| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Department of Radiology, Onze Lieve Vrouw Hospital Aalst, Moorselbaan 164, 9300 Aalst, Belgium. Received December 2, 2003; revision requested February 3, 2004; final revision received May 17; accepted May 27. Address correspondence to K.P.M. (e-mail: mermuyskoen@hotmail.com).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Magnetic resonance (MR) venography has an advantage over other visualization techniques, such as cerebral computed tomographic (CT) venography and digital subtraction angiography, in that it is noninvasive and does not require ionizing radiation exposure or iodinated contrast material administration. Furthermore, MR images can demonstrate the secondary parenchymal changes caused by the venous congestion.
Multiple MR angiographic techniques have been proposed for evaluation of the dural sinuses and the cortical veins (4–18). Techniques of bolus injection of contrast material with subsequent acquisition with a magnetization-prepared rapid gradient-echo sequence have previously been described and have been used with success. Farb et al (12) used gadolinium-enhanced three-dimensional (3D) automatically triggered elliptic centric-ordered MR venography for imaging of the intracranial venous system and proved it to be superior to time-of-flight MR venography. Stevenson et al (19) first described the use of magnetization-prepared rapid acquisition gradient-echo venography in regard to visualization of the dural sinuses and cortical veins. Liang et al (1,20) used the magnetization-prepared rapid acquisition gradient-echo sequence as a means to depict intracranial dural sinus thrombosis and to evaluate the normal structures in the dural sinuses, with success. A slight disadvantage of the magnetization-prepared rapid acquisition gradient-echo and automatically triggered elliptic centric-ordered techniques is the relatively longer acquisition time.
We thought the usefulness of cerebral MR venography with a short acquisition time would be improved by using a volumetric interpolated brain examination sequence with asymmetric k-space sampling and interpolation of image data (14,15,21). This sequence has a much shorter acquisition time than do the magnetization-prepared rapid acquisition gradient-echo and automatically triggered elliptic centric-ordered techniques (1 minute 24 seconds for the volumetric interpolated brain examination vs 6 minutes 43 seconds to 7 minutes 43 seconds for the magnetization-prepared rapid acquisition gradient-echo technique and 4 minutes 38 seconds for the automatically triggered elliptic centric-ordered technique). Rofsky et al (21) described a 3D T1-weighted gradient-recalled echo sequence with interpolation of image data that they referred to as "volumetric interpolated breath-hold examination" for abdominal imaging. This type of sequence can also be used to form rapid images of the brain and has also been described by Wetzel et al (14,15). They found it to be an effective alternative approach to the magnetization-prepared rapid acquisition gradient-echo technique for fast 3D T1-weighted imaging of the brain (14). Thus, the purpose of our study was to prospectively compare, both qualitatively and quantitatively, the quality of images obtained with 3D volumetric interpolated brain examination cerebral venography with that of those obtained with 3D magnetization-prepared rapid acquisition gradient-echo cerebral venography.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
MR Protocol and Parameters
MR imaging was performed with a 1.5-T system (Sonata; Siemens Medical Systems, Erlangen, Germany). Maximum gradient strength was 40 mT/m. A quadrature head coil was used. All patients were examined prospectively with both pre- and postcontrast 3D magnetization-prepared rapid acquisition gradient-echo venography and 3D volumetric interpolated brain examination venography during the same imaging session. First, the precontrast magnetization-prepared rapid acquisition gradient-echo and volumetric interpolated brain examination sequences were performed. Thereafter, a single intravenous dose (0.1 mmol/kg) of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) was administered manually at a rate of 1–2 mL/sec (total dose of 20 mL), and then the postcontrast magnetization-prepared rapid acquisition gradient-echo and volumetric interpolated brain examination sequences were performed in a randomized order. The postcontrast magnetization-prepared rapid acquisition gradient-echo and volumetric interpolated brain examination sequences were performed in succession immediately after contrast agent administration, without reinjection of contrast agent.
