| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology and Research Institute of Radiology, University of
Ulsan College of Medicine, Asan Medical Center, 388-1, Pungnap-2 dong,
Songpa-ku, Seoul 138-736, Korea.
2 Department of Surgery, University of Ulsan College of Medicine, Asan Medical
Center, Seoul 138-736, Korea.
Received October 15, 2006;
accepted after revision January 26, 2007.
Address correspondence to K. W. Kim
(kimkw{at}amc.seoul.kr).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Using hepatic venography as the reference standard, 73 grafts with venographic evaluation in 43 dual-graft LDLT recipients were classified into either a stenosis (n = 39) or a nonstenosis (n = 34) group. CT scans were evaluated for relative attenuation, enhancement pattern, and HV abnormality for each graft. Doppler sonography evaluation of the flow pattern of HVs for each graft was performed. CT and Doppler sonography findings were compared in the stenosis and nonstenosis groups using the independent sample Student's t test and Fisher's exact test. Multifactorial logistic regression analysis was performed to determine the best predictors of the diagnosis of HV stenosis.
RESULTS. Heterogeneous enhancement (p = 0.046), abnormal HV on CT (p = 0.025), and HV wave pattern on Doppler sonography (p = 0.005) were significant findings. The accuracy for the diagnosis of HV stenosis was 60.0% for heterogeneous enhancement, 61.5% for abnormal HV, and 66.2% for a monophasic flow pattern. Heterogeneous enhancement and HV wave pattern were significant independent findings on multifactorial logistic regression analysis. The overall accuracy of the logistic model in the diagnosis of HV stenosis was 71.7%
CONCLUSION. Although CT and Doppler sonography can be helpful in diagnosing HV stenosis, given the low accuracy of individual imaging findings, the diagnosis of HV stenosis should be made cautiously, with both CT and Doppler sonography regarded as complementary examinations.
Keywords: CT • Doppler sonography • hepatic vein stenosis • liver transplantation
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Dual-graft LDLT is a surgical technique that transplants two grafts harvested from two donors into a single recipient. Although various combinations of graft types can be used, two left lobe grafts or a combination of one left lobe and one left lateral segment graft are most commonly used [4, 6]. Because a left lobe or left lateral segment graft usually has a single draining vein, single HV anastomosis is usually performed per each graft for dualgraft LDLT [4]. As in LDLT using a single graft, HV stenosis is one of the important vascular complications of dual-graft LDLT.
HV stenosis after LDLT can cause graft dysfunction and sometimes loss of the graft, especially during the immediate postoperative period [7, 8]. The radiologic diagnosis of HV stenosis after LDLT is important because early and adequate therapeutic intervention such as angioplasty or stenting can safely restore the graft function [7, 9]. Although several previous studies have indicated the efficacy of CT or Doppler sonography in the diagnosis of HV stenosis in LDLT recipients [10-13], these studies have focused on LDLT using a right lobe graft having multiple draining veins; therefore, little has been discovered about the value of CT or Doppler sonography in the diagnosis of HV stenosis in dual-graft LDLT in which a single HV anastomosis is usually performed for each graft.
Furthermore, to our knowledge, the CT findings of HV stenosis have been reported only for segmental venous congestion caused by impaired venous drainage of the anterior segmental veins in LDLT using a right lobe graft [11, 13], and there have been no previous reports regarding the CT findings of main draining vein stenosis in LDLT. Therefore, we performed the study reported here to evaluate the CT and Doppler sonography findings of HV stenosis in dual-graft LDLT with a single draining vein per graft and to determine the value of CT and Doppler sonography in the diagnosis of HV stenosis in dual-graft LDLT.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Forty-three patients were finally included in this study. They were 38 men and five women ranging in age from 18 to 66 years (mean age, 47.3 ± 9.2 years). All study patients underwent dual-graft LDLT using two left lobe grafts (n = 32) or one left lobe and one left lateral segment graft (n = 11). All patients had a single HV anastomosis per graft. This study was a single-institution retrospective study. Our institutional review board approved this study, but patient informed consent was not required for retrospective review of patient medical records and radiologic images.
Hepatic Venography and Corresponding CT and Doppler Sonography
Hepatic venography was performed using a 5-French cobra catheter via a
right internal jugular vein approach. Venous pressure on both sides of the HV
anastomosis was recorded, and the pressure gradient across the anastomosis was
calculated. Six of the 43 patients underwent hepatic venography twice.
