.5 mm in the transaxial plane but only 2±4 mm along the longitudinal axis [2]. The pitch, collimation, reconstruction interval and degree of vascular enhancement determine the quality of the final image, the visualization of small vessels and the smoothness of the vessel outline. Volumetric data acquisition by means of helical CT has other advantages. First, a number of postprocessing techniques are available for displaying the images. The data can be reformatted in non-axial planes (i. e., multiplanar reconstruction, or MPR), allowing a better visualization of vascular abnormalities such as the extent of aortic aneurysms, or the true and false channel in patients with aortic dissection. Moreover, the vessels can be three-dimensionally reconstructed [i. e., using surface shading or maximum intensity projection (MIP)], to better visualize complex vascular abnormalities such as pulmonary arteriovenous malformations. The MIP images can be computed from several angles and displayed as a cine loop to allow a three-dimensional appreciation of the thoracic vessels. Other, less common applications

1100

include virtual intravascular endoscopy [3]. Unfortunately, most of these post-processing procedures remain time-consuming and tedious. Magnetic resonance imaging Magnetic resonance imaging (MRI) made its entry in medicine almost a decade after CT. It combines superior contrast and spatial resolution with multiplanar imaging capability in a fast and totally noninvasive manner. Advantages over CT are that the patient is not exposed to either ionizing radiation or iodinated contrast material. Imaging of the heart and great vessels, however, was impeded by artifacts resulting from cardiac contraction and respiratory motion, by susceptibility artifacts at the lung-heart interface, by the low proton density of the lungs and by the large volume of the thorax. A first step to improving image quality was obtained by coupling data acquisition to the cyclic motion of the heart, which is accomplished most commonly with electrocardiographic (ECG) gating [4]. Another step towards improved image quality was the introduction of segmented MR sequences that could be performed during breathhold, thus significantly reducing the detrimental effects of respiratory motion [5]. More recently the advent of fast gradient systems with ultrafast repetition and echo times (with echo times up to 1 ms) provided an important breakthrough for three-dimensional (3D) MR angiography (MRA) of the chest [6, 7]. The radiologist now has a whole gamut of MR sequences with which to study the great vessels of the chest. The oldest but still very valuable technique is the spin-echo (SE) technique. It is a dark-blood technique that provides not only excellent contrast with the surrounding tissues without the necessity for administering paramagnetic contrast agents, but also good visualization of intraluminal structures such as an intimal flap in patients with an aortic dissection. The absence of signal in the blood vessels and cardiac chambers occurs because between the time of excitation and read-out the excited spins leave the imaging plane and are replaced by fully relaxed spins (i. e., the flow void phenomenon). During read-out, no signal is acquired from the intravascular and intracardiac flowing blood, and the blood appears dark. This principle remains valid in conditions with a sufficiently high blood flow. Problems arise with slow-flowing blood (e. g., during cardiac diastole, in patients with reduced cardiac function or in turbulent flow), which may limit the clinical utility of SE sequences for studying the thoracic vessels. The use of longer echo times and thinner sections may further decrease the intravessel signal. Conventional SE sequences are, however, time-consuming. In patients with slow heart rates these sequences easily exceed 10 min. To reduce the total imaging time, faster alternatives such as turbo SE and HASTE (half Fourier single shot turbo SE) have been developed. These techniques rely on the acquisition of several k-lines (instead of one) per excitation pulse. Turbo SE uses a segmented approach in which filling of k-

