Chest Imaging Ko et al. Aortic Motion Artifacts on MDCT Sheung-Fat Ko 1 Ming-Jeng Hsieh 2 Min-Chi Chen 3 Shu-Hang Ng 4 Fu-Min Fang 1 Chung-Cheng Huang 1 Yung-Liang Wan 4 Tze-Yu Lee 1 Ko S-F, Hsieh M-J, Chen M-C, et al. Received March 22, 2004; accepted after revision August 12, 2004. Supported, in part, by a grant to Sheung-Fat Ko from the National Science Council (Taiwan), grant no. NSC91-2314- B-182A-171. 1 Department of Radiology, Chang Gung Memorial Hospital at Kaohsiung, Chang Gung University, 123 Ta-Pei Rd., Niao-Sung Hsiang, Kaohsiung Hsien 833, Taiwan. Address correspondence to S-F Ko (sfatko@adm.cgmh.org.tw). 2 Department of Cardiovascular and Thoracic Surgery, Chang Gung Memorial Hospital at Kaohsiung, Chang Gung University, Kaohsiung Hsien 833, Taiwan. 3 Department of Public Health and Biostatistics, Chang Gung Memorial Hospital at Kaohsiung, Chang Gung University, Kaohsiung Hsien 833, Taiwan. 4 Department of Radiology, Chang Gung Memorial Hospital at Linkou, Chang Gung University, Taoyuen Hsien, Taiwan. AJR 2005;184:1225 1230 0361 803X/05/1844 1225 American Roentgen Ray Society Effects of Heart Rate on Motion Artifacts of the Aorta on Non-ECG- Assisted 0.5-Sec Thoracic MDCT OBJECTIVE. Our aim was to evaluate the effects of heart rate on aortic motion artifacts on 0.5-sec non ECG-assisted thoracic MDCT. MATERIALS AND METHODS. A total of 124 non ECG-assisted thoracic MDCT scans with satisfactory simultaneous ECG data were reviewed. Scans were grouped according to patient heart rates (beats per minute [bpm]: group A, 46 55; B, 56 65; C, 66 75; D, 76 85; E, 86 95; and F > 95). The groups were compared regarding the presence, locations, and spatial distributions of pulsation artifact, number of slices affected, maximum amplitude of pulsation, continuity of artifact, and the presence of superior vena cava (SVC) pseudoflaps. RESULTS. Of the 124 scans, 114 (91.9%) had aortic motion artifacts, with prevalence ranging from 85.3% (66 75 bpm) to 100% (65 bpm or less). Of the 114 motion artifacts, all affected the ascending aorta, 105 (92.1%) involved the left anterior and right posterior aspects of the aortic circumference, and 106 (93%) were associated with SVC pseudoflaps. Group B had significantly greater numbers of images with artifacts (p < 0.001 0.006), greater artifact amplitudes (p < 0.001 0.002), and a higher continuity trend for the artifacts (p = 0.003 0.194) than did the other five groups. CONCLUSION. Aortic motion artifacts are frequently seen on thoracic MDCT, especially in patients with heart rates of 65 bpm or less. The presence of a SVC pseudoflap is helpful for distinguishing artifacts from dissection. If aortic disease is suspected, then measures to reduce motion artifact, such as ECG-gating, should be considered. T is a commonly used noninvasive C imaging technique for the diagnosis of aortic dissection, but a variety of atypical imaging features, pitfalls, and artifact have been reported [1, 2]. The presence of a hypodense curvilinear interface or a crescentlike thin flap along the wall of the aortic root on CT can simulate life-threatening aortic dissection and is thought to be related to both the pendular and circular aortic wall motions [1 5]. The prevalence of aortic motion artifact on conventional and 1-sec-scanning-time helical CT are reported to be 37% and 57%, respectively, and such artifacts are usually limited to two or three images, with mean maximal amplitudes of 3 4 mm [3 5]. With the development of MDCT technology, simultaneous use of multiple channels allows further reduction of gantry rotation scanning time to 0.5 sec [6 9]. Nevertheless, even with MDCT, aortic motion artifacts remain a potential pitfall. Recently, we encountered a patient who presented with acute severe chest pain, and thoracic MDCT showed a long continuous pseudointimal flap in the ascending aorta, mimicking type A dissection. Subsequent findings of ECG and cardiac catheterization confirmed occlusion of