| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Departments of Radiology (G.B., M.P.F., K.P.) and Surgery (D.A.G.), University of Pittsburgh Medical Center, Presbyterian Hospital, Rm 4660, 200 Lothrop St, Pittsburgh, PA 15213-2582. Received October 6, 2000; revision requested November 16; revision received January 29, 2001; accepted February 26. G.B. supported by the Nicholas Green Fulbright Grant. Address correspondence to M.P.F. (e-mail: federle@pitt.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Among 379 OLTs, PV thrombosis or sclerosis was found at surgery in 39 patients (10.3%). Before OLT, surgical records and CT images were retrospectively reviewed in 35 patients with available CT studies. Diameter of the extrahepatic PV and patency of the PV system were evaluated. Cavernous transformation, calcifications of the venous wall or thrombus, lesions suggestive of tumor, mesenteric varices, edema, or splenorenal shunt were recorded. A nonpaired Student t test and the Fisher exact test were used to analyze the results.
RESULTS: Of 35 patients, 23 (66%) underwent thrombectomy with direct PV-to-PV anastomosis and 12 (34%) had placement of venous grafts or other anastomoses. The extrahepatic PV was 8.2 mm, but it significantly (P 
.05) decreased in patients with splenorenal shunt. In 30 patients, CT depicted thrombosis, PV calcification, or other abnormalities. The thrombus extended to or beyond the confluence of the splenic and superior mesenteric veins in 21 (60%) and 11 (31%) patients, respectively. Eleven patients (31%) had cavernous transformation of the PV; eight (23%), a cordlike sclerotic PV; 19 (54%), a splenorenal shunt; 11 (31%), PV calcification; 17 (49%), mesenteric edema; 14 (40%), mesenteric varices. Patients with a cordlike or calcified PV were significantly (P .05) more likely to require modification of the surgical technique.
CONCLUSION: Before OLT, CT can aid in assessment of PV and associated findings and in surgical management.
Index terms: Liver, transplantation, 761.455 • Portal vein, CT, 957.12912, 957.12915 • Portal vein, thrombosis, 957.751
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
PV thrombosis is no longer considered a contraindication to orthotopic liver transplantation (OLT), but it does have a substantial effect on surgical complexity and perioperative morbidity and mortality rates (4). In addition, PV calcification may be detected preoperatively, and there are anecdotal reports of calcified veins encountered at OLT with difficult (14) or impossible venous anastomosis and fatal consequences (2). Therefore, patients who have PV anomalies recognized at preoperative imaging are often referred to more experienced transplantation centers for optimal care.
The prevalence of PV thrombosis found at OLT ranges from 2.1% to 13.0% in large series (4,15). The role and accuracy of computed tomography (CT) in depicting PV thrombosis and associated anomalies have not been studied extensively. Most published reports stress the role of ultrasonography (US) (16,17) and angiography (18), but these have well-recognized limitations, especially in cases of partial thrombosis, calcification, and cavernous transformation of the PV.
At our transplantation center, all candidates for OLT are screened with helical multiphasic CT to evaluate for signs of cirrhosis, portal hypertension, focal hepatic lesions, extrahepatic diseases, and any hepatic or extrahepatic vascular anomalies that might influence the surgical technique.
The purpose of this investigation was to review the CT findings in patients with surgically proved PV thrombosis or sclerosis and to correlate these findings with the surgical procedure used at OLT.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
All patients had cirrhosis. The causes of cirrhosis were cryptogenic (n = 9), autoimmune (n = 8), and alcoholic (n = 3); hepatitis C (n = 5), hepatitis B (n = 4), and primary biliary cirrhosis (n = 2); and there was one case each of primary sclerosing cholangitis, Wilson disease, 
-1 antitrypsin deficiency, and hemochromatosis. The average time between CT and OLT was 155 days (range, 0–460 days). Seventeen patients underwent CT within 4 months, and 32 patients underwent CT within 10 months prior to OLT.
CT Imaging
All CT scans were obtained at our institution with a helical CT scanner (Hi-Speed Advantage or LightSpeed QX/I; GE Medical Systems, Milwaukee, Wis). In all patients, CT examinations were performed by using a multiphasic CT protocol that included images obtained during the unenhanced hepatic arterial dominant phase and during the PV dominant phase. In the series obtained during the hepatic arterial dominant phase, the scanning delay was 25–35 seconds; at the PV dominant phase, 60–70 seconds following injection of a bolus of contrast material. All patients received 125 or 150 mL of either iothalamate meglumine (Conray 60; Mallinckrodt Medical, St Louis, Mo) or ioversol (Optiray 350; Mallinckrodt Medical) injected intravenously at a rate of 3–5 mL/sec with a power injector (OP 100; Medrad, Pittsburgh, Pa).
