| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Laurie Imaging Center, University Radiology Group, University of Medicine and
Dentistry New Jersey, 141 French St., New Brunswick, NJ 08901.
2
Department of Diagnostic Radiology, Mayo Clinic and Foundation, 200 First St.
S.W., Rochester, MN 55905.
Received August 23, 1999;
accepted after revision June 15, 2000.
Address correspondence to J. K. De Marco.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Sixty-three patients were enrolled. A 2-mL test bolus and commercially available software were used to time the gadolinium bolus. High-resolution contrast-enhanced MR angiography was performed with elliptic centric acquisition.
RESULTS. The average time of bolus arrival was 17.3 sec (range, 12-25 sec). In 60 of the 63 patients, we had excellent or good visualization of the carotid bifurcation using contrast-enhanced MR angiography with little or no venous contamination. Two observers ranked delineation of stenosis and morphology of proximal internal carotid artery and overall diagnostic confidence statistically significantly higher for contrast-enhanced MR angiography compared with MOTSA. Ulceration, length of stenosis, and slow flow distal to a critical stenosis were better depicted with contrast-enhanced MR angiography than with MOTSA. Venetian blind artifact, saturation of slow or in-plane flow, and artifactual narrowing in carotid artery kinks plagued MOTSA but were not seen on contrast-enhanced MR angiography. MOTSA was graded superior to contrast-enhanced MR angiography in nine of 120 carotid bifurcations analyzed.
CONCLUSION. High-resolution carotid contrast-enhanced MR angiography is technically feasible. We found a 95% success rate using commercially available hardware and software. The image quality with carotid contrast-enhanced MR angiography has improved so that it is equal or superior to the longer MOTSA in most patients.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
To overcome limitations in gadolinium-bolus timing while striving for high resolution, we decided to combine two commercially available triggering sequences. MR computer-automated scan technology (SmartPrep; General Electric Medical Systems, Milwaukee, WI) is a new technique for detecting the bolus arrival during a user-defined detection interval by monitoring signal intensity from the tracking volume placed on the vessel of interest and by automatically initiating acquisition of contrast-enhanced studies [9]. We elected to individualize this detection interval for each patient by estimating the time-to-bolus arrival with a small test injection of gadolinium.
In addition to more accurate gadolinium-bolus timing, the centric view-ordering contrast-enhanced MR angiography was replaced with an experimental elliptic centric phase-encoding view-ordering MR angiography [10]. This technique provides a quicker trajectory through the contrast portions of 3D k-space. Elliptic centric phase encoding may therefore provide better separation between the contrast enhancement within the arteries and the veins compared with standard centric ordering, thereby allowing longer acquisitions while still providing excellent arterial enhancement with little venous contrast. Higher resolution in the phase-encoding and slice-encoding directions could be achieved during this longer 3D acquisition. The use of elliptic centric view ordering combined with careful bolus timing may result in higher resolution contrast-enhanced MR angiography with excellent arterial enhancement and less venous contamination compared with standard centric ordering. A version of elliptic MR angiography similar to that described in this study is now commercially available. In addition, other major manufactures are evaluating an implementation of elliptic MR angiography.
We hypothesized that the combination of improved gadolinium-bolus timing and the use of elliptic MR angiography could result in a high success rate of contrast-enhanced carotid MR angiography with little or no venous contamination. We report on an ongoing study to evaluate the technical feasibility of high-resolution contrast-enhanced MR angiography of the carotid arteries with commercially available hardware and software.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Timing Sequence
After the MOTSA MR angiography was completed, a timing sequence consisting
of a T1-weighted spoiled gradient-echo axial image was placed at the level of
the distal common carotid artery near the bifurcation. The timing sequence had
the following parameters: TR/TE, 20/4; flip angle, 70°; field of view, 20
x 15 cm; matrix size, 256 x 128; excitations, one. The slice
thickness was 20 mm. Inferior and superior saturation bands were used to
suppress unwanted flow-related enhancement. Images were obtained every 2 sec
beginning with the bolus injection of 2 mL of gadopentetate dimeglumine
followed by 10 mL of normal saline solution; both solutions delivered a rate
of 3 mL/sec with a power injector (Spectris, Medrad). The technologists
reviewed the data and visually determined the image that showed peak
enhancement. Because the images were obtained every 2 sec, simply multiplying
the image number by two yielded the time-to-peak enhancement. An additional 2
sec were added to the time-to-peak enhancement to insure that the elliptic MR
angiography would begin in the plateau phase of the large injection of
contrast material. This time was a combination of the time-to-peak enhancement
and a delay equal to half the time it takes to inject the large gadolinium
bolus. For simplicity, this time was called the bolus-arrival time. The
maximum detection time for the MR SmartPrep was set to the bolus-arrival time.
