| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 All authors: Department of Diagnostic Radiology, University Hospital of Regensburg, Franz-Josef-Strauss-Allee 11, 93042 Regensburg, Germany.
Received September 1, 2004;
accepted after revision December 6, 2004.
Address correspondence to O. W. Hamer
(o.hamer{at}gmx.de).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. In a prospective study, 34 iliac stenoses in 27 patients were treated by implantation of 35 nitinol stents. MR angiography was performed immediately after stent placement for 32 stents, and both digital subtraction angiography (DSA) and MR angiography were repeated at the 6-month follow-up for 23 stents. Three blinded observers assessed stent patency and the degree of in-stent stenoses on MR angiography and DSA (the standard of reference) images. The difference between the observers' grading of stenoses on DSA and on MR angiography was determined. Statistical analysis was performed using the Student's t test for paired samples.
RESULTS. Stent patency was assessed correctly for all stents and both sets of MR angiography images. Evaluation of DSA 1 images (obtained at end of implantation procedure) revealed that 96.9% of in-stent stenoses were less than 50%. On DSA 2 images (obtained at follow-up), 95.7% of in-stent stenoses were graded as less than 50%. The difference between grading of stenoses on DSA and MR angiography images was 15.0% ± 16.0% (minimum, 0.0%; maximum, 63.3%) for DSA 1 versus MR angiography 1 (statistically significant, p = 0.037) and 9.8% ± 13.5% (minimum, 0.0%; maximum, 63.3%) for MR angiography 2 versus DSA 2 (not statistically significant, p = 0.355).
CONCLUSION. Patency was correctly assessed for all stents on MR angiography. The quality of MR angiography regarding characterization of in-stent stenoses improved with time after stent placement. However, discrepancies of more than 60% between grading of lumen narrowing on DSA and MR angiography images occurred even at the 6-month follow-up. Thus, MR angiography is not yet a reliable technique for characterization of in-stent stenoses.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Many experimental studies for a variety of stent designs and materials have been performed to evaluate the nature and amount of artifacts that impair the quality of MR angiography [5-10]. The studies suggest that nonferromagnetic metals or alloys are more suitable for MRI than ferromagnetic ones because of artifacts that are not as severe. Among others, nitinol, an alloy of nickel and titanium, has been shown to be advantageous regarding MR image quality [7, 11, 12]. A few reports have been published that evaluated MR angiography for monitoring endovascular prostheses in small samples [13-18]. These reports were focused on the assessment of stent patency and did not evaluate whether grading of potential in-stent stenoses is possible.
The purposes of this prospective in vivo study were to evaluate whether MR angiography is suitable for assessing stent patency and grading in-stent stenoses and to examine whether the accuracy of MR angiography changes with time after stent implantation.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
50%
or mean translesional pressure gradient 10 mm Hg) in combination with
symptoms of ischemia in the dependent limb; vascular bypasses further
downstream and stenoses in the feeding iliac vessel (in these cases
intervention was performed after consultation with or on demand of the
vascular surgeons to increase the in-flow in the bypass); and an age of 40
years or older (because the study protocol included follow-up DSA at 6 months
regardless of recurrence of symptoms). Only those lesions that were regarded
as treatable with a 10-mm-diameter stent were enrolled because, during the
period of patient recruitment, this size was the only available stent diameter
at the authors' institution. The stent diameter was oversized by 1-2 mm
compared with the reference vessel, which was defined as a nonpathologic
vessel segment immediately distal to the lesion to treat. The exclusion
criteria included any contraindication to the administration of iodinated
contrast material or anticoagulant therapy; any contraindication to MRI
evaluation, including the administration of gadopentetate dimeglumine; septic
disease or end-stage malignancy; and pregnancy. Patients were given a written consent form approved by the institutional ethics committee. All patients agreed to participate in the study and signed the consent.
Major risk factors for developing vascular disease were smoking in 26/27 patients (96%), hypertension in 18/27 patients (67%), diabetes mellitus in 8/27 patients (30%), and elevated serum cholesterol in 11/27 patients (41%). More than one risk factor was present in 21/27 patients (78%). Regarding clinical staging, 4/27 patients (15%) presented with mild claudication (Rutherford category 2), 14/27 (52%) with moderate claudication (Rutherford category 3), 8/27 (30%) with severe claudication (Rutherford category 4), and 1/27 patients (4%) with focal gangrene (Rutherford category 5) [19].
