In Vitro Evaluation of Current Thoracic Aortic Stent- Grafts for Real-time MR-Guided Placement

62 J ENDOVASC THER EXPERIMENTAL INVESTIGATION In Vitro Evaluation of Current Thoracic Aortic Stent- Grafts for Real-time MR-Guided Placement Holger Eggebrecht, MD 1,2 ; Michael Zenge 2 ; Mark E. Ladd, PhD 2 ; Raimund Erbel, MD 1 ; and Harald H. Quick, PhD 2 Departments of 1 Cardiology and 2 Diagnostic and Interventional Radiology and Neuroradiology, West German Heart Center Essen, University of Duisburg-Essen, Germany. Purpose: To systematically evaluate the magnetic resonance imaging (MRI) characteristics of current thoracic aortic stent-graft devices before, during, and after in vitro deployment as a step toward real-time MRI-guided stent placement. Methods: Six stent-graft devices used for thoracic aortic repair were examined in a dedicated phantom model using a 1.5-T MRI scanner. First, the delivery systems with the mounted stent-graft were examined using real-time fast imaging with steady-state precession (TrueFISP) with Cartesian and radial k-space filling. TrueFISP imaging was subsequently used for real-time monitoring of stent-graft expansion. The deployed stent-grafts were then examined in a water bath containing gadolinium (1:40) with high-resolution T 1 - weighted 3D fast low-angle shot (FLASH) sequences. The images were analyzed for artifacts, radiofrequency caging effects, and device visualization quality. Results: Three delivery systems with mounted stent-grafts did not contain ferromagnetic elements and were well visualized. Imaging with radial k-space filling showed fewer artifacts than Cartesian imaging. Movement of the delivery system and stent-graft expansion of these devices were successfully demonstrated at a rate of up to 6 frames per second. Evaluation of the expanded stent-grafts revealed only minor susceptibility artifacts without relevant signal attenuation in the stent-graft lumen for 5 nitinol-based stent-grafts. Only a stainless steel based stent-graft was associated with severe artifacts, thwarting visualization of its lumen or surroundings. Conclusion: The present study shows that 3 nitinol-based thoracic stent-graft devices are potentially suited for real-time MRI-guided placement with respect to both the delivery system and the stent-graft itself. These observations provide the basis for the evaluation of MRI-guided stent-graft placement in vivo. J Endovasc Ther Key words: magnetic resonance imaging, real-time MRI, MR guidance, thoracic aorta, stent-graft, susceptibility artifact Endovascular stent-graft placement is emerging as a promising alternative to medical therapy and surgery in the treatment of patients with diseases of the descending thoracic aorta. 1,2 Stent-graft implantation is usually performed under fluoroscopy, which has several shortcomings in this setting beyond the ionizing radiation and nephrotoxic contrast material. For example, fluoroscopy does not allow consistent differentiation of the false from Dr. Eggebrecht is recipient of a research grant from the University Duisburg-Essen (IFORES 102). The authors have no commercial, proprietary, or financial interest in any products or companies described in this article. Address for correspondence and reprints: Dr. Holger Eggebrecht, Department of Cardiology, University of Duisburg- Essen, Hufelandstraße 55, 45122 Essen, Germany. Fax: 49-201-723-5480; E-mail: holger.eggebrecht@uni-essen.de 2006 by the INTERNATIONAL SOCIETY OF ENDOVASCULAR SPECIALISTS Available at www.jevt.org

J ENDOVASC THER MR EVALUATION OF THORACIC STENT-GRAFTS 63 Figure 1The stent-grafts examined in this study: (A) Relay, (B) Zenith, (C) TAG, (D) Evita, (E) Talent, and (F) Valiant. the true lumen or adequate visualization of the false lumen. Immediate evaluation of procedural success after stent-graft placement (i.e., thrombosis of the false lumen) is thus not possible. These limitations may be overcome by magnetic resonance imaging (MRI) based procedural guidance. MRI guidance of vascular interventional procedures offers several potential advantages over fluoroscopically-guided techniques, including image acquisition in any desired orientation, superior 3-dimensional (3D) soft tissue contrast with simultaneous visualization of the interventional device, absence of ionizing radiation, and avoidance of nephrotoxic contrast media. 