the left anterior descending artery rather than aortic dissection. In addition, a relatively slow heart rate of 60 beats per minute (bpm) was incidentally found. In contrast to two recent reports concerning a significant linear correlation between heart rate and mean aortic motion artifact score on ECG-assisted MDCT of the aorta with fixed reconstruction points in diastole [10 11], this particular case caused us to postulate that heart rate may affect aortic motion artifacts differently on routine thoracic MDCT, rather than a simple positive linear correlation between heart rate and motion artifact. To our knowledge, the influences of the complete cardiac cycle for different heart rates with regard to motion artifacts using non ECG-assisted MDCT have not been reported. The purpose of our study was to evaluate the effects of heart rate on and the prevalence of aortic motion artifacts on non ECG-assisted thoracic MDCT. AJR:184, April 2005 1225
Ko et al. Materials and Methods From September 2002 to June 2003, excluding those patients with known impaired cardiac function, prior thoracic surgery histories, and suspected large mediastinal or lung lesions, we enrolled in this study a total of 141 consecutive adult patients (77 men and 64 women) who were referred to our department for thoracic CT because of a variety of thoracic disorders. This study was approved by the institutional review board of our hospital. Informed consent was obtained from the patients before the examination. The initial diagnoses of these patients included a suspected solitary lung nodule (n = 31), colon cancer with suspected lung metastasis (n = 25), urogenital tumors or gynecologic malignancies with suspected lung metastasis (n = 22), head and neck cancers with suspected lung metastasis (n = 15), esophageal cancer (n = 15), followup of breast cancer after radiation therapy (n = 10), pneumonia (n = 10), suspected bronchiectasis (n = 8), and a suspicious small mediastinal mass (lesion diameter, < 3 cm) (n = 5). None of the patients had symptoms and history of aortic dissection. MDCT All the patients underwent MDCT using the LightSpeed 4-MDCT scanner (GE Healthcare). ECG leads were attached to each patient s body before scanning and then connected to a vital signs monitoring system (Millennia 3500CT-P, In Vivo Research), which was activated simultaneously with CT. The ECG and heart rates were recorded throughout the entire CT examination. TABLE 1 The patients were scanned in a supine position in a craniocaudal direction with the arm above the head; the entire thorax from the thoracic inlet to the level of the adrenal glands was studied. The scanning protocol was set to slice collimation of 2.5 7.5 mm, 0.5-sec gantry rotation, 130 kv, 160 mas, and a field of view of 35 cm. The technician instructed each patient to take three deep breaths before the examination started and to hold their breaths during scanning. In all patients, venous access was achieved through an 18- or 20-gauge IV catheter placed in an antecubital vein. The contrast agent (100 ml of Optiray [ioversol], 300 mg/ml, Mallinckrodt Medical) was given via a mechanical power injector at a rate of 2 ml/sec with a scanning delay of 25 30 sec. Optimum aortic attenuation was ensured by triggering scanning on the basis of identification of 100 H in the upper descending thoracic aorta. The mean breathhold time was 30 ± 4 [SD] sec. After the examination, axial images were reconstructed in 5-mm-thick slices for analysis. In addition, the raw data were reconstructed in 2-mm-thick slices so that multiplanar reconstructions of the thoracic aorta, if necessary, could be performed with good quality. For each patient, hard copies of the images were obtained at a setting appropriate for vascular visualization (window level, 40 50 H; window width, 350 400 H). In addition, the investigators were free to review all images, to perform multiplanar reconstructions of the aorta (liberally in either