Therefore, 49 hepatic venographies were performed in 43 dual-graft LDLT
recipients. A study coordinator selected a CT or Doppler sonography study
performed nearest to and within 1 week before each of the 49 hepatic
venographies. Forty-two CT and 46 Doppler sonography studies were selected.
Both CT and Doppler sonography studies that were performed within 1 week
before venography were available in 39 venographies, only Doppler sonography
in seven, and only CT in three.
Forty-two CT examinations were performed using either a 4-MDCT scanner (LightSpeed Plus, GE Healthcare) (n = 13) or a 16-MDCT scanner (Somatom Sensation 16, Siemens Medical Solutions) (n = 29). After obtaining the unenhanced CT scan, 150 mL of iopromide (Ultravist 370, Schering) was administered at a flow rate of 3 mL/s using a mechanical injector. Biphasic CT was then performed in the arterial phase and portal venous phase. Delay time for arterial phase scanning was determined by using bolus-tracking methods (SmartPrep, GE Healthcare or CARE-Bolus, Siemens Medical Solutions). Portal venous phase scanning was performed at a fixed delay of 72 seconds. CT parameters for the LightSpeed Plus scanner included detector configuration of 5 mm x 4 (unenhanced) or 2.5 mm x 4 (arterial and portal venous phase), table feed of 15 mm per gantry rotation, gantry rotation time of 0.6 second, 200 effective mAs, 120 kVp, 5-mm slice thickness, and 5-mm interval. CT parameters for the Somatom Sensation 16 scanner included detector configuration of 1.5 mm x 16 (unenhanced) or 0.75 mm x 16 (arterial and portal venous phases), table feed of 24 mm (unenhanced) or 12 mm (arterial and portal venous phases) per gantry rotation, 200 effective mAs, 120 kVp, 5-mm slice thickness, and 5-mm interval.
Forty-six Doppler sonography examinations were performed using a Sequoia 512 scanner (Acuson) with a 1-4-MHz convex probe, an HDI 5000 scanner (Philips Medical Systems) with a 2-4-MHz convex probe, or a LOGIQ 700 scanner (GE Healthcare) with a 2-4-MHz convex probe. Color and spectral Doppler sonograms were obtained of the hepatic arteries, portal veins, and HVs of both grafts. Spectral Doppler wave sonograms of the HVs were obtained within 3 cm of the inferior vena cava.
Image Interpretation
Two board-certified abdominal radiologists retrospectively reviewed the 46
Doppler sonograms and developed a consensus opinion. They evaluated the
Doppler sonograms unaware of the venography and CT findings and assessed the
resistive index of the hepatic artery, the peak flow velocity of the portal
vein in the postanastomotic segment, the portal vein flow pattern, and the HV
flow pattern for each graft. Portal vein flow patterns were classified as
continuous, to-and-fro, and reversed. We classified the HV wave into three
patterns—triphasic, biphasic, and monophasic. A triphasic wave consisted
of two forward peaks with a short period of reversed flow at the end of the
second forward wave. The monophasic wave was defined as the loss of normal
periodic pulsatility of the HV and a continuous flat wave. The biphasic wave
was defined as the loss of reversed flow and obscured two forward flow peaks
but preservation of HV pulsatility.
One month after completing their review of the Doppler sonograms, the same two reviewers similarly performed a retrospective review of 42 CT scans and formed a consensus opinion without any knowledge of the venography or Doppler sonography findings. The relative attenuation of each of the right and left grafts was determined by subjective visual assessment compared with the contralateral graft on unenhanced, arterial phase, and portal venous phase scans. The parenchymal enhancement patterns (homogeneous or heterogeneous) of each graft on the portal venous phase scans was noted. Heterogeneous enhancement was further classified as diffuse heterogeneous enhancement (inhomogeneous enhancement involving most of the graft) or segmental heterogeneous enhancement (segmental hypo- or hyperattenuation). The HVs of each graft were evaluated on the portal venous phase scans and were classified as normal or abnormal (i.e., focal narrowing, invisible, or nonopacified). Focal narrowing was defined as a short segmental slitlike narrowing of the HV anastomotic site with prestenotic dilatation and was considered to be present when the diameter of the narrowest anastomotic segment was less than 2 mm and the ratio of the diameter of the preanastomotic segment to that of the anastomotic segment was more than 2.0. When the HV was seen as a low-attenuating structure relative to the hepatic parenchyma, the HV was considered to be nonopacified. When a reviewer was unable to identify the HV, it was considered to be invisible.