Images were reviewed with a picture archiving and communication system (PACS) workstation (Impax RS3000 1K review station; AGFA Technical Imaging Systems, Richfield Park, NJ) or as hard copy images by two abdominal imaging radiologists (G.B. and M.P.F.) who had knowledge of the surgical diagnosis of PV thrombosis but not of specific surgical and pathologic findings. In cases of interobserver disagreement, final decisions were reached by means of consensus. No attempt was made to test for interobserver differences because disagreements were considered relatively few and minor. Both observers agreed on the number and extent of thrombosed vessels in all patients.
We evaluated the maximum diameter of the extrahepatic main PV and the patency of the main portal, splenic, and superior mesenteric veins. Thrombosis was classified as occlusive or nonocclusive. Cavernous transformation or presence of a previous surgical or spontaneous splenorenal shunt was recorded. We also evaluated for calcifications in the venous wall or in the thrombi within the vessels, mesenteric varices and mesenteric edema, and hepatic lesions suggestive of tumor. In patients with hepatocellular carcinoma, we looked for evidence of PV tumor thrombus, such as distention of the PV lumen, enhancement of the thrombus, or proximity of tumor and thrombosed PV.
The maximum transverse diameter of the PV was measured with calipers from wall to wall in the segment of the extrahepatic PV located between the hepatic artery and the caudate lobe. The vessels were defined as patent if the entire lumen was filled with contrast material–enhanced blood on enhanced images. The splenic vein was evaluated from the confluence to splenic hilum; and the superior mesenteric vein, from the confluence to the right colic branch.
An occlusive thrombus was defined as a nonenhancing filling defect within the lumen of the vessel seen on contrast material–enhanced images. Nonocclusive thrombi were identified when there was contrast-opacified blood adjacent to the existing thrombus, and these thrombi were further divided into those occupying more than 50% of the lumen or less than 50% of the lumen. Cavernous transformation was defined as an abundance of collateral veins in the hepatic hilum that provided an alternative route around a thrombosed segment of the main PV or lobar branches. Acute PV thrombosis was defined as abnormal distention of the PV (>20 mm), with absence of calcification, cavernous transformation, or other well-developed collateral veins. PV was classified as cordlike or diminutive when no patent lumen could be identified and when the vessel could hardly be seen. A splenorenal shunt was defined as a surgical or spontaneous anastomosis of the splenic vein or a perisplenic varix to an enlarged left renal vein. A calcification was defined as an elongated or punctate area of hyperattenuation seen along the venous walls or within the thrombus.
Any discrete hepatic lesion enhancing in the arterial dominant phase was considered hepatocellular carcinoma until proved otherwise. Mesenteric varices were defined as enlarged collateral veins located in the mesentery. Mesenteric edema was diagnosed if the adipose tissue around the mesenteric vessels or their branches had a visually perceptible increased attenuation relative to that of subcutaneous or retroperitoneal fat. Any patent collateral vein, such as the coronary vein or umbilical vein, of a size that would have allowed anastomosis to the PV was noted. All of these findings were tabulated for each patient.
Surgical Report
Two authors, a research associate (K.P.) and a transplant surgeon (D.A.G.), evaluated the surgical reports of the anatomic findings in the PV system, including any portosystemic shunt, and the specific comments of the surgeons on the presence and extent of sclerosis, thrombosis, or friability of the PV. The transplant surgeons were not blinded to the CT results and so their findings were not independent. We recorded the procedures used to deal with PV thrombosis, such as thrombectomy, use of a venous graft to create an anastomosis, or use of a collateral vein. The findings at CT were compared with those from surgery by using consensus among the radiologists (M.P.F., G.B.) and the surgeon (D.A.G.). The surgical pathology report was also reviewed for evidence of hepatocellular carcinoma in the excised liver.
Statistical Analysis
A nonpaired Student t test was used to assess the statistical significance of diameter differences of the main PV between patients with and those without splenorenal shunt. To assess whether there was increased use of interposition grafts in patients with what we assumed were the signs of sclerosis, we used a Fisher exact test to determine the statistical significance of differences in incidence of calcifications, splenorenal shunt, occlusion to or beyond superior mesenteric-splenic venous confluence, cordlike appearance, and cavernous transformation; we used this test in patients who underwent thrombectomy alone and in patients who had a major alteration of the surgical procedure.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
Of the 28 patients with CT signs of thrombosis of the extrahepatic PV, nine had complete thrombosis of the main PV and 19 had partial or nonocclusive thrombus, 11 with thrombus filling less than 50% of the lumen and eight with occlusion of more than 50% (Fig 6). The thrombus extended to the confluence of the splenic and superior mesenteric veins in 21 (60%) of 35 patients and extended into the splenic or superior mesenteric vein in 11 (31%) of 35 (Fig 7). No patients had thrombosis of the splenic or superior mesenteric vein or confluence without involvement of the main PV. Of the 11 patients with thrombus extending beyond the PV confluence, four (36%) had modification of the OLT procedure beyond simple PV thrombectomy and PV-to-PV anastomosis.
