Orthogonal Measurement of Thoracic Aorta Luminal Diameter Using ECG-Gated High- Resolution Contrast-Enhanced MR Angiography

Transcription

1 JOURNAL OF MAGNETIC RESONANCE IMAGING 26: (2007) Original Research Orthogonal Measurement of Thoracic Aorta Luminal Diameter Using ECG-Gated High- Resolution Contrast-Enhanced MR Angiography William R. Bireley II, BS, Lincoln O. Diniz, MD, MPH,* Elliott M. Groves, BS, Karin Dill, MD, Timothy J. Carroll, PhD, and James C. Carr, MD Purpose: To compare orthogonal measurements of the thoracic aortic luminal diameter to standard axial measurements within the same patient population using ECGgated high-resolution contrast-enhanced MR angiography (CE-MRA). Materials and Methods: In all, 45 consecutive patients who had undergone CE-MRA for suspected disease of the thoracic aorta were evaluated retrospectively. Two diameter measurement techniques were performed for each patient s thoracic aorta: standard axial and orthogonal to the aorta. Seven anatomic locations along the thoracic aorta were used for measurement. The data obtained were compared using a paired, two-tailed t-test. Results: We found that the aorta diameter measurements acquired from axial MRA images were significantly greater (P 0.05) than those acquired from images orthogonal to the course of the aorta at six of seven anatomic sites. Overall, standard axial measurements were found to overestimate luminal diameter of the thoracic aorta by 0.24 cm (95% confidence interval [CI]: 0.14, 0.33) compared to orthogonal measurements. 13.3% of the patients were placed into a greater aorta size classification based on the axial versus the orthogonal measurements. Conclusion: Standard axial measurements can overestimate vessel size of the thoracic aorta. ECG-gated highresolution CE-MRA can be used to measure orthogonal diameters of the thoracic aorta. Key Words: MR angiography; thoracic aorta; aortic aneurysms; aortic luminal diameter measurement J. Magn. Reson. Imaging 2007;26: Wiley-Liss, Inc. Department of Radiology, Cardiovascular Imaging, Northwestern University, Chicago, Illinois. Contract grant sponsor: the Feinberg School of Medicine. *Address reprint requests to: L.D., Northwestern University, Cardiovascular Imaging, Department of Radiology, 448 East Ontario St., Ste. 700, Chicago, IL diniz1969@yahoo.com Received June 14, 2006; Accepted June 27, DOI /jmri Published online 26 October 2007 in Wiley InterScience (www. interscience.wiley.com). THORACIC AORTIC ANEURYSMS are a significant cause of mortality worldwide and require regular follow-up (1 3). In previous years, follow-up required digital subtraction angiography (DSA), which is invasive and may result in significant complications. Today, less invasive imaging techniques have become standard using cross-sectional imaging (4,5). Axial CT has routinely been used to evaluate the thoracic aorta (6 8) and normal values of luminal diameter have been generated from CT data (9). More recently, Hager et al (10) obtained normal values for the thoracic aorta diameter using orthogonal CT images that were manually adjusted to be perpendicular to the course of the aorta. The authors compared their results to the normal axial values established by Aronberg et al (9) 18 years earlier. Hager et al suggested that their results indicated that measurements taken from axial images could overestimate the diameter of the ascending aorta by as much as 6 mm, or 21%. Conventional contrast-enhanced MR angiography (CE-MRA) is frequently used to assess the thoracic aorta (11); however, the aortic root and sinuses are commonly degraded by cardiac motion artifact. This precludes accurate measurement of aortic dimensions in this region. With current acceleration techniques, such as parallel imaging (12,13), it may be possible to shorten the acquisition time to such an extent that ECG-gating can be used (14). This may significantly improve the image quality (15) at the aortic root, allowing orthogonal measurements to be accurately obtained. The objective of this study was to compare orthogonal measurements of thoracic aortic luminal diameter to standard axial measurements within the same patients using ECG-gated high-resolution CE- MRA. MATERIALS AND METHODS The study protocol was approved by our Institutional Review Board. Informed consent was not required. The study was compliant with the Health Insurance Portability and Accountability Act. Patients A series of consecutive patients who had undergone ECG-gated CE-MRA for suspected disease of the tho Wiley-Liss, Inc. 1480