Volumetric interpolated brain examination sequence.—The volumetric interpolated brain examination sequence is a 3D gradient-recalled echo sequence with radiofrequency spoiling (3.63/1.51 [repetition time msec/echo time msec], 12° flip angle). Images were acquired in the sagittal direction. One 3D slab consisted of 144 partitions with 1.1-mm thickness and no intersection gap, which resulted in a slab thickness of 159 mm. In-plane field of view was 230 mm (read direction) x 230 mm (phase direction). The image matrix was 256 (read direction) x 192 (phase direction) x 144 (partition direction). Sampling was symmetric in the read and phase directions and asymmetric in the partition direction. Before Fourier transformation, zero interpolation of image data in the partition direction was performed to obtain 144 sections.
Other parameters were as follows: fat suppression, spectral; number of signals acquired, one; bandwidth, 490 Hz/pixel; acquisition time, 1 minute 24 seconds.
Magnetization-prepared rapid acquisition gradient-echo sequence.—The 3D magnetization-prepared rapid acquisition gradient-echo sequence is a small-flip-angle gradient-recalled echo sequence with a 3D Fourier transform acquisition technique that has been implemented with a 180° inversion recovery preparation pulse. The following parameters were used: 6.25/3.93/300 [repetition time msec/echo time msec/inversion time msec], 15° flip angle). Acquisition was performed in the sagittal direction. One 3D slab consisted of 144 partitions with 1.1-mm thickness and no intersection gap, which resulted in a slab thickness of 159 mm. In-plane field of view was 230 (read direction) x 230 (phase direction). The image matrix was 256 (read direction) x 192 (phase direction) x 144 (partition direction). Sampling was symmetric in the read, phase, and partition directions. No zero filling was performed.
Other parameters were as follows: fat suppression, none; number of signals acquired, one; bandwidth, 130 Hz/pixel; acquisition time, 6 minutes 6 seconds.
The sequence parameters were optimized in accordance with the guidelines of the manufacturer of our MR imaging unit.
Postprocessing.—Pre- and postcontrast magnetization-prepared rapid acquisition gradient-echo and volumetric interpolated brain examination images were subtracted on a pixel-by-pixel basis, and maximum intensity projections, also called maximum intensity projection images, were obtained. Postcontrast source images, subtracted images, and maximum intensity projection images were used for analysis. Image examples are displayed in Figures 1 and 2.
|
|
|
|
and SIWM/, respectively, where SIGM is signal intensity in gray matter, is the standard deviation, and SIWM is signal intensity in white matter. Gray matter–white matter, SSS–white matter, SSS–gray matter, GCV–white matter, and GCV–gray matter contrast-to-noise ratios (CNRs) were calculated as (SIWM – SIGM)/, (SISSS – SIWM)/, (SISSS – SIGM)/, (SIGCV – SIWM)/, and (SIGCV – SIGM)/, respectively (14). Qualitative assessment of image quality.—The 3D MR angiograms were generated by using a maximum intensity projection algorithm and were rotated around a vertical axis. The two-dimensional postcontrast source images, subtracted images, and 3D maximum intensity projection angiograms were randomly interpreted by two independent experienced radiologists (P.K.V. and L.V.H., with 15 and 6 years of experience with brain MR imaging, respectively) who were blinded to patient and sequence information. The quality of visualization of the veins was classified as follows: high signal intensity (clearly visible structure) and continuous visualization (no clear interruption along the course of the vessel), low signal intensity (poorly but still visible structure) and continuous visualization, noncontinuous visualization (clear interruption along the course of the vessel), or no visualization. We also compared the images of the cortical veins obtained with the two MR angiographic techniques by determining the numbers of cortical veins and their first-order branches in both hemispheres. We followed the courses of the vessels continuously to avoid mistaking the proximal and distal aspects of vessels for distinct branches (13).
The venous structures that were assessed were the SSS, the inferior sagittal sinus, the straight sinus, the transverse sinus, the sigmoid sinus, the cavernous sinus, the internal cerebral vein, the basal vein of Rosenthal, the GCV, the septal veins, the thalamostriate veins, the vein of Labbé, and the vein of Trollard (22).
