Can Doppler Sonography Discern Between Hemodynamically Significant and Insignificant Portal Vein Stenosis After Adult Liver Transplantation?

Vascular and Interventional Radiology Original Research Mullan et al. Sonography of Portal Vein Stenosis Vascular and Interventional Radiology Original Research Charles P. Mullan 1 Bettina Siewert Robert A. Kane Robert G. Sheiman Mullan CP, Siewert B, Kane RA, Sheiman RG Keywords: Doppler sonography, liver transplantation, peak systolic velocity (PSV), portal vein stenosis DOI:10.2214/AJR.10.4636 Received March 19, 2010; accepted after revision May 1, 2010. 1 All authors: Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, 1 Deaconess Rd., Boston, MA 02215. Address correspondence to C. P. Mullan (cpmullan@hotmail.com). AJR 2010; 195:1438 1443 0361 803X/10/1956 1438 American Roentgen Ray Society Can Doppler Sonography Discern Between Hemodynamically Significant and Insignificant Portal Vein Stenosis After Adult Liver Transplantation? OBJECTIVE. The purpose of our study was to determine whether Doppler sonography, using a strict reference standard, can specifically identify hemodynamically significant portal vein anastomotic stenosis after liver transplantation in adults. MATERIALS AND METHODS. The duplex and color Doppler examinations of 13 consecutive adult patients who underwent portal venography for suspected portal vein stenosis after liver transplantation were retrospectively examined. Peak systolic velocity (PSV) and change in PSV (ΔPSV) along the portal vein were correlated with portal venography. Stenoses above 50% on the basis of strict venographic criteria were considered hemodynamically significant. The Doppler studies before and after intervention were also assessed. Fourteen randomly chosen subjects with transplants without suspicion of portal anastomotic stenosis acted as controls. RESULTS. Six patients had significant portal vein stenosis (> 50%) and seven had stenosis below 50%. PSV and ΔPSV were significantly greater for patients with > 50% stenosis in comparison with those with 50% stenosis and control subjects. Optimal threshold values for PSV and ΔPSV were 80 and 60 cm/s, respectively, with either value alone yielding sensitivity of 100% and specificity of 84% for significant stenosis. Threshold values also included cases of stenosis below 50%. Five of six patients with > 50% stenosis underwent stenting, with poststent PSV and ΔPSV significantly declining to match that of control subjects. Three of seven with stenosis below 50% had stents placed but no significant change in the Doppler examination. CONCLUSION. Doppler threshold criteria reliably exclude those without posttransplantation portal vein stenosis and have high sensitivity for detecting portal stenosis. However, these criteria cannot discern the extent of stenosis. L iver transplantation is the treatment of choice for end-stage liver disease. Vascular complications after transplantation represent a significant cause of morbidity after surgery, potentially leading to graft failure without appropriate treatment. Hepatic arterial insufficiency is the most frequent posttransplantation vascular problem, with hepatic artery thrombosis occurring in 3 9% of patients [1]. Portal vein pathology is a less common cause of graft failure but may also cause significant morbidity [2]. Anastomotic stenosis can occur between the donor and recipient portal veins causing portal hypertension and predisposing to the development of portal vein thrombosis. Published data indicate that portal vein stenosis occurs in 1 2% of adults after orthotopic transplantation [3 5]. Doppler sonography provides anatomic and dynamic data on the patency of graft vessels and is accepted as a screening exam- ination to identify vascular complications in liver transplantation patients before they become clinically symptomatic, particularly in the early postoperative period [6]. There is general acceptance of the clinical role of ultrasound imaging in detecting portal occlusion due to thrombosis in the postoperative period [6 8], but there is conflicting evidence on the clinical significance of ultrasound-documented portal vein patency with elevated flow velocities and perceived intraluminal narrowing. Many published studies have correlated velocities with portal vein stenosis but have used mostly qualitative, nonstandardized reference standards for hemodynamically significant stenosis and fail to present follow-up Doppler data on those cases in which intervention (stenting, dilatation) was performed [6, 8 10]. Confounding this, other published data indicate that narrowing at the portal vein anastomosis and elevated portal vein velocities may be seen on 1438 AJR:195, December 2010