2 Orthogonal Measurement of Thoracic Aorta 1481 racic aorta between December 2004 and June 2005 were included in this study. ECG-gated CE-MRA underwent technical developments over a period of 2 months before it was introduced clinically in our institution, and for this study that commenced in December Only one, the most recent image study, was used of each patient. All the patients had aneurysms of the ascending aorta, two of which were dissecting. Patients were excluded if they had received a graft replacement of the thoracic aorta. The patients ECG-gated CE-MRA images were evaluated retrospectively. A total of 45 patients were included (27 men, 18 women, mean age [SD]). Imaging Technique All imaging was performed on a clinical 1.5T MRI scanner (Avanto, Siemens Medical Systems, Malvern, PA). Patients were placed in the scanner head first in a supine position. A multichannel phased array coil was placed over the chest for signal reception. CE-MRA images were acquired after scout images and localization was performed. The basic MRA technique involved a 3D gradient echo pulse sequence with the following scanning parameters: TR/TE: 2.8/1.4 msec, flip angle: 25, 512 readout, voxel size: mm 3. The gradient echo technique included spoiling. The acquisition time was 20% of the R-R interval. 0.2 mmol/kg of gadolinium-dtpa (Magnevist, Berlex Laboratories, Wayne, NJ) was injected at a rate of 2 ml/sec through a peripheral intravenous cannula during image acquisition and the contrast transit time was calculated using a standard timing bolus acquisition. Generalized autocalibrating partially parallel acquisition (GRAPPA) imaging, with acceleration factor of 2, was utilized to shorten the acquisition time. An asymmetric partial Fourier k-space acquisition scheme with zero filling was used in the phase-encode and slice-encode directions. Cartesian linear reordering was employed. The MR acquisition was synchronized with the ECG tracing so that each partition was collected only during end diastole. Approximately 20% of the RR interval was used for data acquisition and a delay time was manually set at the user interface, depending on the patient s heart rate. The entire MRA acquisition typically lasted 20 seconds. Digital subtraction of 3D datasets was carried out in-line and automatic maximum intensity projection (MIP) postprocessing was utilized. Contiguous breath-hold axial true fast imaging and steady precession (TrueFISP) images (16,17) of the thoracic aorta were also obtained using the following parameters: TR/TE: 3.2/1.6; flip angle 70 ; mm 2 pixels; 5-mm thick slice. Parallel imaging (GRAPPA) was used to shorten the acquisition time per slice to 200 msec. ECG-gating was also used such that each slice was acquired at end diastole. A delay time was manually set at the user interface, depending on the patient s heart rate. Image Analysis Two diameter measurement techniques were performed for each patient s thoracic aorta: standard axial and orthogonal. The measurements were outer wall to outer wall. Seven anatomic locations were used for measurement: aortic annulus, sinuses of Valsalva, sino-tubular junction, mid ascending aorta at the level of the main pulmonary artery, aortic arch immediately proximal to the innominate artery, descending aorta immediately distal to the left subclavian artery, and descending aorta immediately superior to the diaphragm (Fig. 1). Two measurements were made at each site: anterior posterior and lateral, except for the aortic arch axial images. Only one oblique diameter measurement was made for each axial image of the aortic arch due to the limitation of an axial view of the arch. Standard axial diameters of luminal size were measured from axial TrueFISP images in all seven anatomic locations. Corresponding orthogonal measurements of luminal diam- Figure 1. Diagrammatic representation of the thoracic aorta showing the anatomic locations at which orthogonal measurements were obtained. Axial measurements were obtained in the axial plane in the same locations. 1, aortic annulus; 2, sinuses of Valsalva; 3, sino-tubular junction; 4, mid ascending aorta at the level of the main pulmonary artery; 5, aortic arch immediately proximal to the innominate artery; 6, descending aorta immediately distal to the left subclavian artery; 7, descending aorta immediately superior to the diaphragm.

