- Email: [email protected]
MDCT imaging of the aorta and peripheral vessels
European Journal of Radiology 45 (2003) S42 /S49 www.elsevier.com/locate/ejrad
MDCT imaging of the aorta and peripheral vessels Geoffrey D. Rubin * Department of Radiology, Stanford University School of Medicine, Mail Code 5105 GRANT, S072B Stanford, CA 94305-5105, USA
Abstract Since its clinical introduction in 1991, volumetric CT scanning using spiral or helical scanners has resulted in a revolution for diagnostic imaging. Helical CT has improved over the past 8 years with faster gantry rotation, more powerful X-ray tubes, and improved interpolation algorithms, but the greatest advance has been the recent introduction of multi detector-row CT (MDCT) scanners [J. Comput. Assist. Tomogr. 23 (1999) S83]. Currently capable of acquiring four channels of helical data simultaneously, MDCT scanners have achieved the greatest incremental gain in scan speed since the development of helical CT and have profound implications for clinical CT scanning. Fundamental advantages of MDCT include substantially shorter acquisition times, retrospective creation of thinner or thicker sections from the same raw data, and improved three-dimensional (3-D) rendering with diminished helical artifacts. While these features will likely be important to many applications of CT scanning, including the characterization of focal lung and liver lesions through the creation of thin sections retrospectively, the greatest impact has been on CT angiography. The implication for CT angiography is that scans can be performed approximately three-times faster than is possible with the fastest single-detector CT scanner. For example a 1.25 mm nominal thick section (1.6 mm effective section thickness) can be acquired with a table speed of 9.4 mm/s, and a 2.5 mm nominal thick section (3.2 mm effective section thickness) can be acquired with an 18.8 mm/s table speed. The advantages of MDCT for imaging the vascular system can be broken down into three fundamental improvements over single detector-row CT scanners speed (faster), distance (longer), and section thickness (better). The focus of this article will be how multidetector-row CT technology has substantially improved aortoiliac and lower extremity arterial imaging. # 2003 Published by Elsevier Science Ireland Ltd. Keywords: MDCT; Aorta; Peripheral vessels
1. Faster Imaging of the abdominal aortoiliac system is now performed in 20 s using a section thickness that is 34% thinner than our former protocol requiring 50 s with a subsecond single-detector row CT scanner. As a result of the shorter acquisition time, we have reduced our contrast dose from 180 to 80 ml. In fact we can now image the entire thoracoabdominal aorta and iliac arteries in a mere 30/40 s using 2.5-mm section thickness throughout the entire scan volume. We studied 48 patients with aortic aneurysm or dissection who underwent CT angiography with both
* Tel.: /1-650-723-7647; fax: /1-650-725-7296. E-mail address: [email protected] (G.D. Rubin).
MDCT and single detector CT (SDCT) on separate dates. Section thickness (full-width at half-maximum of section profile), scan duration, scan coverage, iodinated contrast dosage and mean aorto-iliac attenuation, measured every 20 mm along the z-axis of each patient, were compared. Data were summarized as the speed (coverage/duration), scan efficiency (speed/section thickness), and contrast efficiency (mean aortic attenuation/ contrast dose). We found that the speed of CTA was 2.5-times faster, scan efficiency was 4.1-times greater, and contrast efficiency was 2.4-times greater with MDCT than SDCT. The average section thickness was 40% thinner (3.2 vs. 5.3 mm), despite a scan duration that was 58% shorter (22 vs. 54 s) with MDCT versus SDCT [2]. Image quality was also substantially improved with MDCT, because a narrower section profile was em-
0720-048X/03/$ - see front matter # 2003 Published by Elsevier Science Ireland Ltd. doi:10.1016/S0720-048X(03)00036-6
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
ployed, and the faster acquisition resulted in greater coherency of the contrast bolus. This is particularly evident in obliquely oriented vessels such as the renal and iliac arteries. Within the chest, artifacts relating to cardiac and arterial pulsation are substantially diminished in part due to the substantial increase in the table travel distance during each cardiac cycle.
2. Longer In addition to improving the quality and efficiency of aortic, CTA, MDCT has enabled the development of fundamentally new applications of CTA, such as the assessment of lower extremity occlusive disease. The principal clinical presentations of lower extremity occlusive disease vary between claudication, non-healing ulceration, and gangrene. Clinically significant lesions could be located anywhere between the aorta to the trio of lower leg vessels, the anterior tibial, peroneal, and posterior tibial arteries. The vessels proximal to the femoral arteries are considered inflow vessels, while those distal are ‘run-off’. CTA has been used to study lower extremity inflow lesions, but the length of the runoff has limited its role for assessing lower extremity occlusive disease as a whole. Multi detector-row CT makes a complete acquisition of lower extremity inflow and run-off possible. Using 2.5-mm collimation and a table speed of 18.8 mm/s (3.2 mm-effective section thickness), a distance of 1300 mm can be covered in less than 70 s. This distance is adequate for imaging from the celiac origin through the mid-foot in the majority of patients. We currently perform this examination using 180 ml of iodinated contrast medium injected at a rate of 3.5 ml/s, which results in a 51-s injection. This amount of contrast medium suffices to image only the celiac to the femoral artery bifurcations, using a single-detector CT scanner. Although the contrast injection is nearly 20 s less than the scan duration, our preliminary experience suggests that excellent arterial opacification is achieved throughout the scan volume. This may be attributable to the substantial delay that tends to occur when blood travels from the aorta through the run-off vessels. Nevertheless, this is an issue that must be closely monitored as this new CT application emerges, particularly given the propensity for substantial variations in flow rates between the two legs, when there are substantial asymmetries of disease involvement. We have studied 16 patients with lower extremity claudication or limb-threatening ischemia had severe lower extremity occlusive disease documented with intra-arterial DSA. All 52 occluded segments and 47 patent segments with ]/50% stenoses were identified and correctly characterized with MDCTA. Patent distal run-off vessels, that were undetected with DSA, were
S43
identified with CTA in six limbs. MDCTA uniquely displayed complex three-dimensional (3-D) relationships between patent vessels, collateral vessels, and mural calcium both in patent and occluded segments. An alternative noninvasive imaging technique for assessing lower extremity arterial disease that requires no ionizing radiation nor iodinated contrast media and has been in routine clinical use for several years is magnetic resonance (MR) angiography [3]. There are many compelling advantages to the use of MR imaging, which in addition to the aforementioned include less primary reconstructions owing to coronal acquisitions and no interference from high signal non-arterial structures analogous to bone in CT angiography. Nevertheless, there are several reasons why CT angiography might be a compelling alternative to MR angiography. Because of an association between lower extremity atherosclerosis and coronary artery disease, implanted pacemakers and defibrillators are common in this patient population [4] and are contraindications to MR imaging. Additionally, other metallic devices such as arterial stents or joint prostheses can result in substantial artifact, hindering arterial evaluation [3]. The characterization of mural calcification is not possible with MR and may have therapeutic relevance[5,6]. Finally, we have achieved voxel dimensions of 0.7 /0.7 /1.25, 3.2 mm that allow us to image the entirety of the lower extremity arterial system in 60 /85 s with 4DRCT. These voxel dimensions are five to fourteen-times smaller than current MR angiographic techniques [7,8] and thus result in a higher spatial resolution acquisition.
3. Better For dedicated applications of CT angiography requiring limited table travel such as renal, pulmonary, mesenteric, cranial, and distal extremity arterial imaging, MDCT allows acquisitions with near isotropic spatial resolution, where the effective section thickness is between 0.75 and 1.6 mm. As a result, MDCT provides imaging of the vasculature with unprecedented volumetric spatial resolution for any imaging modality. This has proved useful to us in a number of instances where either the inferior spatial resolution of MRA or the limited contrast resolution and display of conventional angiography have failed to fully illustrate clinically relevant complex arterial lesions. Examples of these situations will be presented. The advantages of using MDCT for assessing the vasculature are substantial, and we have not found any situations where SDCT would be preferred to MDCT. MDCT allows substantial reductions in iodinated contrast utilization, improved spatial resolution, less pulsation artifacts, and greater coverage than SDCT.