Device Description
Self-expanding nitinol stents (JostentSelfX, Abbott Laboratories) were
implanted in all patients. The stent is laser cut from a slotted nitinol tube
with a strut length of 3.2 mm and a minimal wall thickness of 0.2 mm. The
delivery system consists of inner and outer catheters locked together with a
Tuohy-Borst valve. The stent rests on the inner catheter. Two radiopaque
markers on the inner catheter indicate the proximal and distal ends of the
stent. There are no radiopaque markers on the stent itself.
Stent Implantation and Intraarterial DSA
All patients underwent treatment under local anesthesia in the angiography
suite. Using a retrograde femoral approach, intraarterial diagnostic
angiography was performed in three planar views (posteroanterior and left and
right 35-45° oblique views), and the view that best showed the luminal
narrowing was determined. The length of the stent was selected in such a way
that the entire lesion was covered, with the ends of the stent lying in a
normal portion of the vessel wall. After stent delivery, balloon angioplasty
within the stent was performed for all lesions (balloon diameter, 7-8 mm;
working pressure, 8-12 atm) to create an angiographically normal lumen. The
intraarterial mean translesional pressure gradient was assessed before and
after stent placement. At the end of the procedure, intraarterial DSA
(hereafter referred to as DSA 1) of the stent was obtained in two planar views
(including the view that best showed the stenosis in the preinterventional
DSA) to assess eventual residual stenosis. After stent implantation, patients
were admitted to the hospital, which is routine practice at our institution.
Before hospital discharge (within 1-3 days after stent placement), MR
angiography (hereafter referred to as MR angiography 1) of the pelvis was
performed. Six months after stent placement or in case of a recurrence of
symptoms, each patient was again examined by intraarterial DSA (hereafter
referred to as DSA 2) and MR angiography (hereafter referred to as MR
angiography 2).
MR Angiography Sequence and Image Acquisition
The identical MR angiography protocol was used for both MR angiography 1
and MR angiography 2. The applied sequence was a breath-hold 3D gradient-echo
sequence with radiofrequency spoiling (fast low-angle shot [FLASH]; TR/TE,
4/1.6; 30° excitation angle; 224 x 256 matrix; 306 x 350
field-of-view; 1.6-mm partition thickness [after Fourier interpolation]; 260
Hz/pixel bandwidth; frequency encoding parallel to the main magnetic field).
The patient was positioned in the MR scanner in the supine position. All
images were acquired on a commercial 1.5-T MRI unit (Magnetom Symphony,
Siemens Medical Solutions). A dedicated angiography phased-array coil
(circularly polarized) served as a receive coil. Coronal images were acquired.
First, images before contrast administration were obtained. For determining
the individual MR angiography scanning delay, a test bolus of 2 mL of
gadopentetate dimeglumine (Magnevist, Schering) was administered into an
antecubital vein followed by a saline flush of 20 mL (flow rate, 2.5 mL/sec).
A turbo FLASH sequence providing one image per second was started
simultaneously with the test bolus, and the bolus arrival time was evaluated.
The scanning delay from the start of contrast administration until the start
of MR angiography was set to be identical to the bolus arrival time. A
conventional reordering scheme for the MR angiography sequence was selected.
MR angiography was performed after injection of 25 mL of gadopentetate
dimeglumine followed by a 25-mL saline flush, both at a flow rate of 2.5 mL
sec. All injections were performed using an automatic power injector (Spectris
MR Injector, Medrad). The angiographic data sets were postprocessed by image
subtraction from the unenhanced 3D MR images and application of a standard
maximum-intensity-projection algorithm (22.5° rotational intervals around
the craniocaudal axis covering 180°).