|
|
A splenorenal shunt was recognized at CT in 19 (54%) of 35 patients; all except one shunt apparently arose spontaneously rather than from prior surgery (Fig 4). The diameter of the main PV was smaller in patients with splenorenal shunt. After exclusion of a single patient with an acutely thrombosed PV with a diameter of 27 mm, the mean diameter of the PV in patients with shunts was 6.9 mm, as compared with 11.1 mm in those without a shunt (P .05). Eight (42%) of these 19 patients underwent modification of the OLT procedure beyond thrombectomy.
PV calcification was noted in 11 patients (31%). In nine, the calcification was linear or curvilinear and within the venous wall (Fig 5), while one patient had calcification only within the thrombus (Fig 8), and another had calcification within the PV wall and the thrombus. Seven (64%) of these 11 patients underwent modification of the OLT procedure (P .05).
|
Correlation between Surgical and CT Findings
The surgeons described a cordlike PV in eight patients who required an interposition graft; CT depicted no patent lumen in any of these PVs (Fig 4). The patient who had a fatal cardiac arrest and extensive intraoperative hemorrhage had a calcified PV at CT. In five patients, the surgeons described the PV wall as sclerosed. In these patients, CT depicted PV wall calcification in two, PV thrombus in one, and neither calcification nor thrombus in two.
Three of the eight hepatocellular carcinomas that were present in the excised livers were identified at CT performed before OLT (diameter, 1, 2, and 7 cm), while five tumors were missed (diameter, 0.3, 0.4, 0.6, 1.6, and 2.3 cm). The mean diameter of the PV in these eight patients was 1.1 cm (range, 0.3–1.5 cm), and the PV was never seen to enhance in the arterial phase nor was the tumor in proximity to any of the major PV branches. Therefore, the PV thrombus was judged to be bland rather than tumor thrombus.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
PV thrombosis or sclerosis at the time of OLT makes the surgical procedure more difficult and is associated with increased perioperative morbidity and mortality due to rethrombosis, pancreatitis, and increased blood loss (2,4,8). Preoperative knowledge of PV thrombosis may facilitate surgical strategy and allowing surgeons to anticipate the need for thrombectomy (4), bypass venous grafting (9–11,14), or cavoportal hemitransposition (20). Hoffman et al (14) stated that preoperative knowledge of PV sclerosis may allow the transplant surgeon to begin dissection of the superior mesenteric vein and creation of a venous conduit before extensive dissection of the recipient liver, which minimizes the threat of the anhepatic phase of the OLT procedure.
The most common PV abnormality that we detected at CT was partial or nonocclusive thrombosis within the extrahepatic PV (Fig 6). Such cases are usually amenable to simple thrombectomy, with direct anastomosis of the donor PV to the recipient PV (Table).
Extension of the thrombus to the superior mesenteric-splenic venous confluence or beyond requires more extensive thrombectomy but may still permit PV-to-PV anastomosis in some cases (Fig 7, Table). However, if the entire splanchnic venous system is thrombosed or if the PV is markedly fibrosed, calcified, or narrowed, the OLT must be modified (Fig 3). Extensive dissection of the superior mesenteric vein is known to increase the risk of pancreatitis and excessive blood loss and is discouraged by experienced transplant surgeons (4,8).
We noted a good correlation between the status of the PV described at surgery and that depicted at CT. Some discrepancies are due to changes in the PV between the time of CT and OLT, as described later. CT evidence of diminutive caliber, cavernous transformation, or calcification of the PV (Fig 4) correlated with the surgeon’s observation of a sclerotic, cordlike and friable PV. Patients with any CT findings of a calcified or cordlike PV were significantly (P .05) more likely to require surgical modifications, such as interposition of an iliac venous graft (Figs 2, 4), than were those who had only thrombosis of the PV (Table).
PV calcifications have been noted in prior radiology and surgical publications (14,21–27). Most reports in the radiology literature are descriptive, noting that PV calcification is usually found in the setting of cirrhosis and portal hypertension (21–25). Others have noted that the calcifications may be within the PV wall (intima and media) (21–26), presumably due to a mechanical stress (25), or they may develop within the thrombus (21). There are some reports (13,14) of difficult or impossible portal-systemic shunt procedures or OLTs, and a few cases (2) of fatal hemorrhage in the setting of PV calcifications, but we are unaware of any prior attempts to define the prevalence or clinical implications of PV calcifications.