G. Marchal and J. Bogaert: Imaging of the great vessels of the chest

space is spread over several heart beats, usually during a breath-hold period [8]. The HASTE technique uses a single excitation pulse after which half the k-space is filled (half Fourier approach). The excitation pulse is preceded by a dark-blood preparation pulse. Since the total acquisition window is approximately 300 ms, the HASTE technique can be run without the necessity of breath-holds, which is advantageous in patients unable to hold their breath. Both sequences provide good dark-blood images. Turbo SE has a better spatial resolution than HASTE, while HASTE allows a very fast evaluation of the thoracic vessels (e. g., only 40 heart beats for a set of 20 images). Several bright-blood techniques are also available for study of the thoracic vessels. For this purpose gradient-echo (GE) instead of SE techniques are used. The basic principle is that inflowing blood contains fully relaxed spins while the spins in the surrounding stationary tissues become progressively saturated (i. e., inflow enhancement). Thus, the blood vessels are maximally enhanced when they are transversely sectioned, while the intravessel signal will decrease when they are too long in-plane. The two major applications of GE sequences are cine MRI and time-of-flight (TOF) MRA. Cine MRI provides dynamic information not only on cardiac motion but also on the vascular pulsations and intraluminal blood flow. Images are acquired at different time points (e. g., every 40 ms) during the cardiac cycle. These images can be loaded into an endless cine loop providing dynamic information comparable to realtime imaging techniques such as echocardiography ± but taking into account that the images are acquired over several heart beats. With the advent of segmented acquisition techniques these sequences can be performed in breath-hold [5]. Until recently, the majority of the MRA protocols for the great vessels of the chest were based on two-dimensional (2D) TOF acquisitions [9], while MRA protocols based on phase contrast were less successful in the chest. The long acquisition times of 2D TOF sequences limit their clinical utility, and the images are often degraded by respiratory motion. Furthermore, the intravessel signal is dependent on the blood flow and thus on the cardiac output of the patient. Similarly, while three-dimensional TOF MRA sequences provide excellent angiographic images in other parts of the body, the long acquisition time and the destructive effects of respiration made this sequence useless for the study of the thoracic vasculature. Automatic techniques have therefore been developed to track the position of the diaphragm and then perform acquisitions during specific phases of the respiratory cycle, without any need for patient cooperation (i. e., navigator echo). Respiratory motion is usually frozen, which results in high-quality images. This technique has been developed for applications that are not feasible during a breath-hold (such as three-dimensional TOF MRA), in regular acquisitions where there is a need for better resolution and/or signal-to-noise ratio and in patients who cannot cooperate. The technique is being investigated regarding the visualization of the coronary arteries [10].

G. Marchal and J. Bogaert: Imaging of the great vessels of the chest

Recently, with the development of ultrafast gradients with ultrashort repetition and echo times, ultrafast 3D contrast-enhanced MRA became feasible. The technique is still under development, but it is very likely that it will become the first choice for visualization of the great vessels of the chest [6]. Imaging of the vessels is independent on the intravessel flow but relies on a T1shortening of the blood following a bolus injection of a gadolinium chelate [11], since acquisition on the k-space (especially the central k-lines) has to occur at the time of the first pass of the bolus through the vessel of interest. As with CTA, the use of a trigger system or a test bolus (usually 1±2 ml) is strongly advisable to calculate the peak enhancement time in a vessel. Moreover, since the time between arterial enhancement and venous filling is extremely short, acquisitions have to be well timed to avoid venous overlap. The total acquisition time, which relies mainly on the number of partitions acquired, varies in clinical practice between 7 and 27 s, and can be performed during breath-hold [12]. Furthermore, timeresolved approaches allow visualization of the passage of contrast medium in the thoracic vessels following injection in a cubital vein (e. g., first the pulmonary arteries and veins, followed by the thoracic aorta and its branches, and then, finally, the systemic venous return), and thus reduce the need for proper bolus timing. Although many topics such as total dose, injection scheme and total acquisition time are still under study, this ultrafast 3D angiographic acquisition is quite robust and very promising. Post processing of the 3D data is similar to that for CTA but less time-consuming, and includes MPR, MIP and virtual intra-arterial endoscopy. MPR provides tomographic images of the vessels in virtually any desired plane while MIP provides projections similar to conventional contrast angiography. Rotating these MIP images provides a 3D view of the thoracic vasculature. Virtual intra-arterial endoscopy is a form of virtual reality imaging capable of rendering internal views of the vessel walls. Since the clinically available paramagnetic contrast agents remain at a high concentration within the vessel for only a very short period, gadolinium-based blood pool agents (e. g., polylysine, dextran, albumin) are currently under study. Moreover, substraction techniques may improve the contrast-to-noise ratio and visualization of smaller vessels, as well as removing the venous signal. Third, blood flow can be quantified by using differences in phase angle between intravascular spins and surrounding stationary spins (i. e., velocity mapping). The insertion of a bipolar pulse into any normal sequence induces flow-related phase shifts in the acquired data. This characteristic, which is usually unwanted, is exploited in flow quantification techniques. Magnetic field gradients applied in the through-plane direction require that the image is positioned perpendicular to the vessel. Gradients along the in-plane directions quantify the flow in the plane [13]. Stationary material is represented as mid-gray whilst increasing velocities in either direction are shown in increasing grades of black or white. Measurement of the spatial mean velocity for all pixels

