| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, St. Louis University Hospital, 3635 Vista Ave. at
Grand Blvd., P.O. Box 15250, St. Louis, MO 63110-0250.
2
Present address: Department of Radiology, University of Michigan, B1F510 UH,
1500 E. Medical Center Dr., Ann Arbor, MI 48109-0030.
3
Department of Surgery, St. Louis University Hospital, St. Louis, MO
63110-0250.
Received July 19, 1999;
accepted after revision October 5, 1999.
Address correspondence to J. F. Glockner.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Thirty-eight consecutive three-dimensional gadolinium-enhanced MR angiograms were obtained in 34 patients. Results were retrospectively reviewed and correlated with conventional angiography in 20 of the 38 cases and sonography in 37 of the 38 cases. MR angiograms were evaluated for technical adequacy, vascular patency, and parenchymal abnormalities, and results were compared with angiography and sonography. Conventional angiography and surgery were used as gold standards when available.
RESULTS. Thirty-four (90%) of 38 MR angiograms were technically adequate. Vascular abnormalities were identified in 20 patients, and 19 of these patients subsequently underwent angiography, surgery, or both. There were seven cases of hepatic artery thrombosis; all were detected with MR angiography with no false-positive or false-negative interpretations. Seven patients had moderate to severe hepatic artery stenosis (>50% narrowing as determined by conventional angiography). MR angiography revealed this stenosis in six of the seven patients, with one false-negative and three false-positive interpretations. Portal vein thrombosis was detected in three patients, and portal vein stenosis was detected in two patients.
CONCLUSION. Three-dimensional gadolinium-enhanced MR angiography is useful in the examination of liver transplant patients and offers a noninvasive adjunct in patients with difficult or indeterminate sonographic examinations.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MR Angiography
MR angiography was performed on a Signa 1.5-T scanner (General Electric
Medical Systems, Milwaukee, WI). A body coil or phased array torso coil was
used in all patients. Examination consisted of coronal fast spoiled
gradient-echo scout images followed by axial fast spin-echo T2-weighted images
with spectral fat saturation through the liver and spleen. This imaging was
followed by three-dimensional MR angiography.
A timing bolus was used to calculate the scan delay needed to assure that acquisition of the central portion of K-space corresponded to maximal arterial enhancement. Two milliliters of gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) followed by 14 ml of saline flush was injected at 2 ml/sec. Sagittal spoiled gradient-echo images of the abdominal aorta were obtained at 1-sec intervals for 60 sec, with acquisition of this sequence coinciding with the start of the test bolus injection. A region of interest was then placed in the abdominal aorta adjacent to the celiac axis, and a time-intensity plot was generated to define the time of maximal enhancement. This time was defined as the travel time, and a scan delay was calculated according to a method described by Prince et al. [17].
Three-dimensional MR angiography was performed in the coronal plane using a fast spoiled gradient-echo sequence (TR/TE, 5.7/2 msec; 35° flip angle; and 62-kHz bandwidth). Section thickness varied from 2.4 to 3.8 mm. Thirty-two sections were obtained, with sequence length varying from 21 to 38 sec. In the final eight examinations, zero filling along the z-axis was performed to obtain 64 reconstructed sections overlapping by 50%. This technique was not available for the initial 30 examinations. Field of view varied from 34 to 44 cm. Matrix size was typically 142x256, but varied from 128 x 256 to 162 x 512. All these parameters were chosen according to the size and breath-hold capacity of the patient. Thirty-eight milliliters of IV contrast material was injected at 2 ml/sec with an MR-compatible injector (Spectris; Medrad, Pittsburgh, PA) followed by 15 ml of saline flush at the same rate. At least two additional three-dimensional sequences were performed immediately after acquisition of the arterial phase images to obtain portal venous and hepatic venous phase images. A final gadolinium-enhanced axial T1-weighted spin-echo sequence with spectral fat saturation was performed after completion of the MR angiography. The entire examination required approximately 40 min.
Three-dimensional reconstruction (subvolume maximum-intensity-projection images) and single voxel two-dimensional reformatting in multiple planes were performed on a workstation (Windows Advantage; General Electric Medical Systems). The MR angiograms were interpreted by one radiologist at the workstation, after viewing raw data, reformated images, maximum-intensity-projection images, and the axial T1- and T2-weighted sequences.