3,4 The feasibility of MRIguided vascular interventions has been demonstrated for a variety of procedures, including peripheral and coronary stent placement, 5,6 selective embolization, 7 and implantation of atrial septal closure devices. 8 MRI guidance may be particularly useful for aortic stent-graft placement because it can provide all relevant information for the preinterventional planning of such procedures and can also be used for postinterventional evaluation of treatment success. 9,10 Several studies have suggested that MRI and MR angiography (MRA) techniques may even be superior to standard techniques (i.e., computed tomography [CT]) for detecting endoleaks during follow-up, as well as for evaluation of aneurysm sac exclusion and graft patency. 11,12 An important prerequisite for performing stent-graft placement under MR guidance is MR compatibility of the delivery systems and endografts with respect to device visibility, image artifacts, and postdeployment visualization of the expanded stent-grafts and their lumens. 3,13 In the present study, we evaluated the MRI characteristics of 6 current thoracic aortic stentgrafts in phantom model experiments. METHODS Stent-Graft Devices Five commercially available thoracic stentgrafts (Relay [Bolton Medical, Sunrise, FL, USA], Zenith [Cook, Bloomington, IN, USA], TAG (W.L. Gore & Associates, Flagstaff, AZ, USA], Evita [Jotec, Hechingen, Germany], and Talent [Medtronic Vascular, Santa Rosa, CA, USA]), as well as a prototype device (Valiant; Medtronic Vascular), were examined (Fig. 1). The individual device specifications are given in Table 1. In Vitro Phantom Setup Scanning was performed on a 1.5-T wholebody MRI scanner (Magnetom Sonata, Siemens Medical Solutions, Erlangen, Germany)

64 MR EVALUATION OF THORACIC STENT-GRAFTS J ENDOVASC THER TABLE 1 Characteristics of the Stent-Grafts Assessed Metal Support Markers Graft Fabric Relay Nitinol Platinum Dacron Zenith Stainless steel Gold Dacron TAG Nitinol Gold eptfe Evita Nitinol Nitinol Polyester Talent Nitinol Nitinol Polyester Valiant Nitinol Nitinol Polyester eptfe: expanded polytetrafluoroethylene. equipped with gradients capable of a maximum amplitude of 40 mt/m and a slew rate of 200 T/m s 1. In an initial experiment, each stent-graft was mounted on its delivery system and placed in a water bath phantom consisting of a 1-m-long, 10-cm-wide plastic tube with a central hole for linear catheter manipulation over a fishing cord. The phantom (free from any silicone or glass tubes) was designed to allow assessment of the delivery device and stent-graft artifact properties during manual advancement and subsequent stent-graft deployment. The phantom was placed with its longitudinal axis along the axis of the main magnetic field on top of the spine phased-array radiofrequency (RF) coil with 3 coil elements activated for signal reception (Fig. 2). A 2-element body flex phased-array RF coil was placed on top of the phantom. Real-time Imaging of the Delivery Systems Figure 2Experimental setup. The stent-grafts with their delivery systems were contained in a Plexiglas phantom, which consisted of a 1-m-long, 10-cm-diameter water-filled tube that was placed with its longitudinal axis in parallel to the main magnetic field of the scanner on top of the spine phased-array RF coil. A body phased-array RF coil was placed on top of the phantom. The delivery systems were examined with real-time fast imaging with steady-state precession (TrueFISP) sequences using Cartesian and radial k-space filling (Table 2). For the Cartesian sequence, the frequency encoding direction was chosen parallel to the longitudinal direction of the devices and of the phantom; thus, it was also parallel to the direction of the main magnetic field. First, static images of each delivery system with the mounted stent-graft were obtained. Then, the delivery systems were advanced at 2 cm/s as realtime TrueFISP imaging was performed at 1.5 frames per second (fps) for the