S44
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
3.1. Thoracic aortic aneurysm CTA is a useful tool for diagnosing thoracic aortic aneurysms, determining their extent, and predicting appropriate management [9]. While the diagnosis of aortic aneurysms are readily made from transverse sections, an assessment of the extent of the lesion, particularly when the brachiocephalic branches are involved is facilitated by an assessment of CPRs and SSDs. In general thoracic aortic aneurysms greater than 5 cm are at an increased risk for rupture. Although thoracic aortic aneurysms expand at a slower rate than abdominal aortic aneurysms, surgical repair is contemplated when thoracic aneurysms reach a diameter of 5/6 cm [10,11]. Helical CT can facilitate surgical planning by delineating the extent of the aneurysm and the involvement of aortic branches. In fact Quint and colleagues found that an analysis of transverse sections and multiplanar reformations were 94% accurate with a positive predicative value of 95% and negative predictive value of 93% for successfully predicting the need for hypothermic circulatory arrest. With the exception of the elimination of one false negative result, the addition of multiplanar reformations did not alter the predictions made from the transverse sections alone. 3-D renderings were not evaluated in this study [9]. Due to the tortuosity and curvature of the thoracic aorta, aneurysm sizing is performed most accurately when double-oblique tomograms are generated perpendicular to the aortic flow lumen. The challenge of such an approach is that data concerning the risk of aneurysm rupture and expansion rate are based upon measurements made from transverse sections, where true diameters can be overestimated. Further the measurement technique must be reproducible to assess the rate of aneurysm expansion on sequential studies. Until analysis tools are available to automatically identify the center of the flow channel, create true perpendicular tomograms, and compute accurate cross-sectional areas and mean diameters, the creation of true vessel cross-sections is probably not practical for routine applications unless sizing endoluminal prostheses. The most sensitive measure of aneurysm size, however, is not the determination of true aortic diameters or even cross-sectional area, but aneurysm volume. This approach has substantial drawbacks. While the volumetric data of helical CT should be excellent for determining aneurysm volume, the accurate segmentation of both patent, thrombosed, and atheromatous elements of the aorta must be segmented from the adjacent structures to make this determination. Currently, the only technique available to perform this is a painstaking manual segmentation performed by drawing regions-of-interest around the aorta on each cross-
section. Further, to date aneurysm expansion has been studied primarily in terms of radial expansion of the aorta. While aneurysm volume determination is an attractive measure of aneurysm growth based upon theoretical considerations, data concerning the risk of aneurysm rupture and guidelines for intervention are based upon traditional transverse diameter measurements [10,11]. 3.2. Abdominal aortic aneurysm Aortography has traditionally been obtained prior to resection of abdominal aortic aneurysms to define (a) the relationship of the aneurysm to main and accessory renal, iliac, superior, and inferior mesenteric arteries, (b) extension of the aneurysm into the common, external, or internal iliac arteries to determine the type and length of prosthetic graft utilized, and (c) detect the presence of coexistent iliac or renal occlusive disease [12,13]. Computed tomography has been advocated in the preoperative evaluation of abdominal aortic aneurysm [12]. It is less invasive than angiography, more accurate than conventional angiography for predicting abdominal aortic aneurysm size [14,15] and is superior to angiography in its ability to demonstrate mural thrombus within an aneurysm, inflammatory aneurysms, perianeurysmal blood due to contained rupture [12,14], and co-existent non-vascular abdominal disease [12]. The ability of conventional CT to accurately demonstrate juxta-renal or supra-renal extension of the aneurysm has been controversial. Some authors have reported predictive values as low as 13% for supra-renal extension [16], while others have reported infra-renal localization of the aneurysm neck in as high as 94% [12]. Data such as this must be interpreted carefully with consideration of the pre-test probability that an abdominal aortic aneurysm will be infra-renal in approximately 95% of cases. Conventional CT has been considered inadequate for assessing the position and patency of aortic branch origins and stenoses and detecting accessory renal arteries [12,15 /17]. Identification of these coexistent abnormalities is critical to optimizing surgical repair of abdominal aortic aneurysms. Helical CT overcomes many of the limitations associated with conventional angiography. Specifically, the elimination of ventilatory misregistration enables an accurate depiction of the aneurysm neck relative to aortic branch vessels as well as an evaluation of the aortic branch vessels themselves. In a preliminary study of nine aortic aneurysms, CTA correctly determined the relationship of the proximal aspect of the aneurysm relative to renal artery branches in all cases[18]. Zeman and colleagues studied the effects of overlapping versus non-overlapping reconstruction and fixed versus variable collimation for determining the extent of abdominal aortic aneurysms on spiral CT sections. On axial
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
sections alone, the determination of supra- versus infrarenal extension of the aneurysm was correctly diagnosed in 23 of 23 patients, whereas with non-overlapping reconstruction, there was false positive supra-renal extension in one patient. Finally, Van Hoe and colleagues studied 38 patients with AAA and a high prevalence of juxta- (n /8) and supra-renal (n /7) proximal necks. In comparison to measurements made during open repair, they found that DSA correctly characterized only 12/15 (80%) aneurysms with juxta- or supra-renal necks, while spiral CTA, performed with 2 mm collimation correctly characterized 14/15 (93%) [19]. The greater accuracy of CTA can be attributed to its ability to detect the outer wall of the aorta and not just the flow lumen. Juxta- and supra-renal extension of AAA can be missed on DSA due to the presence of mural thrombus and atheroma at the proximal neck. Regarding distal AAA involvement of the iliac arteries, Zeman et al. found that transverse CTA sections reconstructed with overlapping reconstructions resulted in four of 23 misdiagnoses (three false positive; one false negative) while non-overlapping sections resulted in five of 20; 23 incorrect diagnoses (four false positive iliac involvement; one false negative) [20]. It is likely that these diagnostic errors would have been reduced if multiplanar and 3-D renderings were used in association with the axial sections for interpretation. The ability of CTA to identify renal artery stenosis in the presence of abdominal aortic aneurysms has been assessed in two studies. Using spiral CT with 3-D rendering, four of four greater than 50% diameter renal artery stenoses were detected in nine abdominal aortic aneurysms, with no false positive diagnoses of renal artery stenosis[18]. Of 83 arteries identified in Van Hoe and colleagues series, hemodynamically significant stenoses or occlusion were present in 17 on DSA and detected with 94% sensitivity and 96% specificity with CTA [19]. The influence of optimized CT angiographic technique (small effective section thickness and overlapping reconstructions) has been suggested by Zeman and coworkers, in a study of 23 patients with abdominal aortic aneurysms, seven of which that were associated with greater than 50% renal artery stenosis. Renal artery stenosis was detected by spiral CT in one of two patients imaged with 5 mm collimation, and in four of five patients imaged with 3 mm collimation. Further, stenotic renal arteries were correctly identified in only two of seven patients with stenoses on non-overlapping sections; whereas, on overlapping sections stenoses were correctly identified in five of seven patients. Moreover, none of the four accessory renal arteries present in the aneurysm population was seen with non-overlapping sections, however, two of four were seen with overlapping sections [20].