Sonography
Sonography was performed by an experienced technologist using an HTL 3000
scanner (Advanced Technology Laboratories, Bothell, WA) with probe frequencies
varying from 3 to 7 MHz, depending on body habitus. Gray-scale, color Doppler,
power Doppler, and spectral Doppler images were obtained in all patients. The
hepatic artery, main portal vein and right and left branches, hepatic veins,
and inferior vena cava were examined for patency and characteristics of the
spectral waveform. When possible, both right and left branches of the
intrahepatic artery were evaluated. Sonographers attempted to examine the
hepatic artery throughout its extrahepatic course; however, visualization of
the hepatic artery proximal to the porta hepatis was frequently impossible
because of overlying bowel gas.
Angiography
Angiography was performed from a femoral approach in all patients. The
celiac axis was selected using a 5-French selective visceral catheter, and the
catheter was advanced into the proper or common hepatic artery over a
guidewire. Digital subraction angiography was performed using non-ionic
contrast media and frame rates of 3.8 or 7.5 frames per second. Injection
rates varied, depending on the size of the artery. Typically, 4-6 ml was
injected per second for total volumes of 8-16 ml. One half to 1 mg of IV
glucagon was occasionally administered to reduce bowel gas—related
artifacts. Images were obtained in at least two tangential projections with
emphasis on clearly showing the arterial anastomosis, the intraparenchymal
branches, and the pattern of blood flow.
Image Analysis
Sonograms and MR angiograms were reviewed by one of the authors experienced
in cross-sectional imaging at separate single sessions approximately 1 month
apart. At the time of review, no information regarding the results of the
conventional angiography was available. Conventional angiograms were reviewed
by a radiologist experienced in angiography, again at a single session,
without knowledge of the results of the sonography or MR angiography.
Reviewers were not made unaware of patient name, and therefore a relatively
long interval between evaluation of MR angiograms and sonograms was chosen to
minimize recollection bias. Initial reports from all examinations were then
obtained, and discrepancies between these reports and findings at review
sessions were noted.
Imaging criteria for hepatic artery thrombosis consisted of nonvisualization of the hepatic artery within the hepatic parenchyma on MR angiography, sonography, or conventional angiography. Portal vein thrombosis included both occlusive and nonocclusive thrombi within the main portal vein or proximal branches. Hepatic artery stenosis was suggested sonographically when the hepatic artery resistive index (peak systolic velocity — end diastolic velocity / peak systolic velocity) was less than 0.5, when a parvus-tardus waveform was seen with spectral Doppler sonography, or when the main hepatic artery peak systolic velocity was greater than 200 cm/sec [18, 19]. The vessels evaluated on MR angiography and conventional angiography were the main hepatic artery and anastomosis, celiac axis, main portal vein and right and left branches, splenic vein, superior mesenteric vein, inferior vena cava, and hepatic veins. Vessels were defined as patent, thrombosed, or stenotic, and the stenotic vessels were subjectively graded as mild (<50%), moderate (50-75%), or severe (>75%). Abnormalities of distal intrahepatic arterial branches were noted on conventional angiography and, when visualized, on MR angiography. Accuracy of MR angiography was determined using conventional angiography and surgery as gold standards.
MR angiograms were also evaluated by a radiologist for technical adequacy. Examinations were classified as either adequate or inadequate. An adequate examination allowed clear visualization of the main hepatic artery, main portal vein and proximal right and left branches, hepatic veins, and inferior vena cava.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
The 14 remaining technically adequate examinations revealed no significant vascular abnormality. Within 1 week of the MR angiography, all 14 patients in whom these technically adequate examinations were performed underwent sonography, which either had normal findings or indicated decreased diastolic flow with an elevated resistive index (>0.7) in the hepatic artery.
No significant discrepancies were noted between initial reports and reviewed examinations; however, the extent of hepatic artery stenosis was not quantified on several MR angiography and conventional angiography reports. Nevertheless, no discrepancies existed regarding the presence or absence of vascular thrombosis or stenosis.