Data Analysis
Three radiologists, each with at least 4 years of experience in reviewing
digital subtraction and MR angiograms, evaluated the images. The radiologists
viewed the images independently in six sessions. DSA and MR angiography images
of the same patient were not evaluated in the same session to avoid a learning
effect. Measurements were performed on soft-copy displays at a workstation
(Magic View, Siemens). For the assessment of severity of stenosis, the
observers were free to use the implemented distance measurement software.
On DSA images, the observers were instructed to determine the diameter of the stent lumen at its narrowest location. The degree of lumen narrowing had to be expressed as percentage of the reference vessel, which was defined to be a nonpathologic vessel segment within 2 cm distal to the stent end. In case of a poststenotic dilatation (i.e., if the diameter of the vessel segment distal to the stent exceeded the diameter of the vessel segment proximal to the stent), the diameter of the vessel segment immediately adjacent to the proximal stent end was defined as the reference.
Regarding analysis of MR angiography images, maximum intensity projections and source images were considered. However, the apparent diameter of the stent lumen at its narrowest location was primarily assessed on the source images and was done according to the same strategy as on DSA images. The degree of lumen narrowing was again expressed as a percentage of the reference vessel. The observers were instructed not to rank obvious artifacts, such as bandlike susceptibility artifacts at the stent ends, as stenosis.
For each stent, the mean of the three observers' grading of stenosis was calculated for DSA and MR angiography images. Based on the observers' performances on the DSA images, the frequency of stenoses less than 50% and 50% or greater was determined for DSA 1 and DSA 2.
The presence of bandlike susceptibility artifacts at the stent ends was evaluated by each observer on a 3-point scale (1 = no band artifacts visible; 2 = band artifacts visible but less pronounced so that evaluation of the adjacent stent lumen was not compromised; and 3 = band artifacts pronounced so that evaluation of the adjacent stent lumen was compromised). The frequency of each score given per reviewer was determined. The mean frequency for all three observers was calculated for each score. This was done separately for MR angiography 1 and MR angiography 2.
Statistics
All variables are expressed as mean ± SD or as numbers and
percentages of patients.
Statistical analysis for paired continuous data— Statistical analysis for paired continuous data (grading of in-stent stenoses) was performed using the Student's t test for paired samples.
Statistical analysis for paired categoric data— Statistical analysis for paired categoric data (scoring of bandlike artifacts at the stent ends) was performed using the Wilcoxon's signed rank test. A p value of 0.05 or less was considered statistically significant.
Interobserver agreement—Kappa statistics require that all observers use the same rating categories. In this study, the observers' grading of stenoses was not classified but resulted in a large number of possible responses. That is why kappa statistics were not applicable. Therefore, interobserver agreement regarding evaluation of severity of in-stent stenoses was evaluated by calculating Spearman's correlation coefficient.
Power analysis—We retrospectively calculated the power at
the 
= 0.05 level for the Student's t test for paired samples
to detect a clinically meaningful difference (defined retrospectively as a
difference of 10%) between grading of stenoses on DSA and on MR angiography
images. This was done separately for DSA 1 versus MR angiography 1 and DSA 2
versus MR angiography 2. For determination of the SD relevant for the power
analysis, the absolute difference between the mean of the three observers'
grading of stenosis on DSA and MR angiography images was calculated for every
evaluated stent. The SD of these differences (32 single differences for
analysis DSA 1 vs MR angiography 1 and 23 single differences for DSA 2 vs MR
angiography 2) was used for the power analysis.
All statistics except the power analysis were calculated using the computer program SPSS version 10.0 (Statistical Package for the Social Sciences). The power calculation was performed using software provided by the Department of Statistics of the University of California Los Angeles. The power calculator is available on the Internet at www.stat.ucla.edu/ (accessed November 8, 2004).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
6/34 lesions (18%) were classified as category 1, 22/34 (65%) as category 2, and 6/34 (18%) as category 4c. No lesion met the criteria of category 3, 4a, or 4b.
Follow-Up Data
MR angiography before hospital discharge (MR angiography 1) was performed
in 25 patients with 32 stents. Two patients were discharged without
postprocedural imaging. The mean interval between stent placement and the
6-month follow-up was 197.9 ± 49.1 days (range, 99-258 days).