S45
Helical CT provides the same information regarding aneurysm size and extent of mural thrombus available with conventional CT, however, the volumetric acquisition allows multiplanar and three dimensional renderings to be generated perpendicular to the long axis of the aneurysm, resulting in greater accuracy of the aneurysm size measurements. 3.3. Thoracic aortic dissection The critical clinical issue required of any imaging test applied to a patient suspected of having an aortic dissection is the identification of an intimal flap and its localization to the ascending (type A) or descending (type B) aorta. This fundamental diagnostic feature that determines the need for emergent repair can be addressed by at least four imaging modalities */angiography, CT, MR imaging and transesophageal echocardiography (TEE). The relative accuracy of these modalities has been debated in the medical literature and is confounded by the fact that technical improvements in CT, MR, and TEE have outpaced our ability to compare them in appropriately designed prospective trials. Recent opinion has shifted toward MRI or TEE as the most sensitive tests for aortic dissection [21]. Unfortunately much of this opinion is based upon comparative studies where state-of-the-art MR or TEE is compared with relatively primitive conventional CT technique [22,23]. In 1989, Erbel and colleagues studied 164 consecutive patients with suspected aortic dissection. All patients were studied with transthoracic echocardiography and TEE, 85 patients were studied with CT, and 96 patients were studied with aortography. The technique for CT scanning was 10 mm thick sections acquired at 20 /40 mm intervals through the chest. Details of how iodinated contrast was administered are not given. Not surprisingly CT was found to be less sensitive (77%) than echocardiography (98%) and aortography (89%) in those patients with surgical proof [22]. In 1993 Nienaber and co-workers compared 110 patients with suspected aortic dissection who underwent at least two of three imaging tests*/TEE, CT, or MR. CT was found to have lower sensitivity (93.8%) than TEE (97.7%) and MR (98.3%) and lower specificity (87.1%) than MR (97.8%). For this study, 80 /100 ml of contrast were administered during the course of 5/15 min scans that were performed with section intervals of 20 mm [23]. To date there have been no comparisons of helical CT to either MR or TEE, and while it is tempting to discount these prior studies because of the substantial advances that have occurred in CT imaging, TEE and MR have undergone further improvements as well. TEE has progressively improved with the introduction of biplane and multiplane probes that substantially reduce the blind spots and ambiguous reverberation artifacts that can limit the accuracy of monoplane
S46
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
devices [21]. Newer MR techniques, employing a dynamic intravenous injection of gadolinium coupled with a rapid breath-held gradient echo acquisition have also yielded impressive images that may further the diagnostic accuracy of MR over that of spin echo and cine techniques [24,25]. The one modality that has not substantially improved has been conventional angiography. Currently the primary indication for diagnostic arteriography of acute aortic dissection is in the setting of arrhythmia or ECG abnormalities suggestive of coronary artery involvement and myocardial ischemia. It is likely that in appropriately skilled hands, the accuracy of TEE, CT and MR will be nearly identical for the diagnosis of aortic dissection. Access to these three modalities is another issue, however. Since the patients suffering from acute aortic dissection are typically critically ill and potentially in need of an emergent operation, expediency of diagnosis is important. In a center where experienced cardiologists are available to perform state-of-the-art TEE in the emergency room to identify the presence of an intimal flap in the ascending aorta, this will likely be the preferred first line imaging test. An additional advantage of TEE over CT acutely is its ability to identify aortic valvular insufficiency, which in the setting of acute aortic dissection will indicate the need for emergent valve replacement in addition to aortic repair. When high quality TEE is unavailable, however, in most institutions CT will be the modality that is most accessible and staffed to handle potentially hemodynamically unstable patients. When considering the clinical utility of a diagnostic test, it is useful to consider the likelihood that a test will suggest an alternative diagnosis when the primary diagnosis is not present. In the setting of chronic aortic dissection or acute dissection that does not necessitate emergent operation, other imaging issues emerge that cannot be addressed by TEE. These include the extension of the intimal flap into aortic branches, true lumenal compression by the false lumen limiting blood flow to the abdomen, and the presence of fenestrations that allow communication between the true and false lumen. Clinically relevant aortic branch involvement typically is intraabdominal where mesenteric ischemia, renal insufficiency, and lower extremity claudication indicate extension into mesenteric, renal and iliac arteries, respectively. Intervention, either surgically or using catheter based techniques may be required. CT appears to be excellent for establishing extension of intimal flaps into aortic branches, however, its accuracy has not been established. When considering catheter based interventions, preprocedural CT can be very useful. The simultaneous visualization of all aortic lumena can help to avoid confusion in the angiography suite that results from opacification of only one of three or more lumena in a
complex dissection. The CT scan can also help identify the best route for achieving access to aortic branches. While the superior spatial resolution and improved aortic enhancement provided by helical CT results in substantially better images than conventional CT some pitfalls persist. Pulsation in the ascending aorta can mimic an intimal flap. This artifact tends to be less of a problem with helical CT as the generation of closely overlapping sections in the region of the suspected artifact or intimal flap, displayed sequentially as a cine loop, will usually establish the artifact as rotating or moving relative to the aorta. Examination of wide windows may document extension of artifacts beyond the aortic wall. Also, as previously discussed, the use of retrospective segmented reconstruction of the helical scan data may eliminate motion artifacts observed on standard reconstructions [26]. Finally, differential flow in the true and false lumena can result in the spurious appearance of a thrombosed false lumen when scan delay is based upon a timing study directed to the true lumen of the aorta. In the setting of suspected aortic dissection, bolus timing should be performed just below the aortic arch, where transverse cross-sections of the distal ascending and proximal descending aorta can be evaluated. A region-of interest is placed in both the true and false lumen of and two curves are generated. A delay time that assures some false lumenal opacification is selected and the bolus duration then extended by the number of seconds between the true lumenal peak and the selected delay time. This assures true and false lumenal opacification for the duration of the helical acquisition. Other potential pitfalls of CTA are illustrated by two missed aortic dissections in Quint’s series. One was in a patient with hematoma surrounding the ascending aorta, but no identifiable aortic lesion on CT. At surgery, a short ascending aortic intimal tear was noted with a thrombosed false lumen, that simulates mediastinal hematoma on CT. The other misdiagnosed case of aortic dissection was in a patient with an intimal tear on the underside of the aortic arch associated with partial thrombosis of the false lumen, which mimicked a small penetrating ulcer with associated intramural hematoma on CT [9]. When considering alternative visualization techniques in the setting of aortic dissection, it should not be surprising that MIPs are limited as they will not demonstrate the intimal flap unless it is oriented perpendicular to the MIP. Curved planar reconstructions can be very useful for displaying the flap within the center of the vessel, while shaded surface displays depict the interface if the intimal flap with the aortic wall. The use of raysum projections has also been advocated for intimal flap depiction [27].
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
S47
3.4. Intramural hematoma
3.5. Thoracic aortic trauma
Although initially recognized pathologically in 1920, intramural hematoma has only recently been recognized as a distinct clinical entity from aortic dissection. Several mechanisms have been proposed, including spontaneous rupture of vasa vasorum, intimal fracture of an atherosclerotic plaque, and intramural propagation of hemorrhage adjacent to a penetrating atherosclerotic ulcer. Regardless of the cause, patients with intramural hematomas exhibit signs and symptoms as well as risk profiles that are virtually identical to classic aortic dissection [28]. The CT appearance of intramural hematomas caused by penetrating atherosclerotic ulcers was described in detail by Kazerooni and co-workers, based upon an analysis of conventional CTs in 16 patients. In addition to the visualization of at least one ulcer in 15/16, the intramural hematoma was visualized in all 16, and its subintimal location was confirmed by the observation of displaced intimal calcifications in 13 [29]. While intramural hematomas associated with ulceration tend to predominate in the descending aorta [29], the distribution of all intramural hematomas was 48% ascending aortic, 8% aortic arch, and 44% descending aortic in one series [28]. The advantage of helical CT for the assessment of intramural hematomas hasnot been specifically reported, however, the thinner collimation, volumetric acquisition, and superior aortic enhancement is likely to improve the identification and characterization of small atherosclerotic ulcerations over conventional CT and possibly aortography. In Quint’s series two patients in whom intramural hematomas were identified, but an associated ulceration could not be identified were scored as false negative examinations. Recognizing that intramural hematoma formation has two proposed mechanisms: (1) extension of a penetrating ulcer into the media with subsequent bleeding into the aortic wall and (2) primary disruption of vasa vasorum without penetrating ulcer formation [29,30], the diagnoses in these cases may have been correct given a primary intramural bleed as the mechanism for intramural hematoma formation, which need not be associated with ulceration. One useful observation that may help differentiate intramural hematomas from the thrombosed false lumen of a classical intimal dissection is that the latter tend to longitudinally spiral around the aorta, whereas the former tend to maintain a constant circumferential relationship with aortic wall. Another finding that can be seen in the setting of intramural hematoma is intense enhancement and thickening of the aortic wall external to the hematoma, which may represent adventicial inflammation.
The use of CT scanning for the detection of aortic injury has becomes a controversial topic [31 /33]. The principle application of conventional CT is the detection of mediastinal hemorrhage. A recently published metaanalysis of 18 previously published series of posttraumatic thoracic CT revealed that mediastinal hemorrhage had a specificity of 87.1% and a sensitivity of 99.3% for predicting aortic injury [32]. Further reliance on CT for triaging patients to angiography only when CT was suspicious resulted in an overall cost savings of over $365 000.00 in their own series of 677 trauma patients with chest radiographic abnormalities warranting aortic imaging [32]. While these results are impressive, some have argued that the confident identification of mediastinal hematoma, particularly on unenhanced CT, is extremely difficult [31]. The application of helical CTA to suspected aortic trauma offers a new and important dimension to CT studies in these patients. Although initial reports have not relied upon the use of high resolution helical acquisitions coupled with high flow iodinated contrast injections, the results are encouraging. Gavant and colleagues published the first helical CTA series of aortic injury. Using 7 mm collimation and a contrast medium flow rate of 1.5 /2.0 ml/s, they found that the sensitivity of CT was greater than that of conventional arteriography (100 vs. 94.4%) but the specificity and positive predictive values were less than those of conventional arteriography (81.7 and 47.4%, respectively, for CT, vs. 96.3 and 81% for aortography) in a subset of 127/1518 patients with nontrivial blunt thoracic trauma who under went both CT and aortography. Perhaps the most encouraging result of this study was that no false negative results occurred in the 21 patients with aortic injury [34]. This remains the only published comparative series of helical CTA of blunt thoracic aortic trauma to date. Gavant subsequently described the CT appearance of 38 thoracic aortic or great vessel injuries in 36 patients identified with helical CT and confirmed with aortography or surgery. Six (17%) of these cases were found to have either no or ‘difficult-to-detect’ para-aortic or mediastinal hematoma. Transverse sections showed either an intimal flap or thrombus protruding into the aortic lumen in all cases. Of 28 injuries to the descending aorta, 23 (82%) were associated with a pseudoaneurysm. In subjectively comparing the value of the reconstructed transverse sections to multiplanar reformations and 3-D renderings, the authors felt that the transverse sections were best for depicting the proximal and distal extent of the lesion, and aside from the fact that multiplanar reformations and 3-D renderings ‘portray the thoracic aortic lumen in a familiar light’ did not contribute
S48
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
substantially to the identification and characterization of aortic injury [35]. Some caution should be exercised when interpreting the very impressive reported sensitivity of helical CT for aortic injury. In the aforementioned study, iodinated contrast medium was only administered to patients with evidence for mediastinal hematoma. This of course assumes that the sensitivity of mediastinal hematoma on CT is 100% for detecting aortic injury. Although many authors believe this to be true, the negative predictive value of CT for aortic injury has not been proven [36]. The definitive answer to this issue requires a prospective trial where both CT and angiography are performed in all patients who would undergo aortography on the basis of clinical and chest radiographic criteria. According to Trerotola, this would require approximately 1500 patients, assuming 98% negative predictive value with 98% statistical power [36]. 3.6. Peripheral arteries CT angiography of the lower and upper extremity arteries can be performed exclusively using MDCT. While there are scant clinical reports in the literature, the technique for lower extremity CTA using a 4-row CT scanner is described in [37]. The main difference in using an 8- or 16-row CT scanner is that 1.25-mm or lower section thickness is always used, and the overall scan time is diminished.
