The effect of magnetic resonance imaging on stainless-steel Z-stent–based abdominal aortic prosthesis

Methods 

A total of 550 patients underwent endovascular repair of an infrarenal abdominal aortic aneurysm between June 1996 and December 2005 at the University of California–San Francisco Medical Center by using a stainless-steel Z-stent–based abdominal aortic prosthesis. The stent-graft designs included an aortouni-iliac device (with occluder and femoral-to-femoral bypass), the Zenith bifurcated stent graft, fenestrated Zenith bifurcated stent grafts, and multiply bifurcated stent grafts. Routine patient follow-up included computed tomography (CT) scans and multiview abdominal plain films at 1, 6, and 12 months after surgery and then yearly thereafter. All patients were specifically instructed not to undergo MRI at any site, including brain, chest, abdomen/pelvis, spine, or lower extremity MRI. Despite these instructions, retrospective review identified 22 patients who underwent MRI after stent-graft implantation.

All patients who consented to participate in the study to have their records reviewed were questioned about symptoms of abdominal pain or back pain after the MRI. Pre- and post-MRI CT scans and abdominal radiographs were compared to identify any change in stent-graft structure (stent fracture or barb fracture), position (component separation, twisting, or kinking), or patency.

Results 

Of the 17 patients who consented to have their records reviewed as part of this study, 11 had Zenith bifurcated stent grafts, and 6 had custom-made uni-iliac stent grafts. These patients underwent a total of 20 MRI studies at a mean interval of 669 days (range, 3-2179 days) after stent-graft implantation. Three of these patients underwent two MRIs each during the follow-up period. These studies comprised 10 MRIs of the brain or neck and 10 MRIs of the abdomen, pelvis, or spine. In all cases, the static magnetic field strength was 1.5 T or less. No patients experienced any symptoms of abdominal or back pain during or after the MRI.

The mean aneurysm neck diameter was 24 mm and ranged between 16 and 35 mm. The mean neck length was 31 mm and ranged from 12 to 55 mm. Five of these 17 patients had suboptimal anatomy. Two patients had severe neck angulation (90°), two had neck lengths less than 15 mm, and 1 patient had a neck diameter larger than 28 mm. The mean aneurysm diameter at the time of repair was 62 mm and ranged from 36 to 95 mm. One patient had a ruptured aneurysm at the time of treatment.

Comparison of pre- and post-MRI CT scans (available in 15 of 17 patients) and abdominal radiographs showed no change in the structure, position, or function of the stent graft and no increase in the diameter of the aneurysm in any patient. Eleven of these patients had at least two CT scans after the most recent MRI study. The average follow-up was 899 days (range, 24-1822 days) after MRI.

Discussion 

The presence of a Zenith stent graft is regarded as a contraindication to MRI because the stents are made of 304 stainless steel. The ferromagnetic properties of the Zenith stent graft raise two separate issues: the stent graft may degrade the magnetic resonance images (magnetic resonance compatibility), and the magnetic field may potentially destabilize the prosthesis (magnetic resonance safety). This study focused on the latter issue.

The Food and Drug Administration labeling requirements for magnetic resonance safety are based on the findings of in vitro deflection angle testing, as originally described by New et al.6 In this test, the prosthesis is suspended from a 30-cm-long string in the center of a 1.5-T magnetic field. The angle of deflection from vertical determines magnetic resonance compatibility. An angle of more than 45° indicates that the magnetic field generates more force than the weight of the device, which is considered unsafe. In the case of the Zenith stent graft (unpublished data), the angle was almost 90°, thus indicating a magnetic force 38 times the weight of the prosthesis. However, this force (2.64 N) is far less than the force exerted on the prosthesis by aortic pressure and blood flow all day, every day, for the life of the patient.7 In the same series of tests, the magnetic field generated a torque on the device 10 times greater than the weight of the device. Again, this is far less than the torque generated by hemodynamic forces. Finally, in a cadaver study that specifically examined the caudally directed forces required to dislodge the suprarenal stent component from the aorta, the displacement force was between 23 and 26.5 N.8 This is approximately 10 times the force predicted to result from the MRI scanner. However, no experimental study has evaluated the effects of force applied in any direction but caudal, and magnetic fields produce forces in all directions. In the clinical setting, anatomic factors may influence the interaction between the stent graft and the nondilated segment of aorta between the renal arteries and the aneurysm (neck), but these are of uncertain significance here because both stent-graft designs have barbed suprarenal stents for attachment.

Given these findings, it is hardly surprising that MRI had no apparent effect on the structure, position, or function of the Zenith stent grafts in the patients in this series. Half of these studies were centered on the area containing the stent graft. In these studies, the stent-graft lay in the center of the magnetic field and would have been subjected to the greatest force. Because the magnetic field strength declines with increasing distance from the center of the field, MRI of the head or neck imparts very little force to an abdominal aortic stent graft.

The same is probably true regarding the effect of the stent graft on the quality of MRI (MRI compatibility). One cannot assess the structure, position, or function of the Zenith prosthesis by using MRI.1, 9 There is such significant artifact from the stent graft that an abdominal MRI is of minimal use for imaging the abdominal viscera (Fig 1). Only nitinol-based stent grafts can undergo magnetic resonance–based follow-up.1, 9 However, stent graft–induced imaging artifacts do not affect the entire field, especially in a T1-weighted spin-echo pulse sequence. In our experience, images of the spine are of variable quality (Fig 2), and the presence of the Zenith abdominal aortic stent graft has no effect on images of the head and neck.

View Large Image Download to PowerPoint

Fig 1. Magnetic resonance image of the abdomen in a patient with an abdominal aortic Zenith stent graft.

View Large Image Download to PowerPoint

Fig 2. Magnetic resonance image of the spine in a patient with an abdominal aortic Zenith stent graft.

We hesitate to state that the Zenith stent-graft is MRI safe, but we have concluded, on the basis of the findings of this study, that the benefits of MRI should not be denied to every patient with a Zenith stent graft. The anatomic characteristics of our study cohort are a fairly representative sample, including patients with both straightforward and suboptimal anatomy. The relative merits will vary from case to case. MRI of the head may certainly be indicated in a patient with an abdominal aortic Zenith stent graft. MRI in the setting of a patient with an implanted Zenith stent graft should be followed by imaging of the stent graft either via multiview abdominal radiographs or CT scan to assess its position.

Author contributions