3 1482 Bireley et al. eter were then made from unsubtracted ECG-gated CE- MRA images in the same anatomic locations using a double-oblique multiplanar reformatting (MPR) algorithm. Statistical Analysis Measurements were stored in a database and exported to a statistical software package (SPSS v. 13.0; Chicago, IL). The data obtained from the axial and orthogonal measurement techniques were compared using a paired, two-tailed t-test at each anatomical site, analyzing the anterior posterior and lateral diameters independently. Statistical significance was defined at the 5% level. The paired difference means and 95% confidence interval (CI) of the differences were also determined. To determine if the difference between the axial and orthogonal measurements was clinically significant, patients were placed into four classifications based on their maximal thoracic aortic diameter. Patients were classified once based on axial measurements and again based on orthogonal measurements. The following four simple classifications were used: normal (3.0 cm or less), ectatic ( 3.0 to 4.0 cm), aneurysmal ( 4.0 to 5.0 cm), and surgically indicated (greater than 5.0 cm). The numbers of patients in each category based on axial and orthogonal measurements were then compared. RESULTS Overall, the orthogonal measurements of luminal diameter were found to be significantly different (P 0.05) from the standard axial measurements in six of the seven anatomic sites examined (Table 1). At five of the seven sites a significant difference was found for both the anterior posterior diameter and the lateral diameter. At the annulus, only the lateral diameter was significantly different (P 0.05). No significant difference was found at the aortic arch immediately proximal to Table 2 Number of Patients within Each Aorta Size Classification Based on Orthogonal Measurements and Axial Measurements Measurement Technique Number of Patients Orthogonal Axial Normal ( 3 cm) 6 2 Ectatic (3-4 cm) Aneurysmal (4-5 cm) Surgically indicated ( 5 cm) 6 9 Total patients the innominate artery. However, this may have been affected by the limitation of an axial view of the aortic arch, which will be discussed below. At each site where a statistically significant difference was found the axial measurements were greater than the orthogonal measurements (Table 1). The axial measurements were greater, on average, by 0.19 cm (95% CI: 0.10, 0.29). Excluding the aortic arch, the average difference was 0.24 cm (95% CI: 0.14, 0.33). The greatest differences were at the sinuses of Valsalva and the sino-tubular junction. The axial measurements were greater than the orthogonal measurements of lateral diameter by 0.36 cm (95% CI: 0.21, 0.50) and 0.74 cm (95% CI: 0.61, 0.87) at the sinuses of Valsalva and sino-tubular junction, respectively. When patients were placed into four classifications based on the maximal diameter of their thoracic aorta, more patients fell into the aneurysmal and surgically indicated classifications when based on axial measurements than orthogonal measurements (Table 2). Four fewer patients were categorized as normal and two less as ectatic based on axial versus orthogonal measurements. Three more patients were categorized as aneurysmal and three more as surgically indicated based on axial versus orthogonal measurements. This represents a total of six patients, or 13.3% of the sample population, which had a different aorta size classification be- Table 1 Paired Differences (Axial-Orthogonal in cm) for All Patients (N 45) Aorta Anatomical Site (diameter) 95% Confidence t-test Interval of the Difference Significance Mean Lower Upper (2-tailed) Annulus (ant-post) Annulus (lateral) Sinuses (ant-post) Sinuses (lateral) ST junction (ant-post) ST junction (lateral) Mid ascending (ant-post) Mid ascending (lateral) Arch (oblique; ant-post) Arch (oblique; lateral) Upper descending (ant-post) Upper descending (lateral) Lower descending (ant-post) Lower descending (lateral) Average Average (excluding arch)