If the MR SmartPrep showed the gadolinium bolus before the estimated
bolus-arrival time, the contrast-enhanced MR angiography sequence would be
initiated. If there was no gadolinium-bolus detection, the contrast-enhanced
MR angiography sequence would begin at the estimated bolus-arrival time. The
tracking volume was graphically placed within the mid-to-distal aspect of the
left common carotid artery with a 34-sec phase contrast-enhanced MR
angiography as previously described (De Marco et al., presented at the IX
Interventional Workshop on Magnetic Resonance Angiography, October 1997). The
tracking volume was set to 3- to 5-cm long and 20-mm wide. Before beginning MR
angiography, a right antecubital vein was accessed for injection of gadolinium
contrast media. We constantly flushed the 18-g angiocath (Jelco; Johnson &
Johnson, Arlington, TX) with saline solution, using the power injector.
Elliptic MR Angiographic Sequence
At the same time that the MR SmartPrep sequence was begun, a large
gadopentetate dimeglumine bolus (18 mL) was injected at 3 mL/sec followed by a
30-mL bolus of saline solution. No contrast material reactions or adverse
events occurred during the contrast-enhanced portion of the studies. The MR
SmartPrep sequence would trigger the contrast-enhanced MR angiography sequence
to begin as described previously. The contrast-enhanced MR angiography was a
research sequence with elliptic centric phase-encoding view ordering. As
previously described, the standard nested phase-encoding and slice-encoding
loops were replaced by a single loop (Bernstein et al., presented at the
International Society for Magnetic Resonance in Medicine meeting, April 1998).
The view ordering within this single loop was determined by the distance to
the origin in the kz-ky- plane. Thus elliptic centric is a true
centric k-space ordering for 3D acquisition. When we combined it with careful
bolus timing, this sequence provided an intrinsically high degree of venous
suppression. A longer acquisition with more partitions for improved coverage
or thinner partitions and strong arterial enhancement with little venous
contamination are possible. Contrast-enhanced MR angiography included the
following parameters: TR/TE, 5.6/1.4; bandwidth, ±62.5 MHz; flip angle,
45°; field of view, 22 x 18 cm; coronal partitions, 36, 1.4-mm
thick; matrix size, 256 x 224; excitations, one; scan time, 30 sec. The
resolution in the frequency direction was 220 mm per 256 pixels or 0.86 mm per
pixel. In the phase direction, the resolution was 180 mm per 183 pixels or
0.98 mm per pixel. Note the y-matrix of 183 represented the product
of the prescribed value of 224 multiplied by the field-of-view factor of
18/22. Thus, the actual voxel size was 0.86 (x) x 0.98
(y) x 1.4 mm (z), yielding a voxel volume of 1.18
mm3. Zero filling provided overlapping sync-interpolated voxels and
thereby reduced partial volume artifacts. As shown by Du et al.
[11], this technique can
improve the detectability of small vessels. The reconstruction used zero
filling in all three directions to yield 72 1.4-mm-thick partitions with
0.7-mm overlap and a matrix of 512 x 366.