coronal, coronal oblique, or candy cane view and aortic root long-axis or short-axis view) for delineation of the pulsation artifacts, and to make any measurements using the CT console. Image Evaluation All hard copies were evaluated for motion artifacts by two investigators (each with > 8 years of experience reviewing chest CT scans and who were blinded to the clinical and heart rate data) who worked together by consensus. The absence or presence of aortic motion artifacts throughout the entire thoracic aorta was evaluated first. First, CT was classified as positive for aortic motion artifacts when at least one axial image showed an obvious pseudointimal flap with an enhanced pseudofalse lumen mimicking aortic dissection or an obvious crescent hypodensity along the wall of the aorta mimicking a thrombosed false lumen. Second, for positive cases, the numbers of axial images (5-mm thickness, 360 reconstruction) with aortic motion artifacts were counted. In addition, locations (ascending aorta, arch, or descending thoracic aorta) and the spatial distributions of the aortic artifacts on axial CT images (which were divided into four quadrants: right anterior, right posterior, left anterior, and left posterior) were also recorded. Third, the artifacts were classified as continuous or noncontinuous; continuity of the artifacts was defined as the presence of peripheral crescent hypodensity or a pseudointimal flap in the aorta that could be traced continuously for at least four consecutive axial images. Fourth, the maximum amplitude of the artifact, which was defined as the maximum width Summary of 124 Patients Who Underwent Chest CT with Monitoring of Heart Rate and Assessment of Aortic Motion Artifacts Group A (n =4) Group B (n = 17) Group C (n = 34) Group D (n = 35) Group E (n = 25) Group F (n =9) Sex ratio = M:F 2:2 7:10 20:14 19:16 16:9 6:3 Age (yr) Artifacts Pos:Neg (%) Heart rate a Images with artifacts Artifact max amplitude Continuity of artifact Pos:Neg SVC pseudo flap Pos:Neg 48.8 ± 10.2 49 (37, 60) 51.9 ± 12.1 51 (33, 78) 56.5 ± 15.1 57 (28, 80) 54.2 ± 13.9 52 (19, 82) 56.5 ± 14.3 56 (25, 77) 52.8 ± 10.2 56 (32, 78) 4:0 (100%) 17:0 (100%) 29:5 (85.3%) 31:4 (88.6%) 24:1 (96%) 9:0 (100%) 51.3 ± 2.8 51.5 (48, 54) 3.5 ± 1.3 3.5 (2, 5) 3.3 ± 1.7 3.5 (1, 5) 61.2 ± 3.3 61 (56, 65) 6.9 ± 3.3 7 (3, 16) 5.8 ± 2.3 5 (3, 11) 70.1 ± 3.2 69.5 (66, 75) 3.8 ± 2.2 4 (0, 9) 3.1 ± 2.0 3 (0, 8) Note. Pos = positive, Neg = negative, min = minimum, max = maximum, SVC = superior vena cava a Beats per minute. 79.7 ± 2.5 79 (76, 85) 3.4 ± 2.0 3 (0, 8) 2.8 ± 1.8 3 (0, 7) 89.1 ± 2.6 88 (86, 95) 4.4 ± 2.1 4 (0, 8) 3.4 ± 2.1 3 (0, 9) 100.3 ± 4.5 100 (96, 109) 4.2 ± 1.5 4 (2, 6) 3.1 ± 1.4 3 (1, 5) 2:2 13:4 12:22 11:24 10:15 4:5 3:1 16:1 28:6 29:6 21:4 9:0 1226 AJR:184, April 2005
Aortic Motion Artifacts on MDCT between the outer and pseudointimal flap, was measured using the electronic calipers on the CT console. Finally, the absence or presence of a pseudoflap in the periphery of the superior vena cava (SVC) was also recorded. Statistical Analysis In addition to the determination of overall prevalence, location, and spatial distribution of aortic motion artifacts on MDCT with 0.5-sec gantry rotation, the study population was subdivided into six groups to analyze the effect of heart rates: group A, 46 55 bpm; group B, 56 65 bpm; group C, 66 75 bpm; group D, 76 85 bpm; group E, 86 95 bpm; and group F, greater than 95 bpm. The differences in the absence or presence of aortic motion artifacts, continuity of the artifact and presence or absence of a SVC pseudoflap for the six groups were analyzed using Fisher s exact test. Continuous variables, including numbers of images with artifacts and maximal amplitudes among the six groups, were analyzed using the Mann-Whitney test. Analyses were performed using SAS software (SAS Institute); and because multiple groups were involved, a p value of less than 0.01 was considered statistically significant. A Fig. 1. 