Sonography of Portal Vein Stenosis ultrasound examinations in the early postoperative period without evidence of subsequent graft dysfunction [7]. Our goal was to attempt to clarify the role of Doppler sonography for detecting hemodynamically significant portal vein stenosis after liver transplantation using a strict reference standard for stenosis on the basis of venography. Changes in the Doppler examination before and after intervention also were examined to assess the validity of our reference standard. Materials and Methods The study was approved by our hospital institutional review board and informed consent from patients was not required. The duplex and color Doppler sonography examinations of all adult patients with allograft liver transplantation who underwent portal venography from October 2003 to September 2009 for potential portal vein anastomotic stenosis were assessed retrospectively. Study inclusion required the ultrasound examination and corresponding venogram to be within 1 week of each other, the ultrasound obtained a minimum of 1 week after liver transplantation (to avoid postoperative hyperdynamic state), and postvenography or postintervention Doppler studies also available. Patients with current or previous documentation of portal vein occlusion were not considered. Thirteen consecutive patients were identified from our interventional radiology database who fulfilled these criteria. Indications for venography included a question of a portal venous anastomotic stenosis based on cross-sectional imaging performed for other reasons, clinical suspicion of portal hypertension, or elevated transaminases and alkaline TABLE 1: Disease Before Liver Transplantation Disease phosphatase in the setting of imaging-confirmed patency of the biliary tree and hepatic artery and veins. The mean time from transplantation to portal venography was 218.5 ± 68.0 days. Over the time course of this study, these 13 patients represented 7.1% of all 184 liver transplantations performed at our institution. Eleven of the 13 patients had orthotopic liver transplantation, and two patients received living donor grafts. All patients had direct anastomosis between the donor and recipient portal vessels, without the use of interposition grafts. The mean age at time of transplantation was 54.2 ± 8.3 years, with the study group consisting of four women and nine men. The distribution of underlying disease leading to liver transplantation is shown in Table 1. One patient had previously undergone liver transplantation and had retransplantation for hepatic dysfunction due to recurrent hepatitis C. Because all our liver transplantation patients undergo routine ultrasound surveillance (immediately postoperative, at 3 months, then every 6 months), 14 additional subjects with allograft liver transplants and no clinical or radiographic suspicion of portal anastomotic stenosis were randomly chosen as controls. The control subjects underwent Doppler sonography examination during No. of Patients Hepatitis C alone 4 Hepatitis C and hepatocellular carcinoma 4 Alcohol-related liver disease 3 Hepatitis B and hepatocellular carcinoma 1 Primary biliary cirrhosis 1 Total 13 the same 6-year period as the patients. The control group consisted of four women and 10 men, with a mean age of 59.0 ± 10.5 years and mean time from transplantation to ultrasound examination of 182.1 ± 42.1 days. Doppler Sonography Sonography examinations included gray-scale assessment of the allograft for mass lesions, biliary ductal dilation, echotexture, and signs of portal hypertension such as varices and ascites. Duplex and color Doppler examinations of the portal vein, splenic and superior mesenteric veins, hepatic artery, and hepatic veins were performed and included angle-corrected peak systolic velocities (PSVs) along the course of all vessels. Specifically, Doppler assessment of the portal vein included PSVs along the course of the native portal vein, anastomosis, and donor portal vein to the liver hilum (Fig. 1). Patients fasted for 6 hours before ultrasound examination according to our protocol. Doppler sonography examinations were carried out on one of two units (IU-22 or ATL 5000, Philips Healthcare) by registered technologists with gates, pulse repetition frequency, and angle of insonation optimized for each patient. All Doppler studies A Fig. 1 Doppler sonography study of 52-year-old man performed 72 days after orthotopic liver transplantation. Subsequent portal venography confirmed > 50% stenosis. A, Hepatopetal flow with normal peak systolic velocity is seen in main portal vein proximal to anastomotic stenosis (arrow). B, Disturbed flow is seen at site of anastomotic narrowing between native and donor vessels, with angle-corrected peak systolic velocity of 206 cm/s. B AJR:195, December 2010 1439

Mullan et al. erature search revealed no definitive pressure gradient with high sensitivity to diagnose posttransplantation portal stenosis in an adult population. Fig. 2 Portal venography in 52-year-old man with hemodynamically significant portal vein stenosis after orthotopic liver transplantation. A, Venogram shows > 50% stenosis at anastomosis between recipient and donor portal veins and some reflux into splenic vein. Splenic varices were also present (not shown). B, Venogram shows stent deployed across anastomotic narrowing. C, Venogram obtained after stenting shows resolution of stenosis. were reviewed in consensus by three radiologists, each with a minimum of 10 years of experience in abdominal and Doppler ultrasound. Reviews were carried out in a blinded