Evaluation of MR Angiographic Sequences
Global maximum intensity projections (MIP) of the contrast-enhanced MR
angiograms were presented to two observers who evaluated them for arterial
enhancement, venous signal intensity, and motion artifact, using a
modification of a previously described protocol
[7]. The evaluation criteria
for arterial enhancement of the contrast-enhanced MR angiograms were the
following: excellent (intense intraarterial enhancement), good (more than
adequate for diagnosis), fair (adequate for diagnosis), and poor (not adequate
for diagnosis). The following were the criteria for venous signal intensity:
none, mild with barely detectable venous signal intensity, moderate with
noticeable venous signal intensity, and severe with venous signal intensity
equal to or greater than arterial signal intensity. The following criteria for
motion artifact were used: none, minimal artifact; moderate artifact but
diagnostic quality; and severe artifact rendering the examination
nondiagnostic. After reviewing the global MIP separately, during a second
session the reviewers reached a consensus grading for each criterion.
We directly compared the MOTSA and contrast-enhanced MR angiography in all 63 patients by presenting four images from selective subvolume MIP targeted to the right and left carotid bifurcations separately. The four images were the anterior, 30° and 60° ipsilateral oblique, and the lateral projections. They were magnified so that the resulting images had identical scaling. Both observers independently ranked the MOTSA and contrast-enhanced MR angiography in an unblinded fashion on the basis of the following criteria: delineation of the stenotic segment, delineation of the morphology of the proximal internal carotid artery, intravascular signal intensity, and diagnostic confidence with a three-point scale as has been previously described [8]. Later, these same selective subvolume MIP images targeted to the bifurcation from the contrast-enhanced MR angiography and MOTSA MR angiography were compared with the following established evaluation criteria: contrast-enhanced MR angiography markedly superior to MOTSA, contrast-enhanced MR angiography superior to MOTSA, comparable quality, contrast-enhanced MR angiography inferior to MOTSA, and contrast-enhanced MR angiography markedly inferior to MOTSA [7]. The two observers reviewed the two MR angiograms side by side and reached a consensus grading.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
With careful placement of the tracking volume within the mid aspect of the common carotid artery, we found that MR SmartPrep revealed the arrival of the large contrast bolus earlier than predicted by the timing run in 19 (30%) of the examinations and triggered the contrast-enhanced MR angiography earlier than predicted by the bolus technique alone. The elliptic contrast-enhanced MR angiography sequence was triggered on average 1.7 ± 0.8 sec (SD) earlier with minimal and maximal values of 1 and 4 sec respectively. In all but one patient, the resulting contrast-enhanced MR angiography was graded as excellent. In the remaining patient, the contrast-enhanced MR angiography was good. No or little venous enhancement was detected, and no portion of the arterial anatomy was obscured.
Both observers separately graded the arterial enhancement of the contrast-enhanced MR angiograms highly and then later in a second session by consensus. In 90% of the examinations, contrast-enhanced MR angiography was graded as either excellent (n = 45) or more than adequate for diagnosis (n = 12). Three additional contrast-enhanced MR angiograms were graded as adequate for diagnosis. In the three remaining patients, contrast-enhanced MR angiography was begun too early with little or no intravascular contrast enhancement noted and was graded as poor.
In no case did examples of venous enhancement obscure arterial anatomy. In the 19 patients in whom there was mild venous enhancement, it mostly occurred at the level of the jugular bulb and was easily removed with targeted MIP imaging. There were no examples of motion that could be detected on any of the contrast-enhanced MR angiographic examinations. By comparison, four of the 63 MOTSA MR angiograms were sufficiently degraded by motion to limit their interpretability.
Two observers ranked the MOTSA MR angiography and contrast-enhanced MR angiography in the 63 patients using four criteria: delineation of stenosis, delineation of the morphology of the internal carotid artery, intravascular signal intensity, and diagnostic confidence. Identical projections with identical magnification from the MOTSA MR angiography and contrast-enhanced MR angiography were presented to the observers. The results are summarized in Table 1. Both observers ranked delineation of stenosis, internal carotid artery morphology, and diagnostic confidence higher for contrast-enhanced MR angiography than for MOTSA MR angiography. Both the Wilcoxon's signed rank test (p < 0.001) and McNemar test (p < 0.001) showed these differences to be significant for both observers. There was no statistically significant difference in intravascular signal intensity between the two methods of MR angiography for either observer.
