British Journal of Radiology 74 (2001),147-152 © 2001 The British Institute of Radiology
Gadolinium enhanced three-dimensional magnetic resonance portography with subtraction
N Nishida, MD
G Sakai, MD
A Morimoto, MD
M Isota, MD
T Kaminou, MD
T Matsuoka, MD
K Nakamura, MD
and
R Yamada, MD
Osaka City University Medical School, Department of Radiology, 1-4-3, Asahi-machi, Abeno-ku, Osaka City 545-8585, Japan
 |
Abstract
|
|---|
The aim was to evaluate a subtraction technique for breath-hold gadolinium enhanced three-dimensional magnetic resonance portography (3D-MRP). 26 patients with gastric and/or duodenal varices related to portal hypertension were investigated by 3D-MRP with two phase acquisitions. A partial volume maximum intensity projection (MIP) image after subtracting selective arterial phase images from subsequent portal venous phase images (subtraction 3D-MRP) was compared with the partial volume MIP without subtraction (non-subtraction 3D-MRP) to assess visualization of the portal vein and its collaterals. Subtraction 3D-MRP depicted excellent visualization of the portal vein, although this was not significantly better than non-subtraction 3D-MRP. However, subtraction 3D-MRP gave superior visualization of portal collaterals, with effective suppression of arterial and renal signal intensities, compared with non-subtraction 3D-MRP (p<0.001).
|
Introduction
|
|---|
Gadolinium enhanced three-dimensional magnetic resonance angiography (3D-MRA) is useful in obtaining high quality portal vein images [1–3]. However, the arterial signal and enhancement of adjacent organs such as the kidney, pancreas and intestine can obscure visualization of the portal vein and portal collateral systems. The presence of the signal from arterial enhancement can easily be avoided and selective venous phase images obtained by appropriate timing of data acquisition and the subtraction of selective arterial phase images from subsequent arterial venous images [4–6]. We describe the usefulness of breath-hold gadolinium enhanced 3D magnetic resonance portography (3D-MRP) with subtraction of selective arterial phase images from subsequent arterial venous phase images.
|
Materials and methods
|
|---|
26 patients who presented with gastric and/or duodenal varices related to portal hypertension were evaluated by gadolinium enhanced 3D-MRP. All examinations were performed on a Signa 1.5 T imaging system (General Electric Medical System, Milwaukee, WI) using a four channel, body phased array coil. In all cases, two phase acquisitions were obtained using an unchanged fast imaging sequence with steady-state precession in a central k-space order. A coronal 3D dynamic gadolinium enhanced spoiled gradient echo sequence was prescribed with the following parameters: repetition time 4.6–6.1 ms, echo time 1.2–1.3 ms, inversion time 17 ms, flip angle 30°, 16–20 partitions, 4 mm effective section thickness, matrix 256x160 pixels, and 36x28 cm field of view. Image acquisition time ranged from 15–20 s. 20 ml gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) was used for all examinations, followed by 20 ml of saline solution through a 22 G venous catheter in an antecubital vein. Contrast medium was injected by an MR-compatible power injector (Sonicshot 50; Nemoto Kyorindo, Japan), with an injection flow rate of 3 ml s-1.
Smartprep (GE Medical Systems, Milwaukee, WI) was used to automatically detect arrival of contrast medium in the aorta, thus allowing scan initiation at the appropriate time. The first-phase arterial images were acquired 5 s after the arrival of contrast medium in the abdominal aorta. The second-phase portal venous images were obtained with a 15 s interval for breathing after completion of the first-phase acquisition. Before the examination, patients were instructed in breath-holding in the same position.
Selective portal venous images were acquired by subtracting the arterial phase from the portal venous phase image using standard software.
Individual sections covering the volume of interest were post-processed using a partial volume maximum intensity projection (MIP) technique. The volume of interest spanned from the umbilicus ventrally to the hilum of the left kidney dorsally. A partial volume MIP image of portal venous phase images (non-subtraction 3D-MRP) and a partial volume MIP image after subtracting the arterial phase from subsequent portal venous phase images (subtraction 3D-MRP) were obtained.
Image quality of non-subtraction and subtraction 3D-MRP was graded by two experienced radiologists (NN, AM). Depiction of the portal vein together with its first branches in the liver and portal collateral systems, including left gastric vein, retrogastric vein, gastrorenal shunt and paraoesophageal vein, was evaluated with a four-point scale for ranking of image quality: 1, non-diagnostic; 2, good; 3, marginally adequate; 4, excellent. Discrepancies between observers were resolved by consensus. The statistical significance of the difference in these ratings was tested by the Wilcoxon signed-rank test.
|
Results
|
|---|
First-phase 3D-MRA image acquisition depicted only the arterial system. However, both arterial and venous systems were visualized with the second acquisition. The process of subtracting the first-phase images from the second-phase images was performed on a satellite console within 5 min.