fashion, with the reviewers unaware of which patients did (study subjects) and did not (control subjects) undergo portal venography. Reviewers recorded PSV and change in PSV (ΔPSV) across the anastomosis from donor to recipient portal vein. Biliary tree, hepatic artery, and hepatic vein assessment was also performed to exclude any abnormality that could affect portal venous flow. All postvenography and postintervention ultrasound examinations were reviewed 2 weeks later to avoid bias. Portal Venography The patients underwent portal venography via ultrasound-guided percutaneous entrance into a peripheral right anterior portal branch. Initial access was achieved with an Accustick system (Boston Scientific). Once access to a patient s main portal vein was achieved, the Accustick system was exchanged for a 5-French Sos catheter (AngioDynamics), and venography was performed in two orthogonal views, each at 4 frames per second during injection of 30 ml of nonionic contrast material (ioversol, Optiray 320, Mallinckrodt) at a A rate of 15mL/s. Additionally, pullback pressures were usually obtained from the recipient portal vein through to the bifurcation of the donor portal vein. We could find no accepted grading of portal stenosis based on pressure measurements and venographic imaging criteria in the literature. Hence, for this study, portal stenosis was graded as greater or less than 50% or normal using the following criteria: Greater than 50% stenosis required a reduction of > 50% of the diameter of the smaller of the portal veins (either the donor or the recipient portal vein) in at least one projection, slow antegrade flow on digital subtraction angiography, and visualization of varices on venography. Stenosis 50% was defined as a diameter reduction 50% or less of the smaller of the donor or recipient portal vein in both projections or absence of both varices and reversed or delayed antegrade flow. Diameters were determined by caliper measurements obtained from our archival system (Centricity, GE Healthcare) calibrated using the 5-French Sos catheter as the reference. Measurements were performed by an individual with more than 20 years of experience in interventional procedures. Although portal vein pressures were obtained, they were not by themselves used to define portal venous stenosis primarily because our lit- B C Stent Placement The decision for portal stent placement did not result from imaging findings alone but rather was based on discussion with the surgeon who performed the transplantation in conjunction with venographic findings and the clinical features (abnormal liver function tests or other clinical findings suggestive of portal hypertension, such as encephalopathy). Placement of the stent required exclusion of all other possible causes of hepatic dysfunction, such as arterial and venous stenosis and biliary dilatation. The stent placement was performed using the access created for venography. An 8-French Bright Tip sheath (Cordis) was placed via the right anterior portal branch. Only Wallstents were placed, ranging from 8 to 12 mm depending on the size of the donor or recipient portal veins. Stent length was chosen to bridge the surgical anastomosis without crossing the superior mesenteric vein. After stent placement, venography and pressure measurements across the stent were repeated (Fig. 2). The 8-French sheath was removed and its intrahepatic course embolized with gel foam pledgets. The patients underwent baseline poststent sonography usually within 24 hours of stent placement and enrolled in the standard follow-up protocol previously described. Poststent venography was also assessed in comparison with prestent imaging to evaluate whether there was resolution of varices and improved venous flow. Statistical Analysis The patients were divided into five groups: group A and group B were those with > 50% or 50% stenosis of the portal vein, respectively, as defined by the reference standard (venography), group C was the control group, and groups D and E consisted of the subjects in groups A and B, respectively, who had undergone portal vein stenting. The portal PSV and ΔPSV were recorded for all subjects, and portal venous pressure gradients were obtained for those undergoing venography. PSV and ΔPSV between all five groups were compared using one-way analysis of variance with post hoc two-tail pairwise comparisons. Additionally, portal pressure gradients before and after stenting were compared among all groups using the Student s t test. Results Doppler sonography review indicated patent portal veins in all subjects and no subjects were found to have ductal dilation, abnormal arterial inflow, or hepatic venous outflow, which 1440 AJR:195, December 2010