|
During a second session, both observers directly compared the targeted subvolume MIP images of each carotid bifurcation in all 63 patients. Of the 126 carotid bifurcations evaluated, the contrast-enhanced MR angiography was graded as far superior (n = 8), superior (n = 37), or equivalent (n = 66) to MOTSA MR angiography in most patients. The MOTSA MR angiography was superior (n = 9) or far superior (n = 6) to contrast-enhanced MR angiography in a minority of patients. The latter category represented the three patients in whom the timing of the contrast bolus failed. Similar results were seen when contrast-enhanced MR angiography and MOTSA MR angiography were compared by grade of stenosis (Table 2). Contrast-enhanced MR angiography was superior to MOTSA MR angiography in depicting ulcerations and length of stenosis, in revealing slow flow distal to a critical stenosis, in being insensitive to venetian blind artifact [13], and in not overestimating stenosis at the level of kinks (Figs. 1A,1B,1C,2A,2B,3A,3B). Shortcomings of elliptic contrast-enhanced MR angiography occurred with unsuccessful timing in three of the 63 patients. There were five patients (9/10 carotid bifurcations) in whom MOTSA MR angiograms were graded as superior to contrast-enhanced MR angiograms (Fig. 4A,4B). In one of these five patients, the contrast-enhanced MR angiography was triggered slightly early and resulted in a typical edge-enhancement artifact.
|
|
|
|
|
|
|
|
|
|
Overall the mid and distal aspects of both common carotid arteries, the carotid bifurcations, and the remaining cervical portions of the internal carotid arteries to near the skull base were well visualized. These contrast-enhanced MR angiograms were graded as good or excellent in 60 of the 63 patients. No portions of these vessels were excluded from contrast-enhanced MR angiography. The coverage afforded by elliptic contrast-enhanced MR angiography was sufficient to include the mid and distal common carotid artery and the cervical portion of the internal carotid artery.
The first 41 patients were imaged with one neurovascular coil. In 16 patients, there was signal dropout of the inferior aspect of the field of view. The proximal aspects of both common carotid arteries could not be well seen. There have been modifications to this coil to improve the homogeneity of the signal intensity across the field of view, but we did not have access to these modifications. It was not our practice to attempt to include the aortic arch or takeoff of the great vessels when we were using this coil. The last 22 patients were scanned with the second neurovascular coil. No signal dropout was seen across the field of view, except in one patient in whom there was mild drop-off in signal intensity that did not limit our ability to see the proximal common carotid artery. We did attempt to position the contrast-enhanced MR angiographic volume to include the aortic arch and takeoff of the great vessels. In three patients, the technologists did not center sufficiently low enough to include the aortic arch. In one patient, the proximal right common carotid artery had a large anterior kink that carried it out of the contrast-enhanced MR angiographic acquisition volume. In the remaining 18 patients, the aortic arch and the origin and proximal aspects of both common carotid arteries were well visualized along with the mid and distal aspects of the common carotid artery, carotid bifurcation, and cervical portions of both internal carotid arteries. No significant stenosis of the common carotid arteries was seen. However, two high-grade stenoses in the proximal left subclavian artery were noted (Fig. 5A,5B).
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Elliptic centric phase ordering provided a quicker trajectory through the contrast portions of 3D k-space. This speed provided a better separation between the contrast enhancement with the arteries and the veins in the neck compared with the standard centric ordering. Longer MR angiographic sequences were possible with elliptic centric phase ordering while still providing excellent separation between arterial and venous enhancement. Higher resolution in the phase-encoding and slice-encoding directions was achieved during this longer 3D acquisition. The use of elliptic centric view ordering combined with careful bolus timing resulted in higher resolution contrast-enhanced MR angiography with excellent arterial enhancement and less venous contamination compared with standard centric ordering.