Classification of Grafts According to Hepatic Venography
After reviewing the CT scans and Doppler sonograms, the two reviewers
interpreted the hepatic venograms. During a hepatic venography session,
venographic examination of the HV was performed for both grafts (n =
24) or for only a single graft (n = 25) depending on the patient's
condition. As a result, 49 venography sessions provided venographic
examinations of 73 graft HVs. Using the hepatic venogram as the reference
standard, 73 venographically evaluated grafts were divided into either
stenotic or nonstenotic groups according to the presence or absence of HV
anastomotic stenosis on venography. The criteria for HV anastomotic stenosis
were a pressure gradient across the HV anastomosis of 6 mm Hg or more or
contrast stagnation in the HV with severe narrowing at the anastomosis seen on
venography. Among 73 grafts, corresponding Doppler sonograms were available in
68 grafts and corresponding CT scans in 65 grafts.
Statistical Analysis
The Kolmogorov-Smirnov test was used to determine the normality of the
continuous values. Unifactorial analysis was performed by comparing the CT
findings and the Doppler sonography findings between the stenotic and
nonstenotic groups using the independent sample Student's t test and
Fisher's exact test. For all CT and Doppler sonography findings that were
statistically significant on unifactorial analysis, the sensitivity,
specificity, and accuracy of each finding in the diagnosis of HV stenosis were
calculated on a per-graft basis with a 95% CI. To determine the best
predictors of the diagnosis of HV stenosis, multifactorial backward stepwise
logistic regression analysis was performed, including variables with a
p value less than 0.25 in the unifactorial analysis. A logistic model
for the diagnosis of HV stenosis was obtained using resultant independent
significant variables, and the diagnostic accuracy of the logistic model was
then calculated. For all statistical analysis, a p value less than
0.05 was considered statistically significant. Statistical analyses were
performed using the SPSS version 11.5 software package.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Unifactorial Analysis
Table 1 summarizes the
results of the comparison of the CT findings in the stenotic and nonstenotic
groups. Relative graft attenuation showed no statistically significant
differences between the two groups in any of the three phases (i.e.,
unenhanced, arterial, and portal venous phases). The parenchymal enhancement
pattern on portal venous phase was significantly different between the two
groups (p = 0.046), with the heterogeneous enhancement pattern being
seen more often in the stenotic group (n = 13 [37.1%]) (Figs.
1A,
1B,
1C,
2A,
2B,
2C,
3A,
3B,
3C, and
3D) than in the nonstenotic
group (n = 4 [13.3%]). All grafts with a heterogeneous enhancement
pattern on portal venous phase scans showed diffuse heterogeneous enhancement
(Figs. 1A,
1B,
1C,
2A,
2B,
2C,
3A,
3B,
3C, and
3D). Segmental heterogeneous
enhancement was not seen in any graft. An abnormal HV on portal venous phase
scans was significantly more common in the stenotic group (n = 14
[40.0%]) (Figs. 1A,
1B,
1C,
2A,
2B,
2C,
3A,
3B,
3C,
3D,
4A,
4B, and
4C) than in the nonstenotic
group (n = 4 [13.3%]) (p = 0.025). Although invisible or
nonopacified HVs were seen in the stenotic group (n = 7 [20.0%]) and
in the nonstenotic group (n = 4 [13.3%]), focal narrowing was seen
only in the stenotic group (n = 7 [20.0%]) (Figs.
3A,
3B,
3C,
3D,
4A,
4B, and
4C). Among the Doppler
sonography findings, only the HV wave pattern showed a statistically
significant difference between the two groups (p = 0.005)
(Table 2.). A monophasic flow
pattern was more common in the stenotic group (n = 19 [50.0%]) (Figs.
1A,
1B,
1C,
3A,
3B,
3C,
3D,
4A,
4B, and
4C) than in nonstenotic group
(n = 6 [20.0%]), whereas a triphasic flow pattern was more common in
the nonstenotic group (n = 15 [50.0%]) than in the stenotic group
(n = 6 [15.8%]).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
For the three findings that were statistically significant in the unifactorial analysis, the respective sensitivity, specificity, and accuracy in the diagnosis of HV stenosis were as follows: for heterogeneous enhancement on portal venous phase scans, 37.1% (95% CI: 21.4-55.1%), 86.7% (69.3-96.2%), and 60.0% (47.8-71.0%); for abnormal HVs, 40.0% (24.0-57.9%), 86.7% (69.3-96.2%), and 61.5% (49.4-72.4%); for the monophasic flow pattern, 50.0% (33.4-66.6%), 80.0% (61.4-92.3%), and 66.2% (51.3-73.7%). The triphasic flow pattern was a finding suggesting the absence of HV stenosis with sensitivity, specificity, and accuracy excluding HV stenosis of 50.0% (31.3-68.7%), 84.2% (68.7-94.0%), and 69.1% (57.3-78.9%), respectively.