Cartesian and 6 fps for the radial TrueFISP sequence. For both sequences, images were displayed without detectable image reconstruction delay. Subsequently, stent-graft expansion and deployment were recorded using radial True- FISP sequences. High-Resolution Imaging of the Expanded Stent-Grafts In a second step, the expanded stent-grafts without their delivery systems were examined hanging freely in a plastic container with water and gadolinium (Gd-DTPA) at a concentration of 1:40. This concentration was chosen based on the results of prior in vitro solution series, which approximated the intra-arterial signal intensity of contrast-enhanced 3D MRA of the aorta in patients. Imaging of the stentgrafts was performed using high-resolution

J ENDOVASC THER MR EVALUATION OF THORACIC STENT-GRAFTS 65 TABLE 2 Characteristics of the MRI Sequences Evaluated Cartesian TrueFISP Radial TrueFISP T 1 -weighted 3D FLASH Repetition time, ms 3.6 3.6 4.5 Echo time, ms 1.8 1.8 1.76 Flip angle, 70 70 30 Matrix, pixels 192192 192192 51220440 Field of view, mm 2 400400 400400 250100 Slice thickness, mm 12 12 2 Image acquisition time, s 0.69 0.17 90 (40 slices) Bandwidth, Hz/pixel 1530 610 470 TrueFISP: real-time steady-state free precession sequences, FLASH: fast low-angle shot. T 1 -weighted 3D fast low-angle shot (FLASH) imaging (Table 2). Minimum intensity projections (MIP) were derived from the 3D FLASH datasets. Image Data Analysis Image quality was evaluated with respect to (1) static visualization of the delivery system, (2) visualization of delivery systems under motion, (3) monitoring of stent-graft expansion, and (4) visualization of the deployed stent-graft. Image quality assessment was performed by a radiologist and a physicist (H.E. and H.H.Q., respectively) in consensus using a 3-point grading scale: 1 indicated severe artifacts with poor visualization of delivery system/stent-graft, 2 denoted minor artifacts allowing fair visualization of delivery system/stent-graft, and 3 referred to excellent visualization of the delivery system/stent-graft with delineation of device details (e.g., stentgraft markers). Assessment of potential signal shielding from the inside of the stent due to RF caging was expressed as percentage signal loss within the stent-graft lumen on the T 1 -weighted 3D FLASH scan with respect to the undisturbed signal in the water/gd-dtpa solution outside of the stent-graft. For this purpose, signal intensity was measured at regions of interest (ROI) within the stent lumen, as well as outside of the stent-graft in coronal and axial orientations. Figure 3Visualization of the TAG delivery system with the mounted stent-graft in static position with real-time TrueFISP with (A) Cartesian and (B) radial k-space filling. (C E) Visualization of catheter movement (2 cm/s) using real-time TrueFISP (5 fps) with radial k-space filling. RESULTS Real-time Imaging of the Delivery Systems Two delivery systems each with their mounted stent-grafts (TAG and Valiant) did not contain ferromagnetic elements and were well visualized by real-time MRI (Figs. 3 5). The Evita delivery system was suitable for MRI examination only after device modification (exchange of a ferromagnetic screw at the catheter tip), whereas the remaining delivery devices (Relay, Talent, and Zenith) caused large susceptibility artifacts and were thus not suited for MRI examination. Back and forth movement of the 3 suitable delivery systems

66 MR EVALUATION OF THORACIC STENT-GRAFTS J ENDOVASC THER Figure 4Visualization of the modified Evita delivery system with the mounted stent-graft in static position with real-time TrueFISP with (A) Cartesian and (B) radial k-space filling. (C-E) Visualization of catheter movement (2 cm/s) using real-time TrueFISP (5 fps) with radial k-space filling. in the phantom model was successfully monitored with real-time TrueFISP imaging (Figs. 3 5). The susceptibility artifacts of the stentgrafts enabled clear determination of the position of the loaded stent-graft in relation to the delivery system. Subsequent expansion of the stent-grafts was also successfully visualized under real-time conditions (radial sequence only; Figs. 