References [1] Rubin GD, Shiau MC, Schmidt AJ, et al. CT angiography: historical perspective and new state-of-the-art using multi detector-row helical CT. J Comput Assist Tomogr 1999;23(Suppl. 1):S83 /90. [2] Rubin GD, Shiau MC, Leung AN, Kee ST, Logan LJ, Sofilos MC. Aorta and iliac arteries: single versus multiple detector-row helical CT angiography. Radiology 2000;215:670 /6. [3] Rofsky NM, Adelman MA. MR angiography in the evaluation of atherosclerotic peripheral vascular disease. Radiology 2000;214:325 /38. [4] Leng GC, Lee AJ, Fowkes FG, et al. Incidence, natural history and cardiovascular events in symptomatic and asymptomatic peripheral arterial disease in the general population. Int J Epidemiol 1996;25:1172 /81. [5] Ascer E, Veith FJ, Flores SA. Infrapopliteal bypasses to heavily calcified rock-like arteries. Management and results. Am J Surg 1986;152:220 /3. [6] Misare BD, Pomposelli FB, Jr, Gibbons GW, Campbell DR, Freeman DV, LoGerfo FW. Infrapopliteal bypasses to severely calcified, unclampable outflow arteries: 2-year results. J Vasc Surg 1996;24:6 /15 (Discussion 15 /16). [7] Ho KY, Leiner T, de Haan MW, Kessels AG, Kitslaar PJ, van Engelshoven JM. Peripheral vascular tree stenoses: evaluation with moving-bed infusion-tracking MR angiography [see comments]. Radiology 1998;206:683 /92. [8] Ruehm SG, Hany TF, Pfammatter T, Schneider E, Ladd M, Debatin JF. Pelvic and lower extremity arterial imaging: diag-
[9]
[10]
[11]
[12]
[13]
[14] [15]
[16]
[17] [18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
nostic performance of three-dimensional contrast-enhanced MR angiography. Am J Roentgenol 2000;174:1127 /35. Quint LE, Francis IR, Williams DM, et al. Evaluation of thoracic aortic disease with the use of helical CT and multiplanar reconstructions: comparison with surgical findings. Radiology 1996;201:37 /41. Masuda Y, Takanashi K, Takasu J, Morooka Y, Inagaki Y. Expansion rate of thoracic aortic aneurysms and influencing factors. Chest 1992;102:461 /6. Dapunt L, Galla JD, Sadeghi AM, et al. The natural history of thoracic aortic aneurysms. J Thorac Cardiovasc Surg 1994;107:1323 /33. Papanicolaou N, Wittenberg J, Ferrucci JT, et al. Preoperative evaluation of abdomial aortic aneurysms by computed tomography. Am J Roentgenol 1986;146:711 /5. LaRoy LL, Cormier PJ, Matalon TAS, Patel SK, Turner DA, Silver B. Imaging of abdominal aortic aneurysms. Am J Roentgenol 1989;152:785 /92. Bandyk DF. Preoperative imaging of aortic aneurysms. Surg Clin North Am 1989;69:721 /35. Pavone P, Di Cesare E, Di Renzi P, et al. Abdominal aortic aneurysm evaluation: Comparison of US, CT, MRI, and angiography. Magn Reson Imaging 1990;8:199 /204. Vowden P, Wilkinson MB, Ausobskky JR, Kester RC. A comparison of three imaging techniques in the assessment of an abdominal aortic aneurysm. J Cardiovasc Surg 1989;30:891 /6. Crawford JL, Stowe CL, Safi HJ, Hallman CH, Crawford ES. Inflammatory aneurysms of the aorta. J Vasc Surg 1985;2:113 /24. Rubin GD, Walker PJ, Dake MD, et al. 3D spiral CT angiography: an alternative imaging modality for the abdominal aorta and its branches. J Vasc Surg 1993;18:656 /66. Van Hoe L, Baert AL, Gryspeerdt S, et al. Supra and juxtarenal aneurysms of the abdominal aorta: preoperative assessment with thin-section spiral CT. Radiology 1996;198:443 /8. Zeman RK, Silverman PM, Berman PM, Weltman D, Davros WJ, Gomes MN. Abdominal aortic aneurysms: evaluation with variable-collimation helical CT and overlapping reconstruction. Radiology 1994;193:555 /60. Cigarroa JE, Isselbacher EM, DeSanctis RW, Eagle KA. Medical progress. Diagnostic imaging in the evaluation of suspected aortic dissection: old standards and new directions. Am J Roentgenol 1993;161:485 /93. Erbel R, Daniel W, Visser C, Engberding R, Roelandt J, Rennollet H. Echocardiography in diagnosis of aortic dissection. Lancet 1989;4:457 /61. Nienaber CA, Kodolitsch Y, Nicolas V, et al. The diagnosis of thoracic aortic dissection by noninvasive imaging procedures. New Engl J Med 1993;328:1 /9. Prince MR, Narasimham DL, Jacoby WT, et al. Three-dimensional gadolinium-enhanced MR angiography of the thoracic aorta. Am J Roentgenol 1996;166:1387 /97. Krinsky GA, Rofsky NM, DeCorato DR, et al. Thoracic aorta: comparison of gadolinium-enhanced three-dimensional MR angiography with conventional MR imaging. Radiology 1997;202:183 /93. Posniak HV, Olson MC, Demos TC. Aortic motion artifact simulating dissection on CT scans: elimination with reconstructive segmented images. Am J Roentgenol 1993;161:557 /8. Zeman RK, Berman PM, Silverman PM, et al. Diagnosis of aortic dissection: value of helical CT with multiplanar reformation and three-dimensional rendering. Am J Roentgenol 1995;164:1375 / 80. Nienaber CA, Kodolitsch Y, Petersen B, et al. Intramural hemorrhage of the thoracic aorta: diagnostic and therapeutic implications. Circulation 1995;92:1465 /72. Kazerooni EA, Bree RL, Williams DM. Penetrating atherosclerotic ulcers of the descending thoracic aorta: evaluation with CT
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
[30] [31] [32]
[33]
and distinction from aortic dissection. Radiology 1992;183:759 / 65. Gore I. Pathogenesis of dissecting aneurysm of the aorta. Arch Pathol Lab Med 1952;53:142 /53. Raptopoulos V. Chest CT for aortic injury: maybe not for everyone. Am J Roentgenol 1994;162:1053 /5. Mirvis SE, Shanmuganathan K, Miller BH, White CS, Turney SZ. Traumatic aortic injury: diagnosis with contrast-enhanced thoracic CT */5 year experience at a major trauma center. Radiology 1996;200:413 /22. Brasel KJ, Weigelt JA. Blunt thoracic aortic trauma */a cost / utility approach for injury detection. Arch Surg 1996;131:619 /26.
S49
[34] Gavant ML, Manke PG, Fabian T, Flick PA, Graney MJ, Gold RE. Blunt traumatic aortic rupture: detection with helical CT of the chest. Radiology 1995;197:125 /33. [35] Gavant ML, Flick P, Manke P, Gold RE. CT aortography of thoracic aortic rupture. Am J Roentgenol 1996;166:955 /61. [36] Trerotola SO. Can helical CT replace aortography in thoracic trauma. Radiology 1995;197:13 /5. [37] Rubin GD, Schmidt AJ, Logan LJ, Sofilos MC. Multi-detector row CT angiography of lower extremity arterial inflow and runoff: initial experience. Radiology 2001;221:146 /58.