When reviewing the global MIP images of all the neck vessels, we found that 60 of the 63 patients had excellent (n = 51) or good (n = 9) visualization of the carotid arteries with contrast-enhanced MR angiography with little or no venous contamination. In the remaining three patients, the contrast-enhanced MR angiography was triggered too early with little contrast enhancement in either the arteries or veins. This 95% success rate is an improvement over the 81% success rate we previously reported using MR SmartPrep alone with conventional centric view ordering (De Marco et al., International Workshop on Magnetic Resonance Angiography, October 1997). Previously, 38% of the contrast-enhanced MR angiograms showed moderate or severe venous contamination that at least partially obscured arterial anatomy distal to the carotid bifurcation. Therefore, the overall success rate of contrast-enhanced MR angiography in displaying arterial anatomy from the mid common carotid artery to the upper cervical portion of the internal carotid artery had been 62%. None of the current 63 contrast-enhanced MR angiographic examinations had moderate or severe venous contamination. The overall success rate for contrast-enhanced MR angiography from June to August 1998 was 95%. We believe the reliable venous suppression that is now possible with this new technique compared with conventional centric view ordering is even more impressive, considering that the imaging time has increased from 12 to 30 sec. This success illustrates the excellent venous suppression possible with elliptic centric view ordering and careful bolus timing.
Contrast-enhanced carotid MR angiography has been shown to be sensitive to slow flow, providing excellent coverage of the carotid artery and avoiding venetian blind artifact. This improvement makes 30-sec contrast-enhanced carotid MR angiography an excellent replacement for 6-8 min two-dimensional time-of-flight MR angiography for imaging throughout the neck. Further, contrast-enhanced MR angiography was shown to be superior to MOTSA MR angiography in depicting outpouchings consistent with ulcerated plaque at the carotid bifurcation, in depicting the overall length of stenosis, and in depicting complex vascular anatomy such as the vessel diameter at the level of kinks in the internal carotid artery. Contrast-enhanced carotid MR angiography is less sensitive to motion than MOTSA MR angiography. Statistically, the grading of carotid disease was significantly better with contrast-enhanced MR angiography compared with MOTSA MR angiography. However, in a few cases MOTSA MR angiography depicted the carotid bifurcation stenosis better. This difference was most likely due to relative decrease in signal-to-noise ratio. By decreasing the bandwidth from ±62 to ±32 MHz, the signal-to-noise ratio will improve by approximately 40%. This change may increase the quality of contrast-enhanced MR angiography at the cost of increasing the examination time from 30 to 44 sec. Preliminary data suggest that venous contamination will not obscure carotid arterial anatomy even with a 44-sec scan with elliptic contrast-enhanced MR angiography.
The major pitfall of the elliptic centric phase-encoding view ordering is its sensitivity to accurate triggering during the arrival of the gadolinium bolus. If the sequence is initiated early, there may be no visualization of the gadolinium within the carotid arteries. This problem occurred in three of the 63 patients in this study. If the elliptic centric phase-encoding view-ordering contrast-enhanced MR angiography is begun just slightly early, a typical edge-enhancement artifact will occur. The edge of the artery is markedly enhanced, whereas the central portion is only minimally enhanced. This pitfall was seen in one patient.
Accurate timing of the large gadolinium bolus is paramount in achieving consistent high-quality contrast-enhanced MR angiograms. One approach is to negate the need for timing by taking advantage of the high temporal resolution that is possible with contrast-enhanced MR angiography [6, 14]. It is possible to prescribe 3D contrast-enhanced MR angiography through the carotid bifurcation that can be performed every 5-9 sec. However, this technique cannot match the higher resolution possible with a single contrast-enhanced MR angiography acquisition as described previously. Higher resolution time-resolved contrast-enhanced MR angiography with partial k-space updating and more frequent sampling of the low k-space frequencies has been reported [8]. Currently, the offline processing required with this technique limits its clinical application. Real-time MR fluoroscopic triggering combined with elliptic centric phase-encoding view ordering has been shown to provide high-quality venous-suppressed contrast-enhanced MR angiography with high resolution [7] using a similar single-pass elliptic centric phase-encoding view-ordering contrast-enhanced MR angiography as described in this study. The hardware-modified MR imaging necessary to implement fluoroscopic triggering limits its current clinical applicability.