Arterial Complications
Hepatic artery thrombosis was identified in seven examinations in six
patients (Fig. 2). All cases of
hepatic artery thrombosis were confirmed angiographically or surgically, with
no false-positive or false-negative findings. Concurrent sonography was
performed in these six patients, also without false-positive or false-negative
findings.
|
MR angiography suggested moderate to severe hepatic artery stenosis (>50% narrowing) in nine patients. Findings were confirmed with angiography in six patients (Fig. 3A,3B). Angiographic findings were negative in three patients (Fig. 4A,4B). One MR angiogram had a false-negative finding for hepatic artery stenosis. There was discrepancy in the estimation of the degree of stenosis in one patient: MR angiography suggested moderate stenosis and conventional angiography revealed severe stenosis. In addition, there were discrepancies involving the hepatic artery distal to the anastomosis or the peripheral branches of the hepatic artery in two patients with hepatic artery stenosis. In one patient, MR angiography suggested that the hepatic artery and proximal branches were diffusely abnormal and narrowed distal to the anastomotic stenosis, and angiography revealed only a focal anastomotic stenosis. In one patient, stenoses of the proximal right and left hepatic arteries identified on conventional angiography were not seen prospectively on MR angiography. In retrospect, these stenoses were present but were difficult to identify.
|
|
|
Sonographic correlation was obtained in all patients with confirmed stenotic lesions on conventional angiography and in all patients in whom MR angiography suggested hepatic artery stenosis. Sonographic findings were positive in four patients with hepatic artery stenosis and negative in three patients. There was one false-positive sonographic finding. In one patient with a false-negative sonographic finding, the MR angiographic finding was also false-negative. In two other patients with sonographic false-negative findings, the MR angiographic findings were positive. Of the three false-positive MR angiographic findings for hepatic artery stenosis, sonographic findings were negative in two patients and positive in the third patient. A summary of the data in patients with suspected or confirmed hepatic artery stenosis is provided in Table 1.
|
Venous Complications
Portal vein thrombosis was identified in three patients, all verified with
either angiography or contrast-enhanced CT
(Fig. 5). One patient with
nonocclusive thrombus in the portal confluence had a negative sonographic
finding. Moderate to severe portal vein stenosis was suggested in two
patients. Angiography revealed severe stenosis in one patient (Fig.
6A,6B),
and no significant stenosis in the second patient. MR angiography revealed
narrowing of the intrahepatic inferior vena cava and hepatic veins in two
patients, both of whom had concurrent hepatic artery thrombosis.
|
|
|
Miscellaneous Findings
Hepatic artery aneurysms or pseudoaneurysms were identified in two
patients, one confirmed with angiography and one with surgery
(Fig. 7). Bilomas and biliary
dilatation were seen in three patients
(Fig. 8), a focal hepatic
infarct in one patient, and a subcapsular fluid collection in one patient, all
of whom had either hepatic artery thrombosis or stenosis. Recurrent hepatoma
was identified in one patient with hepatic artery stenosis and portal vein
thrombosis. Celiac axis stenosis was identified in two patients and was
verified with angiography in one of the two.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The ability of duplex sonography to detect these complications has been extensively studied, with widely varying results. A study by Flint et al. [20] reported a sensitivity of 92% for the detection of hepatic artery thrombosis with an 8% false-negative rate. However, several reports have noted a high incidence of both false-positive and false-negative findings for hepatic artery thrombosis, with false-negative rates as high as 50% in children, presumably caused by the development of extensive arterial collaterals [19, 21]. Similarly, the reported success of duplex sonography in detecting hepatic artery stenosis also varies widely. Dravid et al. [22] detected 22% of extrahepatic arterial stenoses with sonography (this included stenoses <50% of the diameter of the normal lumen and stenoses of the celiac axis). Defrancq et al. [23] reported a sensitivity of 60% for detection of hepatic artery stenosis; Dodd et al. [18], a sensitivity of 73%; and Abbasoglu et al. [24], a sensitivity of 85%.
Our results indicate that gadolinium-enhanced three-dimensional MR angiography is a useful technique for evaluation of liver transplant patients. Technical success was achieved in 34 (89%) of the 38 examinations in a population that included many patients within the first or second postoperative week and others ill enough to be admitted to an intensive care unit. This examination is more demanding for patients than sonography or CT; however, most patients were able to cooperate for the 40-min examination. Breath-holding typically ranged from 20 to 30 sec, and these could be adjusted downward in hypoxic patients by making compromises in anatomic coverage or spatial resolution.