4 Orthogonal Measurement of Thoracic Aorta 1483 Figure 2. Average paired differences between axial and orthogonal measurements for each classification group. A trend of increasing difference was observed but was not statistically significant. tween the axial and orthogonal measurement techniques. A general trend of increasing difference between axial and orthogonal measurements was observed as thoracic aorta diameter increased (Fig. 2). However, this trend was not statistically significant. Figure 4. Orthogonal section of the aorta at the level of the sinuses of Valsalva (AP diameter 4.87 mm, LAT diameter 4.89 mm). Note differences between the measurements obtained from the axial (Fig. 3) and orthogonal images. DISCUSSION In this study we used high-resolution CE-MRA with ECG-gating to measure the orthogonal dimensions of the thoracic aorta. The size of the thoracic aorta is a strong predictor of rupture, dissection, and mortality (1,18). Regular follow-up is therefore indicated for patients with enlarged aortas (1 3). To date, CT angiography has been routinely used to evaluate patients with thoracic aortic aneurysms (6 8). CT, particularly with the increasing use of multidetector CT, provides high spatial resolution and, when ECG-gating (19) is used, it can provide accurate orthogonal measurements in the thoracic aorta. However, CT involves significant radiation exposure (20 24), which can become cumulative as these patients are followed at 6 12 monthly intervals. The radiation dose is even higher when ECG-gating is used. CT angiography also uses iodinated contrast, which has a small, but real, risk of renal dysfunction (25). MRA is also used frequently to evaluate thoracic aortic disease (4,11) in many institutions. MRA has the advantage of eliminating exposure to frequent ionizing radiation (20 24) and potential nephrotoxic contrast agents. Conventional MRA has lower spatial resolution than CT. Furthermore, conventional MRA has long acquisition times, precluding the use of ECG-gating. This can result in significant image blurring in the proximal thoracic aorta due to cardiac motion artifact. Therefore, it Figure 3. A 61-year-old female with a fusiform ascending aortic aneurysm. ECG-gated high-resolution CE-MRA at the level of the sinuses of Valsalva. Axial image (anteroposterior [AP] diameter 52.5 mm, lateral [LAT] diameter 51.2 mm). Figure 5. A 69-year-old male preoperative for coronary artery bypass graft. ECG-gated high-resolution CE-MRA at the level of the sinuses of Valsalva. Axial image (AP diameter 29.3 mm, LAT diameter 35.2 mm).

5 1484 Bireley et al. Figure 6. Orthogonal section of the aorta at the level of the sinuses of Valsalva (AP diameter 27.3 mm, LAT diameter 28.6 mm). Note differences between the measurements obtained from the axial (Fig. 5) and orthogonal images. would be impossible to accurately measure orthogonal dimensions, particularly at the aortic root. In this study we imaged the thoracic aorta using CE-MRA with ECG-gating. With the development of parallel imaging techniques (12,13,26), it is now feasible to significantly shorten acquisition times for MRA. GRAPPA (27), with an acceleration factor of 2, and partial Fourier (28) was used in the phase-encoding direction to reduce the acquisition time per partition. With short scan times, it becomes possible to use ECG-gating to delay the acquisition period to diastole. This can result is substantial improvements in image quality, particularly at the aortic root (15,29), which is most affected by cardiac motion. With sharper images it becomes possible to more accurately measure orthogonal dimensions as proximal as the aortic annulus. Sharper images visually enhanced by contrast also allowed outer wall to outer wall measurements of axial and orthogonal diameters of the thoracic aorta. We compared orthogonal images obtained using ECG-gated MRA to axial images obtained using the TrueFISP pulse sequence. TrueFISP is frequently used to depict thoracic aortic morphology (16) and was therefore considered a suitable comparative technique. Each TrueFISP image was ECG-gated to diastole, similar to the MRA. The acquisition time per slice was also similar to the MRA. We found that the aorta diameter measurements acquired from axial TrueFISP images were significantly greater (P 0.05) than those acquired from images orthogonal to the course of the aorta at six of seven anatomic sites. At five of the seven sites the axial measurements were significantly greater (P 0.05) than the orthogonal measurements for both the anterior posterior and lateral diameter. At the annulus, only the lateral diameter was significantly greater (P 0.05). There was no significant difference found at the aortic arch immediately