MDCT imaging of the aorta and peripheral vessels Geoffrey D. Rubin * Department of Radiology, Stanford University School of Medicine, Mail Code 5105 GRANT, S072B Stanford, CA 94305-5105, USA
Abstract Since its clinical introduction in 1991, volumetric CT scanning using spiral or helical scanners has resulted in a revolution for diagnostic imaging. Helical CT has improved over the past 8 years with faster gantry rotation, more powerful X-ray tubes, and improved interpolation algorithms, but the greatest advance has been the recent introduction of multi detector-row CT (MDCT) scanners [J. Comput. Assist. Tomogr. 23 (1999) S83]. Currently capable of acquiring four channels of helical data simultaneously, MDCT scanners have achieved the greatest incremental gain in scan speed since the development of helical CT and have profound implications for clinical CT scanning. Fundamental advantages of MDCT include substantially shorter acquisition times, retrospective creation of thinner or thicker sections from the same raw data, and improved three-dimensional (3-D) rendering with diminished helical artifacts. While these features will likely be important to many applications of CT scanning, including the characterization of focal lung and liver lesions through the creation of thin sections retrospectively, the greatest impact has been on CT angiography. The implication for CT angiography is that scans can be performed approximately three-times faster than is possible with the fastest single-detector CT scanner. For example a 1.25 mm nominal thick section (1.6 mm effective section thickness) can be acquired with a table speed of 9.4 mm/s, and a 2.5 mm nominal thick section (3.2 mm effective section thickness) can be acquired with an 18.8 mm/s table speed. The advantages of MDCT for imaging the vascular system can be broken down into three fundamental improvements over single detector-row CT scanners speed (faster), distance (longer), and section thickness (better). The focus of this article will be how multidetector-row CT technology has substantially improved aortoiliac and lower extremity arterial imaging. # 2003 Published by Elsevier Science Ireland Ltd. Keywords: MDCT; Aorta; Peripheral vessels
1. Faster Imaging of the abdominal aortoiliac system is now performed in 20 s using a section thickness that is 34% thinner than our former protocol requiring 50 s with a subsecond single-detector row CT scanner. As a result of the shorter acquisition time, we have reduced our contrast dose from 180 to 80 ml. In fact we can now image the entire thoracoabdominal aorta and iliac arteries in a mere 30/40 s using 2.5-mm section thickness throughout the entire scan volume. We studied 48 patients with aortic aneurysm or dissection who underwent CT angiography with both
* Tel.: /1-650-723-7647; fax: /1-650-725-7296. E-mail address: [email protected] (G.D. Rubin).
MDCT and single detector CT (SDCT) on separate dates. Section thickness (full-width at half-maximum of section profile), scan duration, scan coverage, iodinated contrast dosage and mean aorto-iliac attenuation, measured every 20 mm along the z-axis of each patient, were compared. Data were summarized as the speed (coverage/duration), scan efficiency (speed/section thickness), and contrast efficiency (mean aortic attenuation/ contrast dose). We found that the speed of CTA was 2.5-times faster, scan efficiency was 4.1-times greater, and contrast efficiency was 2.4-times greater with MDCT than SDCT. The average section thickness was 40% thinner (3.2 vs. 5.3 mm), despite a scan duration that was 58% shorter (22 vs. 54 s) with MDCT versus SDCT [2]. Image quality was also substantially improved with MDCT, because a narrower section profile was em-
0720-048X/03/$ - see front matter # 2003 Published by Elsevier Science Ireland Ltd. doi:10.1016/S0720-048X(03)00036-6
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
ployed, and the faster acquisition resulted in greater coherency of the contrast bolus. This is particularly evident in obliquely oriented vessels such as the renal and iliac arteries. Within the chest, artifacts relating to cardiac and arterial pulsation are substantially diminished in part due to the substantial increase in the table travel distance during each cardiac cycle.
2. Longer In addition to improving the quality and efficiency of aortic, CTA, MDCT has enabled the development of fundamentally new applications of CTA, such as the assessment of lower extremity occlusive disease. The principal clinical presentations of lower extremity occlusive disease vary between claudication, non-healing ulceration, and gangrene. Clinically significant lesions could be located anywhere between the aorta to the trio of lower leg vessels, the anterior tibial, peroneal, and posterior tibial arteries. The vessels proximal to the femoral arteries are considered inflow vessels, while those distal are ‘run-off’. CTA has been used to study lower extremity inflow lesions, but the length of the runoff has limited its role for assessing lower extremity occlusive disease as a whole. Multi detector-row CT makes a complete acquisition of lower extremity inflow and run-off possible. Using 2.5-mm collimation and a table speed of 18.8 mm/s (3.2 mm-effective section thickness), a distance of 1300 mm can be covered in less than 70 s. This distance is adequate for imaging from the celiac origin through the mid-foot in the majority of patients. We currently perform this examination using 180 ml of iodinated contrast medium injected at a rate of 3.5 ml/s, which results in a 51-s injection. This amount of contrast medium suffices to image only the celiac to the femoral artery bifurcations, using a single-detector CT scanner. Although the contrast injection is nearly 20 s less than the scan duration, our preliminary experience suggests that excellent arterial opacification is achieved throughout the scan volume. This may be attributable to the substantial delay that tends to occur when blood travels from the aorta through the run-off vessels. Nevertheless, this is an issue that must be closely monitored as this new CT application emerges, particularly given the propensity for substantial variations in flow rates between the two legs, when there are substantial asymmetries of disease involvement. We have studied 16 patients with lower extremity claudication or limb-threatening ischemia had severe lower extremity occlusive disease documented with intra-arterial DSA. All 52 occluded segments and 47 patent segments with ]/50% stenoses were identified and correctly characterized with MDCTA. Patent distal run-off vessels, that were undetected with DSA, were
S43
identified with CTA in six limbs. MDCTA uniquely displayed complex three-dimensional (3-D) relationships between patent vessels, collateral vessels, and mural calcium both in patent and occluded segments. An alternative noninvasive imaging technique for assessing lower extremity arterial disease that requires no ionizing radiation nor iodinated contrast media and has been in routine clinical use for several years is magnetic resonance (MR) angiography [3]. There are many compelling advantages to the use of MR imaging, which in addition to the aforementioned include less primary reconstructions owing to coronal acquisitions and no interference from high signal non-arterial structures analogous to bone in CT angiography. Nevertheless, there are several reasons why CT angiography might be a compelling alternative to MR angiography. Because of an association between lower extremity atherosclerosis and coronary artery disease, implanted pacemakers and defibrillators are common in this patient population [4] and are contraindications to MR imaging. Additionally, other metallic devices such as arterial stents or joint prostheses can result in substantial artifact, hindering arterial evaluation [3]. The characterization of mural calcification is not possible with MR and may have therapeutic relevance[5,6]. Finally, we have achieved voxel dimensions of 0.7 /0.7 /1.25, 3.2 mm that allow us to image the entirety of the lower extremity arterial system in 60 /85 s with 4DRCT. These voxel dimensions are five to fourteen-times smaller than current MR angiographic techniques [7,8] and thus result in a higher spatial resolution acquisition.
3. Better For dedicated applications of CT angiography requiring limited table travel such as renal, pulmonary, mesenteric, cranial, and distal extremity arterial imaging, MDCT allows acquisitions with near isotropic spatial resolution, where the effective section thickness is between 0.75 and 1.6 mm. As a result, MDCT provides imaging of the vasculature with unprecedented volumetric spatial resolution for any imaging modality. This has proved useful to us in a number of instances where either the inferior spatial resolution of MRA or the limited contrast resolution and display of conventional angiography have failed to fully illustrate clinically relevant complex arterial lesions. Examples of these situations will be presented. The advantages of using MDCT for assessing the vasculature are substantial, and we have not found any situations where SDCT would be preferred to MDCT. MDCT allows substantial reductions in iodinated contrast utilization, improved spatial resolution, less pulsation artifacts, and greater coverage than SDCT.