Breath-hold images were successfully obtained in all patients except one. The portal vein and portal collateral systems were depicted with good or excellent quality with both subtraction and non-subtraction 3D-MRP. Subtraction 3D-MRP demonstrated no significant improvement in visualization of the portal vein. Average scores for image quality in visualization of the portal vein were 3.5±0.58 for non-subtraction 3D-MRP and 3.5±0.65 for subtraction 3D-MRP (p>0.9). However, 3D-MRP with subtraction showed superior image quality for visualizing the portal collateral systems than images without subtraction (Figures 1
and 2). Average scores for image quality in visualization of portal collateral systems were 2.5±0.71 for non-subtraction 3D-MRP and 3.5±0.76 for subtraction 3D-MRP, which are significantly different (p<0.001).
View larger version (80K):
[in this window]
[in a new window]
|
Figure 1. 49-year-old man with gastric and duodenal varices due to liver cirrhosis. (a) A partial maximum intensity projection (MIP) image of portal venous phase image (non-subtraction three-dimension magnetic resonance portography (3D-MRP)) demonstrates excellent visualization of the portal vein. The gastric varices (arrowheads) and duodenal varices (arrows) are not clearly visualized owing to enhancement of the aorta and right kidney. (b) A partial MIP image after subtraction of the arterial phase data set from the portal venous phase data set (subtraction 3D-MRP) demonstrates effective suppression of the aortic signal intensity and the right kidney enhancement, and superior visualization of portal collateral venous systems compared with (a). The portal vein is clearly depicted. (c,d) Percutaneous transhepatic portal venogram and (e) retrograde balloon occluded venogram from the left adrenal vein also shows the portal collateral venous systems.
|
|
View larger version (106K):
[in this window]
[in a new window]
|
Figure 2. 33-year-old man with duodenal varix due to liver cirrhosis. (a) Non-subtraction three-dimension magnetic resonance portography (3D-MRP) image. (b) Subtraction 3D-MRP demonstrates effective suppression of the aortic signal intensity and the enhancement of pancreas and intestine. Superior visualization of the duodenal varix (arrows) compared with (a). (c) Percutaneous transhepatic portal venogram also shows the duodenal varix.
|
|
In four patients who could not hold their breath in the same position between each acquisition, the source images after the subtraction process showed that the signal of the aorta was completely eliminated although adjacent organ enhancement from, for instance, the kidney was not eliminated perfectly. Subtraction 3D-MRP consequently revealed superior visualization of portal collateral systems compared with non-subtraction, even with such imperfect subtraction conditions (Figure 3).
View larger version (79K):
[in this window]
[in a new window]
|
Figure 3. 52-year-old man with gastric varix due to liver cirrhosis. (a) Source image of portal venous phase. (b) Source image after subtraction of an arterial phase image from a portal venous phase image shows that the signal of the aorta is completely eliminated, but enhancement of the left kidney is not entirely eliminated. (c) Non-subtraction three-dimension magnetic resonance portography (3D-MRP) image. (d) Subtraction 3D-MRP demonstrates effective suppression of the arterial signal and the enhancement of adjacent organs, even in these imperfect subtraction conditions. The portal collateral venous systems (arrows) are more clearly visualized compared with (c). (e) Retrograde balloon occluded venogram from the left adrenal vein also shows the gastric varix.
|
|
|
Discussion
|
|---|
Several methods are available for treating oesophageal and gastric varices including endoscopy, surgical portosystemic shunt and intravascular interventional techniques such as percutaneous portosystemic shunt and embolisation [7]. Precise assessment of the portal vein and collateral pathways is potentially helpful for treatment and planning purposes.
Gadolinium enhanced 3D-MRA using the T1 shortening effect can avoid motion artefact by breath-holding, while the strong intravascular signal intensity provides excellent visualization of portal venous systems with tortuous, turbulent, stagnant or even retrograde flow [8]. A partial volume MIP technique is helpful for visualizing the portal collateral systems with their complex and tortuous course [9]. However, the arterial signal and enhancement of adjacent organs such as kidney, pancreas and intestine may appear in the portal venous phase images when using a partial volume MIP technique, and this can prevent visualization of portal collateral systems. It is difficult to display the entire portal collaterals when the partial volume MIP image is produced to exclude enhancement of these other organs.