1101

in a region of interest of known area enables the calculation of the instantaneous flow volume at any point in the cardiac cycle. The flow volume per heart beat can be obtained by integrating the instantaneous flow volumes of all frames throughout the cardiac cycle. Velocity mapping has been validated in vitro and in vivo, and is extremely accurate and reproducible [14±16]. It now represents the gold standard for in vivo flow measurements [17]. Pathology of the thoracic aorta Coarctation of the thoracic aorta An extensive description of all congenital abnormalities of the thoracic aorta is beyond the scope of this paper. They include developmental abnormalities of the aortic valve (e. g., bicuspid valve), aneurysms of the sinus of Valsalva, aortic arch abnormalities (e. g., right-sided aortic arch, double aortic arch) and coarctation of the aorta. Coarctation of the aorta is the third most common congenital malformation of the cardiovascular system [18], and consists of a narrowing of the proximal descending thoracic aorta, most often located just distal to the left subclavian artery, opposite the insertion of the ductus arteriosus. The degree of collateral blood supply, via the subclavian and intercostal arteries, depends on the severity of the aortic coarctation. The utility of MRI for evaluating coarctation of the aorta has been well established [19, 20]. Thin-section SE imaging in the oblique sagittal plane can be used to determine the exact site and degree of narrowing and also allows visualization of the extent of coarctation and the presence of collateral vessels [21]. Cine MRI and velocity mapping can be used to gain insight into the functional significance of the aortic coarctation. Cine MRI may reveal abnormal flow patterns caused by the aortic coarctation as a signal void originating from the site of narrowing [22]. Velocity mapping can be applied to measure the highest velocity across the stenosed vessel. However MRI is at present not an adequate technique for studies of coarctation in infants and young children, as a result of mismatch between the size of the region of interest and feasible slice thickness in this age group. The contribution of new techniques, such as ultrafast 3D contrast-enhanced MRA, in the diagnosis of aortic coarctation are under study (Fig. 1). Aortic aneurysm Several entities may cause an aneurysmal dilatation of the thoracic aorta. Often the dilatations have a preferential location: the aortic root is widened in patients with Marfan's disease, cystic medianecrosis preferentially involves the ascending aorta, involvement of the descending thoracic aorta or the entire thoracic aorta is frequent in atherosclerosis, and saccular aneurysms located within the aortic arch are often posttraumatic.

1102

a

G. Marchal and J. Bogaert: Imaging of the great vessels of the chest

b

Fig. 1 a, b. Coarctation of the thoracic aorta imaged by ultrafast 3D contrast-enhanced MRA. a MPR image in the oblique sagittal plane. b MIP image. There is extensive narrowing of the proximal descending aorta. Note the presence of a slight poststenotic widening and several collateral vessels

rysms, while combination with cine MRI allows definition of aortic valve disease; the ability to use contrast enhancement, if branch vessel anatomy is unclear, provides an extra degree of reliability [7, 23]. There is some dispute about the relative merits of CTA and MRA in evaluating the thoracic aorta. Both are extremely accurate and versatile. CTA requires the use of iodinated contrast agents, which are more toxic than gadolinium-based agents, and is currently unable to assess aortic regurgitation. In addition, some branch vessel anatomy may be obscured by calcifications in the vessel wall. Fig. 2. Posttraumatic saccular aneurysm of the thoracic aorta imaged by CTA. The saccular aneurysm arises from the left lateral border of the aortic arch. Note the presence of a small mural thrombus, and the compression on the adjacent pulmonary vessels

In most patients, CT and MRI are appropriate methods for investigating thoracic aortic aneurysms. Diagnostic accuracy is necessary and is provided by the ability not only to define clearly the full diameter of the aortic aneurysm but also to assess the amount of thrombus within it and its craniocaudal extent (Fig. 2). Many thoracic aneurysms extend into the abdomen and hence the ability to image the entire aorta (which may not be available with echocardiography) is important. The precise relationship of the aneurysm to the aortic branch vessels is crucial when planning the surgical approach and site of cross-clamping. In ascending aortic disease, assessment of aortic valve morphology, aortic regurgitation, and the relationship of aneurysmal disease to the coronary arteries is required. MRI and MRA can provide all of this information relatively rapidly and reliably, although it may be necessary to use a combination of imaging protocols. SE sequences are accurate for assessing the dimensions of aortic aneu-