proximal to the innominate artery, but this may have been due to the limitation of the axial view of the aortic arch. An axial image of the aortic arch does not allow an anterior posterior and lateral measurement to be made, but only allows one oblique diameter measurement. One oblique measurement of the aortic arch from an axial image for each patient was compared to both the anterior posterior and lateral diameters obtained from an orthogonal image. Thus, the aortic arch orthogonal measurements were not compared to true corresponding axial measurements. Due to the geometry of the aorta and its position in the chest, it is not surprising to find that the axial measurements were greater than the orthogonal measurements. What was unknown, however, is the magnitude of size overestimation by axial measurements, whether the overestimation is statistically significant, and whether it is clinically significant by potentially altering patient management. Standard axial measurements may overestimate luminal diameter by up to 0.74 cm (95% CI: 0.61, 0.87 cm; at the sino-tubular junction) compared to orthogonal measurements. Our results are similar to those obtained by Hager et al (10), whose results suggested that measurements taken from axial images could overestimate the diameter of the ascending aorta by as much as 6 mm or 21%. Furthermore, for patients with larger thoracic aortic aneurysms ( 5.0 cm. n 6), axial measurements may overestimate the sino-tubular junction by as much as 0.93 cm (95% CI: 0.39, 1.48). Segments of the aorta that run more perpendicular to axial cross sections had smaller differences between axial and orthogonal measurements. For example, the paired differences for the annulus and lower descending thoracic aorta were 0.07 cm anterior posterior (95% CI: 0.06, 0.19), 0.14 cm lateral (95% CI: 0.02, 0.26); and 0.19 cm anterior posterior (CI: 0.14, 0.24), and 0.12 cm lateral (CI: 0.06, 0.18), respectively. The bulk of the aorta is not truly perpendicular with any axial cross section, which can explain why there is at least some difference, on average, at every measured site. We determined whether or not there could be clinical significance for the difference between the axial and orthogonal measurement techniques. We placed each patient s aorta twice into four simple clinical classifications based on the maximum diameter of their aorta, once based on axial measurements and once based on orthogonal measurements (Table 2). We based our classifications on what is commonly used in practice. More specifically, the aortic diameter threshold that can be an indication for surgery can range between 4.5 and 5.5 cm, depending on a patient s other risks (30). We selected 5.0 cm as an average threshold; in addition, patients with an aortic diameter larger than 5 cm have a significantly higher rate of aneurysm expansion (31). Moreover, diameter values in the literature obtained by CT and MRI are comparable because when patients are measured with both techniques the results correlate with each other (32,33). We did not have enough patients with CT angiography studies of the thoracic aorta to make a statistically meaningful comparison between orthogonal measurements obtained using ECG-gated CE-MRA and axial measurements obtained using CT angiography. In our patient population three additional patients aortas were classified as aneurysmal and three additional as surgically indicated based on axial versus orthogonal measurements. This represents a total of six patients, or 13.3% of the sample population, which could have been given more severe diagnosis and/or treatment if based on axial measurements rather than the more accurate orthogonal measurement technique. In conclusion, we have demonstrated that ECG-gated high-resolution CE-MRA can be used to measure or-

