Real-Time Magnetic Resonance-Guided Endovascular Repair of Experimental Abdominal Aortic Aneurysm in Swine P R E A V A B R R B E e r l t r e d i b s d i c D B S 0 w a Journal of the American College of Cardiology Vol. 45, No. 12, 2005 © 2005 by the American College of Cardiology Foundation ISSN 0735-1097/05/$30.00 P RECLINICAL RESEARCH eal-Time Magnetic Resonance-Guided ndovascular Repair of Experimental bdominal Aortic Aneurysm in Swine enkatesh K. Raman, MD,* Parag V. Karmarkar, MSC,*‡ Michael A. Guttman, MSC,† lexander J. Dick, MD,* Dana C. Peters, PHD,† Cengizhan Ozturk, MD, PHD,* reno S. S. Pessanha, MD,* Richard B. Thompson, PHD,† Amish N. Raval, MD,* anil DeSilva, MBBS, PHD,* Ronnier J. Aviles, MD,* Ergin Atalar, PHD,‡ Elliot R. McVeigh, PHD,† obert J. Lederman, MD* ethesda and Baltimore, Maryland OBJECTIVES This study tested the hypotheses that endografts can be visualized and navigated in vivo solely under real-time magnetic resonance imaging (rtMRI) guidance to repair experimental abdominal aortic aneurysms (AAA) in swine, and that MRI can provide immediate assessment of endograft apposition and aneurysm exclusion. BACKGROUND Endovascular repair for AAA is limited by endoleak caused by inflow or outflow malappo- sition. The ability of rtMRI to image soft tissue and flow may improve on X-ray guidance of this procedure. METHODS Infrarenal AAA was created in swine by balloon overstretch. We used one passive commercial endograft, imaged based on metal-induced MRI artifacts, and several types of homemade active endografts, incorporating MRI receiver coils (antennae). Custom interactive rtMRI features included color coding the catheter-antenna signals individually, simultaneous multislice imaging, and real-time three-dimensional rendering. RESULTS Eleven repairs were performed solely using rtMRI, simultaneously depicting the device and soft-tissue pathology during endograft deployment. Active devices proved most useful. Intrapro- cedural MRI provided anatomic confirmation of stent strut apposition and functional corrobo- ration of aneurysm exclusion and restoration of laminar flow in successful cases. In two cases, there was clear evidence of contrast accumulation in the aneurysm sac, denoting endoleak. CONCLUSIONS Endovascular AAA repair is feasible under rtMRI guidance. Active endografts facilitate device visualization and complement the soft tissue contrast afforded by MRI for precise positioning and deployment. Magnetic resonance imaging also permits immediate post- procedural anatomic and functional evaluation of successful aneurysm exclusion. (J Am Coll ublished by Elsevier Inc. doi:10.1016/j.jacc.2005.03.029 Cardiol 2005;45:2069 –77) © 2005 by the American College of Cardiology Foundation t s M p c p n e f d r p g c e M A ndovascular repair is an alternative to open surgery that is merging as an elective treatment for abdominal aortic aneu- ysm (AAA) (1–3). An important complication of endovascu- ar AAA repair is endoleak, a persistent systemic communica- ion with the aneurysm sac that risks continued expansion and upture. Of the types described by White et al. (4), type I ndoleak results from an incomplete seal at the proximal or istal attachment site of the endograft. This has been reported n up to 25% of cases (5– 8) in older series, and may be caused y device misposition, stent malapposition, and device under- izing (9) or oversizing (10). Better soft tissue visualization and epiction of complex three-dimensional (3D) anatomy by nteractive magnetic resonance imaging (MRI) may limit this omplication. From the *Cardiovascular Branch and the †Laboratory of Cardiac Energetics, ivision of Intramural Research, National Heart, Lung, and Blood Institute, ethesda, Maryland; and the ‡Department of Radiology, Johns Hopkins University chool of Medicine, Baltimore, Maryland. Supported by NIH Z01-HL005062- 1CVB (to Dr. Lederman). Drs. Raman and Karmarkar contributed equally to this ork. c Manuscript received May 24, 2004; revised manuscript received February 20, 2005, ccepted March 1, 2005. Magnetic resonance imaging may be equivalent or superior o X-ray computed tomography for procedure planning and urveillance after endograft placement (11–13). Real-time RI (rtMRI) can guide clinical invasive procedures (14) and re-clinical interventions, such as transcatheter repair of intra- ardiac shunt (15,16) and endomyocardial injection of thera- eutic agents (17). An MRI permits 3D tissue and hemody- amic characterization, creating opportunities to improve ndovascular aneurysm treatment and to limit procedural ailure from type I endoleak. We hypothesize that: 1) en- ografts can be visualized and navigated in vivo solely under tMRI to repair experimental AAA in swine; 2) MRI can rovide useful intraprocedural information about anatomy and uide device placement; and 3) MRI can provide immediate onfirmation of successful endograft apposition and aneurysm xclusion. ETHODS nimal preparation and aneurysm model. Animal proto- ols were approved by the National Heart, Lung, and Blood I s v t P A t ( p a 4 V u o 2 s 2 i a r r C e p r ( e r i a i l S i t t f d a 8 r Y n P d d T m o e X l t r ( fi l t ( s E a e a a i s [ fi Q i F F e t d s ( 2070 Raman et al. JACC Vol. 45, No. 12, 2005 Real-Time MRI AAA Repair June 21, 2005:2069 –77 nstitute Animal Care and Use Committee. Eleven York- hire swine (Animal Biotech Industries, Danboro, Pennsyl- ania) or National Institutes of Health mini-swine (Na- ional Institutes of Health Veterinary Resource Program, oolesville, Maryland) weighing 60 to 85 kg were studied. nesthesia was induced with ketamine/xylazine and main- ained with inhaled isoflurane. Nonferrous 12-F sheaths Check Flo II, Cook, Bloomington, Indiana) were placed ercutaneously in the femoral artery. Animals underwent nticoagulation with a heparin 100-IU/kg bolus dose and a 0- to 60-IU/kg/h infusion. We modified an acute nonsurgical model of AAA (18). essels were sized by X-ray digital subtraction aortography sing a marker pigtail catheter. Single, double, or triple verlapping balloons (XXL, 14- to 18-mm diameter � 0-mm length, Boston Scientific/Medi-Tech, Natick, Mas- achusetts; and AgilTrac, 10- to 14-mm diameter � 0-mm length, Guidant, Menlo Park, California) were nflated for up to 10 min within the infrarenal aorta to chieve an overstretch ratio of at a least 2:1. This was epeated until the dilated segment was at least 1.5 times the eference diameter or until dissection or rupture. onstruction of endograft devices. Different self- xpanding endograft designs were tested, including one assive device and three different active designs. Passive efers to device visibility based on susceptibility artifacts dark spots on MRI) generated by intrinsic magnetic prop- rties of the device. Active refers to incorporation of an MRI eceiver coil (antenna, electrically connected to the scanner) nto the catheter, which is sensitive to signal only from djacent tissue and is used to create bright spots on the MR mages. The passive device was an unmodified 10 � 100 mm iliac imb of a commercial nitinol endograft (Vanguard, Boston cientific, Natick, Massachusetts). A quarter-wave 0.030- nch active guidewire was used to enhance visualization of he passive endograft. All active devices were built by one of he investigators (P.V.K.). Endografts were constructed rom 0.009-inch gold-plated nitinol wire in a Z-stent esign, lined with thin expanded polytetrafluoroethylene nd having diameters of 10 to 14 mm and lengths of 60 to 0 mm. Endografts were mounted on 5-F 65-cm nonfer- ous Kumpe catheters (Angiodynamics, Queensbury, New ork), acting as delivery shafts, and crimped inside 10-F Abbreviations and Acronyms 3D � three dimensional AAA � abdominal aortic aneurysm MRA � magnetic resonance angiography MRI � magnetic resonance imaging rFOV � reduced field of view SSFP � steady state free precession MRI TE � echo time TR � repetition time ylon sheaths (Fast-Cath, Daig, Minnetonka, Minnesota). t olyimide tubing improved the stiffness of the Kumpe elivery shaft and the ease of sheath retraction for endograft eployment (Fig. 1). Schematics for the active devices are shown in Figure 2. he passive endograft is depicted in Figure 2A. An active- arker/passive-stent design (Fig. 2B) incorporated two pposed solenoid coils built into the delivery shaft at each dge of the stent, analogous to radiopaque end markers on -ray systems. Another active-stent design incorporated a oopless antenna. The delivery shaft acted as ground, and he Z-stent served as antenna whip (region of antenna eceiving signal), connected to the shaft by a wire (19 –21) Fig. 2C) and designed to detach after deployment. The nal active-stent/active-marker system combined the loop- ess coil endograft with a separate loop coil wound around he delivery device’s distal cone just beyond the stent itself Fig. 2D). Decoupling circuitry was housed in external hielded boxes. NDOGRAFT PHANTOM, HEATING, AND IN VIVO TESTING. All ctive devices were imaged in saline phantoms before in vivo xperiments to confirm signal and appearance. Heating of the endograft was tested in vitro in a poly- crylamide gel phantom and in vivo in one additional nimal. Conditions were intended to exaggerate heating, ncluding continuous steady-state free precession (SSFP) canning with a high flip angle (alpha � 75°, repetition time TR] � 3.1 ms). Temperature was measured using four beroptic thermistor probes (Umi-4, Fiso Technologies, uebec, Canada) placed along the length of the device, ncluding the tip. UNCTIONAL AND ANATOMIC MRI TECHNIQUES. Func- igure 1. Active-stent endograft. (A) Components of homemade active ndograft, constructed from 0.009-inch nitinol wire and expanded poly- etrafluoroethylene graft material. (B) Completed one-channel active-stent evice mounted on a 5-F catheter and constrained within a 10-F nylon heath (arrowhead). Matching tuning circuitry is housed in a separate box arrow). ional MRI techniques included dynamic magnetic reso- n I f G m 8 fi d j ( r u t 2 � v o a w p d 5 b S h 2 0 ( a m 6 s l s r fl e t s 3 r 1 c m v R r t F H d s d 2071JACC Vol. 45, No. 12, 2005 Raman et al. June 21, 2005:2069 –77 Real-Time MRI AAA Repair ance angiography (MRA) and phase-contrast imaging. mmediately after aneurysm creation, animals were trans- erred to a co-located 1.5-T MRI scanner (Signa CV/I, eneral Electric, Waukesha, Wisconsin, or Sonata, Sie- ens, Erlangen, Germany) for imaging using 4- or -channel phased-array surface coils (NovaMedical, Wake- eld, Massachusetts, or Siemens). Contrast-enhanced igital-subtraction MRA was performed with systemic in- ection of 0.1 to 0.2 mmol/kg gadopentate dimeglumine Magnevist, Berlex, Wayne, New Jersey) with a 3D adiofrequency-spoiled gradient echo (SPGR) acquisition sing the following parameters: repetition time (TR)/echo ime (TE) 6.7/1.2 ms, flip angle 45°, matrix 512 � 192 � 4, field-of-view 36 � 27 � 8.2 cm, receiver bandwidth 62.5 kHz, voxel size 0.7 � 1.4 � 3.4 mm. Mask, arterial, enous (after 60 s), and late phases (after 5 min) were btained to identify slow contrast accumulation in the neurysm sac. A 3D, low-flip-angle fast-gradient-echo scan ith and without fat saturation was run with thin axial artitions delineating aortic anatomy before and after en- ograft deployment to assess stent strut apposition (TR/TE .8/1.2 ms, flip angle 15°, matrix 256 � 160 � 24, andwidth 490 Hz/pixel, resolution 1.1 � 1.7 � 4 mm). tent strut apposition was also confirmed on axial cuts using igure 2. Schematic of endograft designs. (A) Unmodified commercial de omemade endograft device with active opposed loop solenoid coils as m evice with active stent connected to delivery system shaft by a detachable tent inactive. (D) Homemade endograft device with active stent as desc elivery shaft just beyond the distal stent edge. igh-resolution rectilinear trajectory SSFP (matrix 256 � t 56, field-of-view 20 � 20 cm, in-plane resolution 0.8 � .8 mm, slice thickness 6 mm) and reduced field-of-view rFOV) radial trajectory SSFP (22) (TR/TE 4.5/2.3 ms, flip ngle 60°, 32 projections with 17 interleaves, regridded to atrix 128 � 128, field-of-view 12 � 12 cm, slice thickness mm, bandwidth 560 Hz/pixel), as well as black-blood fast pin echo (TR/TE 1200/100 ms, flip angle 180°, echo train ength 16, matrix 384 � 256, field-of-view 30 � 20 cm, lice thickness 5 mm, bandwidth �62.5 kHz, in-plane esolution 0.8 � 0.8 mm). Phase-contrast imaging was used qualitatively to assess ow perturbations within the aneurysm before and after ndograft deployment. Flow velocities were assessed along hree axes using in-plane and through-plane phase-contrast cans (TR/TE 5.0/3.0 ms, matrix 192 � 96, field-of-view 2 � 24 cm, slice thickness 5 mm, bandwidth � 62.5 kHz, esolution 2.5 � 2.5 mm, through-plane velocity encoding 50 cm/s, in-plane velocity encoding 80 cm/s). Phase- ontrast data were represented with through-plane flow apped as color, and in-plane flow was mapped as velocity ectors (MATLAB, Mathworks, Natick, Massachusetts) (23). eal-time MRI and interventional procedure. The tMRI for procedural guidance required several modifica- ions to commercial hardware and software, including ex- imaged on the basis of intrinsic magnetic susceptibility (signal void). (B) s delineating proximal and distal stent edges. (C) Homemade endograft e that, after deployment and removal of the delivery catheter, renders the in (C) and second active marker composed of a multilooped coil on the vice arker cabl ribed ernal image reconstruction and in-room display, as previ- o w f c i s a t p s s m u n s a r p a b a c u i i o s r s C d t s n i R A f P s p t p u E h e d r t s a l m a t t c h f e R t l S c p a t s d s a g t w a v ( d p i i t n p r P c a m d i l d t r t t d a 2072 Raman et al. JACC Vol. 45, No. 12, 2005 Real-Time MRI AAA Repair June 21, 2005:2069 –77 usly described (17,24,25). Interactive rtMRI user interfaces ere customized with the addition of several useful features or rectilinear SSFP imaging (25,26): individual receiver hannel gain, coloring, and highlighting for use with active ntravascular devices; preparatory saturation pulses to negate ignal from fat or specific spatial regions; simultaneous cquisition and display of multiple image slices; and real- ime 3D rendering of multislice data (25). An interactive oint marking system was implemented allowing user- elected reference points on individual slices to be repre- ented on the 3D rendering. This was useful, for example, to ark visceral artery ostia. A parallel real-time environment sed radial k-space trajectories, a data-undersampling tech- ique that optimizes temporal resolution without sacrificing patial resolution (22). This interface was also customized to llow interactive overlay of previously acquired angiographic oadmaps. Typical real-time SSFP imaging used the following arameters: TR/TE 3.8/1.8 ms, matrix 192 � 128, flip ngle 60°, slice thickness 8 mm, field-of-view 36 � 24 cm, andwidth �64 to 128 kHz. Image position could be djusted interactively by drawing new prescriptions on the urrent image, pushing or pulling through parallel planes by ser-specified gap thickness, and rotating vertically or hor- zontally around the image center by user-specified angle ncrements. These yielded four to eight frames/s, depending n whether image acceleration techniques such as echo haring were used. The imaging latency, or time to acquire, econstruct, and display images to the operator using this ystem is approximately 250 ms. orroboration of aneurysm exclusion. After endograft eployment and post-procedural MR anatomic and func- ional imaging as described for baseline studies, X-ray ubtraction angiography was repeated. Animals were eutha- ized, and the abdominal aorta was excised for visual nspection in six. ESULTS neurysm model. Balloon overstretch dilation of the in- rarenal aorta was performed successfully in all 11 animals. erforation/rupture of the aorta occurred in two, but both urvived for the duration of the nonsurvival experimental rotocol. At the widest, dilated segments were up to twice he reference aortic diameters. In all cases, aneurysms ersisted throughout the experimental protocol, up to 6 h, sually without significant recoil. ndograft testing. In vitro all active devices showed en- anced signal in the immediately surrounding region. How- ver, the receiver coil on the active-marker/passive-stent evice (Fig. 2B) coupled inductively with the nitinol stent, esulting in increased signal along the entire device rather han discretely at each end-marker coil. There was marked ignal drop-off within 2 to 3 mm of the distal end of the ctive-stent devices (Fig. 2C), a known limitation of the oopless antenna design (20). The active-stent/active- 2 arker design (Fig. 2D) overcame this limitation by placing loop coil (27) to act as an edge marker distal to the stent. Heating was only noted at the tip. The maximum emperature increase was 2.0°C in the in vitro static phan- om, and 2.2°C in vivo. Signal-to-noise profiles of the final endograft design, ombining an active stent with active edge markers, was igher than for all previous designs, and contrast-to-noise or the final device was different primarily at the endograft dges (data not shown). eal-time MR guidance of endograft procedures. Real- ime imaging with SSFP provided adequate temporal reso- ution for device navigation, positioning, and deployment. patial resolution was sufficient to visualize important vis- eral and branch artery origins necessary for precise device lacement. In one of the animals with aortic rupture, the rapidly ccumulating retroperitoneal hematoma obscured impor- ant anatomy. This problem was circumvented by thick- lice, real-time angiography with a hand injection of ilute (30 mM) gadolinium contrast during image acqui- ition. Active devices were easily visualized within the orta using color highlighting and individual channel ain adjustment. Multislice imaging provided reference coronal and sagit- al slices of the infrarenal aorta with renal and iliac vessels hile allowing interactive adjustment of axial slices to verify nd adjust device position. The slices were displayed indi- idually and in combination after real-time 3D rendering Figs. 3A and 3B). A point marking system allowed precise elineation of important anatomic references, including roximal and distal aneurysm extent, with display on the 3D mage to guide device positioning and deployment (Fig. 3A). The balance between spatial resolution, the number of mage slices, and temporal resolution was adjusted itera- ively, in combination with temporal filtering and multipla- ar imaging. Optimal anatomic guidance seemed to be rovided by multiple (3–5) non-orthogonal slices, each fully efreshed only approximately once per second. erformance of different MRI-endograft designs. The ommercial endograft system (Fig. 2A, one tested) was visu- lized passively by its marked susceptibility artifact, but this ade it very difficult to differentiate the stent itself from the elivery shaft. We attempted to improve device visualization by nserting a quarter-wave active antenna-guidewire through the umen of the endograft system. This enhanced local signal but id not satisfactorily delineate the device. The active-marker/passive-stent device (Fig. 2B, one ested) showed bright signal along the length of the device ather than discretely at the markers only. The distal edge of he stent could not definitively be identified. Additionally, he tip of the delivery system, extending 2 cm beyond the istal active marker, was not conspicuous because of volume veraging. The loopless coil design of the active-stent device (Fig. C, two tested) provided good signal except along the distal s t a c m o t t o u w M d T a t s r s e r p l l o c a c t l F i o b “ , I � i 2073JACC Vol. 45, No. 12, 2005 Raman et al. June 21, 2005:2069 –77 Real-Time MRI AAA Repair everal millimeters of the actual endograft, reducing opera- or confidence during positioning and deployment. The final device iteration combined an active stent with n active distal loop coil at the distal tip of the delivery atheter (Fig. 2D, seven tested). This design produced the ost reproducible signal pattern and provided satisfactory perator confidence in device position. The two procedural failures were attributed to the fragility of hese homemade prototypes, built primarily for imaging rather han mechanical characteristics. Of the two procedural failures, ne was related to shifting of the self-expanding endograft during nsheathing, and the other was related to migration during ithdrawal of the detachable antenna connection. R assessment of procedural success. Nine of 11 en- ograft procedures were successful under rtMRI guidance. he two failures were identified using first-pass MRA. No ttempt was made to correct acute endoleaks using adjunc- igure 3. Real-time multislice imaging and three-dimensional rendering d maging multislice axial, sagittal, and coronal images shown simultaneously n the right integrates multislice information. Positioning three axial slice ifurcation, respectively, allows simultaneous capture of the most important bird’s eye” view of the aorta. Orientation markers indicate: S � superior ndicates aneurysm. ive balloon or stent devices. c Stent strut apposition at the proximal and distal target egments of the aorta was convincingly shown by high- esolution axial SSFP, fast spin echo, and 3D gradient echo cans (Fig. 4). Repeat MRA in 9 of 11 cases showed both aneurysm xclusion by the endograft (Fig. 5) and patency of the enal arteries. Iliac arteries were also patent by angiogra- hy. Successful exclusion was further corroborated by ack of contrast accumulation in the aneurysm sac during ate-phase angiographic and real-time SSFP scans. In the ther two cases, MR contrast-angiography revealed pro- edural failure with evidence of gadolinium within the neurysm sac. High-resolution phase-contrast studies with vector- and olor-flow mapping showed reduction of in-plane and hrough-plane turbulence consistent with restoration of aminar flow (Fig. 6). In the two procedural failures, phase endograft positioning. In the left column, real-time magnetic resonance tate precise device positioning. Concomitant three-dimensional rendering he caudal renal artery origin, the middle of the aneurysm, and the aortic my for device placement. The coronal and sagittal slices provide an overall inferior, A � anterior, P � posterior, R � right, L � left. Blue arrow uring facili s at t anato ontrast imaging was not specifically conducted to iden- t c X a d p f i F tent f exclu F m n 2074 Raman et al. JACC Vol. 45, No. 12, 2005 Real-Time MRI AAA Repair June 21, 2005:2069 –77 ify flow jets associated with endoleaks identified by ontrast MRA. -ray angiography and pathology. Digital subtraction ngiography corroborated MRA findings in all nine cases igure 4. Stent strut apposition. (A) Fast spin echo image shows nitinol s rom stent struts). (B) Spin-echo axial image at level of aneurysm, showing igure 5. Magnetic resonance angiogram (maximum-intensity projectio agnetic resonance angiography shows infrarenal abdominal aortic aneurysm a itinol endograft (causing luminal artifact). Orientation markers as in Figure 3. eemed successful by MRI. Direct inspection and alpation of the resected abdominal aorta in six success- ul cases confirmed that the renal arteries were not nvolved. well apposed to target proximal infrarenal aorta (arrows show signal void ded sac (dashed outline). Orientation markers as in Figure 3. fore and after endograft delivery. (A) Conventional contrast-enhanced n) be fter balloon overstretch. (B) Abdominal aortic aneurysm is excluded by D W p s M a s E e t e s d d a B i t p i c s m c d c o a p p u r t r ( i c t t r w a g o t r c a e c c a s e S c m m s n i u c e p d m t l r g t F d p e fl c d i a 2075JACC Vol. 45, No. 12, 2005 Raman et al. June 21, 2005:2069 –77 Real-Time MRI AAA Repair ISCUSSION e showed the feasibility of endovascular repair of AAA erformed entirely under real-time MRI guidance in a wine model. Functional imaging with dynamic contrast RA and phase-contrast techniques complemented the natomic imaging for immediate assessment of procedural uccess. ndograft devices and MRI. Our approach to devices ncompassed both passive and active designs for visualiza- ion and tracking. The unmodified, commercial passive ndograft was difficult to differentiate from the delivery ystem, limiting operator confidence in positioning and eployment. Moreover, signal voids generated by passive evices cannot be readily distinguished from signal attenu- tion of other causes or from volume averaging (28). ecause active devices function as receiver coils, the result- ng local signal enhancement markedly facilitates visualiza- ion (Figs. 3 and 4) (21,26,27,29). Mahnken et al. (30) ositioned passive commercial endografts below renal arter- A P RL A P RL igure 6. Phase-contrast flow assessments before and after endograft eployment. Axial overlays of in-plane (vector flow map) and through- lane (color map) flow within ruptured experimental aneurysm. Before ndovascular repair, there is marked turbulence and evidence of retrograde ow (blue) within the aneurysm (dashed outline). The vena cava is ollapsed in this hemorrhagic state. Laminar flow is restored after en- ograft (dashed outline) deployment. Solid lines border vena cava, dentified on cine loops by constant nonpulsatile flow. Orientation markers s in Figure 3. es in healthy swine under MRI guidance, but also con- c luded that dedicated MRI endograft designs might be uperior. We constructed several different types of home- ade active endograft systems, of which the final two- hannel conductively coupled active-marker/active-stent evice provided the most useful signal profile. The rtMRI with SSFP provided excellent intrinsic blood ontrast within the aorta and its branch vessels, allowing ptimal slice selection for the procedure. Unlike coronary nd intracardiac procedures, in which ultrafast imaging is a rerequisite because of cardiac and respiratory motion, eripheral endovascular repair under MR guidance may be ndertaken with slower frame rates that allow better spatial esolution. Radial k-space acquisition seems well suited to ranscatheter aneurysm therapy because of advantages in FOV imaging and because of intrinsic edge enhancement 22). Customized features on the interactive real-time nterface proved very useful during the interventional pro- edure, particularly in channel coloring and gain adjustment o highlight and contrast the device from surrounding issue. Multislice imaging with simultaneous real-time 3D endering facilitated device positioning and deployment ith complementary orthogonal and/or oblique views. Our bility to precisely position and deploy the endografts was ood and was limited only by the mechanical shortcomings f our homemade devices. Post-deployment scanning combined anatomic and func- ional assessment of procedural success. High-resolution, FOV SSFP, fast spin echo and 3D gradient echo scans onvincingly showed stent strut apposition to the proximal nd distal target aortic segments. Dynamic contrast- nhanced MRA during arterial, venous, and late phases orroborated aneurysm exclusion in successful cases and learly showed contrast accumulation in the aneurysm sac fter procedural failures. Phase-contrast scans visually howed restoration of laminar flow in the grafted segment. These studies were readily completed within minutes of ndograft delivery. tudy limitations and clinical implications for endovas- ular repair. Although we successfully excluded aneurys- al segments immediately in all cases in which the home- ade device functioned appropriately, our experience hould be qualified. We used a model of AAA in healthy, onatherosclerotic swine. Even after balloon overstretch njury, the course of the infrarenal aorta remained essentially nchanged. This and more complicated surgical models annot represent the complex 3D anatomy and tortuosity ncountered clinically in degenerative arterial segments in atients with AAA, in whom the expected benefits of etailed soft tissue contrast and 3D representation by MRI ay be more dramatic. In particular, safe traversal of ortuous iliac artery segments, one of the procedural chal- enges of AAA endografting, might be simplified by using tMRI to visualize device-related anatomic distortion and to uide operator adjustments. Furthermore, we used a simple ubular