Follow-up, including DSA 2 and MR angiography 2, was available in 19 patients
with 23 stents. In one additional patient, only DSA was performed at the
6-month follow-up so that no comparison with MR angiography was possible. Data
of this patient were excluded from statistical analysis of the follow-up.
Seven patients were not available for follow-up. These patients were located,
but six of them refused to return because they were feeling well. One patient
had severe lung disease and was therefore not able to come to the hospital. In
the entire series, no MR angiography-related side effects were seen.
Severity of In-Stent Stenoses on DSA and MR Angiography Images
The means of the observers' assessments of in-stent stenoses on images of
DSA 1 compared with MR angiography 1 and DSA 2 compared with MR angiography 2
are shown in Figures 1A and
1B for all stenoses.
|
|
The mean of stenoses as assessed by the observers on MR angiography images was 18.3% ± 23.2% for MR angiography 1 and 17.9% ± 22.4% for MR angiography 2. The difference of grading of stenoses as determined by the observers on DSA 1 and MR angiography 1 images averaged 15.0% ± 16.0% (minimum, 0.0%; maximum, 63.3%). The Student's t test for paired samples revealed the differences to be statistically significant (p = 0.037). The degree of stenosis was overestimated on MR angiography 1 images for 15 stents and underestimated for 11 stents. For six stents, the observers' measurements of stenosis on images of DSA 1 and MR angiography 1 were equal.
The average difference between grading of stenoses assessed on DSA 2 and MR angiography 2 images was 9.8% ± 13.5% (minimum, 0.0%; maximum, 63.3%). The difference was not statistically significant (p = 0.355) (Figs. 2A, 2B, and 2C). The degree of stenosis was overestimated on MR angiography 2 images for 10 stents and underestimated for nine stents. For four stents, the observers' grading of stenosis on images of DSA 2 and MR angiography 2 was equal. The ranking of bandlike artifacts at the stent ends revealed that for MR angiography 1 the observers ranked 58.3% of evaluated stents as meeting the criteria of category 1, 31.3% as category 2, and 4.2% as category 3. For MR angiography 2, the respective values were 65.2%, 26.1%, and 0.0%. According to the Wilcoxon's signed ranked test, the comparison between the rankings on MR angiography 1 versus MR angiography 2 images was not statistically significant but showed a statistical trend (p = 0.074).
|
|
|
Interobserver Agreement
Regarding the evaluation of the degree of in-stent stenoses on DSA 1 and
DSA 2 images, Spearman's correlation coefficient was 0.67-0.84, revealing a
statistically significant positive relationship between the observers'
measurements (p < 0.0005). The only exception was the correlation
of two of the observers for evaluation of DSA 2 with Spearman's correlation
coefficient being 0.35 (p = 0.089).
As for the evaluation of the degree of instent stenoses on MR angiography 1 and 2 images, Spearman's correlation coefficient was 0.76-0.94 for all observers, revealing a statistically significant positive relationship (p < 0.0005).
Retrospective Power Analysis
The power to detect a difference of 10% in grading of stenoses on DSA and
MR angiography images was 96.5% for DSA 1 versus MR angiography 1 and 96.4%
for DSA 2 versus MR angiography 2.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In a prospective study, we evaluated a self-expanding nitinol stent that was implanted in 27 patients with chronic iliac stenoses. Three-dimensional MR angiography was performed within 3 days after stent placement and at follow-up 99-258 days later. Each time DSA was performed as the reference standard. According to this reference, the degree of residual and recurrent stenoses was low, most being less than 50%, and no stent occlusion occurred. This might be expected because balloon angioplasty in the stent had been performed after stent delivery, guaranteeing a high rate of technical success; and patency rates after iliac stenting are generally high [23-27]. On both MR angiography 1 and MR angiography 2, observers correctly identified patency of all stents. The assessment of severity of in-stent stenoses on DSA 1 and MR angiography 1 images (immediately after the procedure) differed significantly (Fig. 1A). The observers tended to overestimate rather than underestimate the degree of stenosis. Interestingly, observer performance was significantly better for MR angiography images at follow-up (MR angiography 2), no longer showing any statistically significant discrepancies from the standard of reference (DSA 2) (Fig. 1B). These numbers are promising in terms of MR angiography developing into a useful follow-up technique after stent implantation. However, even for MR angiography 2 there were single cases in which grading of stenosis differed by as much as 63% from the reference standard when an artifact was misinterpreted as stenosis. The magnitude of these artifacts depends on the orientation of the stent axis relative to the main magnetic field, B0, and direction of the frequency-encoding gradient and on flow-related dephasing [5-9, 28]. These conditions are difficult to predict. Thus, applying MR angiography as a routinely used follow-up technique after stent implantation cannot yet be recommended.
To our knowledge, the interesting phenomenon of an improving observer
performance with time after stent placement has not yet been described in
literature. The study was adequately powered so that differences in grading of
stenoses, particularly on DSA 2 versus MR angiography 2 images, would have
been detected. The sequence parameters and the protocol of contrast
administration were kept constant for all MR angiography examinations,
eliminating the imaging protocol as a possible reason for improved image
quality on MR angiography 2 versus MR angiography 1 images. However, we
observed that the in-stent signal was more homogeneous and the delineation of
the inner contour of the stents was better on MR angiography 2 than on MR
angiography 1 images (Figs. 3A,
3B,
3C, and
3D). In conjunction with the
dynamic observed for the bandlike artifacts at the stent ends that were ranked
as less pronounced on MR angiography 2 than on MR angiography 1 images, the
stent-related artifacts might have changed with time after stent placement. In
our opinion, decreasing susceptibility artifacts are most likely the reason
for the improved observer performance. Susceptibility artifacts are caused by
differences between the magnetic susceptibility of the stent material and the
surrounding tissue and lead to local field inhomogeneities. These field
inhomogeneities interfere with the imaging gradient fields, resulting in local
image distortions [29,
30]. The larger the
susceptibility gradient between the stent material and human tissue, the more
the magnetic field lines are distorted and the larger are the susceptibility
artifacts. Histologic studies have shown a complete covering of the stent
struts by tissue after a minimum of 1 month
[31]. A change in intima
distribution or in coating of the stent material over time might decrease the
susceptibility gradient between the stent material and blood and thus reduce
the artifact size. Further studies are needed to systematically evaluate
whether stent-related artifacts become less pronounced with time after stent
placement and influence image quality of MR angiography. One limitation of our
study was the small number of severe stenoses ( 50%), a consequence of the
high technical success rates and patency rates that are commonly seen in iliac
stenting. Further study of larger samples, including longer follow-up periods,
is warranted to evaluate the accuracy of MR angiography for the detection and
grading of severe stenoses in iliac vessels.
|
|
|
|
Our results might have been influenced by a changing orientation of the stent axis in relation to the main magnetic field because, as mentioned previously, the magnitude of susceptibility artifacts is a function of the orientation of the stent axis relative to the main magnetic field, B0, and the frequency-encoding gradient. After stent implantation, variations in the angulation or course of the stented vessel could occur, leading to a more parallel orientation of the stented vessel segment in relation to B0. However, the first MR angiography (MR angiography 1) was performed at least 24 hr and up to 3 days after stent placement. It can be assumed that a potential rearrangement of the vessel and stent orientation took place during this period and thus did not alter the imaging conditions between MR angiography 1 and MR angiography 2.
In conclusion, evaluation of MR angiography images after implantation of self-expandable nitinol stents in iliac arteries revealed that patency was correctly assessed for all stents on MR angiography. Although grading of in-stent stenoses on MR angiography was promising, MR angiography still does not to seem to be a reliable technique for the characterization of in-stent stenoses because in single cases discrepancies of more than 60% occurred between the grading of lumen narrowing on DSA and the grading on MR angiography images. Interestingly, observer performance was improved on MR angiography images at the 6-month follow-up compared with images acquired immediately after stent placement. This phenomenon needs to be further evaluated.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  O W Hamer, I Borisch, C Paetzel, W R Nitz, J Seitz, S Feuerbach, and N Zorger In vitro evaluation of stent patency and in-stent stenoses in 10 metallic stents using MR angiography Br. J. Radiol., August 1, 2006; 79(944): 636 - 643. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