S44
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
3.1. Thoracic aortic aneurysm CTA is a useful tool for diagnosing thoracic aortic aneurysms, determining their extent, and predicting appropriate management [9]. While the diagnosis of aortic aneurysms are readily made from transverse sections, an assessment of the extent of the lesion, particularly when the brachiocephalic branches are involved is facilitated by an assessment of CPRs and SSDs. In general thoracic aortic aneurysms greater than 5 cm are at an increased risk for rupture. Although thoracic aortic aneurysms expand at a slower rate than abdominal aortic aneurysms, surgical repair is contemplated when thoracic aneurysms reach a diameter of 5/6 cm [10,11]. Helical CT can facilitate surgical planning by delineating the extent of the aneurysm and the involvement of aortic branches. In fact Quint and colleagues found that an analysis of transverse sections and multiplanar reformations were 94% accurate with a positive predicative value of 95% and negative predictive value of 93% for successfully predicting the need for hypothermic circulatory arrest. With the exception of the elimination of one false negative result, the addition of multiplanar reformations did not alter the predictions made from the transverse sections alone. 3-D renderings were not evaluated in this study [9]. Due to the tortuosity and curvature of the thoracic aorta, aneurysm sizing is performed most accurately when double-oblique tomograms are generated perpendicular to the aortic flow lumen. The challenge of such an approach is that data concerning the risk of aneurysm rupture and expansion rate are based upon measurements made from transverse sections, where true diameters can be overestimated. Further the measurement technique must be reproducible to assess the rate of aneurysm expansion on sequential studies. Until analysis tools are available to automatically identify the center of the flow channel, create true perpendicular tomograms, and compute accurate cross-sectional areas and mean diameters, the creation of true vessel cross-sections is probably not practical for routine applications unless sizing endoluminal prostheses. The most sensitive measure of aneurysm size, however, is not the determination of true aortic diameters or even cross-sectional area, but aneurysm volume. This approach has substantial drawbacks. While the volumetric data of helical CT should be excellent for determining aneurysm volume, the accurate segmentation of both patent, thrombosed, and atheromatous elements of the aorta must be segmented from the adjacent structures to make this determination. Currently, the only technique available to perform this is a painstaking manual segmentation performed by drawing regions-of-interest around the aorta on each cross-
section. Further, to date aneurysm expansion has been studied primarily in terms of radial expansion of the aorta. While aneurysm volume determination is an attractive measure of aneurysm growth based upon theoretical considerations, data concerning the risk of aneurysm rupture and guidelines for intervention are based upon traditional transverse diameter measurements [10,11]. 3.2. Abdominal aortic aneurysm Aortography has traditionally been obtained prior to resection of abdominal aortic aneurysms to define (a) the relationship of the aneurysm to main and accessory renal, iliac, superior, and inferior mesenteric arteries, (b) extension of the aneurysm into the common, external, or internal iliac arteries to determine the type and length of prosthetic graft utilized, and (c) detect the presence of coexistent iliac or renal occlusive disease [12,13]. Computed tomography has been advocated in the preoperative evaluation of abdominal aortic aneurysm [12]. It is less invasive than angiography, more accurate than conventional angiography for predicting abdominal aortic aneurysm size [14,15] and is superior to angiography in its ability to demonstrate mural thrombus within an aneurysm, inflammatory aneurysms, perianeurysmal blood due to contained rupture [12,14], and co-existent non-vascular abdominal disease [12]. The ability of conventional CT to accurately demonstrate juxta-renal or supra-renal extension of the aneurysm has been controversial. Some authors have reported predictive values as low as 13% for supra-renal extension [16], while others have reported infra-renal localization of the aneurysm neck in as high as 94% [12]. Data such as this must be interpreted carefully with consideration of the pre-test probability that an abdominal aortic aneurysm will be infra-renal in approximately 95% of cases. Conventional CT has been considered inadequate for assessing the position and patency of aortic branch origins and stenoses and detecting accessory renal arteries [12,15 /17]. Identification of these coexistent abnormalities is critical to optimizing surgical repair of abdominal aortic aneurysms. Helical CT overcomes many of the limitations associated with conventional angiography. Specifically, the elimination of ventilatory misregistration enables an accurate depiction of the aneurysm neck relative to aortic branch vessels as well as an evaluation of the aortic branch vessels themselves. In a preliminary study of nine aortic aneurysms, CTA correctly determined the relationship of the proximal aspect of the aneurysm relative to renal artery branches in all cases[18]. Zeman and colleagues studied the effects of overlapping versus non-overlapping reconstruction and fixed versus variable collimation for determining the extent of abdominal aortic aneurysms on spiral CT sections. On axial
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
sections alone, the determination of supra- versus infrarenal extension of the aneurysm was correctly diagnosed in 23 of 23 patients, whereas with non-overlapping reconstruction, there was false positive supra-renal extension in one patient. Finally, Van Hoe and colleagues studied 38 patients with AAA and a high prevalence of juxta- (n /8) and supra-renal (n /7) proximal necks. In comparison to measurements made during open repair, they found that DSA correctly characterized only 12/15 (80%) aneurysms with juxta- or supra-renal necks, while spiral CTA, performed with 2 mm collimation correctly characterized 14/15 (93%) [19]. The greater accuracy of CTA can be attributed to its ability to detect the outer wall of the aorta and not just the flow lumen. Juxta- and supra-renal extension of AAA can be missed on DSA due to the presence of mural thrombus and atheroma at the proximal neck. Regarding distal AAA involvement of the iliac arteries, Zeman et al. found that transverse CTA sections reconstructed with overlapping reconstructions resulted in four of 23 misdiagnoses (three false positive; one false negative) while non-overlapping sections resulted in five of 20; 23 incorrect diagnoses (four false positive iliac involvement; one false negative) [20]. It is likely that these diagnostic errors would have been reduced if multiplanar and 3-D renderings were used in association with the axial sections for interpretation. The ability of CTA to identify renal artery stenosis in the presence of abdominal aortic aneurysms has been assessed in two studies. Using spiral CT with 3-D rendering, four of four greater than 50% diameter renal artery stenoses were detected in nine abdominal aortic aneurysms, with no false positive diagnoses of renal artery stenosis[18]. Of 83 arteries identified in Van Hoe and colleagues series, hemodynamically significant stenoses or occlusion were present in 17 on DSA and detected with 94% sensitivity and 96% specificity with CTA [19]. The influence of optimized CT angiographic technique (small effective section thickness and overlapping reconstructions) has been suggested by Zeman and coworkers, in a study of 23 patients with abdominal aortic aneurysms, seven of which that were associated with greater than 50% renal artery stenosis. Renal artery stenosis was detected by spiral CT in one of two patients imaged with 5 mm collimation, and in four of five patients imaged with 3 mm collimation. Further, stenotic renal arteries were correctly identified in only two of seven patients with stenoses on non-overlapping sections; whereas, on overlapping sections stenoses were correctly identified in five of seven patients. Moreover, none of the four accessory renal arteries present in the aneurysm population was seen with non-overlapping sections, however, two of four were seen with overlapping sections [20].