41-year-old woman with nasopharyngeal carcinoma who underwent thoracic MDCT with 0.5-sec scanning time because of suspicious lung nodule and ECGrevealed heart rate of approximately 60 beats per minute. A, Twelve consecutive axial MDCT images show presence of pseudointimal flap in left anterior and right posterior aspects of aortic circumference, from aortic root to level of distal ascending aorta. Note pseudointimal flap in superior vena cava. B, Oblique coronal reconstruction of MDCT image reveals long continuous pseudointimal flaps (arrows) due to aortic motion artifacts in ascending aorta, mimicking aortic dissection. Results Seventeen of 141 patients were excluded from the study because of improper body motions or unsatisfactory ECG records during their thoracic MDCT examinations or both. A total of 124 patients (70 men and 54 women, age range, 19 79.8 years; mean age, 54.5 years) were available for study. Table 1 summarizes their sex, age, heart rate, the presence or absence of aortic motion artifacts, numbers of images, maximal amplitudes and continuity of the aortic motion artifacts, and the presence of SVC pseudoflaps for each group. There were no significant differences in age and sex among the six groups. Of the 124 cases reviewed, 114 (91.9%) were positive for aortic motion artifacts on MDCT. Among these 114 cases, all scans showed variable degrees of aortic motion artifacts in the aortic root and ascending aorta (Fig. 1), and three also had coexistent minimal artifacts in the aortic arch and descending thoracic aorta. In addition, aortic motion B Fig. 2. 45-year-old woman with esophageal carcinoma and ECG-revealed heart rate of approximately 78 beats per minute. Axial thoracic MDCT scan reveals pseudointimal flap in left anterior quadrant (solid arrow) of ascending aorta. Note presence of pulsation artifact (open arrow) in pulmonary artery. AJR:184, April 2005 1227
Ko et al. Fig. 3. 54-year-old man with suspected lung metastasis from colon carcinoma and ECG-revealed heart rate of approximately 86 beats per minute. Axial thoracic MDCT scan reveals crescent hypodense artifact along anterior wall and pseudointimal flap along posterior wall (arrows) of ascending aorta. Note enlarged azygoesophageal lymph node. Fig. 4. 62-year-old woman with cervical cancer with suspected lung metastasis and ECG-revealed heart rate of approximately 75 beats per minute. Axial thoracic MDCT scan reveals pseudointimal flap (arrow) in right anterior aspect of ascending aorta. Note enlarged lymph nodes in bilateral hilar and azygoesophageal regions. Fig. 5. 48-year-old woman who underwent right mastectomy due to breast cancer with ECG-revealed heart rate of approximately 79 beats per minute. Axial MDCT scan shows pulsation artifacts in all four quadrants of ascending aorta (arrows). artifacts were simultaneously seen in the left anterior and right posterior quadrants of the aortic circumferences in most (105/114, 92.1%) of the cases. In the remaining nine cases, four had artifacts solely in the left anterior quadrant (Fig. 2), three had artifacts in the midline anterior and posterior aspects of the ascending aorta (Fig. 3), one had an artifact in the right anterior quadrant (Fig. 4), and one had artifacts in all four quadrants (Fig. 5). On the other hand, concerning the presence or absence of pulsation artifacts, there were no significant differences among groups A F (p = 0.156 to > 0.999, Fisher s exact test). Among the 114 positive cases, group B cases had significantly greater numbers of images that showed aortic motion artifacts (mean number of images affected, 6.9 vs 3.4 4.4; p = < 0.001 to 0.006, Mann-Whitney test) and greater maximal amplitude (mean maximal amplitude, 5.8 vs 2.8 3.4 mm; p = < 0.001 to 0.002, Mann-Whitney test) than the other five groups. Group B also had higher probability of continuous artifacts than groups C and D (p = 0.003 0.008, Fisher s exact test) but showed no significant differences from