6 Orthogonal Measurement of Thoracic Aorta 1485 thogonal diameters in the thoracic aorta. In this study, similar to other studies with CT, we have shown that axial measurements are consistently larger than orthogonal measurements. We found that this could result in a potential change in clinical management in up to 13% of cases. REFERENCES 1. Davies RR, Goldstein LJ, Coady MA, et al. Yearly rupture or dissection rates for thoracic aortic aneurysms: simple prediction based on size. Ann Thorac Surg 2002;73:17 27; discussion Hirose Y, Hamada S, Takamiya M, et al. Aortic aneurysms: growth rates measured with CT. Radiology 1992;185: Pitt MP, Bonser RS. The natural history of thoracic aortic aneurysm disease: an overview. J Card Surg 1997;12: Hartnell GG. Imaging of aortic aneurysms and dissection: CT and MRI. J Thorac Imaging 2001;16: Tatli S, Yucel EK, Lipton MJ. CT and MR imaging of the thoracic aorta: current techniques and clinical applications. Radiol Clin North Am 2004;42: Chow LC, Rubin GD. CT angiography of the arterial system. Radiol Clin North Am 2002;40: Napoli A, Fleischmann D, Chan FP, et al. Computed tomography angiography: state-of-the-art imaging using multidetector-row technology. J Comput Assist Tomogr 2004;28(Suppl 1):S Rubin GD. CT angiography of the thoracic aorta. Semin Roentgenol 2003;38: Aronberg DJ, Glazer HS, Madsen K, et al. Normal thoracic aortic diameters by computed tomography. J Comput Assist Tomogr 1984;8: Hager A, Kaemmerer H, Rapp-Bernhardt U, et al. Diameters of the thoracic aorta throughout life as measured with helical computed tomography. J Thorac Cardiovasc Surg 2002;123: Prince MR, Yucel EK, Kaufman JA, et al. Dynamic gadoliniumenhanced three-dimensional abdominal MR arteriography. J Magn Reson Imaging 1993;3: Sodickson DK, Manning WJ. Simultaneous acquisition of spatial harmonics (SMASH): fast imaging with radiofrequency coil arrays. Magn Reson Med 1997;38: Sodickson DK, McKenzie CA, Ohliger MA, et al. Recent advances in image reconstruction, coil sensitivity calibration, and coil array design for SMASH and generalized parallel MRI. Magma 2002;13: Spinc le P, Cheng L, Ersoy H, et al. Motion artifact suppression in ECG ordered breath hold 3D contrast enhanced magnetic resonance angiography. In: Proc ISMRM 2005;3: Carr JC, Sandman DN, O Hara K, et al. ECG-triggered dynamic time-resolved magnetic resonance angiography of the thoracic aorta. In: Proc ISMRM 2005;13: Pereles FS, McCarthy RM, Baskaran V, et al. Thoracic aortic dissection and aneurysm: evaluation with nonenhanced true FISP MR angiography in less than 4 minutes. Radiology 2002;223: Carr JC, Simonetti O, Bundy J, et al. Cine MR angiography of the heart with segmented true fast imaging with steady-state precession. Radiology 2001;219: Elefteriades JA. Natural history of thoracic aortic aneurysms: indications for surgery, and surgical versus nonsurgical risks. Ann Thorac Surg 2002;74:S ; discussion S Roos JE, Willmann JK, Weishaupt D, et al. Thoracic aorta: motion artifact reduction with retrospective and prospective electrocardiography-assisted multi-detector row CT. Radiology 2002;222: Hidajat N, Maurer J, Schroder RJ, et al. Radiation exposure in spiral computed tomography. Dose distribution and dose reduction. Invest Radiol 1999;34: Huda W, Scalzetti EM, Roskopf M. Effective doses to patients undergoing thoracic computed tomography examinations. Med Phys 2000;27: Mayo JR, Jackson SA, Muller NL. High-resolution CT of the chest: radiation dose. Am J Roentgenol 1993;160: McGhee PL, Humphreys S. Radiation dose associated with spiral computed tomography. Can Assoc Radiol J 1994;45: McNitt-Gray MF. Radiation issues in computed tomography screening. Semin Roentgenol 2003;38: Weisbord SD, Palevsky PM. Radiocontrast-induced acute renal failure. J Intensive Care Med 2005;20: Weiger M, Pruessmann KP, Kassner A, et al. Contrast-enhanced 3D MRA using SENSE. J Magn Reson Imaging 2000;12: Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson Med 2002;47: Xu Y, Haacke EM. Partial Fourier imaging in multi-dimensions: a means to save a full factor of two in time. J Magn Reson Imaging 2001;14: Groves EM, Bireley W, Dill K, Carroll TJ, Carr JC. Quantitative analysis of ECG-gated high-resolution contrast-enhanced MR angiography of the thoracic aorta. Am J Roentgenol 2007;188: Zipes DP, Braunwald E. Braunwald s heart disease: a textbook of cardiovascular medicine. Philadelphia: Elsevier Saunders; Dapunt OE, Galla JD, Sadeghi AM, et al. The natural history of thoracic aortic aneurysms. J Thorac Cardiovasc Surg 1994;107: ; discussion Dinsmore RE, Liberthson RR, Wismer GL, et al. Magnetic resonance imaging of thoracic aortic aneurysms: comparison with other imaging methods. Am J Roentgenol 1986;146: Hager A, Kaemmerer H, Leppert A, et al. Follow-up of adults with coarctation of the aorta: comparison of helical CT and MRI, and impact on assessing diameter changes. Chest 2004;126: