| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Institute of Diagnostic Radiology, University Hospital Basel, Petersgraben 4,
4031 Basel, Switzerland.
2 Biocenter, University of Basel, 4056 Basel, Switzerland.
3 Department of Angiology, University Hospital Basel, Petersgraben 4, 4031
Basel, Switzerland.
Received April 13, 2006;
accepted after revision July 31, 2006.
Address correspondence to S. Potthast
(spotthast{at}uhbs.ch).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Intraarterial MR aortography was performed at 1.5 T on nine patients with peripheral arterial occlusive disease and in an aortic phantom with pulsatile flow. A 3D fast low-angle shot MRI sequence was used for standard technique (acquisition time, 20 seconds) and for parallel acquisition technique (acquisition time, 14 seconds). In all patients, a pigtail catheter was left in the suprarenal position after digital subtraction angiography. Contrast-enhanced intraarterial MR aortography was performed after automated injection of 50 mmol/L gadoterate dimeglumine at an injection rate of 4 mL/s. Contrast-to-noise ratio (CNR) and image quality were evaluated in both imaging series at different locations. In an aortic phantom with pulsatile flow, CNR was determined 1, 30, and 60 cm distal to the catheter tip with standard and parallel acquisition techniques.
RESULTS. In all patients, intraarterial MR aortography was feasible with both acquisition techniques. No significant difference in CNR or image quality was observed in the patient study. Similar results were calculated for the pulsatile aortic flow phantom at all locations.
CONCLUSION. Intraarterial MR aortography is feasible with parallel acquisition technique without a significant loss of CNR. This technique reduces contrast agent consumption approximately 30% owing to an approximately 30% reduction in acquisition time.
Keywords: angiography • arteriography • digital subtraction angiography • MR aortography • MR contrast agents • MR technique
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
An important precondition for localizing stenosis and occlusion in MR angiography is to achieve a high contrast-to-noise ratio (CNR) in the vessel lumen. Theoretic and animal models and initial investigations in patients have shown encouraging results in the development of high-spatial-resolution 3D contrast-enhanced MR angiography with intraarterial contrast injection [1, 5-13]. Because of the short half-life of gadolinium chelates in the circulating blood, several injections are mandatory during endovascular interventions.
Protocols for low-dose injection of gadolinium chelates are necessary in order not to exceed the daily dose limit 0.3 mmol/L/kg body weight set by the U.S. Food and Drug Administration. The total amount of injected gadolinium chelates is based on the concentration of injected contrast agent, the injection rate, and the duration of injection [14]. Various technical possibilities are conceivable for reducing the amount of contrast agent per injection.
Because of the paramagnetic properties of gadolinium-based contrast agents, blood gadolinium concentration must meet a target range during MR angiography data acquisition to achieve adequate arterial enhancement [5-7, 14, 15]. A low-dose injection protocol for intraarterial 3D contrast-enhanced MR angiography of the infrainguinal arteries has been suggested [1] that assures acceptable CNR values. In theoretic models the injected gadolinium chelate concentration and injection rates varied and were validated successfully in patients. In accordance with this protocol, an injection protocol for intraarterial MR aortography in humans was derived with an injection rate of 4 mL/s and a gadolinium chelate concentration of 50 mmol/L, which is the lowest acceptable dose for obtaining an adequate CNR [12]. Because of the high aortic blood flow rate, approximately 30 mL/s, only a limited number of intraaortic injections can be performed to meet the maximum dose approved by the Food and Drug Administration, even with low-dose protocols. Usually, however, complex endovascular interventions such as endovascular aortal stenting require several injections.
Another option for reducing the amount of gadolinium injected is to use faster acquisition techniques, such as the parallel acquisition techniques sensitivity encoding [16, 17] and generalized autocalibrating partially parallel acquisition (GRAPPA) [18]. These techniques accelerate MRI data acquisition by a factor of two or more. In GRAPPA, a k-space-based reconstruction algorithm, additional reference k-lines are used for self-cal-ibration of coil sensitivities for each image, avoiding artifacts due to patient motion. Simplified, the reduction in acquisition time with GRAPPA is possible because missing k-space lines are generated from a series expansion of reduced-k-space raw data sets in two coils with distinct sensitivities: one homogeneous volume coil and one coil with a constant sensitivity gradient [19]. The CNR is reduced by a factor of at least the square root of two in accelerated acquisition.
In this study, we evaluated CNR and image quality in different vessel diameters by using 3D fast low-angle shot (FLASH) gradientecho MR angiography with standard acquisition and with a parallel acquisition technique (GRAPPA) after intraarterial contrast injection in patients and in an aortic flow phantom.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Patient Study
The study was performed in accordance with the guidelines issued by the
local institutional review board. Written informed consent was obtained from
all patients before inclusion. Exclusion criteria for this study consisted of
all generally accepted contraindications to MRI, such as pacemakers,
implantable cardioverter defibrillators, and claustrophobia.
Inclusion criteria were arterial occlusive disease in patients who underwent endovascular intervention with retrograde percutaneous puncture of the common femoral artery. Nine patients (five men and four women, 56-90 years old) enrolled in the study. Seven patients had arterial occlusive disease of the iliac axis, proximal femoral arteries, or the subclavian artery; one patient had renal artery stenosis on the left side; and one patient had high-grade stenosis of a femorocrural bypass on the right side (Table 1).
|
Retrograde percutaneous puncture of the common femoral artery was performed under radiographic guidance. For reference purposes, diagnostic angiography of the aortic runoff was performed in anteroposterior projection for placement of a catheter (Omniflush, AngioDynamics) in the suprarenal position at the level of the first lumbar vertebra. After angiography and angioplasty, a 6-French introducer sheath was left in the common femoral artery, and the tip of a catheter (Omniflush) was placed at the same suprarenal position. A permanent saline flush guaranteed patency of the device during transport from the angiography suite to the MR unit. Patients underwent imaging in the supine position, entering the scanner feet first to facilitate communication and catheter handling.
In the standard protocol, acquisition time was 20.3 seconds, resulting in a total of 81.2 mL of contrast agent (8.1 mL stock solution). For parallel acquisition technique, acquisition time was 14 seconds, resulting in a total of 56 mL of contrast agent (5.6 mL stock solution). Scan delay between the two acquisition techniques was at least 7 minutes.
Gadoterate dimeglumine (Dotarem, Guerbet) distributed in full-strength concentration of 500 mmol/L was used as the contrast agent. The stock solution of gadoterate dimeglumine was diluted with physiologic saline solution to obtain a gadolinium concentration of 50 mmol/L. This solution was administered automatically with a power injector (Spectris, Medrad) at an injection rate of 4 mL/s during acquisition.
Aortic Flow Phantom
The aortic flow phantom was an open system with a 60-L container that
served as the reservoir (Fig.
1). An acrylic tube with an inner diameter of 12 mm represented
the human infrarenal aorta. The insertion sheath was attached to the tube via
Y-connector, and a catheter (Omniflush) was inserted. The container was filled
with a 1:4,000 solution of diluted gadolinium and physiologic saline solution
to achieve the T1-weighted effect of blood. This solution was pumped through
the tube with a pulsatile flow of 30 mL/s generated by a motorized pump (Cobe
Laboratories) and flowed back into the reservoir. A flow rate of 30 mL/s was
chosen because the human aortic blood flow rate is 
30 mL/s. The tube was
placed in a box filled with 100 mL of gel consisting of 0.02 mL of gadoterate
dimeglumine (Dotarem) and 2.2 mL of ferumoxsil (Lumiren, Guerbet), 100 mL of
hydroxyethyl cellulose 3%, and 0.01% benzalkonium chloride simulating the T1-
and T2-weighted effects of muscle tissue. As in the patient study, a contrast
solution with a concentration of 50 mmol/L gadoterate dimeglumine was
administered at an injection rate of 4 mL/s during data acquisition.
|
|
Data Analysis
Patient study—Maximum intensity projections of all patients
were calculated from the contrast-enhanced images. On unsubtracted source
images, signal intensity was measured in five selected regions with
homogeneous signal intensity within the vessel (SIblood):
infrarenal aorta (SIaorta), both renal arteries (SIra),
and both common iliac arteries (SIciar/SIcial).
Soft-tissue signal intensity was measured in gluteal muscle
(SItissue), and the SD of background noise was measured in the air
(SDair) outside the body (Fig.
2).
The sizes of the regions of interest were not changed in the different regions and ranged from five to 125 pixels. The average of arterial signal intensity (SIblood) and CNR in the arterial lumen was calculated according to the following equation: CNR = (SIblood - SItissue)/SDair. The CNR in the infrarenal aorta and the common iliac arteries and the average CNR of the arteries were calculated for the standard technique and for parallel acquisition MR angiography.
Intraindividual CNRs were compared and statistically analyzed with Student's t test. The level of statistically significant difference was set at p < 0.05. Two observers with extensive experience in digital subtraction angiography and MR angiography used consensus to evaluate on film the diagnostic quality of the images obtained with standard technique and with parallel acquisition MR angiography. Both observers were blinded to imaging protocol and diagnosis. Image quality for evaluation of the renal arteries, iliac arteries, and infrarenal aorta was subjectively graded 1, excellent; 2, good; 3, fair; or 4, poor.
Aortic flow phantom—In the aortic phantom of pulsatile flow, intraluminal signal intensity was measured 1, 30, and 60 cm distal to the tip of the catheter for regions of interest along the tube. CNR was computed according to the equation CNR = (SIblood - SItissue)/SDair, where SIblood is the signal intensity of the tube lumen, SItissue the signal intensity of the muscle-equivalent fluid, and SDair the SD of air measured outside the box in each acquisition. CNR evaluation was performed for standard acquisition and parallel acquisition MR angiography.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Figures 3A, 3B, 3C through 5 are examples of maximum-intensity-projection images of three patients obtained with low-dose intraarterial 3D contrast-enhanced MR angiography with the standard technique (acquisition time, 20.3 seconds), parallel acquisition MR angiograms with an acceleration factor of 2 (acquisition time, 14 seconds), and the corresponding digital subtraction angiograms. Figure 3A, 3B, 3C shows the images of a patient with high-grade stenosis of the left external iliac artery, which was treated with a stent. Figure 4A, 4B, 4C shows images acquired after percutaneous transluminal angioplasty of both external iliac arteries. The digital subtraction angiogram shows a lower-pole artery on the right side, which was well shown with standard MR technique and with parallel acquisition MR angiography. Figure 5A, 5B, 5C shows high-grade stenosis of the common iliac artery on the right side. After angioplasty, runoff in the periphery was well visualized with intraarterial MR angiography. All parallel acquisition MR angiograms were acquired after standard acquisitions; therefore the ureters show contrast enhancement in the parallel acquisition series.
|
|
|
|
|
|
|
|
|
Table 2 shows the CNR of each patient at each location with standard technique and with parallel acquisition MR angiography. Table 3 summarizes mean CNR measured from standard 3D contrast-enhanced MR angiograms and from parallel acquisition MR angiograms at all locations. CNR with standard acquisition technique ranged from 31.5 in the renal arteries to 117.0 in the aorta. With parallel acquisition technique, CNR ranged from 32.0 in the renal arteries to 125.5 in the aorta. The SD was very high for the standard technique, ranging from 20.5 to 60.6, and was even higher for parallel acquisition MR angiography, ranging from 25.0 to 87.6. Student's t test showed no significant difference in mean CNR in comparisons of standard acquisition technique and parallel acquisition MR angiography for the aorta (p = 0.63), renal arteries (p = 0.88), iliac arteries (p = 0.25), and overall arteries (p =0.98).
|
|
Aortic Flow Phantom
Mean CNR and SD calculated 1, 30, and 60 cm distal to the tip of the
catheter from the standard 3D FLASH MR angiograms (acquisition time, 28
seconds) and from parallel acquisition MR angiograms (acquisition time, 19
seconds) are shown in Figure
6. Mean CNR ranged from 230 to 585 with the standard technique and
from 290 to 717 with parallel acquisition technique. At all three locations,
an increase in CNR of 15-26% was calculated for MR angiograms acquired with
parallel acquisition technique.
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A reduction in duration of bolus administration during data acquisition was been shown in a phantom model with use of partial k-space acquisition [5, 20, 21]. In the model, coverage of 50% of the k-space is sufficient to gain an adequate signal-to-noise ratio (SNR) in major arteries, such as the aorta and iliac arteries. For smaller vessels, such as the renal arteries, data acquisition of the outer k-space lines is mandatory for an adequate SNR.
MR angiographic techniques without contrast enhancement, such as time-of-flight MR angiography and 3D true fast imaging with steady-state free precession MR angiography, may be an option for reducing the amount of contrast agent. During interventions, however, time-of-flight MR angiography can be too time consuming because acquisition lasts for approximately 2-4 minutes. True fast imaging with steady-state free precession MR angiography may be an interesting option for larger vessels.