groups A, E, and F (p = 0.029 0.194, Fisher s exact test). There were no significant intergroup differences among groups A, C, D, E, and F for image numbers with artifacts, continuity, and maximal amplitude of the aortic artifacts. Of the 114 patients with aortic motion artifacts, 106 (93%) had associated SVC pseudoflaps (Table 1), which might be a useful finding for differentiating motion artifacts from true aortic dissection flaps. There were no significant differences for occurrence of SVC pseudoflaps among groups A F (p = 0.352 to > 0.999, Fisher s exact test). Discussion CT is an important tool for the evaluation of acute aortic dissection, which requires prompt and accurate diagnosis to initiate appropriate surgical intervention or medical treatment [1, 2]. A confident diagnosis of aortic dissection relies on the visualization of an intimal flap in the thoracic aorta and the presence of enhanced or thrombosed true and false and lumens. However, several artifacts and pitfalls, including aortic and cardiac motion artifacts, streak artifacts, and pericardial structures (e.g., pericardial recess, left brachiocephalic vein, thickened pleura, thymus, and atelectasis) adjacent to the aorta, may cause misinterpretation of the 1228 AJR:184, April 2005
Aortic Motion Artifacts on MDCT images, leading to false-negative or falsepositive diagnoses [1, 2]. On conventional CT with 1.2- to 5-sec scanning times, aortic motion artifacts are unusual because several cardiac cycles are averaged during each data acquisition [3, 4]. Advances in CT technology allow shortening of the gantry rotation time to 1 sec. However, instead of reducing artifacts, Burns et al. [3] and Duvernoy et al. [4] reported incremental prevalence of aortic motion artifacts of 18% and 34%, respectively, for 1-sec incremental thoracic CT. Using 1-sec helical CT, Qanadli et al. [5] reported that the prevalence of aortic motion artifacts was even higher (57%) for data acquired in a helical manner, because there is less averaging of the aortic motion during image reconstruction [5] and thus the aortic motion artifacts are more apparent. MDCT technology allows further reduction of scanning time to 0.5 sec, which has a great impact on thoracic examinations by markedly reducing respiratory artifacts and drastically improving scanning speed and temporal resolution, without any drawbacks to image quality [6 9]. Nevertheless, such fast scanning techniques have not solved the problem of aortic motion artifacts [10]. As shown in our study, aortic motion artifacts were frequently seen on thoracic MDCT, with an overall prevalence of 91.9%; thus, familiarity with these artifacts can help avoid false-positive interpretations, especially in patients with clinical presentations of severe chest pain. In concurrence with prior reports [1 5], our study also showed that aortic motion artifacts predominately affected the aortic root and ascending aorta. The left anterior and right posterior quadrants of the aortic circumferences were simultaneously affected in 105 (92.1%) of 114 patients. The greater the distance away from the heart, the lesser the influence of cardiac and aortic motions; thus, motion artifacts in the aortic arch and descending thoracic aorta were rarely seen among our patients [1 5, 10]. These findings are consistent with the hypothesis that both physiologic pendular motion of the heart and circular aortic motion during systole and diastole are both important factors in the generation of motion artifacts. The distensibility of the aorta, which might be reflected by the severity of the artifact, has also been reported as a factor affecting motion for aortic motion artifacts, which were less marked among elderly patients [12]. Furthermore, we believe that the motion of the aorta may possibly be affected, to a certain degree, by the cardiovascular status, the presence of a mediastinal mass or mediastinal adhesion, due to previous surgery, that might limit the