6 8). Figure 5Visualization of the Valiant delivery system with the mounted stent-graft in static position with real-time TrueFISP with (A) Cartesian and (B) radial k-space filling. (C-E) Visualization of catheter movement (2 cm/s) using real-time TrueFISP (5 fps) with radial k-space filling. All 5 expanded nitinol-based stent-grafts were well visualized by MRI and did not produce disturbing susceptibility artifacts. Selected slices of the high-resolution T 1 -weighted 3D FLASH datasets demonstrated the individual amount of RF signal shielding among the different stent-grafts (Fig. 9). There was no relevant signal attenuation within the lumen of the 5 nitinol-based stent-grafts (Fig. 9). Three stent-grafts (Relay, TAG, and Talent) showed slight regional signal enhancement due to RF resonance effects at longitudinal stent struts (Fig. 9). Quantitative assessment of RF signal caging showed that the percentage attenuation was around 0%, i.e., basically no signal attenuation was observed for the nitinol-based stent-grafts (Table 3). Small variations in the attenuation score might be due to a different location of the measurement ROIs relative to the position of the surface coils (inherent spatial signal variations). Only High-Resolution Imaging of the Expanded Stent-Grafts Figure 6Visualization of the TAG stent-graft deployment using real-time TrueFISP (5 fps) with radial k-space filling (A). The graft is released (B D) in a fraction of a second by retracting a restraining string that prevents the graft from unfolding.

J ENDOVASC THER MR EVALUATION OF THORACIC STENT-GRAFTS 67 Figure 7Visualization of the Evita stent-graft deployment using real-time TrueFISP (5 fps) with radial k-space filling (A). The graft is deployed (B D) by retracting the sheath of the delivery system. the stainless steel based Zenith stent-graft caused severe susceptibility artifacts, which did not allow visualization of the stent-graft lumen or its adjacent area. The MIPs derived from the high-resolution T 1 -weighted 3D FLASH scans demonstrated that artifacts from nitinol-based stent-grafts most likely do not compromise the depiction of the tissue directly surrounding the stent-graft (Fig. 10). Again, the stainless steel based stent-graft produced large artifacts, which hindered visualization of the stent-graft lumen and the surrounding tissue. Image Data Analysis Figure 8Visualization of the Valiant stent-graft deployment using real-time TrueFISP (5 fps) with radial k-space filling (A). The graft is slowly released (B D) by retracting a catheter sheath. The position of the retracting catheter sheath is clearly defined by the partly deployed stent-graft (arrows). Comparison of image quality revealed that static imaging with radial k-space filling generally displayed the delivery systems with less pronounced artifacts than imaging using Cartesian k-space filling (Table 4). With radial k-space filling, the artifacts were constrained to the actual device. Images acquired with the radial sequence were thus rated better than images acquired with the Cartesian acquisition scheme (Table 4). Imaging of the delivery systems under motion was associated with less image quality than static imaging independent of radial or Cartesian k-space filling; however, visualization of the position of the loaded stent-graft in relation to the delivery system was still sufficient for device positioning. Here, the higher temporal resolution of the radial sequence (6 fps) compared to the Cartesian sequence (1.5 fps) provided much better depiction of the moving devices, which again resulted in a higher image quality rating for the radial sequence. Owing to the rapid expansion mechanism of the TAG device, its deployment was less well monitored than for the Relay and Valiant stent-grafts. (Figs. 6 8) DISCUSSION The present study evaluated the applicability of 6 current thoracic aortic stent-graft devices for real-time MRI-guided placement. Three nitinol-based stent-grafts in combination with their delivery systems (TAG, Valiant, and with restrictions Evita) were found to be suitable for MRI-based procedural guidance using passive device tracking. Movement and subsequent stent-graft expansion of these devices were well visualized using real-time TrueFISP im-