Multifactorial Logistic Regression Analysis
Variables with a p value less than 0.25 in the unifactorial
analysis included heterogeneous enhancement (p = 0.046), abnormal HV
(p = 0.025), HV wave pattern (p = 0.005), and portal vein
velocity (p = 0.118). Among these CT and Doppler sonography findings,
multifactorial backward stepwise logistic regression analysis showed that
heterogeneous enhancement on portal venous phase CT and the HV wave pattern on
Doppler sonography were independent significant variables (p = 0.028
and 0.002, respectively). With the resultant logistic model of y = 1
/ (1+EXP(1.991+1.503x A – 1.105 x B),
where y is the possibility of HV stenosis, A is the presence
or absence of heterogeneous enhancement (0 = absent, 1 = present), and
B is the HV wave pattern (1 = monophasic, 2 = biphasic, and 3 =
triphasic). The overall accuracy of the logistic model in the diagnosis of HV
stenosis was 71.7% when a y value (the possibility of HV stenosis) of
0.5 or higher was considered as the positivity for HV stenosis. This model
correctly diagnosed the presence or absence of HV stenosis in 43 of 60
grafts.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our study showed several CT and Doppler sonography findings that would be helpful in the diagnosis of HV stenosis. Heterogeneous parenchymal enhancement and an abnormal HV on portal venous phase CT scans and monophasic flow pattern of the HV on Doppler sonography were suggestive of HV stenosis, whereas the triphasic flow pattern of the HV was the finding suggesting the absence of HV stenosis. Regarding the parenchymal enhancement pattern, heterogeneous enhancement seen in our study was completely different from that seen in a previous study of right lobe LDLT [11] that reported segmental hyper- or hypoattenuation as the CT finding of HV congestion.
In our study, all grafts with a heterogeneous enhancement pattern showed inhomogeneous enhancement involving most of the graft. Segmental heterogeneous enhancement was not seen in any graft in our study. This difference can be explained by the presence of multiple draining HVs in the right lobe graft. In right lobe LDLT, stenosis of a single HV may cause venous congestion in the area drained by the stenotic HV, which would result in a segmental attenuation difference. However, owing to the presence of a single draining vein per each graft in most dual-graft LDLTs, HV stenosis may result in venous congestion of the entire graft, which would manifest as a heterogeneous enhancement pattern similar to the findings of passive hepatic congestion [14] or the acute stage of Budd-Chiari syndrome [15].
Although there were several suggestive CT and Doppler sonography findings of HV stenosis, it should be noted that the sensitivities of these findings in the diagnosis of HV stenosis were disappointing. Although the monophasic flow pattern was the most sensitive of these findings, the sensitivity of the monophasic flow pattern was only 50% in the diagnosis of HV stenosis. The results of our study are completely different from those from the study of Ko et al. [10] of right lobe LDLT, which showed that the monophasic flow pattern of HV was a sensitive but not a specific indicator of right HV stenosis. The reasons for the discrepancies between the two studies are not completely understood; however, several factors can be considered. We used a lower cutoff value of the pressure gradient (6 mm Hg) than did Ko et al. (10 mm Hg) to determine the presence of HV stenosis on hepatic venography because a pressure gradient of 6 mm Hg or more is reported to be clinically and hemodynamically significant [9, 16] and is regarded as the indicator for stent insertion [17]. Furthermore, because we frequently performed venographic examinations for both grafts during a single venography session, venography may have identified mild unexpected HV stenosis of the contralateral graft other than the graft suspected of having HV stenosis on CT or Doppler sonography. Therefore, because our study might include a milder degree of HV stenosis in the stenosis group than did the study of Ko et al., this might result in a low sensitivity of the monophasic flow pattern and other imaging findings suggestive of HV stenosis in our study.