The impetus for this study was to design an imaging protocol that achieved the highest possible resolution contrast-enhanced MR angiography with techniques that could be implemented immediately on our clinical MR scanner and could allow us rapid acquisition and postprocessing locally to meet our busy clinical practice. We combined a timing bolus and SmartPrep in the hopes of achieving accurate triggering of contrast-enhanced MR angiography without waiting for a hardware-modified MR scanner capable of fluoroscopic triggering. Elliptic centric phase-encoding view ordering could be implemented on our clinical MR scanner without modification. This advantage allowed us the opportunity to evaluate high-resolution contrast-enhanced MR angiography without waiting for faster postprocessing of the time-resolved contrast-enhanced MR angiography. Since this study concluded, one manufacturer (General Electric Medical Systems) provides elliptic centric phase-encoding view ordering in its current software release.
Combining SmartPrep with the information from a test bolus does more than simply add automation to the triggering of the carotid contrast-enhanced MR angiography. It adds a second way to detect the arrival of the large gadolinium bolus. Indeed 30% of the time in this study, the SmartPrep technique revealed the gadolinium bolus and triggered the contrast-enhanced MR angiography earlier than predicted by the bolus technique alone. All the resulting contrast-enhanced MR angiograms showed good or excellent intraarterial contrast enhancement with little or no venous contamination. It is possible that in some cases, triggering the contrast-enhanced MR angiography later as predicted by the test-bolus information might have resulted in significant venous contamination. This question will be part of an ongoing investigation. For the present, our current combination of SmartPrep and the timing run to trigger the contrast-enhanced MR angiography along with the new elliptic sequence has resulted in a robust technique to generate consistent high-quality high-resolution carotid MR angiography.
At this time we recommend replacing two-dimensional time-of-flight MR angiography from the aortic arch to the skull base with contrast-enhanced carotid MR angiography. With clinically available improvements in coil design, signal dropout inferiorly near the aortic arch is no longer a problem. Although no significant stenosis was seen in the takeoff of the common carotid arteries from the aortic arch in this study, two examples of significant stenoses of the left subclavian artery were noted. These examples suggest that contrast-enhanced MR angiography will make an excellent screening sequence to examine the carotid arteries from the aortic arch to the upper neck. Contrast-enhanced MR angiography and MOTSA MR angiography have been shown to be complementary in depicting carotid bifurcation disease. With further improvements in signal-to-noise ratio and resolution, contrast-enhanced MR angiography may replace MOTSA MR angiography for depicting carotid bifurcation stenosis.
In conclusion, rapid high-resolution carotid contrast-enhanced MR angiography with consistent venous suppression is technically feasible and with the addition of improved triggering and data acquisition is now successful in most patients. A hardware-modified MR scanner or data-intensive postprocessing is not required with this technique. The image quality has also improved to the point that it is equal to or superior to the longer MOTSA MR angiography in most patients. With at least one major MR manufacturer providing elliptic centric phase-encoding view ordering, similar high-resolution carotid contrast-enhanced MR angiography will soon be routinely possible with commercially available equipment.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Alvarez-Linera, J. Benito-Leon, J. Escribano, J. Campollo, and R. Gesto Prospective Evaluation of Carotid Artery Stenosis: Elliptic Centric Contrast-Enhanced MR Angiography and Spiral CT Angiography Compared with Digital Subtraction Angiography AJNR Am. J. Neuroradiol., May 1, 2003; 24(5): 1012 - 1019. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. A. Willinek, J. Gieseke, R. Conrad, H. Strunk, R. Hoogeveen, M. von Falkenhausen, E. Keller, H. Urbach, C. K. Kuhl, and H. H. Schild Randomly Segmented Central k-Space Ordering in High-Spatial-Resolution Contrast-enhanced MR Angiography of the Supraaortic Arteries: Initial Experience Radiology, November 1, 2002; 225(2): 583 - 588. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Wutke, W. Lang, C. Fellner, R. Janka, C. Denzel, M. Lell, W. Bautz, and F. A. Fellner High-Resolution, Contrast-Enhanced Magnetic Resonance Angiography With Elliptical Centric k-Space Ordering of Supra-aortic Arteries Compared With Selective X-Ray Angiography Stroke, June 1, 2002; 33(6): 1522 - 1529. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