may alter portal venous flow. Six of 13 subjects had > 50% portal venous stenosis by venography (group A) and seven had 50% stenosis (group B). On gray-scale sonography, some expected element of relative narrowing at the portal anastomosis existed in all 13 patients. Eight stents were placed without complication, five in group A and three in group B. All patients in group B and four of the six patients in group A had received orthotopic liver transplantation, with two patients having received living donor transplants. One patient with stenosis > 50% developed hepatic failure that was clinically determined to be due to recurrence of Hepatitis C, and portal vein stenting therefore was not performed in this subject. Stenting was performed in three subjects in group B despite stenosis below 50% solely at the request of the referring surgeon to exclude stenosis as a contributing factor for graft dysfunction. Peak Systolic Velocity Group A (> 50% stenosis) showed a significant difference in mean PSV (149.67 ± 16.78 [standard error] cm/s) compared with all other groups (Table 2). Group B ( 50% stenosis) had a mean PSV that was significantly different from the control group C but not significantly different from subjects who had undergone stenting (groups D and E). Therefore, those subjects with a portal vein stenosis below 50% who underwent stenting did not alter their Doppler examination with respect to portal PSV after stent placement. Applying a PSV threshold of 80 cm/s to all subjects in groups A, B, and C led to sensitivity of 100% and specificity of 84% for the detection of hemodynamically significant portal vein stenosis (stenosis > 50%). All false-positive cases were within group B. Change in Peak Systolic Velocity (ΔPSV) The mean ΔPSV traversing the portal anastomosis was 122.83 ± 13.67 cm/s for group A, which was significantly different from all other groups (Table 3). Similar significant results were noted for group B (ΔPSV, 64.43 ± 12.56 cm/s) except when compared with the same subject group after undergoing portal venous stenting (p = 0.11), i.e., those with a stenosis below 50% had no significant change in ΔPSV after stenting. Applying a ΔPSV threshold of 60 cm/s to all subjects in groups A, B, and C again led to sensitivity of 100% and specificity of 84% for the detection of portal vein stenosis > 50%. Increasing or decreasing this threshold value led to a decline in sensitivity or specificity, respectively. Sonography of Portal Vein Stenosis TABLE 2: Peak Systolic Velocity of Portal Vein Measured on Doppler Sonography Group Parameters No. of Patients All false-positive cases were identical to those previously noted, and thus no additional sensitivity or specificity was achieved by a combination of PSV and ΔPSV thresholds. Poststenting PSV and ΔPSV in group D (stenting with portal stenosis > 50%) were significantly lower (p = 0.037 and 0.019, respectively) than their prestent values whereas those in group E (stenting with portal stenosis 50%) showed no significant difference compared with their prestent values (Table 4). These results were not unexpected, given the results of the multivariate analysis and again suggest that stenting of patients with portal vein stenosis below 50% fails to alter the Doppler examination. Looking at patients who had undergone stenting overall (n = 8) and separating by extent of stenosis or combined as a single group, PSV and ΔPSV became similar to the control group after stenting (p. 0.84 and 0.81 for group C vs groups D and E, respectively). Pressure Measurements Pressure measurements were available in five of six subjects in group A (pressure gradient for one subject was not formally documented) and in all seven subjects in group B. In group A, the mean pressure gradient was 6.4 ± 2.9 mm Hg and was not significantly different (p = 0.76) from group B, which had a mean pressure gradient of 1.6 ± 0.4 mm Hg. All poststent pressure gradients were below 3 mm Hg. Overall, stent placement did not Peak Systolic Velocity (cm/s) A Stenosis > 50% before stent 6 149.7 ± 16.8 Significance Compared With Group A (p) B Stenosis 50% before stent 7 102.1 ± 15.5 0.0460 C Control subjects 14 50.4 ± 11.0 < 0.0001 D Stenosis > 50% after stent 5 54.6 ± 18.4 0.0006 E Stenosis 50% after stent 3 56.7 ± 23.7 0.0032 TABLE 3: Change in Peak Systolic Velocity ( PSV) Across Portal Vein Group Parameters No. of Patients PSV (cm/s) A Stenosis > 50% before stent 6 122.8 ± 13.6 Significance Compared With Group A (p) B Stenosis 50% before stent 7 64.4 ± 12.6 0.0036 C Control subjects 14 24.9 ± 8.9 < 0.0001 D Stenosis > 50% after stent 5 10.2 ± 14.9 < 0.0001 E Stenosis 50% after stent 3 26.7 ± 19.2 0.0003 significantly affect portal venous pressures regardless of whether subjects had a portal stenosis above or below 50%. Image Changes Before and After Stenting After stenting, no apparent stenosis was visualized on venography in all eight cases. In cases where varices existed, these resolved after stent placement. Additionally, brisk antegrade flow was noted in all cases after stenting. Discussion Although portal vein pathology is an uncommon cause of graft failure, patient survival is lower for portal vein thrombosis than hepatic artery thrombosis after transplantation [11]. Portal vein complications are more likely to occur in pediatric patients and recipients of living donor liver transplants. Studies have shown that 2 22% of patients receiving living donor transplants will develop portal vein stenosis [8, 10, 12]. This is due to discrepancy between the sizes of the vessels to be anastomosed and the potential need for venous extension grafts [10, 13]. Adult patients undergoing orthotopic transplantation also have an increased risk of complications if the portal vein anastomosis is at the confluence of the superior mesenteric vein and splenic vein or an interposition graft is required [4]. At our institution, on the basis of venography, six patients developed portal vein stenosis > 50% during a 6-year period (3.3% of all transplants, four in orthotopic transplants and two in living-related transplants) AJR:195, December 2010 1441