Statistical Analysis
Quantitative.—SNRs and CNRs were compared for the magnetization-prepared rapid acquisition gradient-echo and volumetric interpolated brain examination postcontrast source images by means of a paired two-tailed Student t test (14). Calculations were performed with statistical software (MedCalc 5; MedCalc Software, Mariakerke, Belgium).
Qualitative.—Differences in the number of cortical veins visualized with the two MR angiographic techniques were analyzed by using the Wilcoxon matched-pairs signed rank test. The Wilcoxon matched-pairs signed rank test also was used to analyze the differences in the quality of visualization classifications for the cerebral veins and dural sinuses (13). The 
2 test was used in the assessment of every venous structure to evaluate differences in the number of venous structures associated with a specific quality of visualization classification. Statistical software (Prism; Graphpad Software, San Diego, Calif) was used to perform the calculations.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Quantitative Assessment of Image Quality
There was a significant difference in the mean SNR in the SSS and GCV for images obtained with the volumetric interpolated brain examination sequence and those obtained with the magnetization-prepared rapid acquisition gradient-echo sequence of 51.6 for SSS and 58.4 for GCV, with P <.001 and a 95% confidence interval of 33.31 to 70.07 for SSS and 38.8 to 78.1 for GCV.
Results for SNR are graphically displayed in Figures 3 and 4.
|
|
|
|
|
|
|
|
The number of visualized venous structures and their associated visualization classifications are tabulated in Table 2.
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Intraarterial digital subtraction angiography is still considered the standard for demonstration of cerebral vasculature, but it is time-consuming and invasive (4,10). Cerebral CT venography, which is usually quickly available and easily interpretable, has been proved to be at least equivalent to MR venography in the diagnosis of dural sinus thrombosis, but iodinated contrast material and ionizing radiation are needed (23). MR imaging has become a standard imaging modality in patients with cerebral vascular lesions because of its inherent soft-tissue contrast multiplanar imaging capabilities and its capability for visualization of flowing blood (10). With most MR venography methods, either two-dimensional and 3D phase-sensitive techniques or time-of-flight techniques are used (24).
Time-of-flight imaging has several drawbacks: It is time-consuming. Regions of slow or turbulent blood flow might cause attenuated vascular signal and lead to misinterpretation. Substances with a short T1, such as blood breakdown products in the thrombus, can create a high signal that mimics blood flow (4,25,26). In two-dimensional time-of-flight imaging, there can be a loss of signal in areas of blood flow parallel to the measurement plane, in regions of vessel confluence, and in areas with a dural arteriovenous fistula (27). Therefore, several acquisition planes are often needed (24,26). With 3D time-of-flight imaging, a higher spatial resolution and a high SNR can be obtained, but the technique is more time-consuming and more sensitive to motion. The saturation problems caused by slow blood flow in 3D time-of-flight imaging can be addressed with parametric optimization with smaller flip angles, use of multiple overlapping thin slabs, or intravenous injection of a gadolinium-based contrast agent.
At phase-contrast MR angiography, the measurable phase shift acquired by protons moving through a magnetic field is used. Because the phase shift is mathematically proportional to blood flow velocity and has a positive or negative value that is based on polarity of blood flow along the interrogated axis, it can provide information about both the blood velocity and blood flow direction. Although excellent background suppression is a major advantage, and qualitative determination of blood velocities may be possible, phase-contrast MR angiography requires long imaging times to obtain high-spatial-resolution images and it is more sensitive to signal loss caused by turbulence or intravoxel dephasing (25).
