| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Medical University of Vienna, Waehringer Guertel
18-20, Vienna A-1090, Austria.
2 Department of Internal Medicine IV, Gastroenterology and Hepatology,
University of Vienna, Austria.
Received February 12, 2004;
accepted after revision July 19, 2004.
Address correspondence to T. R. Bader.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Perfusion of the liver parenchyma was measured in 24 healthy volunteers and 41 patients with liver cirrhosis using dynamic single-section CT. Seventeen patients underwent TIPS placement, and CT measurements were repeated within 7 days. CT scans were obtained at a single level comprising the liver, spleen, aorta, and portal vein. Scans were obtained over a period of 88 sec (one baseline scan followed by 16 scans every 2 sec and eight scans every 7 sec) beginning with the injection of a contrast agent bolus (40 mL at 10 mL/sec). Parenchymal and vascular contrast enhancement was measured with regions of interest, and time–density curves were obtained. These data were processed with a pharmaco-dynamic fitting program (TopFit), and the arterial and portal venous component and the total perfusion of the hepatic parenchyma were calculated (milliliters of perfusion per minute per 100 mL of tissue).
RESULTS. Mean normal values for hepatic arterial, portal venous, and total perfusion were 20, 102, and 122 mL/min per 100 mL, respectively. In patients with cirrhosis before TIPS, mean hepatic arterial, portal venous, and total perfusion was 28, 63, and 91 mL/min per 100 mL, respectively, which was statistically significant for all values (p < 0.05). After TIPS, hepatic perfusion increased to a mean value of 48, 65, 113 mL/min per 100 mL for arterial (p < 0.01), portal venous, and total (p = 0.011) perfusion, respectively.
CONCLUSION. In patients with cirrhosis, the hepatic arterial perfusion increased, whereas portal venous and total perfusion decreased compared with that of healthy volunteers. TIPS placement caused a statistically significant increase of the hepatic arterial and total hepatic perfusion. The portal venous parenchymal perfusion remained unchanged.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It has been shown that the measurement of perfusion of the liver parenchyma can be successfully accomplished with dynamic single-section CT [5, 6]. This technique has been shown to be effective in measuring hepatic perfusion in liver transplantation and in patients with metastases [7–9]. It allows the separate quantification of the arterial and portal venous components of hepatic parenchymal perfusion, enabling the assessment of normal physiology and liver disease [10, 11]. The purpose of this study was to evaluate the arterial, portal venous, and total liver perfusion in patients with cirrhosis before and after TIPS placement, using dynamic single-section CT, and to compare these results with normal values.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Study Population
There were 41 patients with liver cirrhosis and 24 healthy volunteers who
were examined with dynamic single-section CT of the liver. The study group
comprised 41 patients with liver cirrhosis, who were evaluated for TIPS
placement. Among these, a total of 17 patients underwent the TIPS procedure
and a second dynamic single-section CT examination within 2–7
days after TIPS placement. The other 24 patients did not receive a TIPS
placement because of clinical considerations, such as hepatic encephalopathy.
The control group of 24 healthy volunteers (16 men and eight women; mean age,
60.5 ± 12.7 years [SD]; age range, 32–81 years) had no
history of liver disease. All patients of the healthy control group underwent
sonography of the liver, including color-coded Doppler sonography of the
portal vein before CT examination, to exclude any incidental liver
abnormality.
In these 17 patients (12 men and five women; mean age, 53.7 ± 11.9 years; age range, 22–68 years) who underwent the TIPS procedure, the indication for this intervention was variceal bleeding (11 patients) or refractory ascites (six patients). Liver cirrhosis in these patients was caused by hepatitis C in five patients and by hepatitis B in one patient and was alcohol-related in nine patients and cryptogenic in two patients. According to the Child-Pugh classification, five patients were classified as Child A, seven as Child B, and one as Child C.
TIPS Placement
Under fluoroscopic guidance, a side-to-side portacaval shunt was implanted
in an appropriate branch of the intrahepatic portal system. The shunt
diameters varied between 8 and 14 mm, with a 12-mm diameter chosen in 60% of
all patients. The portal venous pressure gradient was reduced to below 12 mm
Hg in all patients after TIPS placement. Mean portal pressure decreased from
22 ± 5.6 to 11 ± 3.3 mm Hg (p < 0.05). No clinical
complication was observed during a 7-day follow-up period after TIPS
placement.
CT Data Acquisition
Dynamic single-section CT was performed on a Somatom Plus–4 scanner
(Siemens), as described by Blomley et al.
[6] and Bader et al.