68 MR EVALUATION OF THORACIC STENT-GRAFTS J ENDOVASC THER Figure 9Single coronal slices and corresponding axial slices from the high-resolution T 1 - weighted 3D FLASH sequences: (A) Relay, (B) Zenith, (C) TAG, (D) Evita, (E) Talent, and (F) Valiant. The stainless steel based Zenith stent-graft shows severe susceptibility artifacts, rendering visualization of the stent lumen and stent vicinity impossible. All the other 5 nitinolbased stent-grafts allow detailed evaluation of the stent lumen. Arrows point to slight regional signal enhancement due to RF resonance effects. TABLE 3 RF Signal Attenuation Caused by the Stent-Grafts Lumen Relay 2%0% Zenith NA TAG 5%0% Evita 3%0% Talent 1%0% Valiant 2%3% NA: RF signal attenuations could not be quantified due to severe artifact. aging with a temporal resolution of 6 fps. Moreover, our study demonstrated that expanded nitinol-based stent-grafts were not associated with relevant susceptibility artifacts or relevant RF signal loss within the stent-graft lumen. The only stainless steel based stentgraft was unsuitable for MRI visualization. Collection of precise 3D information about soft tissue anatomy and reliable visualization of catheter instruments in relation to the surrounding anatomy represent the prerequisite for safe and successful performance of vascular interventions under MRI guidance. 14 However, in contrast to ultrasound, fluoroscopy, or CT, visualization of interventional instruments by MRI has proven to be difficult. 15 The optimal MRI technique to visualize intravascular instruments would be characterized by high spatial and temporal resolution and should also provide a high-contrast instrument signature, making it easy to pick out the instrument in the MR image. 16 A number of approaches have been developed for depicting vascular instruments by MRI. Active device tracking is accomplished by incorporating small receiver coils into the tip of a catheter device. The use of active catheter tracking techniques may, however, be potentially hazardous for the patient as long as no safety-tested and commercially available active devices are available. In contrast, passive device tracking by MRI uses susceptibility artifacts of the interventional instrument

J ENDOVASC THER MR EVALUATION OF THORACIC STENT-GRAFTS 69 Figure 10Minimum intensity projections from the high-resolution T 1 -weighted 3D FLASH sequences: (A) Relay, (B) Zenith, (C) TAG, (D) Evita, (E) Talent, and (F) Valiant. Except for (B), all stent-grafts allow detailed evaluation of the lumen and surroundings with the high-resolution 3D FLASH sequence. for visualization. Passive tracking requires no hardware or instrument modifications and thus appears to be particularly promising in terms of potential clinical applications. However, passive visualization requires MR-compatible instruments with adequate susceptibility artifacts. The artifact has to allow clear visualization of the interventional device within the aortic lumen, but should not obscure relevant parts of the field of view. 13 Therefore, the artifact ideally should be as small as possible and confined to the device geometry while providing high image contrast against the background. 17 So far, few studies have addressed the suitability of aortic stent-grafts for MRI applica- TABLE 4 Image Quality Assessment With 3-Point Grading* Visualization of Delivery System Cartesian Radial Monitoring of Delivery System Movement Monitoring of Stent-Graft Expansion Visualization of the Expanded Stent-Graft Relay NA NA NA NA 3 Zenith NA NA NA NA NA TAG 2 3 2 3 3 Evita 2 3 2 3 3 Talent NA NA NA NA 3 Valiant 3 3 3 3 3 NA: examination of the delivery system/stent-graft not possible due to ferromagnetic components. * 1: severe artifacts with poor visualization of delivery system/stent-graft, 2: minor artifacts allowing fair visualization of delivery system/stent-graft, 3: excellent visualization of the delivery system/stent-graft with delineation of device details (e.g., stent-graft markers). Evaluation after device modification compared to the standard device.