Both three-dimensional MR angiography and sonography revealed all cases of hepatic artery thrombosis, assuming that patients with missed diagnoses of hepatic artery thrombosis would have been detected on the basis of deteriorating clinical status—an assumption that may not always be correct. Detection of hepatic artery stenosis was more difficult. True sensitivity and specificity cannot be determined in our study because patients with normal findings on MR angiography and sonography did not undergo conventional angiography, except in one patient in whom hepatic artery stenosis was subsequently revealed. Limited clinical follow-up of this group (1-9 months) has not revealed development of hepatic artery thrombosis or stenosis. If there were no additional false-negative findings in our series, the sensitivity of MR angiography for revealing moderate to severe hepatic artery stenosis (>50%) is six (86%) of the seven patients, with a specificity of 86%.
The number of false-positive MR angiographic findings in our study is surprisingly high. One false-positive MR angiographic finding may have been caused by susceptibility artifact from a surgical clip immediately adjacent to the hepatic artery anastomosis, which was identified on subsequent angiography. Subsequent biopsy in this patient revealed no evidence of rejection or recurrent hepatitis. Two of the false-positive MR angiographic findings were found in patients with no diastolic flow on sonography, and MR angiography revealed a diffuse beaded appearance of the hepatic artery with relatively poor visualization of the distal artery (Fig. 4A,4B). Concurrent biopsy in these patients revealed acute rejection and recurrent hepatitis in one patient and acute rejection and recurrent hepatitis with bridging fibrosis and nodularity in the second patient. The findings on MR angiography may have been related to elevated end-organ resistance, so that relatively little contrast material flowed into the artery and its branches. Additional contributing factors could include pseudostenosis of the celiac trunk from the arcuate ligament of the diaphragm on end-inspiration.
|
In addition to the false-positive MR angiographic findings, discrepancies involved the proximal branches of the hepatic artery. In one patient with anastomotic stenosis, proximal stenoses of the right and left hepatic arteries were not prospectively identified on MR angiography, although, retrospectively, they can be seen. A second patient with a severe stenosis of the main hepatic artery had diffusely narrowed arteries distal to the stenosis on MR angiography, but conventional angiography showed a normal appearance of the hepatic artery and its branches distal to the anastomosis. This was important information in that it allowed for surgical revision and bypass of the stenosis. In general, MR angiography provided relatively poor visualization of distal hepatic artery branches compared with conventional angiography.
Our results are comparable with the more successful sonographic studies we have mentioned; however, we did not achieve the success of Stafford-Johnson et al. [16], who examined 13 liver transplant patients with three-dimensional gadolinium MR angiography and detected three cases of hepatic artery thrombosis and three cases of hepatic artery stenosis without any false-positive or false-negative findings.
Three-dimensional MR angiography remains a relatively new technique, and the optimal parameters for a particular examination are not yet clearly defined. Images may be improved with thinner sections and shorter TRs, resulting in faster scanning times and fewer motion artifacts from incomplete breath-holding. We have noticed a qualitative improvement in image quality in the eight examinations in which zero-fill interpolation with overlapping reconstructions was performed, and we now use this technique in all patients. A recent study of three-dimensional MR angiography in patients without liver transplants showed improved image quality with fat saturation [25]. The optimal contrast dose is also a matter of dispute. Excellent results have been achieved in many vessels using a single dose of gadolinium [14]. A double dose is costly but seemed a more prudent alternative to us because the interval of arterial enhancement is prolonged, and there is more room for error in scan timing. We used a test bolus in all patients to calculate the scan delay. This was important in achieving optimal results because the contrast travel time varied widely, from 10 sec in patients with central lines to 29 sec in a patient with congestive heart failure. Scan delay in all patients was adjusted to optimize the arterial phase. We always obtained good visualization of the portal system, hepatic veins, and inferior vena cava simply by acquiring two additional three-dimensional sequences; however, in cases in which there is a specific question regarding the portal vein or inferior vena cava, better results may be achieved by using the test bolus technique to optimize the venous phase.