[8]. On axial scout scans, a
level was determined that comprised the liver, the portal vein, the aorta, and
the spleen. After TIPS placement, a level through the liver just below the
shunt was selected. The CT protocol comprised 25 dynamic scans (120 kV, 125
mAs, 8-mm collimation), which were obtained at a single level with no table
feed. Iopromide ([40 mL, 300 mg I/mL] Ultravist, Schering) was injected with
an automatic injection pump at a flow rate of 10 mL per second via an
antecubital vein (17-gauge needle). The contrast agent was immediately
followed by a 40-mL saline solution at the same flow rate. For this purpose,
saline solution was layered above the contrast agent in the injection
syringe.
The initial scan of the serial CT scans was an unenhanced baseline scan (Fig. 1A). Consecutive scans were obtained every 2 sec between 7 and 39 sec after the beginning of the contrast injection. Then, scans were obtained every 7 sec until 88 sec, resulting in a total of 25 scans. Patients were advised to breath-hold for as long as possible and to breathe as calmly and shallowly as possible when they could not hold their breaths any longer, to avoid major respiratory artifacts.
|
CT Data Analysis
On each CT scan, the densities of the liver, spleen, aorta, and portal vein
were measured in Hounsfield units. For measurements of the liver and spleen
parenchyma, irregular regions of interest (ROI) were drawn to encompass as
much parenchyma as possible with the exclusion of the large vessels and
imaging artifacts (e.g., metal artifacts). Circular ROIs were used for the
aorta and portal vein (Figs.
1B,
1C, and
2A,
2B). Care was taken to avoid
partial volume effects. If the portal vein was not visualized on single
images, these data points were excluded and extrapolated by the fitting
procedure. The attenuation values of the initial unenhanced baseline scan were
subtracted from each subsequent measurement, giving relative measures of
contrast enhancement [8]. These
attenuation values were plotted along the time axis, giving
time–density curves of relative enhancement of the liver,
aorta, portal vein, and spleen.
|
|
|
|
The original CT data were processed with a fitting procedure using a
pharmacokinetic program [12]
with a linear two-compartment model comprising the vascular and interstitial
space and describing dose-proportional pharmacokinetics suitable for contrast
material distribution [12,
13]. Further processing was
performed with commercially available software (Excel 97 for Windows NT,
Microsoft). The duration of the hepatic arterial phase (HAP) was calculated
according to the following equation
[10,
14]:
 | (1) |
tHAP is the duration of HAP,
tlag is the lag time of the portal vein, and
tlag AO is the lag time of the aorta. Lag time is the time
between the beginning of the contrast agent injection and the arrival of the
contrast agent at the site of measurement
[10,
14].
The cutoff point between the arterial and the portal venous phase, which is
the lag time of the hepatic portal venous phase (HPP), was calculated as
follows:
 | (2) |
The perfusion (P) (blood flow per minute per volume unit of
tissue) of the liver and the spleen was calculated according to the following
equation [5,
6,
15]:
 | (3) |
Statistical Analysis
Statistical analysis was performed with commercially available software
SPSS version 11.5 (Statistical Package for the Social Sciences) for Windows
(Microsoft). All results were statistically described (mean, confidence
interval, SD, range, level of confidence at 95%). A p value of less
than 0.05 was considered significant. To rule out significant differences in
the distribution of sex and age in the study group and the control group, we
used the Student's t test and the Fisher's exact test. The arterial,
portal venous, total hepatic perfusion values and the splenic perfusion of the
control group and the study group before TIPS were compared using the
two-tailed unpaired Student's t test. The perfusion values of the
study group with cirrhosis, before TIPS and after TIPS, were analyzed for
statistically significant differences using the two-tailed paired Student's
t test (level of significance, p < 0.05).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
In healthy volunteers, the hepatic arterial perfusion was statistically significantly lower than that in patients with cirrhosis (p = 0.038), whereas the portal venous perfusion was statistically significantly higher than that in patients with cirrhosis (p < 0.01). The total hepatic perfusion was statistically significantly higher in healthy control subjects than in patients with cirrhosis (p = 0.026) before TIPS implantation (Table 1 and Fig. 3). After TIPS placement, the mean arterial hepatic perfusion increased from 27.9 ± 22.6 to 48 ± 33.3 mL/min per 100 mL, which was statistically highly significant (p < 0.01) (Table 2 and Fig. 4). After TIPS placement, total liver perfusion increased statistically significantly from a mean value of 90.8 ± 28.1 to 113 ± 38.7 mL/min per 100 mL (p = 0.011). There was no statistically significant difference (p > 0.05) compared with healthy control subjects. Mean portal venous perfusion increased from 62.9 ± 23 to 64.6 ± 23 mL/min per 100 mL, which was not statistically significant (p > 0.05) (Table 2 and Fig. 4). Mean splenic perfusion in healthy volunteers was 169 ± 72.8 mL/min per 100 mL. In patients with cirrhosis, the mean splenic perfusion increased from 102 ± 37.6 mL/min per 100 mL before TIPS to 182.7 ± 105.2 mL/min per 100 mL after TIPS placement, which was statistically significant (p < 0.01).