We performed two-phase examinations for arterial and portal venous phase images, and tailored subtraction 3D-MRP for obtaining selected portal venous phase images that result from the subtraction of arterial phase images from portal venous phase images. Although subtraction 3D-MRP demonstrated no significant improvement in visualization of the portal vein compared with non-subtraction 3D-MRP, it did show superior visualization of portal collateral systems, with the effective suppression of arterial signal and enhancement of adjacent organs. Subtraction 3D-MRP thus provides a precise and easy visualization of the entire portal collateral system.
For subtraction images to display only the portal venous system, the appropriate timing of image acquisition should be determined to obtain images of the arterial system alone. The Smartprep system is useful in deciding the exact starting time of scan acquisition [10]. The first-phase images were acquired 5 s after arrival of contrast medium in the abdominal aorta, which was adequate timing for obtaining the arterial phase images alone.
The breath between image acquisitions for the arterial and portal venous phases may lead to different breath-hold positions, although the changes may be subtle. Even in imperfect subtraction conditions due to different breath-hold positions between each acquisition, excellent suppression of arterial signal and enhancement of adjacent organs was obtained after post-processing by using a partial volume MIP technique, because most of the signal enhancement was suppressed. Consequently, the portal collateral systems were more clearly visualized in subtraction 3D-MRP than in non-subtraction 3D-MRP. While the technique of subtracting arterial phase images from portal venous phase images was valuable, the original source images should be carefully assessed because subtraction is an artificial technique.
In conclusion, gadolinium enhanced subtraction 3D-MRP depicted excellent but equivalent visualization of the portal vein and superior visualization of portal collateral systems compared with non-subtraction 3D-MRP.
Received for publication July 10, 2000.
Revision received October 18, 2000.
Accepted for publication October 25, 2000.
|
References
|
|---|
-
Leung DA, McKinnon GC, Davis CP, Pfammatter T, Krestin GP, Debatin JF. Breath-hold, contrast-enhanced, three-dimensional MR angiography. Radiology 1996;200:569–71.[Abstract/Free Full Text]
-
Yamashita Y, Mitsuzaki K, Miyazaki T, Namimoto T, Sumi S, Urata J, et al. Gadolinium-enhanced breath-hold three-dimensional MR angiography of the portal vein: value of magnetization-prepared rapid acquisition gradient echo sequence. Radiology 1996;201:283–8.[Abstract/Free Full Text]
-
Shirkhoda A, Konez O, Shetty AN, Bis KG, Ellwood RA, Kirsch MJ. Mesenteric circulation: three-dimensional MR angiography with a gadolinium-enhanced multiecho gradient-echo technique. Radiology 1997;202:257–61.[Abstract/Free Full Text]
-
Lebowitz JA, Rofsky NM, Krinsky GA, Weinreb JC. Gadolinium-enhanced body MR venography with subtraction technique. AJR 1997;169:755–8.[Free Full Text]
-
Okumura A, Watanabe Y, Dohke M, Ishimori T, Amoh Y, Oda K. Contrast-enhanced three-dimensional MR portography. Radiographics 1999;19:973–87.[Abstract/Free Full Text]
-
Watanabe Y, Dohke M, Okumura A, Amoh Y, Ishimori T, Oda K, et al. Dynamic subtraction contrast-enhanced MR angiography: technique, clinical applications, and pitfalls. Radiographics 2000;20:135–52.[Abstract/Free Full Text]
-
Koito K, Namieno T, Nagakawa T, Morita K. Balloon-occluded retrograde transvenous obliteration for gastric varices with gastrorenal or gastrocaval collaterals. AJR 1996;167:1317–20.[Abstract/Free Full Text]
-
Prince MR, Narasimham DI, Stanley JC, Chenevert TL, Williams DM, Marx MV, et al. Breath-hold gadolinium-enhanced MR angiography of the abdominal aorta and its major branches. Radiology 1995;197:785–92.[Abstract/Free Full Text]
-
Leyendecker JR, Rivera E, Washburn WK, Johnson SP, Diffin DC, Eason JD. MR angiography of the portal venous system: techniques, interpretation, and clinical applications. Radiographics 1997;17:1425–43.[Abstract]
-
Prince MR, Chenevert TI, Foo TKF, Londy FJ, Ward JS, Maki JH. Contrast-enhanced abdominal MR angiography: optimization of imaging delay time by automating the detection of contrast material arrival in the aorta. Radiology 1997;203:109–14.[Abstract/Free Full Text]
Top
Abstract
Introduction