In our series of 379 OLTs (with available CT findings), we found 11 (2.9%) patients with CT evidence of PV calcifications within the wall (Fig 5), the thrombus (Fig 8), or both. Seven of these patients required interposition venous grafts, but four who had thin mural calcifications underwent uneventful production of PV-to-PV anastomosis. One patient with PV calcifications died of cardiac arrest following extensive PV dissection and blood loss during OLT (Fig 5).
Transplant surgeons generally discourage the use of splenectomy or the creation of a splenorenal shunt for control of variceal hemorrhage (4). These are known to increase the incidence of PV thrombosis before and after OLT, as well as other complications such as hepatic encephalopathy (5). The TIPS has largely supplanted other surgical interventions for control of complications of portal hypertension (28). Nevertheless, patients with cirrhosis often develop spontaneous splenorenal shunts, and these can be recognized at CT (Fig 4). Because splanchnic venous return is diverted, PV flow diminishes. We noted a significantly (P .05) decreased diameter of the PV in patients with a spontaneous shunt, as compared with other patients (Fig 4). When splenorenal shunts are accessible during OLT, they are often ligated to increase blood flow to the transplanted liver.
We believe that CT can accurately depict the status of the PV in candidates for OLT. Attention to the CT technique is important, as has been noted by other investigators (29). Our protocol calls for unenhanced images obtained through the liver, followed by both arterial dominant and venous dominant phases during the administration of a 125–150-mL bolus of contrast medium at 4–5 mL/sec. All clinically important criteria, including the presence and extent of PV thrombosis, calcifications, and shunt, should be evident, especially in the venous dominant phase of enhancement.
In a group of 25 patients who had cirrhosis and were being evaluated for TIPS rather than OLT, Kuszyk et al (29) used CT, including multiplanar and three-dimensional reconstructions, to diagnose PV thrombosis in 10 patients. Using angiography, portography, or duplex US as confirmatory studies, they concluded that CT was sensitive and specific for the diagnosis of PV thrombosis.
Silverman et al (30) used magnetic resonance (MR) angiography to diagnose PV thrombosis in 16 patients among 102 candidates for OLT and confirmed high accuracy of MR angiography, with surgical findings as the standard of proof. More recently, Kreft et al (31) also confirmed the accuracy of MR angiography in depicting PV thrombosis in 35 patients with portal hypertension by using intraarterial digital subtraction angiography as the standard of proof.
Other investigators have reported the use of US (16,17) and angiography (18) for the evaluation of PV status in candidates for OLT. In our experience, we have noted some problems with these modalities, including the failure to depict partial thrombosis, calcification, or shunts, and the false-positive US and angiographic diagnosis of PV thrombosis due to stagnant or hepatofugal flow. Because the transplant surgeons were not blinded to the CT findings in patients in our study, the CT and surgical findings are not independent, and this could affect the validity of our findings. Although we could not determine the absolute or relative accuracy of CT in depicting PV thrombosis in this investigation, this would be an important contribution for future investigators.
Although we have been able to correlate the CT and surgical findings in a substantial number of patients who have undergone OLT, our study is limited by the long interval between the CT and OLT in some patients. During this interval, five patients developed PV thrombosis. Three of these underwent a TIPS procedure after CT and before OLT. Because it is important to know the presence and extent of PV thrombosis before OLT, we recommend assessment or reassessment of the candidate with CT shortly before OLT.
In summary, PV thrombosis or sclerosis occurs in approximately 10% of patients who have advanced cirrhosis and undergo OLT. CT can aid in the assessment of PV occlusion, calcifications, and shunts and in the development of surgical strategy to deal with these anomalies.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, M.P.F., G.B.; study concepts and design, M.P.F.; literature research, G.B.; clinical studies, M.P.F., G.B., K.P.; data acquisition, G.B., K.P.; data analysis/interpretation, G.B., M.P.F., D.A.G.; statistical analysis, G.B.; manuscript preparation, definition of intellectual content, and editing, G.B., M.P.F.; manuscript revision/review, M.P.F.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Z. Koc, L. Oguzkurt, and S. Ulusan Portal Venous System Aneurysms: Imaging, Clinical Findings, and a Possible New Etiologic Factor Am. J. Roentgenol., November 1, 2007; 189(5): 1023 - 1030. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Bryce, B. M. Yeh, A. Qayyum, P. Pacharn, N. M. Bass, Y. Lu, and F. V. Coakley CT Signs of Hepatofugal Portal Venous Flow in Patients with Cirrhosis Am. J. Roentgenol., December 1, 2003; 181(6): 1629 - 1633. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