Mullan et al. TABLE 4: Change in Velocity Parameters After Portal Vein Stenting Portal Vein Stenosis > 50% 50% No. of patients 5 3 Mean reduction in PSV (cm/s) 87.4 19.7 Significance of PSV reduction (p) 0.037 0.488 Mean reduction in PSV (cm/s) 109.2 21.7 Significance of PSV reduction (p) 0.019 0.429 Note PSV = peak systolic velocity, PSV = change in PSV. whereas seven patients (all orthotopic transplantations) developed portal vein stenosis 50%. All 13 patients had abnormal serum biochemical findings, including elevated transaminases and alkaline phosphatase and ultrasound confirmation of patency of other transplanted vessels. Compared with control subjects, Doppler sonography showed significantly elevated PSV and ΔPSV across the portal vein anastomosis in all six patients with hemodynamically significant portal vein stenosis. We point out that our control group of patients had a mean PSV of 50.4 cm/s and a mean ΔPSV across the portal venous system of 24.9 cm/s, consistent with normal Doppler indices described by other authors [9, 10]. Exceeding 80 cm/s for PSV would have identified all patients with portal vein stenosis > 50% on portal venography while including four of seven with a stenosis below 50% but including no control patients. Exceeding 60 cm/s for ΔPSV would have similar results. Overall, both these proposed threshold values had sensitivity of 100% and specificity of 84% for the detection of hemodynamically significant portal vein stenosis. The specificity of both parameters declined to 64% when looking only at all 13 subjects with stenosis, regardless of grade. To date, few studies have focused on developing Doppler criteria for portal vein stenosis in adult transplantation patients. In a study by Chong et al. [9], a PSV threshold of 125 cm/s was 73% sensitive and 95% specific for the detection of portal vein stenosis on the basis of 11 adult transplantation patients. In our group of patients, a PSV threshold of 125 cm/s would have sensitivity of 67% and specificity of 87.5% for hemodynamically significant portal vein stenosis. A proposed explanation for this variance and the overlap of our subjects with stenosis above and below 50% may be the presence of varices. Varices from significant portal hypertension actually decrease the volumetric flow of blood through the portal vein. Because volumetric flow is directly related to average velocity and crosssectional area, decreasing flow will directly decrease the PSV (in an ideal system in which flow is laminar in a tube, PSV is about twice the average velocity) [14]. Because all of the cases with significant stenosis had varices, velocities may have been misleadingly low, leading to our lower threshold values. Additionally, the Doppler criteria of Chong et al. are based primarily on pressure gradients at venography, with no formal imaging criteria [9]. Thus, we do not know the extent of varices in the subjects reported by Chong et al. but believe that in the case of portal stenosis in transplant recipients, lower Doppler threshold criteria should exist to maximize sensitivity despite sacrificing specificity. We found the PSV and ΔPSV of the patients with 50% stenosis to be significantly higher than the control group. These findings seem to indicate that abnormal Doppler flow may be observed in the portal vein in the absence of hemodynamically significant stenosis. This lack of ability of Doppler sonography to discern the extent of portal anastomotic stenosis in our study may be explained by a number of factors. There are data showing that patients with advanced liver disease may have high flow velocities and volumetric flow in the portal vein for up to 2 years after successful orthotopic liver transplantation [15]. This may be related to long-standing splanchnic hyperemia due to portal hypertension before surgery [15, 16]. Furthermore, the absence of varices as a criterion for stenosis below 50% may have resulted in elevated velocities. Portal vein stenosis is amenable to percutaneous interventional therapy. Although there are no published data to indicate that early intervention for posttransplantation portal vein stenosis improves long-term outcome, several studies show that endoluminal therapy can be performed successfully in both adults and children [3, 12, 17 21]. The decision to perform portal vein intervention must be made after careful correlation of clinical and radiologic data because stenosis that is not causing graft dysfunction may resolve spontaneously over time [22]. The limited data available suggest that stenting of portal vein stenosis after liver transplantation provides more