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Consistent with the method described by Bader et al. [10], we used a preprocessing curve-fitting procedure to attenuate irregularities of the time–density curves that are predominantly caused by patient motion due to breathing. This procedure applies an exponential function that describes the distribution of a contrast agent after peripheral injection and corrects for recirculation. Another advantage of the fitting procedure was that it allowed the assessment of complete time– density curves even if singular data points were missing (e.g., due to motion artifacts).
A small amount of contrast agent and a high injection flow rate are necessary to achieve a tight bolus, which is mandatory for the separate measurement of the arterial and the portal venous components of hepatic perfusion. Patients with cirrhosis showed increased arterial hepatic perfusion compared with healthy patients; this finding is in accord with prior reports [6, 22–24]. Both the portal venous perfusion and the total hepatic perfusion of patients with cirrhosis were statistically significantly reduced compared with those of control subjects.
Liver cirrhosis causes regional hepatic and general hemodynamic changes. Multiple intrahepatic shunts between branches of the hepatic artery, the portal vein, and the hepatic veins lead to arterialization of the liver perfusion. Splanchnic nitric oxide overproduction leads to vasodilation of the splanchnic and systemic vasculature and, consequently, to a hyperdynamic circulatory state, whereas an increase of the intrahepatic vascular resistance is believed to result from stellate cell contraction, which is mainly caused by a decrease in intrahepatic nitric oxide production by endothelins [25–28].
A large portion of the hepatic blood bypasses the sinusoids through numerous intrahepatic shunts [29] and additional perihepatic venous collaterals. A compensatory sequela is increased arterial blood flow in hepatic sinusoids, which is also fostered by the hepatic artery buffer response [30].
After TIPS, the portacaval pressure gradient was reduced to values below 12 mm Hg in all patients, which is generally considered sufficient for the prevention of recurrent variceal bleeding and reduction of recurrent ascites and meets the criteria for a successful TIPS procedure [27, 31]. The portal venous perfusion after TIPS placement remained almost unchanged. The total hepatic perfusion increased statistically significantly because of the increase in hepatic arterial perfusion.
We believe that the increased arterial hepatic perfusion is caused by an increase in the systemic hyperdynamic circulatory state after TIPS placement due to an increased cardiac output and a decreased systemic vascular resistance. These hemodynamic changes after TIPS placement have been described by Lotterer et al. [27] as short-term effects of TIPS, which do not persist longer than a few weeks after the procedure. The underlying physiologic mechanisms of these hemodynamic changes cannot be fully explained as yet. It is assumed that redistribution of blood volume from the portal to the systemic venous system directly to the right heart caused by the TIPS, combined with parahepatic shunting of vasoactive substances and reduction of ascites, might cause this phenomenon [27].
The increase in the splenic perfusion after TIPS placement observed in our study may, therefore, be a consequence of the increased systemic vascular blood. However, this increase could also be explained by the decrease in splenic vein congestion, which leads to increased arterial inflow into the spleen.
One limitation of this study was the relatively small number of TIPS patients, which precludes detailed analyses according to Child-Pugh groups. Liver function was not routinely assessed (e.g., prothrombin ratio); this omission precludes correlations of perfusion values with synthetic liver function. Liver cirrhosis causes inhomogeneity of the liver parenchyma that might lead to local variations in parenchymal perfusion. The dynamic singlesection CT examination comprised only one representative level of the liver, and the level of the imaging series of the pre- and post-TIPS CT studies showed slight variations in alignment (2 cm difference at most). We believe that these variations of the scanning levels do not significantly affect the results of this study because cirrhosis is a diffuse liver disease that involves the entire organ, and inhomogeneity of the parenchyma is present in all levels of the liver. Furthermore, variations in local perfusion are averaged out by the large size of the volume of measured liver parenchyma on pre- and post-TIPS scans.
In summary, in patients with cirrhosis, hepatic arterial perfusion is increased and portal venous and total perfusions are significantly decreased. After TIPS, hepatic artery and total perfusions increase further, whereas the portal venous perfusion of the liver parenchyma remains unchanged. This technique might be beneficial for the follow-up of diffuse liver disease and the follow-up of patients after TIPS.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  Y. Tsushima, K. Endo, C. Weidekamm, and T. Bader Portal Perfusion Measurement on Dynamic CT in Patients with Liver Cirrhosis Am. J. Roentgenol., September 1, 2005; 185(3): 813 - 813. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