distensibility of the aorta. In this study, patients in groups A F were of similar age ranges with no significant cardiovascular disease, large mediastinal mass, or thoracic surgery history. There have been several studies of ECG-assisted coronary MDCT angiography that focused on reconstruction windows and the effects of heart rate; a negative linear correlation between the heart rate and image quality was found [11, 13, 14]. Use of ECG-assisted thoracic MDCT aortography, with a significant reduction in motion artifacts in the ascending aorta, has also been reported [10]. Morgan- Hughes et al. [11] reported that to maximize the advantage of ECG-assisted MDCT aortic or cardiac images, the patient s heart rate should be slow. Higher rates might produce more artifacts in these studies, for they were performed with image reconstruction at several fixed points in diastole. Nonetheless, these studies had limitations because the aortic motion artifacts were not assessed for the complete cardiac cycle [11, 13, 14]. In reality, aortic motion, which mirrors cardiac motion, continues during diastole. In contrast, for conventional thoracic CT, the aortic motion continues throughout each cardiac cycle, and the severity of the artifact depends on how homogeneous the outcome of averaging is when images are reconstructed at different cardiac cycles. Thus, theoretically, more averaging of different cardiac cycles results in less aortic artifact [3, 4]. On the other hand, with non ECG-assisted thoracic MDCT with a 0.5-sec gantry rotation time, the generation of aortic motion artifact relies on intricate interferences among heart rate; aortic motion; and simultaneous, rapid, helical data volume acquisition by multiple-detector rows. However, previously, to our knowledge, no studies have been published that have evaluated the influences of the complete cardiac cycle for different heart rates with regard to motion artifacts using non ECG-assisted MDCT. Instead of finding a significant linear correlation for heart rate and aortic motion artifacts on ECG assisted MDCT [10, 11], our study showed that a linear relationship was not present on non ECG-assisted MDCT. Nonetheless, the motion artifacts appeared more prominent (with more images and greater amplitude) in group B patients with heart rates ranging from 56 to 65 bpm, whereas there were no significant differences between group A and groups C F with heart rates greater than 65 bpm. On conventional CT and 1- sec helical CT, the motion artifacts are usually limited to one to three contiguous axial images, with a mean amplitude of 3.5 and 4 mm in the proximal ascending aorta, respectively [3 5]. In our study, the motion artifacts in group B and the other five groups (groups A and C F) involved approximately seven images with a mean maximal amplitude of 5.8-mm, and three or four images with a mean maximal amplitude of 2.8- to 3.4-mm, respectively. In addition, there was a higher probability of continuous artifacts (> 4 contiguous axial images) in group B, and a long, continuous pseudointimal flap with pseudolumens could also be clearly seen on multiplanar reconstructed images. We postulate that for a 0.5- sec scanning time, there might be a greater chance for two gantry rotations to match potentially a cardiac cycle with a heart rate of approximately 60 bpm; under such circumstances, reconstructed CT images might encompass full systole to diastole for each heart beat, resulting in more obvious motion artifacts during image reconstruction; and occasionally, generation of a long continuous pseudointimal flap mimicking type A aortic dissection might occur. For groups A and C F, there was less likelihood of such matching of gantry rotation time and heart rate. Various methods of reduction or suppression of heart pulsation and aortic motion artifacts, including extended volume coverage with omission of systole, 180 reconstruction, and reconstruction centered at 50 75% of the R- R interval, have