Aortic dissection Aortic dissection may be a life-threatening condition, necessitating a rapid diagnosis, close monitoring and often surgical intervention. Because patients frequently present acutely, the choice of imaging technique is often determined more by what is most readily available than by what is theoretically the best possible technique. In experienced hands, the accuracy of CTA, transesophageal echocardiography and MRI are comparable. However, the need for close hemodynamic monitoring and hypotensive medication may make imaging by MRI impossible. In these circumstances, transesophageal echocardiography or CTA are more appropriate [24, 25]. MRI is suitable for patients with acute dissection who are hemodynamically stable [26] or patients with chronic dissection or requiring postoperative evaluation following repair of dissection [27±29]. Both CTA and MRI accurately demonstrate acute and chronic dissections and postoperative complications such as extensive dissection, increasing diameter of the residual aorta, development of false channels, and anastomotic aneurysms (Fig. 3) [30, 31]. Contrast-enhanced 3D MRA aids understanding of the relationship of false lumens and branch vessels to the aorta [7, 23, 32]. In the

G. Marchal and J. Bogaert: Imaging of the great vessels of the chest

a

b

Fig. 3 a±c. Stanford type B aortic dissection imaged by CTA. The spiral CT mode allows study of the thoraco-abdominal aorta during maximal enhancement following bolus injection of contrast medium. a Axial CT image at the level of the left main coronary artery. There is an intimal flap in the descending aorta. Both channels are open. The true channel is located anterolaterally, while the larger false channel is located posteromedially. b Axial CT image at the level of the superior mesenteric artery. The superior mesenteric artery is supplied by the true channel. c Axial CT image infrarenally. Aneurysmal dilatation of the infrarenal aorta with mural thrombus and re-entry occurs at this position

past CTA has been reported to be superior to MRI, but the appropriate MRA approach was not used in this comparison [33]. With SE MRI alone diagnostic errors can occur related to slow flow mimicking thrombus, chemical shift artifacts, respiratory artifacts, signal from adjacent veins (i. e., the superior vena cava), tumors, aortic ulcers, etc. [34, 35]. The use of MRA, contrast-enhanced MRA or velocity mapping has largely decreased these errors. A distinct advantage of MRI over CT is the possibility of evaluating aortic valve integrity by means of cine MRI or velocity mapping. Furthermore, the relationship of the dissection with the coronary arteries can be better evaluated on MRI. Occasionally, early dissection may appear as an intramural hematoma without a distinct flap or tear. In these cases CT or MRI will only show wall thickening [36]. In this situation, high signal within the wall on MRI usually helps in the correct diagnosis [37, 38].

a

1103

b

c Aortic trauma Early diagnosis of injuries of the thoracic aorta after blunt chest trauma is imperative because, if untreated, this condition is almost always lethal [39]. The frontal chest radiograph is a highly effective screening test to exclude aortic rupture, with a very high negative predictive value when the findings are normal. However, chest radiographs with abnormal findings are much less helpful. Moreover, positioning of an unstable patient to allow a perfect erect chest radiograph is untenable [40]. While contrast aortography is still considered the gold standard for detecting aortic laceration, many investigators suggest that CTA and transesophageal echocardiography can be used to screen for aortic rupture. CT and CTA are helpful not only for detecting mediastinal hematoma but also for depicting the aortic flap and pseudoaneurysm formation [41]. Further clinical studies are, however, necessary to prove the value of helical CT is this area. Whether there is a role for MRA, especially the contrast-enhanced fast MRA acquisitions, is unknown since for most patients there is no time for an additional MR examination. Infectious and inflammatory aortic diseases Infectious diseases of the thoracic aorta are rare conditions that may occur as a complication of bacterial endocarditis, aortic surgery, and mediastinal sepsis. Infection

Fig. 4 a, b. Thrombus in the right pulmonary artery imaged by CTA. a Axial CT image at the level of the right pulmonary artery. b Axial CT image at the level of the left atrium