70 MR EVALUATION OF THORACIC STENT-GRAFTS J ENDOVASC THER tions. 3,13,18 Most studies focused on the evaluation of expanded stent-grafts in terms of MRI-based surveillance. 13,18 Only a single study so far has addressed the MRI suitability of both stent-grafts and delivery systems, which is the prerequisite for MRI-guided stent placement. Using various real-time gradientecho MRI sequences of static devices, Mahnken et al. 3 examined bifurcated stent-graft systems designed for abdominal aortic (AA) repair. In contrast, the present study evaluated the MRI visibility of the stent-graft systems and endograft deployment process with both static and real-time imaging under device motion. Furthermore, the present study examined tubular stent-grafts designed for thoracic aortic repair. These differ significantly from their AA counterparts with respect to both the endoprostheses and delivery systems. Similar to the results of Mahnken et al., 3 both the nitinolbased TAG stent-graft and the Valiant prototype (in combination with their delivery systems) were found to be potentially suitable for MRI guidance using passive catheter visualization. Movement of these delivery systems and subsequent stent-graft expansion was well visualized using real-time TrueFISP imaging. Again, in concordance with Mahnken s results, real-time imaging with radial k-space filling showed better image quality than with Cartesian k-space filling. The Evita stent-graft was of limited suitability only after modification of the delivery system, whereas the remaining 3 devices tested (Relay, Zenith, and Talent) contained ferromagnetic parts and were thus not suitable for MRI visualization. To achieve high instrument to background contrast in the dynamic phantom experiments, a fast steady-state free precession (SSFP or TrueFISP) sequence was employed for data acquisition. 19 Such steady-state sequences, when used with high excitation flip angles, provide high blood signal even without administration of a contrast agent. To ensure high temporal resolution for visualization of the dynamic process of catheter movement and subsequent stent-graft deployment, TrueFISP imaging with radial k- space filling provided 6 images per second, which proved sufficient for direct control of catheter movement and monitoring of stentgraft deployment. With respect to MRI surveillance after stentgraft placement, this study demonstrates that the tested nitinol-based stent-grafts did not produce relevant susceptibility artifacts, which allowed visualization of both the aortic lumen and the tissue surrounding the aorta. Quantitative assessment of RF caging basically showed no signal attenuation for the nitinolbased stent-grafts. Postinterventional MRA assessment and follow-up of the implanted stent-grafts with 3D contrast-enhanced FLASH sequences thus will not be impaired by stentgraft related artifacts. Only the stainless steel based Zenith stent-graft produced severe image artifacts, rendering MRI-based follow-up of this stent-graft useless. 18 It is of even greater concern that these artifacts also significantly limit the diagnostic value of future MRI scans performed for other diagnostic purposes (e.g., MR evaluation of cardiac function). Moreover, this stainless steel based stent-graft exerted significant ferromagnetic attraction force, which potentially may impose a risk of stentgraft migration while inside or in the vicinity of the MR scanner. Another potential safety hazard associated with MRI examinations is the possibility of localized increases in the radiofrequency (RF) Specific Absorption Rate (SAR) near metallic implants. The local electric field can be amplified, especially if the implants are composed of long conducting structures that potentially can couple significantly with the RF energy of the body coil. This might lead to excessive local tissue heating if standing resonating waves are generated. 20 22 These resonance effects, however, occur only when the length of wire-like structures or longitudinal implants exceeds one half of the RF wavelength, i.e., around 26 cm for 1.5 T in human tissue. 20 For implants that are small compared to the RF wavelength (e.g., stents, coils, clips, endografts, etc.), RF resonance is considered unlikely. All endografts used in this study were 20 cm in length; consequently, no significant heating of these devices is expected. Accordingly, we have not carried out any experiments with regard to RF heating of tissue surrounding the endografts. Conclusion The present study shows that 3 commercially available nitinol-based stent-grafts in