Another option for reducing acquisition time and therefore duration of
bolus administration is the use of parallel acquisition technique, as in our
patient study. Visualization scoring showed no difference in the visualization
of arteries with standard and parallel techniques. In theory, parallel
acquisition technique is used at the cost of CNR. The CNR of parallel
acquisition technique, CNR-PAT, can be described in the following way: CNR -
PAT = CNRno - PAT/g
R, where PAT is parallel acquisition
technique, CNRno is CNR with standard technique, g is a spatially
dependent factor that depends on exact coil element geometry
[16], and R is the
acceleration factor (in our investigation, R = 2). R results from the
reduced number of excitations of the image plane. For GRAPPA, the CNR of the
autocalibration lines can be incorporated into the final images. Therefore,
the effective R with GRAPPA is less than 2. With development of array coils
specifically designed for parallel acquisition technique, the g-factor should
approach its ideal of 1 more closely
[16,
22].
In contrast to the theoretically expected decrease in CNR for parallel acquisition MR angiography, comparable CNRs were obtained for standard and parallel acquisition MR angiography in our patient study when conventional evaluations of CNR were performed. Furthermore, higher CNR was obtained for parallel acquisition MR angiography in the aortic flow phantom. Because the contrast agent was administered for the entire acquisition time and the injection rate was not changed in the two acquisition protocols, variation of gadolinium chelate concentration during MRI can be excluded. Recent articles [22-25] have dealt with parallel acquisition technique and analyzed its influence on CNR. Although all of the studies were based on use of IV contrast agents, some articles describe a reduction in SNR and CNR [22, 25], whereas others show no significant decrease in SNR or CNR with use of parallel acquisition technique [23, 24].
The main reason for different results for SNR and CNR in parallel acquisition technique is that noise varies greatly across the field of view when the technique is used. This variation is described with the aforementioned g-factor. Therefore, SNR or CNR that relies on measurements of noise from regions of interest in the signal-free background, air, are inappropriate for parallel acquisition technique.
There are three main ways to directly estimate noise from reconstructed images [26]. The first is estimation of background noise from a noiseonly region manually selected within the field of view. This method may be most commonly used [27, 28]. The second way is to estimate the noise on the basis of temporal differentiation [29, 30]. The third is to estimate noise after spatial differentiation [31]. Because of lack of homogeneity of the field of view, the first approach to CNR evaluation is not appropriate for parallel acquisition MR angiography, but it is the most often described in the literature [27, 28]. The second approach to noise estimation is based on repetitive measurements to subtract images [29]. This method is appropriate for noise evaluation in parallel acquisition MR angiography, but it cannot be performed on patients because repetitive measurements are impossible, for example, because of the large amount of contrast agent administered. The third approach, estimation of noise based on spatial differentiation, is limited to large regions with only little signal inhomogeneity. Therefore, this approach is limited to standard MR angiography [31].
Our study had several limitations. First, the image analysis did not include the diagnostic value of intraarterial MR aortography performed with standard technique and with parallel acquisition MR angiography. This topic is the subject of ongoing research. Second, our aortic flow phantom and patient studies were limited to the larger arteries, but most endovascular interventions are performed in the arteries of the lower extremity. Further phantom and human studies are necessary to investigate parallel acquisition techniques and the possibility of contrast agent saving in smaller vessels. Third, our results were limited to only a small number of patients.
Major drawbacks of MR endovascular interventions are the poor visibility of conventional interventional devices and safety concerns with respect to conductive guidewires. The development of nonconductive wires affords the opportunity for future work. An MRI-compatible polymer-based guidewire has been introduced [4].
In conclusion, preliminary results show that MR angiography with parallel acquisition technique such as GRAPPA is feasible for intraarterial 3D contrast-enhanced MR aortography without a significant loss of CNR in patients. The 30% reduction in acquisition at R = 2 reduces the amount of contrast agent administered to approximately the same degree. Therefore, parallel acquisition technique may be a valuable tool for reducing the amount of contrast agent used. Because endovascular interventional procedures require repeated injections of contrast agent, it is mandatory that adequate arterial enhancement be obtained with only low-dose intraarterial injections. Higher parallel acquisition technique factors may enable further reduction in the amount of contrast agent needed and therefore result in cost saving.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