Recently, Reichenbach et al (7,8) and Tan et al (9) showed that small veins in the brain can be depicted by using a blood flow–compensated 3D radiofrequency-spoiled gradient-echo sequence (high-resolution blood oxygen level–dependent venography). This method is based on the T2* shortening in deoxygenated blood as compared with oxygenated blood. High-resolution blood oxygen level–dependent venography can be used to show cerebral arteriovenous malformations and cerebral venous thrombosis with high resolution (7,10,17). Reichenbach et al (8) believe this technique to be superior in resolution on images of small venous vessels when it is compared with two-dimensional time-of-flight venography or with contrast-enhanced MR venography. Two major limiting factors, however, have prevented the acceptance of high-resolution blood oxygen level–dependent venography in daily clinical practice. First, it is time-consuming and therefore more susceptible to patient motion. Second, because of the long echo time, the image quality is compromised by a reduction of the SNR and an increased sensitivity to susceptibility artifacts from air-tissue interfaces (17). Lin et al (17) improved the use of high-resolution blood oxygen level–dependent venography by using a T1-reducing contrast agent and therefore obtaining a shorter echo time and repetition time. To our knowledge, however, high-resolution blood oxygen level–dependent venography has not yet been compared with other MR venography techniques in the evaluation of the dural sinuses.
Our study limitations are the following: Volumetric interpolated brain examination venography has not yet been evaluated for the detection of venous thrombosis, volumetric interpolated brain examination venography has not yet been compared with 3D fast low-angle shot MR angiography, our study population consisted of only 28 patients who were considered as a healthy population, and there is no standard of reference for this study.
Finally, the advantage of 3D volumetric interpolated brain examination MR venography over 3D magnetization-prepared rapid acquisition gradient-echo venography is the much shorter acquisition time, which makes it easier to use in patients with impaired consciousness. The disadvantages of volumetric interpolated brain examination venography are as follows: Contrast material is used. On the maximum intensity projection images, arteries and soft tissues may be superimposed on the venous structures, thus decreasing the visualization (as seen in Fig 2). The sequence has to be performed twice, which doubles the total imaging time (2 minutes 48 seconds). As shown, there are significantly lower (P < .001) SNR and CNR when this technique is compared with the magnetization-prepared rapid acquisition gradient-echo technique, although this decrease in these ratios has no statistically significant (P > .05) effect on the subjective visualization classification of the dural sinuses and cortical veins. In our study, we showed that volumetric interpolated brain examination venography may be a valuable alternative technique in evaluation of normal venous cerebral structures compared with 3D magnetization-prepared rapid acquisition gradient-echo venography.
In conclusion, as shown by the significantly lower (P < .001) CNR and SNR for volumetric interpolated brain examination venography, there is a trend toward a decrease in quality of visualization of the venous cerebral structures. This trend was expected, as it is an inherent effect of the volumetric interpolated brain examination venographic technique. Since there was no statistically significant difference (P > .05) in the subjective visualization classification of the venous structures when this technique was compared with magnetization-prepared rapid acquisition gradient-echo venography, 3D volumetric interpolated brain examination cerebral venography seems promising as a more time-efficient alternative for qualitative assessment of cerebral veins and dural sinuses than is 3D magnetization-prepared rapid acquisition gradient-echo venography.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, K.P.M.; study concepts, P.K.V.; study design, K.P.M.; literature research, K.P.M.; clinical studies, K.P.M.; data acquisition, P.C.; data analysis/interpretation, P.K.V., L.V.H.; statistical analysis, P.K.V.; manuscript preparation and final version approval, K.P.M.; manuscript definition of intellectual content, L.V.H.; manuscript editing, K.P.M., P.C.; manuscript revision/review, P.K.V., L.V.H.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Tomasian, N. Salamon, M.S. Krishnam, J.P. Finn, and J.P. Villablanca 3D High-Spatial-Resolution Cerebral MR Venography at 3T: A Contrast-Dose-Reduction Study AJNR Am. J. Neuroradiol., February 1, 2009; 30(2): 349 - 355. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. H. Hu, N. G. Campeau, J. Huston III, D. G. Kruger, C. R. Haider, and S. J. Riederer High-Spatial-Resolution Contrast-enhanced MR Angiography of the Intracranial Venous System with Fourfold Accelerated Two-dimensional Sensitivity Encoding Radiology, June 1, 2007; 243(3): 853 - 861. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