endograft, although most clinical devices are bifur- ated to cover the aorta and both iliac limbs. w D s w m r t r a p d t d s M d p i c b g a c f a h g a b M c i M s fi a C T o u d t a d r g R C I N t R 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2076 Raman et al. JACC Vol. 45, No. 12, 2005 Real-Time MRI AAA Repair June 21, 2005:2069 –77 Our experience suggests that an active endograft design ould be most appropriate for further clinical development. espite iterative prototype testing, the optimal device de- ign remains undetermined. Additional issues include hether an endograft system with associated invasive equip- ent (e.g., guidewires) made from nonferrous materials etains adequate mechanical characteristics to traverse tor- uous and/or calcified iliac arteries. Inductive heating during adiofrequency excitation is an important safety consider- tion for long conductive catheter devices in MRI. Our rototype devices showed minimal local heating because of ecoupling circuitry to limit radiofrequency energy deposi- ion. This is amenable to further optimization (31,32). Although AAA endografting has emerged over the past ecade as a viable alternative to open surgery, failure modes uch as malapposition could conceivably be improved using RI for on-line planning, procedural guidance, and imme- iate post-procedural assessment. The rtMRI might im- rove procedural success by improving soft-tissue and 3D maging during the procedure. The SSFP provides intrinsic ontrast for excellent visualization of the aorta and major ranch vessels without exogenous agents. Substituting adolinium-based MR contrast for iodinated radiocontrast gents might reduce contrast nephropathy in this at-risk linical population (33). Magnetic resonance imaging may acilitate device selection by improving the measurement ccuracy of aneurysm neck dimension, in which oversizing as been associated with endograft displacement and mi- ration (10). Furthermore, aneurysm anatomy may be ltered by the use of a stiff guidewire and introduction of a ulky, high-profile endograft delivery system. Interactive RI may offer better delineation of this complex intrapro- edural anatomy than conventional X-ray projection imag- ng. Finally, combined on-line anatomic and functional RI may be useful for guiding and evaluating procedural uccess. The chief obstacle to clinical translation of our ndings is the unavailability of clinical-grade commercial ctive MRI endografts. ONCLUSIONS hese experiments show that successful endovascular repair f experimental AAA in swine can be conducted solely nder rtMRI. Endografts can be built to be conspicuous uring the procedure. Anatomic and complementary func- ional MRI can guide graft deployment and can immedi- tely assess procedural success. Although significant future evelopment of active devices is required, we believe this epresents an attractive application for cardiovascular MR- uided intervention. eprint requests and correspondence: Dr. Robert J. Lederman, ardiovascular Branch, Clinical Research Program, Division of ntramural Research, National Heart, Lung, and Blood Institute, ational Institutes of Health, Building 10, Room 2c713, Be- hesda, Maryland 20892-1538. E-mail: Lederman@nih.gov. EFERENCES 1. Buth J. Endovascular repair of abdominal aortic aneurysms. Results from the EUROSTAR registry. EUROpean collaborators on stent- graft techniques for abdominal aortic aneurysm repair. Semin Interv Cardiol 2000;5:29 –33. 2. Matsumura JS, Brewster DC, Makaroun MS, Naftel DC. A multi- center controlled clinical trial of open versus endovascular treatment of abdominal aortic aneurysm. J Vasc Surg 2003;37:262–71. 3. Zarins CK, White RA, Moll FL, et al. The AneuRx stent graft: four-year results and worldwide experience 2000. J Vasc Surg 2001; S33:135– 45. 4. White GH, Yu W, May J, Chaufour X, Stephen MS. Endoleak as a complication of endoluminal grafting of abdominal aortic aneurysms: classification, incidence, diagnosis, and management. J Endovasc Surg 1997;4:152– 68. 5. May J, White GH, Waugh R, et al. Adverse events after endoluminal repair of abdominal aortic aneurysms: a comparison during two successive periods of time. J Vasc Surg 1999;29:32–7; discussion 38 –9. 6. Veith FJ, Baum RA, Ohki T, et al. Nature and significance of endoleaks and endotension: summary of opinions expressed at an international conference. J Vasc Surg 2002;35:1029 –35. 7. Alric P, Hinchliffe RJ, Wenham PW, Whitaker SC, Chuter TA, Hopkinson BR. Lessons learned from the long-term follow-up of a first-generation aortic stent graft. J Vasc Surg 2003;37:367–73. 8. Tzortzis E, Hinchliffe RJ, Hopkinson BR. Adjunctive procedures for the treatment of proximal type I endoleak: the role of peri-aortic ligatures and Palmaz stenting. J Endovasc Ther 2003;10:233–9. 