Multifactorial logistic regression analysis showed that heterogeneous enhancement on portal venous phase scanning and an HV wave pattern on Doppler sonography were significant independent variables in the diagnosis of HV stenosis. However, the accuracy of each finding was disappointing for the diagnosis of HV stenosis because it was only 60.0% for heterogeneous enhancement and 66.2% for the monophasic flow pattern. The overall accuracy of the logistic model (using these two findings) in the diagnosis of HV stenosis was slightly higher (71.7%).
Given the low accuracy of individual CT and Doppler sonography findings in the diagnosis of HV stenosis, we assume that the diagnosis of HV stenosis in dual-graft LDLT recipients should not be made on a single finding. Furthermore, it should be emphasized that these findings were reported to occur in various pathologic conditions other than HV stenosis [18-20]. Heterogeneous parenchymal enhancement has been reported in fatty liver disease, massive hepatic necrosis, and acute rejection [18]. A dampened hepatic venous wave can also be seen in patients with liver cirrhosis [19], hepatic steatosis, and obesity [20]. Therefore, to improve diagnostic accuracy, both CT and Doppler sonography findings should be taken into consideration and be regarded as complementary examinations in the diagnosis of HV stenosis after dual-graft LDLT. Furthermore, because of the low sensitivity of CT and Doppler sonography in the diagnosis of HV stenosis, hepatic venography should be performed to exclude HV stenosis when it is clinically suspected, even in the absence of any suggestive CT or Doppler sonography findings of HV stenosis.
In addition to HV flow pattern that we evaluated in our study, there are several other sonography findings that may be helpful in the diagnosis of HV stenosis after LDLT, which include reduced peak flow velocity of the HV, high-flow jet at the anastomosis, and morphologic stenosis of the anastomotic site on color Doppler sonography [21, 22]. A sonographic contrast agent may improve the visualization of the HV anastomotic site and thus help in the morphologic evaluation of the anastomotic site [23], especially in cases with a poor sonic window, in which the direct visualization of HV anastomosis is challenging.
Several limitations of our study should be mentioned. First, there may have been a selection bias in our study population. Because we included only dual-graft LDLT recipients who had undergone hepatic venography, most of our study population consisted of patients who were clinically or radiologically suspected of having HV stenosis. On the other hand, most of the dual-graft LDLT recipients who actually did not have HV stenosis and were not suspected of having HV stenosis clinically or radiologically might have been excluded from our study population. Second, intrinsic limitations of a retrospective study such as our study should be noted. In our study, the practical impact of CT and Doppler sonography on the diagnosis and management of HV stenosis could not be evaluated. Therefore, it is still uncertain whether the information provided by CT and Doppler sonography can alter the management plan and clinical course of LDLT recipients with HV stenosis. For this purpose, we think a further prospective trial is required.
Third, although previous research has reported that patient respiration has a significant effect on the spectral Doppler wave of the HV (Lee SS et al., presented at the 2006 annual meeting of the World Federation for Ultrasound in Medicine and Biology), we did not regulate patient respiration while acquiring the spectral Doppler wave of the HV. Because it has been reported that breath-hold at inspiration can make otherwise pulsatile HV flow monophasic in right lobe LDLT recipients without postoperative complications (Lee SS et al, 2006 WFUMB meeting), our lack of regulation of patient respiration may have reduced the specificity of the monophasic flow pattern in the diagnosis of HV stenosis.
Finally, although we selected a pressure gradient of 6 mm Hg as a cutoff value to determine the presence of HV stenosis on hepatic venography based on its clinical and hemodynamic significance shown in previous studies [9, 16], there is still no consensus regarding the degree of HV pressure gradient that is hemodynamically significant and requires therapeutic intervention. Until now, variable cutoff values of pressure gradient have been used for the diagnosis of HV stenosis ranging from 3 mm Hg [24] to 10 mm Hg [10]. We think fixed criteria for significant HV stenosis are required to make the results of preceding and future studies regarding HV stenosis reproducible. A further large prospective study may be required to deal with this issue.
In conclusion, heterogeneous parenchymal enhancement and an abnormal HV on portal venous phase CT and the monophasic flow pattern of HVs on Doppler sonography are suggestive findings of HV stenosis in dual-graft LDLT. However, the role of CT or Doppler sonography in the diagnosis of HV stenosis is limited after dual-graft LDLT. Because of the low accuracy of individual CT and Doppler sonography findings in the diagnosis of HV stenosis, the diagnosis of HV stenosis should be made cautiously, using both CT and Doppler sonography as complementary examinations.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