S45
Helical CT provides the same information regarding aneurysm size and extent of mural thrombus available with conventional CT, however, the volumetric acquisition allows multiplanar and three dimensional renderings to be generated perpendicular to the long axis of the aneurysm, resulting in greater accuracy of the aneurysm size measurements. 3.3. Thoracic aortic dissection The critical clinical issue required of any imaging test applied to a patient suspected of having an aortic dissection is the identification of an intimal flap and its localization to the ascending (type A) or descending (type B) aorta. This fundamental diagnostic feature that determines the need for emergent repair can be addressed by at least four imaging modalities */angiography, CT, MR imaging and transesophageal echocardiography (TEE). The relative accuracy of these modalities has been debated in the medical literature and is confounded by the fact that technical improvements in CT, MR, and TEE have outpaced our ability to compare them in appropriately designed prospective trials. Recent opinion has shifted toward MRI or TEE as the most sensitive tests for aortic dissection [21]. Unfortunately much of this opinion is based upon comparative studies where state-of-the-art MR or TEE is compared with relatively primitive conventional CT technique [22,23]. In 1989, Erbel and colleagues studied 164 consecutive patients with suspected aortic dissection. All patients were studied with transthoracic echocardiography and TEE, 85 patients were studied with CT, and 96 patients were studied with aortography. The technique for CT scanning was 10 mm thick sections acquired at 20 /40 mm intervals through the chest. Details of how iodinated contrast was administered are not given. Not surprisingly CT was found to be less sensitive (77%) than echocardiography (98%) and aortography (89%) in those patients with surgical proof [22]. In 1993 Nienaber and co-workers compared 110 patients with suspected aortic dissection who underwent at least two of three imaging tests*/TEE, CT, or MR. CT was found to have lower sensitivity (93.8%) than TEE (97.7%) and MR (98.3%) and lower specificity (87.1%) than MR (97.8%). For this study, 80 /100 ml of contrast were administered during the course of 5/15 min scans that were performed with section intervals of 20 mm [23]. To date there have been no comparisons of helical CT to either MR or TEE, and while it is tempting to discount these prior studies because of the substantial advances that have occurred in CT imaging, TEE and MR have undergone further improvements as well. TEE has progressively improved with the introduction of biplane and multiplane probes that substantially reduce the blind spots and ambiguous reverberation artifacts that can limit the accuracy of monoplane
S46
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
devices [21]. Newer MR techniques, employing a dynamic intravenous injection of gadolinium coupled with a rapid breath-held gradient echo acquisition have also yielded impressive images that may further the diagnostic accuracy of MR over that of spin echo and cine techniques [24,25]. The one modality that has not substantially improved has been conventional angiography. Currently the primary indication for diagnostic arteriography of acute aortic dissection is in the setting of arrhythmia or ECG abnormalities suggestive of coronary artery involvement and myocardial ischemia. It is likely that in appropriately skilled hands, the accuracy of TEE, CT and MR will be nearly identical for the diagnosis of aortic dissection. Access to these three modalities is another issue, however. Since the patients suffering from acute aortic dissection are typically critically ill and potentially in need of an emergent operation, expediency of diagnosis is important. In a center where experienced cardiologists are available to perform state-of-the-art TEE in the emergency room to identify the presence of an intimal flap in the ascending aorta, this will likely be the preferred first line imaging test. An additional advantage of TEE over CT acutely is its ability to identify aortic valvular insufficiency, which in the setting of acute aortic dissection will indicate the need for emergent valve replacement in addition to aortic repair. When high quality TEE is unavailable, however, in most institutions CT will be the modality that is most accessible and staffed to handle potentially hemodynamically unstable patients. When considering the clinical utility of a diagnostic test, it is useful to consider the likelihood that a test will suggest an alternative diagnosis when the primary diagnosis is not present. In the setting of chronic aortic dissection or acute dissection that does not necessitate emergent operation, other imaging issues emerge that cannot be addressed by TEE. These include the extension of the intimal flap into aortic branches, true lumenal compression by the false lumen limiting blood flow to the abdomen, and the presence of fenestrations that allow communication between the true and false lumen. Clinically relevant aortic branch involvement typically is intraabdominal where mesenteric ischemia, renal insufficiency, and lower extremity claudication indicate extension into mesenteric, renal and iliac arteries, respectively. Intervention, either surgically or using catheter based techniques may be required. CT appears to be excellent for establishing extension of intimal flaps into aortic branches, however, its accuracy has not been established. When considering catheter based interventions, preprocedural CT can be very useful. The simultaneous visualization of all aortic lumena can help to avoid confusion in the angiography suite that results from opacification of only one of three or more lumena in a
complex dissection. The CT scan can also help identify the best route for achieving access to aortic branches. While the superior spatial resolution and improved aortic enhancement provided by helical CT results in substantially better images than conventional CT some pitfalls persist. Pulsation in the ascending aorta can mimic an intimal flap. This artifact tends to be less of a problem with helical CT as the generation of closely overlapping sections in the region of the suspected artifact or intimal flap, displayed sequentially as a cine loop, will usually establish the artifact as rotating or moving relative to the aorta. Examination of wide windows may document extension of artifacts beyond the aortic wall. Also, as previously discussed, the use of retrospective segmented reconstruction of the helical scan data may eliminate motion artifacts observed on standard reconstructions [26]. Finally, differential flow in the true and false lumena can result in the spurious appearance of a thrombosed false lumen when scan delay is based upon a timing study directed to the true lumen of the aorta. In the setting of suspected aortic dissection, bolus timing should be performed just below the aortic arch, where transverse cross-sections of the distal ascending and proximal descending aorta can be evaluated. A region-of interest is placed in both the true and false lumen of and two curves are generated. A delay time that assures some false lumenal opacification is selected and the bolus duration then extended by the number of seconds between the true lumenal peak and the selected delay time. This assures true and false lumenal opacification for the duration of the helical acquisition. Other potential pitfalls of CTA are illustrated by two missed aortic dissections in Quint’s series. One was in a patient with hematoma surrounding the ascending aorta, but no identifiable aortic lesion on CT. At surgery, a short ascending aortic intimal tear was noted with a thrombosed false lumen, that simulates mediastinal hematoma on CT. The other misdiagnosed case of aortic dissection was in a patient with an intimal tear on the underside of the aortic arch associated with partial thrombosis of the false lumen, which mimicked a small penetrating ulcer with associated intramural hematoma on CT [9]. When considering alternative visualization techniques in the setting of aortic dissection, it should not be surprising that MIPs are limited as they will not demonstrate the intimal flap unless it is oriented perpendicular to the MIP. Curved planar reconstructions can be very useful for displaying the flap within the center of the vessel, while shaded surface displays depict the interface if the intimal flap with the aortic wall. The use of raysum projections has also been advocated for intimal flap depiction [27].
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
S47
3.4. Intramural hematoma
3.5. Thoracic aortic trauma
Although initially recognized pathologically in 1920, intramural hematoma has only recently been recognized as a distinct clinical entity from aortic dissection. Several mechanisms have been proposed, including spontaneous rupture of vasa vasorum, intimal fracture of an atherosclerotic plaque, and intramural propagation of hemorrhage adjacent to a penetrating atherosclerotic ulcer. Regardless of the cause, patients with intramural hematomas exhibit signs and symptoms as well as risk profiles that are virtually identical to classic aortic dissection [28]. The CT appearance of intramural hematomas caused by penetrating atherosclerotic ulcers was described in detail by Kazerooni and co-workers, based upon an analysis of conventional CTs in 16 patients. In addition to the visualization of at least one ulcer in 15/16, the intramural hematoma was visualized in all 16, and its subintimal location was confirmed by the observation of displaced intimal calcifications in 13 [29]. While intramural hematomas associated with ulceration tend to predominate in the descending aorta [29], the distribution of all intramural hematomas was 48% ascending aortic, 8% aortic arch, and 44% descending aortic in one series [28]. The advantage of helical CT for the assessment of intramural hematomas hasnot been specifically reported, however, the thinner collimation, volumetric acquisition, and superior aortic enhancement is likely to improve the identification and characterization of small atherosclerotic ulcerations over conventional CT and possibly aortography. In Quint’s series two patients in whom intramural hematomas were identified, but an associated ulceration could not be identified were scored as false negative examinations. Recognizing that intramural hematoma formation has two proposed mechanisms: (1) extension of a penetrating ulcer into the media with subsequent bleeding into the aortic wall and (2) primary disruption of vasa vasorum without penetrating ulcer formation [29,30], the diagnoses in these cases may have been correct given a primary intramural bleed as the mechanism for intramural hematoma formation, which need not be associated with ulceration. One useful observation that may help differentiate intramural hematomas from the thrombosed false lumen of a classical intimal dissection is that the latter tend to longitudinally spiral around the aorta, whereas the former tend to maintain a constant circumferential relationship with aortic wall. Another finding that can be seen in the setting of intramural hematoma is intense enhancement and thickening of the aortic wall external to the hematoma, which may represent adventicial inflammation.
The use of CT scanning for the detection of aortic injury has becomes a controversial topic [31 /33]. The principle application of conventional CT is the detection of mediastinal hemorrhage. A recently published metaanalysis of 18 previously published series of posttraumatic thoracic CT revealed that mediastinal hemorrhage had a specificity of 87.1% and a sensitivity of 99.3% for predicting aortic injury [32]. Further reliance on CT for triaging patients to angiography only when CT was suspicious resulted in an overall cost savings of over $365 000.00 in their own series of 677 trauma patients with chest radiographic abnormalities warranting aortic imaging [32]. While these results are impressive, some have argued that the confident identification of mediastinal hematoma, particularly on unenhanced CT, is extremely difficult [31]. The application of helical CTA to suspected aortic trauma offers a new and important dimension to CT studies in these patients. Although initial reports have not relied upon the use of high resolution helical acquisitions coupled with high flow iodinated contrast injections, the results are encouraging. Gavant and colleagues published the first helical CTA series of aortic injury. Using 7 mm collimation and a contrast medium flow rate of 1.5 /2.0 ml/s, they found that the sensitivity of CT was greater than that of conventional arteriography (100 vs. 94.4%) but the specificity and positive predictive values were less than those of conventional arteriography (81.7 and 47.4%, respectively, for CT, vs. 96.3 and 81% for aortography) in a subset of 127/1518 patients with nontrivial blunt thoracic trauma who under went both CT and aortography. Perhaps the most encouraging result of this study was that no false negative results occurred in the 21 patients with aortic injury [34]. This remains the only published comparative series of helical CTA of blunt thoracic aortic trauma to date. Gavant subsequently described the CT appearance of 38 thoracic aortic or great vessel injuries in 36 patients identified with helical CT and confirmed with aortography or surgery. Six (17%) of these cases were found to have either no or ‘difficult-to-detect’ para-aortic or mediastinal hematoma. Transverse sections showed either an intimal flap or thrombus protruding into the aortic lumen in all cases. Of 28 injuries to the descending aorta, 23 (82%) were associated with a pseudoaneurysm. In subjectively comparing the value of the reconstructed transverse sections to multiplanar reformations and 3-D renderings, the authors felt that the transverse sections were best for depicting the proximal and distal extent of the lesion, and aside from the fact that multiplanar reformations and 3-D renderings ‘portray the thoracic aortic lumen in a familiar light’ did not contribute
S48
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
substantially to the identification and characterization of aortic injury [35]. Some caution should be exercised when interpreting the very impressive reported sensitivity of helical CT for aortic injury. In the aforementioned study, iodinated contrast medium was only administered to patients with evidence for mediastinal hematoma. This of course assumes that the sensitivity of mediastinal hematoma on CT is 100% for detecting aortic injury. Although many authors believe this to be true, the negative predictive value of CT for aortic injury has not been proven [36]. The definitive answer to this issue requires a prospective trial where both CT and angiography are performed in all patients who would undergo aortography on the basis of clinical and chest radiographic criteria. According to Trerotola, this would require approximately 1500 patients, assuming 98% negative predictive value with 98% statistical power [36]. 3.6. Peripheral arteries CT angiography of the lower and upper extremity arteries can be performed exclusively using MDCT. While there are scant clinical reports in the literature, the technique for lower extremity CTA using a 4-row CT scanner is described in [37]. The main difference in using an 8- or 16-row CT scanner is that 1.25-mm or lower section thickness is always used, and the overall scan time is diminished.