1104

can also occur in a severely atherosclerotic aorta. Both MRI and CT will show the abnormalities associated with any coexistent mediastinitis, and perivalvular abscesses or mycotic aneurysms [42]. In postoperative patients clear differentiation from surgery-related changes may be difficult [43]. CT and MRI are also useful for assessing complications of inflammatory aortic syndromes (such as Takayasu's arteritis) in which contrast angiography may be particularly hazardous. Both techniques allow evaluation of the location and extent of the inflammatory wall thickening and the results of corticosteroid therapy. Pulmonary artery embolism Several studies in literature have shown the usefulness of CTA for visualizing the pulmonary arteries and detecting abnormalities such pulmonary artery embolism (Fig. 4) [44±48]. Not infrequently central pulmonary embolism may be an unsuspected finding on CT [49, 50]. In a study by Remy-Jardin et al. [51, 52] a 100 % sensitivity and a 96 % specificity for the detection of the main, lobar and segmental pulmonary arteries was found. With CTA, endoluminal clots can be depicted in second- to fourth-division pulmonary arteries. The feasibility of different MRA techniques for depicting pulmonary embolism has been evaluated by several groups [53±55]. Probably the most promising results were recently published by Meany and coworkers [56]. They compared ultrafast contrast-enhanced 3D MRA with conventional pulmonary angiography for the diagnosis of pulmonary embolism. Overall, 21 of the 22 sites of embolism identified by conventional angiography were also identified by MRA. If those results can be confirmed, MRA can be considered a noninvasive alternative for the diagnosis of pulmonary embolism without the need for ionizing radiation or iodinated contrast material. References 1. Kalender WA, Polacin A (1991) Physical performance characteristics of spiral CT scanning. Med Phys 18: 910±915 2. Bluemke DA, Chambers TP (1995) Spiral CT angiography: an alternative to conventional angiography. Radiology 195: 317±319 3. Rubin GD, Dake MD, Semba CP (1995) Current status of three-dimensional spiral CT scanning for imaging the vasculature. Radiol Clin North Am 33: 51±70 4. Lanzer P, Botvinick EH, Schiller NB, et al (1984) Cardiac imaging using gated magnetic resonance. Radiology 150: 121±127 5. Atkinson DJ, Edelman RR (1991) Cineangiography of the heart in a single breath hold with a segmented turboflash sequence. Radiology 178: 357±360 6. Leung DA, Debatin JF (1997) Three-dimensional contrast-enhanced magnetic resonance angiography of the thoracic vasculature. Eur Radiol 7: 981±989 7. Prince MR (1994) Gadolinium-enhanced MR aortography. Radiology 191: 155±164 8. Seelos KC, von Smekal SA, Vahlensieck M, Gieseke J, Reiser M (1993) Cardiac abnormalities: assessment with T2-weighted turbo spin-echo MR imaging with electrocardiogram gating at 0.5 T. Radiology 189: 517±522