been described [11, 13 15]. However, use of such ECG-assisted MDCT methods may not be practically applied for all routine thoracic examinations, especially for examining patients in the emergency department. Recognition of aortic artifacts on routine thoracic MDCT without ECG-gating is important. Consistent with previous reports, our study also confirmed that aortic motion artifacts typically occurred in the left anterior and right posterior aspects of the ascending aorta [3 5, 13 15]. The low attenuation line extending beyond the wall of the aorta is also reported to be a useful finding to suggest a motion artifact [16], but this finding was not clearly seen among our patients. Nonetheless, 93% of cases with aortic motion artifacts were associated with a SVC pseudoflap in our experience; this CT feature is useful for differentiating motion artifact from true intimal flaps. In conclusion, aortic motion artifacts were frequently seen in the ascending aorta on 0.5-sec non ECG-assisted thoracic MDCT. The artifacts appeared more prominent in group B patients with heart rates of 56 65 bpm. If aortic disease is AJR:184, April 2005 1229
Ko et al. suspected, measures to reduce motion artifact, such as ECG-gating, should be considered. References 1. Fisher ER, Stern EJ, Gidwing JD, Otto CM, Johmson JA. Acute aortic dissection: typical and atypical imaging features. RadioGraphics 1994;14:1263 1271 2. Batra P, Bigoni B, Manning J, et al. Pitfalls in the diagnosis of thoracic aortic dissection at CT angiography. RadioGraphics 2000;20:309 320 3. Burns MA, Molina PL, Gutierrez FG, Sagel SS. Motion artifact simulating aortic dissection on CT. AJR 1991;157:465 467 4. Duvernoy O, Coulden R, Ytterberg C. Aortic motion: a potential pitfall in CT imaging of dissection in the ascending aorta. J Comput Assist Tomogr 1995;19:569 572 5. Qanadli SD, Hajjam ME, Mesurolle B, et al. Motion artifacts of the aorta simulating aortic dissection on spiral CT. J Comput Assist Tomogr 1999;23:1 6 6. Hu H, He H, Fox S, et al. Imaging characteristics and trade-off selections of four-slice CT: theoretical and experimental results. (abstr) Radiology 1998;209(P):283 7. Klingenbeck-Regn K, Schaller S, Flohr T, Ohnesorge B, Kopp AF, Baum U. Subsecond multislice computed tomography: basic and applications. Eur J Radiol 1999;31:110 124 8. Hui H, Pan T, Shen Y. Multislice helical CT: image temporal resolution. IEEE Eng Med Biol Mag 2000;19:384 390 9. Fuchs TO, Kachelriess M, Kalender WA. Systemic performance of multislice spiral computed tomography. IEEE Eng Med Biol Mag 2000;19:63 70 10. Roos JE, Willmann JK, Weishaupt D, Lachat M, Marincek B, Hilfiker PR. Thoracic aorta: motion artifact reduction with retrospective and prospective electrocardiography-assisted multi-detector row CT. Radiology 2002;222:271 277 11. Morgan-Hughes GJ, Owens PE, Marshall AJ, Roobottom CA. Thoracic aorta at multi-detector row CT: motion artifact with various reconstruction windows. Radiology 2003;228:583 588 12. Set PA, Lomas DJ, Maskell GF, Flower CD, Dixon AK. Artifacts in the ascending aorta on computed tomography: another measure of aortic distensibility. Eur J Radiol 1993;16:107 111 13. Flohr T, Prokop M, Becker C, et al. A retrospectively ECG-assisted multislice spiral CT scan and reconstruction technique with suppression of heart pulsation artifacts for cardio-thoracic imaging with extended volume coverage. Eur Radiol 2002;12:1497 1503 14. Hong C, Becker CR, Nuber A, et al. ECG-assisted reconstructed multi-detector row CT coronary angiography: effect of varying trigger delay on image quality. Radiology 2001;220:712 717 15. Loubeyre P, Angelie E, Grozel F, Abidi H, Minh VAT. Spiral CT artifact that simulates aortic dissection: image reconstruction with uses of 180 and 360 linear-interpolation algorithms. Radiology 1997;205:153 157 16. Mukher JI, Varma P, Stark P. Motion artifact simulating aortic dissection on CT. (letter) AJR 1992;159:674 1230 AJR:184, April 2005