Sonography has proven extremely useful in examining liver transplant patients, with high sensitivity for detection of arterial thrombosis and good sensitivity for detection of hepatic artery stenosis. However, sonography is sometimes technically challenging, and it is often difficult to visualize the arterial and portal venous anastomoses directly. MR angiography is a useful adjunct to sonography in these difficult cases. MR angiography provides visualization of the arterial supply to the liver from the celiac axis to the proximal intrahepatic branches. The anastomosis is almost always clearly seen. Likewise, the portal venous and inferior vena caval anastomoses are well visualized, and the splenic and superior mesenteric veins are also clearly identified. This is important in complex cases of partial portal vein thrombosis. The anatomic presentation of vascular images in multiple planes is useful for surgical planning when complications are identified and surgical revision is contemplated.
MR angiography is not ready to replace diagnostic angiography, at least in our hands. We have had a few false-positive findings for hepatic artery stenosis, one false-positive finding for portal venous stenosis, and one false-negative finding for hepatic artery stenosis. Accuracy of MR angiography will likely improve as more rapid scanning techniques are applied, allowing thinner sections and shorter breath-holds. Currently, we use MR angiography as a problem-solving technique in patients who have ambiguous results on sonography, who have a discrepancy between clinical and sonographic findings, or who are poor candidates for conventional angiography.
In summary, three-dimensional MR angiography is a useful technique in the evaluation of potential vascular complications of liver transplantation. Arterial and venous anatomy is well visualized, and most complications can be confidently diagnosed. MR angiography is a useful adjunct to sonography and offers a noninvasive alternative to conventional angiography.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. F. Hanna, N. Kased, S. W. Kwan, A. C. Gamst, A. C. Santosa, T. Hassanein, and C. B. Sirlin Double-Contrast MRI for Accurate Staging of Hepatocellular Carcinoma in Patients with Cirrhosis Am. J. Roentgenol., January 1, 2008; 190(1): 47 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. H. M. Caiado, R. Blasbalg, A. S. Z. Marcelino, M. da Cunha Pinho, M. C. Chammas, C. da Costa Leite, G. G. Cerri, A. C. de Oliveira, T. Bacchella, and M. C. C. Machado Complications of Liver Transplantation: Multimodality Imaging Approach RadioGraphics, September 1, 2007; 27(5): 1401 - 1417. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Hong, A. Y. Kim, H. J. Kim, K. W. Kim, S. G. Lee, P. N. Kim, M.-G. Lee, and H. K. Ha Inferior phrenic arterial bleeding after adult liver transplantation: incidence, clinical manifestations, and predictive CT features. Am. J. Roentgenol., July 1, 2006; 187(1): W15 - W19. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Berrocal, M. Parron, A. Alvarez-Luque, C. Prieto, and M. L. Santamaria Pediatric liver transplantation: a pictorial essay of early and late complications. RadioGraphics, July 1, 2006; 26(4): 1187 - 1209. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. S. Kim, T. K. Kim, D. J. Jung, J. H. Kim, I. Y. Bae, K.-B. Sung, P. N. Kim, H. K. Ha, S. G. Lee, and M.-G. Lee Vascular Complications After Living Related Liver Transplantation: Evaluation with Gadolinium-Enhanced Three-Dimensional MR Angiography Am. J. Roentgenol., August 1, 2003; 181(2): 467 - 474. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A S Shaw, S M Ryan, R C Beese, S Norris, M Bowles, M Rela, and P S Sidhu Liver transplantation Imaging, August 1, 2002; 14(4): 314 - 328. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. V. Pandharipande, V. S. Lee, G. R. Morgan, L. W. Teperman, G. A. Krinsky, N. M. Rofsky, M.-C. Roy, and J. C. Weinreb Vascular and Extravascular Complications of Liver Transplantation: Comprehensive Evaluation with Three-Dimensional Contrast-Enhanced Volumetric MR Imaging and MR Cholangiopancreatography Am. J. Roentgenol., November 1, 2001; 177(5): 1101 - 1107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Quiroga, M. C. Sebastia, C. Margarit, L. Castells, R. Boye, and A. Alvarez-Castells Complications of Orthotopic Liver Transplantation: Spectrum of Findings with Helical CT RadioGraphics, September 1, 2001; 21(5): 1085 - 1102. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Karlik How to Develop and Critique a Research Protocol Am. J. Roentgenol., June 1, 2001; 176(6): 1375 - 1380. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