G. Marchal and J. Bogaert: Imaging of the great vessels of the chest 9. Seelos KC, Von SA, Steinborn M, et al (1994) MR angiography of the heart and thoracic vessels: application of fast ECG-gated techniques with multiplanar reconstruction capability. Radiologe 34: 454±461 10. Muller M, Fleisch M, Kroeker R, Chatterjee T, Meier B, Vock P (1997) Proximal coronary artery stenosis: three-dimensional MRI with fat saturation and navigator echo. J Magn Reson Imaging 7: 644±651 11. Prince MR, Yucel EK, Kaufman JA, Harrison DC, Geller SC (1993) Dynamic gadolinium-enhanced three-dimensional abdominal MR arteriography. J Magn Reson Imaging 3: 877±881 12. Hatabu H, Gaa J, Kim D, et al (1996) Pulmonary perfusion: qualitative assessment with dynamic contrast-enhanced MRI using ultra-short TE and inversion recovery turboFLASH. Magn Reson Med 36: 503±508 13. Lee V, Spritzer C, Carroll B, et al (1997) Flow quantification using fast cine phase-contrast MR imaging, conventional cine phase-contrast MR imaging, and Doppler sonography: in vitro and in vivo validation. AJR 169: 1125±1131 14. Bogren H, Klipstein R, Firmin D, et al (1989) Quantification of antegrade and retrograde blood flow in the human aorta by magnetic resonance velocity mapping. Am Heart J 174: 1214±1222 15. Firmin DN, Nayler GL, Klipstein RH, Underwood SR, Rees R, Longmore DB (1987) In vivo validation of MR velocity imaging. J Comput Assisted Tomogr 11: 751±756 16. Meier D, Maier S, Boesiger P (1988) Quantitative flow measurements on phantoms and on blood vessels with MR. Magn Reson Med 8: 25±34 17. Kondo C, Caputo GR, Semelka R, Foster E, Shimakawa A, Higgins CB (1991) Right and left ventricular stroke volume measurements with velocity-encoded cine MR imaging: in vitro and in vivo validation. AJR 157: 9±16 18. Steffens J, Bourne M, Sakuma H, O'Sullivan M, Higgins C (1994) Quantification of collateral blood flow in coarctation of the aorta by velocity encoded cine magnetic resonance imaging. Circulation 90: 937±943 19. Bank ER (1993) Magnetic resonance of congenital cardiovascular disease: an update. Radiol Clin North Am 31: 553±572 20. Von Schulthess G, Higashino S, Higgins S, Didier D, Fisher M, Higgins C (1986) Coarctation of the aorta: MR imaging. Radiology 158: 469±474 21. Glazer HS, Gutierrez FR, Levitt RG, Lee JK, Murphy WA (1985) The thoracic aorta studied by MR imaging. Radiology 157: 149±155 22. Simpson I, Chung K, Glass R, Sahn D, Sherman F, Hesselink J (1988) Cine magnetic resonance imaging for evaluation of anatomy and flow relations in infants and children with coarctation of the aorta. Circulation 78: 142±148 23. Krinsky G, Rofsky N, DeCorato D, et al (1997) Thoracic aorta: gadolinium-enhanced three-dimensional MR angiography with conventional MR imaging. Radiology 202: 183±193 24. Cigarroa JE, Isselbacher EM, DeSantis RW, Eagle KA (1993) Diagnostic imaging in the evaluation of suspected aortic dissection. N Engl J Med 328: 35±43 25. Mast H, Gordon D, Kantor A (1991) Pitfalls in diagnosis of aortic dissection by angiography: algorithmic approach utilizing CT and MRI. Comp Med Imag Graphics 15: 431±440 26. Nienaber CA, von Kodolitsch Y, Nicolas V, et al (1993) The diagnosis of thoracic aortic dissection by noninvasive imaging procedures. N Engl J Med 328: 1±9 27. Bogaert J, Meyns B, Rademakers F, et al (1997) Follow-up of aortic dissection: contribution of MR angiography for evaluation of the abdominal aorta and its branches. Eur Radiol 7: 695±702 28. Laissy J, Blanc F, Soyer P, et al (1995) Thoracic aortic dissection: diagnosis with transesophageal echocardiography versus MR imaging. Radiology 194: 331±336 29. White RD, Ullyot DJ, Higgins CB (1988) MR imaging of the aorta after surgery for aorta dissection. AJR 150: 87±92