References [1] Rubin GD, Shiau MC, Schmidt AJ, et al. CT angiography: historical perspective and new state-of-the-art using multi detector-row helical CT. J Comput Assist Tomogr 1999;23(Suppl. 1):S83 /90. [2] Rubin GD, Shiau MC, Leung AN, Kee ST, Logan LJ, Sofilos MC. Aorta and iliac arteries: single versus multiple detector-row helical CT angiography. Radiology 2000;215:670 /6. [3] Rofsky NM, Adelman MA. MR angiography in the evaluation of atherosclerotic peripheral vascular disease. Radiology 2000;214:325 /38. [4] Leng GC, Lee AJ, Fowkes FG, et al. Incidence, natural history and cardiovascular events in symptomatic and asymptomatic peripheral arterial disease in the general population. Int J Epidemiol 1996;25:1172 /81. [5] Ascer E, Veith FJ, Flores SA. Infrapopliteal bypasses to heavily calcified rock-like arteries. Management and results. Am J Surg 1986;152:220 /3. [6] Misare BD, Pomposelli FB, Jr, Gibbons GW, Campbell DR, Freeman DV, LoGerfo FW. Infrapopliteal bypasses to severely calcified, unclampable outflow arteries: 2-year results. J Vasc Surg 1996;24:6 /15 (Discussion 15 /16). [7] Ho KY, Leiner T, de Haan MW, Kessels AG, Kitslaar PJ, van Engelshoven JM. Peripheral vascular tree stenoses: evaluation with moving-bed infusion-tracking MR angiography [see comments]. Radiology 1998;206:683 /92. [8] Ruehm SG, Hany TF, Pfammatter T, Schneider E, Ladd M, Debatin JF. Pelvic and lower extremity arterial imaging: diag-
[9]
[10]
[11]
[12]
[13]
[14] [15]
[16]
[17] [18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
nostic performance of three-dimensional contrast-enhanced MR angiography. Am J Roentgenol 2000;174:1127 /35. Quint LE, Francis IR, Williams DM, et al. Evaluation of thoracic aortic disease with the use of helical CT and multiplanar reconstructions: comparison with surgical findings. Radiology 1996;201:37 /41. Masuda Y, Takanashi K, Takasu J, Morooka Y, Inagaki Y. Expansion rate of thoracic aortic aneurysms and influencing factors. Chest 1992;102:461 /6. Dapunt L, Galla JD, Sadeghi AM, et al. The natural history of thoracic aortic aneurysms. J Thorac Cardiovasc Surg 1994;107:1323 /33. Papanicolaou N, Wittenberg J, Ferrucci JT, et al. Preoperative evaluation of abdomial aortic aneurysms by computed tomography. Am J Roentgenol 1986;146:711 /5. LaRoy LL, Cormier PJ, Matalon TAS, Patel SK, Turner DA, Silver B. Imaging of abdominal aortic aneurysms. Am J Roentgenol 1989;152:785 /92. Bandyk DF. Preoperative imaging of aortic aneurysms. Surg Clin North Am 1989;69:721 /35. Pavone P, Di Cesare E, Di Renzi P, et al. Abdominal aortic aneurysm evaluation: Comparison of US, CT, MRI, and angiography. Magn Reson Imaging 1990;8:199 /204. Vowden P, Wilkinson MB, Ausobskky JR, Kester RC. A comparison of three imaging techniques in the assessment of an abdominal aortic aneurysm. J Cardiovasc Surg 1989;30:891 /6. Crawford JL, Stowe CL, Safi HJ, Hallman CH, Crawford ES. Inflammatory aneurysms of the aorta. J Vasc Surg 1985;2:113 /24. Rubin GD, Walker PJ, Dake MD, et al. 3D spiral CT angiography: an alternative imaging modality for the abdominal aorta and its branches. J Vasc Surg 1993;18:656 /66. Van Hoe L, Baert AL, Gryspeerdt S, et al. Supra and juxtarenal aneurysms of the abdominal aorta: preoperative assessment with thin-section spiral CT. Radiology 1996;198:443 /8. Zeman RK, Silverman PM, Berman PM, Weltman D, Davros WJ, Gomes MN. Abdominal aortic aneurysms: evaluation with variable-collimation helical CT and overlapping reconstruction. Radiology 1994;193:555 /60. Cigarroa JE, Isselbacher EM, DeSanctis RW, Eagle KA. Medical progress. Diagnostic imaging in the evaluation of suspected aortic dissection: old standards and new directions. Am J Roentgenol 1993;161:485 /93. Erbel R, Daniel W, Visser C, Engberding R, Roelandt J, Rennollet H. Echocardiography in diagnosis of aortic dissection. Lancet 1989;4:457 /61. Nienaber CA, Kodolitsch Y, Nicolas V, et al. The diagnosis of thoracic aortic dissection by noninvasive imaging procedures. New Engl J Med 1993;328:1 /9. Prince MR, Narasimham DL, Jacoby WT, et al. Three-dimensional gadolinium-enhanced MR angiography of the thoracic aorta. Am J Roentgenol 1996;166:1387 /97. Krinsky GA, Rofsky NM, DeCorato DR, et al. Thoracic aorta: comparison of gadolinium-enhanced three-dimensional MR angiography with conventional MR imaging. Radiology 1997;202:183 /93. Posniak HV, Olson MC, Demos TC. Aortic motion artifact simulating dissection on CT scans: elimination with reconstructive segmented images. Am J Roentgenol 1993;161:557 /8. Zeman RK, Berman PM, Silverman PM, et al. Diagnosis of aortic dissection: value of helical CT with multiplanar reformation and three-dimensional rendering. Am J Roentgenol 1995;164:1375 / 80. Nienaber CA, Kodolitsch Y, Petersen B, et al. Intramural hemorrhage of the thoracic aorta: diagnostic and therapeutic implications. Circulation 1995;92:1465 /72. Kazerooni EA, Bree RL, Williams DM. Penetrating atherosclerotic ulcers of the descending thoracic aorta: evaluation with CT
G.D. Rubin / European Journal of Radiology 45 (2003) S42 /S49
[30] [31] [32]
[33]
and distinction from aortic dissection. Radiology 1992;183:759 / 65. Gore I. Pathogenesis of dissecting aneurysm of the aorta. Arch Pathol Lab Med 1952;53:142 /53. Raptopoulos V. Chest CT for aortic injury: maybe not for everyone. Am J Roentgenol 1994;162:1053 /5. Mirvis SE, Shanmuganathan K, Miller BH, White CS, Turney SZ. Traumatic aortic injury: diagnosis with contrast-enhanced thoracic CT */5 year experience at a major trauma center. Radiology 1996;200:413 /22. Brasel KJ, Weigelt JA. Blunt thoracic aortic trauma */a cost / utility approach for injury detection. Arch Surg 1996;131:619 /26.
S49
[34] Gavant ML, Manke PG, Fabian T, Flick PA, Graney MJ, Gold RE. Blunt traumatic aortic rupture: detection with helical CT of the chest. Radiology 1995;197:125 /33. [35] Gavant ML, Flick P, Manke P, Gold RE. CT aortography of thoracic aortic rupture. Am J Roentgenol 1996;166:955 /61. [36] Trerotola SO. Can helical CT replace aortography in thoracic trauma. Radiology 1995;197:13 /5. [37] Rubin GD, Schmidt AJ, Logan LJ, Sofilos MC. Multi-detector row CT angiography of lower extremity arterial inflow and runoff: initial experience. Radiology 2001;221:146 /58.