J ENDOVASC THER MR EVALUATION OF THORACIC STENT-GRAFTS 71 combination with their delivery systems (TAG, Valiant, and with restrictions Evita) are readily suitable for real-time MRI-based procedural guidance using passive device visualization. These observations provide the basis for the evaluation of MRI-guided aortic stent-graft placement in vivo. Acknowledgments: The authors thank Dr. Peter Speier, PhD, of Siemens Medical Solutions, Erlangen, Germany, for providing support with the interactive real-time Works-in-Progress imaging protocols. REFERENCES 1. Fattori R, Napoli G, Lovato L, et al. Descending thoracic aortic diseases: stent-graft repair. Radiology. 2003;229:176 183. 2. Eggebrecht H, Herold U, Kuhnt O, et al. Endovascular stent-graft treatment of aortic dissection: determinants of post-interventional outcome. Eur Heart J. 2005;26:489 497. 3. Mahnken AH, Chalabi K, Jalali F, et al. Magnetic resonance-guided placement of aortic stents grafts: feasibility with real-time magnetic resonance fluoroscopy. J Vasc Interv Radiol. 2004; 15:189 195. 4. Manke C, Nitz WR, Lenhart M, et al. Magnetic resonance monitoring of stent deployment: in vitro evaluation of different stent designs and stent delivery systems. Invest Radiol. 2000;35: 343 351. 5. Spuentrup E, Ruebben A, Schaeffter T, et al. Magnetic resonance guided coronary artery stent placement in a swine model. Circulation. 2002;105:874 879. 6. Buecker A, Adam GB, Neuerburg JM, et al. Simultaneous real-time visualization of the catheter tip and vascular anatomy for MR-guided PTA of iliac arteries in an animal model. J Magn Reson Imaging. 2002;16:201 208. 7. Fink C, Bock M, Umathum R, et al. Renal embolization: feasibility of magnetic resonanceguidance using active catheter tracking and intraarterial magnetic resonance angiography. Invest Radiol. 2004;39:111 119. 8. Buecker A, Spuentrup E, Grabitz R, et al. Magnetic resonance-guided placement of atrial septal closure device in animal model of patent foramen ovale. Circulation. 2002;106:511 515. 9. Engellau L, Albrechtsson U, Dahlstrom N, et al. Measurements before endovascular repair of abdominal aortic aneurysms. MR imaging with MRA vs. angiography and CT. Acta Radiol. 2003;44:177 184. 10. Thurnher SA, Dorffner R, Thurnher MM, et al. Evaluation of abdominal aortic aneurysm for stent-graft placement: comparison of gadolinium-enhanced MR angiography versus helical CT angiography and digital subtraction angiography. Radiology. 1997;205:341 352. 11. Ayuso JR, de Caralt TM, Pages M, et al. MRA is useful as a follow-up technique after endovascular repair of aortic aneurysms with nitinol endoprostheses. J Magn Reson Imaging. 2004; 20:803 810. 12. Ersoy H, Jacobs P, Kent CK, et al. Blood pool MR angiography of aortic stent-graft endoleak. AJR Am J Roentgenol. 2004;182:1181 1186. 13. Hilfiker PR, Quick HH, Debatin JF. Plain and covered stent-grafts: in vitro evaluation of characteristics at three-dimensional MR angiography. Radiology. 1999;211:693 697. 14. Ladd ME, Quick HH, Debatin JF. Interventional MRA and intravascular imaging. J Magn Reson Imaging. 2000;12:534 546. 15. Quick HH, Kuehl H, Kaiser G, et al. Interventional MRA using actively visualized catheters, TrueFISP, and real-time image fusion. Magn Reson Med. 2003;49:129 137. 16. Quick HH, Zenge MO, Kuehl H, et al. Interventional magnetic resonance angiography with no strings attached: wireless active catheter visualization. Magn Reson Med. 2005;53:446 455. 17. Hilfiker PR, Quick HH, Schmidt M, et al. In vitro image characteristics of an abdominal aortic stent graft: CTA versus 3D MRA. MAGMA. 1999;8:27 32. 18. van der Laan MJ, Bartels LW, Bakker CJ, et al. Suitability of 7 aortic stent-graft models for MRI-based surveillance. J Endovasc Ther. 2004;11:366 371. 19. Hunold P, Maderwald S, Eggebrecht H, et al. Steady-state free precession sequences in myocardial first-pass perfusion MR imaging: comparison with TurboFLASH imaging. Eur Radiol. 2004;14:409 416. 20. Konings MK, Bartels LW, Smits HF, et al. Heating around intravascular guidewires by resonating RF waves. J Magn Reson Imaging. 2000; 12:79 85. 21. Ladd ME, Quick HH. Reduction of resonant RF heating in intravascular catheters using coaxial chokes. Magn Reson Med. 2000;43:615 619. 22. Nitz WR, Oppelt A, Renz W, et al. On the heating of linear conductive structures as guide wires and catheters in interventional MRI. J Magn Reson Imaging. 2001;13:105 114.