G. Marchal and J. Bogaert: Imaging of the great vessels of the chest 30. Di Cesare E, Di Renzi P, Pavone P, Marsili L, Castaldo F, Passariello R (1991) Postsurgical follow-up of aortic dissections by MRI. Eur J Radiol 13: 27±30 31. Hartnell GG, Finn JP, Zenni M, et al (1994) MR imaging of the thoracic aorta: comparison of spin-echo, angiographic and breath-hold techniques. Radiology 191: 697±704 32. Loubeyre P, Delignette A, Bonefoy L, Douek P, Amiel M, Revel D (1996) Magnetic resonance imaging evaluation of the ascending aorta after graft-inclusion surgery: comparison between an ultrafast contrast-enhanced MR sequence and conventional cine-MRI. J Magn Reson Imaging 6: 478±483 33. Sommer B, Fehske W, Holznecht N, et al (1996) Aortic dissection: a comparative study of diagnosis with spiral CT, multiplanar transesophageal echocardiography, and MR imaging. Radiology 199: 347±352 34. Petasnick JP (1991) Radiologic evaluation of aortic dissection. Radiology 180: 297±305 35. Solomon SL, Brown JJ, Glazer HS, Mirowitz SA, Lee JKT (1990) Thoracic aortic dissection: pitfalls and artifacts in MR imaging. Radiology 177: 223±228 36. Wolff KA, Herold CJ, Tempany CM, Parravano JG, Zerhouni EA (1991) Aortic dissection: atypical patterns seen at MR imaging. Radiology 181: 489±495 37. Murray J, Manisali M, Flamm S, et al (1997) Intramural hematoma of the thoracic aorta: MR image findings and their prognostic implications. Radiology 204: 349±355 38. Nienaber CA, von Kodolitsch Y, Petersen B, et al (1995) Intramural hemorrhage of the thoracic aorta: diagnostic and therapeutic implications. Circulation 92: 1465±1472 39. Raptopoulos V, Sheiman RG, Phillips DA, Davidoff A, Silva WE (1992) Traumatic aortic tear: screening with chest CT. Radiology 182: 667±673 40. Trerotola SO (1995) Can helical CT replace aortography in thoracic trauma? Radiology 197: 13±15 41. Gavant ML, Menke PG, Fabian T, Flick PA, Graney MJ, Gold RE (1995) Blunt traumatic aortic rupture: detection with helical CT of the chest. Radiology 197: 125±133 42. Guinet C, Buy J, Ghossain M, et al (1992) Aortic anastomotic pseudoaneurysms: US, CT, MR and angiography. J Comput Assist Tomogr 16: 182±188 43. Rofsky NM, Weinreb JC, Grossi EA, et al (1993) Aortic aneurysm and dissection: normal MR imaging and CT findings after surgical repair with the continuous-suture graft-inclusion technique. Radiology 186: 195±201

1105 44. Remy Jardin M, Remy J, Cauvain O, Petyt L, Wannebroucq J, Beregi JP (1995) Diagnosis of central pulmonary embolism with helical CT: role of two-dimensional multiplanar reformations. AJR 165: 1131±1138 45. Remy Jardin M, Remy J, Deschildre F, et al (1996) Diagnosis of pulmonary embolism with spiral CT: comparison with pulmonary angiography and scintigraphy. Radiology 200: 699±706 46. Tapson VF (1997) Pulmonary embolism: new diagnostic approaches (editorial; comment). N Engl J Med 336: 1449±1451 47. van Rossum AB, Pattynama PM, Ton ER, et al (1996) Pulmonary embolism: validation of spiral CT angiography in 149 patients. Radiology 201: 467±470 48. van Rossum AB, Treurniet FE, Kieft GJ, Smith SJ, Schepers Bok R (1996) Role of spiral volumetric computed tomographic scanning in the assessment of patients with clinical suspicion of pulmonary embolism and an abnormal ventilation/perfusion lung scan. Thorax 51: 23±28 49. Verschakelen JA, Vanwijck E, Bogaert J, Baert AL (1993) Detection of unsuspected central pulmonary embolism with conventional contrast-enhanced CT. Radiology 188: 847±850 50. Winston CB, Wechsler RJ, Salazar AM, Kurtz AB, Spirn PW (1996) Incidental pulmonary emboli detected at helical CT: effect on patient care. Radiology 201: 23±27 51. Remy Jardin M, Remy J, Wattinne L, et al (1992) Central pulmonary thromboembolism: diagnosis with spiral volumetric CT with three single-breath-hold technique. Radiology 185: 381±387 52. Remy Jardin M, Remy J, Artaud D, Deschildre F, Fribourg M, Beregi JP (1997) Spiral CT of pulmonary embolism: technical considerations and interpretive pitfalls. J Thorac Imaging 12: 103±117 53. Erdman WA, Peshock RM, Redman HC, et al (1994) Pulmonary embolism: comparison of MR images with radionuclide and angiographic studies. Radiology 190: 499±504 54. Grist TM, Sostman HD, MacFall JR, et al (1993) Pulmonary angiography with MR imaging: preliminary clinical experience. Radiology 189: 523±530 55. Loubeyre P, Revel D, Douek P, et al (1994) Dynamic contrastenhanced MR angiography of pulmonary embolism: comparison with pulmonary angiography. AJR 162: 1035±1039 56. Meany JFM, Weg JG, Chenevert TL, Stafford-Johnson D, Hamilton BH, Prince MR (1997) Diagnosis of pulmonary embolism with magnetic resonance angiography. N Engl J Med 336: 1422±1427

Notice to authors European Radiology encourages all authors to submit manuscripts printed on both sides of the page. This will save paper and reduce the cost of postage.