9. Tummala S, Powell A. Imaging of endoleaks. Tech Vasc Interv Radiol 2001;4:208 –12. 0. Badran MF, Gould DA, Raza I, et al. Aneurysm neck diameter after endovascular repair of abdominal aortic aneurysms. J Vasc Interv Radiol 2002;13:887–92. 1. 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;205:341–52. 2. Ludman CN, Yusuf SW, Whitaker SC, Gregson RH, Walker S, Hopkinson BR. Feasibility of using dynamic contrast-enhanced mag- netic resonance angiography as the sole imaging modality prior to endovascular repair of abdominal aortic aneurysms. Eur J Vasc Endo- vasc Surg 2000;19:524 –30. 3. Cejna M, Loewe C, Schoder M, et al. MR angiography vs CT angiography in the follow-up of nitinol stent grafts in endoluminally treated aortic aneurysms. Eur Radiol 2002;12:2443–50. 4. Razavi R, Hill DL, Keevil SF, et al. Cardiac catheterisation guided by MRI in children and adults with congenital heart disease. Lancet 2003;362:1877– 82. 5. 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;106:511–5. 6. Rickers C, Jerosch-Herold M, Hu X, et al. Magnetic resonance image-guided transcatheter closure of atrial septal defects. Circulation 2003;107:132– 8. 7. Dick AJ, Guttman MA, Raman VK, et al. Magnetic resonance fluoroscopy allows targeted delivery of mesenchymal stem cells to infarct borders in swine. Circulation 2003;108:2899 –904. 8. Hallisey MJ. 1997 SCVIR Gary J. Becker Young Investigator Award paper. A transluminally created abdominal aortic aneurysm model. J Vasc Interv Radiol 1997;8:305–12. 9. Quick HH, Ladd ME, Nanz D, Mikolajczyk KP, Debatin JF. Vascular stents as RF antennas for intravascular MR guidance and imaging. Magn Reson Med 1999;42:738 – 45. 0. Ocali O, Atalar E. Intravascular magnetic resonance imaging using a loopless catheter antenna. Magn Reson Med 1997;37:112– 8. 1. Ladd ME, Zimmermann GG, Quick HH, et al. Active MR visual- ization of a vascular guidewire in vivo. J Magn Reson Imaging 1998;8:220 –5. 2. Peters DC, Lederman RJ, Dick AJ, et al. Undersampled projection reconstruction for active catheter imaging with adaptable temporal reso- lution and catheter-only views. Magn Reson Med 2003;49:216 –22. 3. Thompson RB, McVeigh ER. High temporal resolution phase con- trast MRI with multiecho acquisitions. Magn Reson Med 2002;47: 499 –512. 2 2 2 2 2 2 3 3 3 3 2077JACC Vol. 45, No. 12, 2005 Raman et al. June 21, 2005:2069 –77 Real-Time MRI AAA Repair 4. Guttman MA, McVeigh ER. Techniques for fast stereoscopic MRI. Magn Reson Med 2001;46:317–23. 5. Guttman MA, Lederman RJ, Sorger JM, McVeigh ER. Real-time volume rendered MRI for interventional guidance. J Cardiovasc Magn Reson 2002;4:431– 42. 6. Lederman RJ, Guttman MA, Peters DC, et al. Catheter-based endomyocardial injection with real-time magnetic resonance imaging. Circulation 2002;105:1282– 4. 7. Serfaty JM, Yang X, Aksit P, Quick HH, Solaiyappan M, Atalar E. Toward MRI-guided coronary catheterization: visualization of guiding catheters, guidewires, and anatomy in real time. J Magn Reson Imaging 2000;12:590 – 4. 8. Bakker CJ, Bos C, Weinmann HJ. Passive tracking of catheters and guidewires by contrast-enhanced MR fluoroscopy. Magn Reson Med 2001;45:17–23. 9. Atalar E, Bottomley PA, Ocali O, et al. High resolution intravascular MRI and MRS by using a catheter receiver coil. Magn Reson Med 1996;36:596 – 605. 0. Mahnken AH, Chalabi K, Jalali F, Gunther RW, Buecker A. Magnetic resonance-guided placement of aortic stents grafts: feasibil- ity with real-time magnetic resonance fluoroscopy. J Vasc Interv Radiol 2004;15:189 –95. 1. Yeung CJ, Susil RC, Atalar E. RF heating due to conductive wires during MRI depends on the phase distribution of the transmit field. Magn Reson Med 2002;48:1096 – 8. 2. Susil RC, Yeung CJ, Atalar E. Intravascular extended sensitivity (IVES) MRI antennas. Magn Reson Med 2003;50:383–90. 3. Spinosa DJ, Kaufmann JA, Hartwell GD. Gadolinium chelates in angiog- raphy and interventional radiology: a useful alternative to iodinated contrast media for angiography. Radiology 2002;223:319 –25; discussion 326 –7. Real-Time Magnetic Resonance-Guided Endovascular Repair of Experimental Abdominal Aortic Aneurysm in Swine METHODS Animal preparation and aneurysm model Construction of endograft devices ENDOGRAFT PHANTOM, HEATING, AND IN VIVO TESTING FUNCTIONAL AND ANATOMIC MRI TECHNIQUES Real-time MRI and interventional procedure Corroboration of aneurysm exclusion RESULTS Aneurysm model Endograft testing Real-time MR guidance of endograft procedures Performance of different MRI-endograft designs MR assessment of procedural success X-ray angiography and pathology DISCUSSION Endograft devices and MRI Study limitations and clinical implications for endovascular repair CONCLUSIONS REFERENCES