durable vessel patency in comparison with endoluminal procedures for nontransplant-related portal vein pathology [23, 24]. The procedure appears to be of low risk because we encountered no immediate or delayed complications related to stent placement. Additionally, no patients proceeded to shunt placement or surgical repair. Strict venographic criteria were used as our reference standard in deciding the hemodynamic significance of portal vein stenosis. We did not use portal pressure measurements as criteria for defining portal anastomotic stenosis. This may be viewed as a limitation, particularly in light of some investigators recommending its usage in addition to the severity of anatomic narrowing [10, 12]. However, our position that portal pressures can be misleading is supported by the presence of a portal pressure gradient above 5 mmhg in only two of our six patients with stenosis above 50%, despite all having varices and sluggish antegrade blood flow. We also point out that in those studies using pressure gradients to support portal venous stenosis, pressure measurements logically were only carried out on those with high suspicion for portal stenosis and who underwent direct portography. Gradients for those considered to have no or insignificant stenoses were not obtained. Thus, pressure measurements in these studies were obtained in a preselected population. In fact, one set of investigators defined a significant stenosis on a qualitative basis only, i.e., when venography indicated a morphologic stenosis that necessitated stent placement [9]. Thus, their reported pressure gradients resulted from patients who first met this qualitative definition of stenosis and exactly how these gradients impacted any decision making is unclear. As alluded to earlier, the presence of portosystemic collateral vessels can impact portal flow, and we believe it can cause pressure measurements to underestimate the degree of portal vein stenosis. Support for this comes from the work of Lee et al. [10] who noted normal portal velocities in a case of significant portal stenosis but with extensive collaterals. Additionally, Ko et al. [23] reported clinical improvement after stent placement in patients after liver transplantation, with preprocedure gradients below 6 mm Hg. Finally, the confounding effect of collateral circulation decreasing venous pressure gradients in the face of significant venous narrowing has been documented in the left renal vein in the Nutcracker syndrome [25]. 1442 AJR:195, December 2010

Sonography of Portal Vein Stenosis For the purposes of this study, we defined significant portal anastomotic stenosis on venography as luminal narrowing greater than 50% relative to the smaller portion of the portal vein with slow antegrade flow and visualization of varices irrespective of the pressure measurement. The resolution of a significant difference in Doppler parameters in subjects in group A (patients with > 50% stenosis) before stenting compared with after stenting (group D) and control subjects would support our definition. That the Doppler parameters of patients with a stenosis below 50% showed no difference after stenting also concurs with our venographic definition of stenosis. We acknowledge an additional study limitation in that the insignificant change in PSV and ΔPSV after stenting in those with venographic stenosis below 50% may be the result of an underpowered comparison because only three patients in group B underwent stenting (primarily to exclude stenosis as a contributing factor to their graft dysfunction). Unfortunately, the low frequency of portal anastomotic stenosis after liver transplantation does not allow a robust overall assessment of Doppler sonography at a single institution but requires a multicenter study. Finally, we offered no clinical follow-up on those who underwent venography whether stented or not. Our goal was not to assess the clinical utility of portal stenting in transplant patients but solely to see if ultrasound could be used to screen for stenosis and help determine who should proceed to venography. The true clinical impact of portal stenting alone in our study population is difficult to discern given the varied medical management these patients received in addition to stenting. Although liver function tests normalized in most of our subjects after stenting, the contribution of portal vein intervention to this improvement cannot be reliably determined. In conclusion, in patients in whom all other causes of hepatic graft dysfunction have been excluded, using a PSV threshold of 80 cm/s or a ΔPSV threshold of 60 cm/s, Doppler sonography had sensitivity of 100% and specificity of 84% for the detection of portal vein stenosis > 50%. These values are based on correlation with objective and reproducible findings on venography and take into account the effects of collateral vessels present as a result of portal hypertension. Unfortunately, varices directly impact the specificity of these values by lowering portal volumetric flow and velocities. After stenting, normal velocity measurements can be expected within the portal vein. The specificity of Doppler sonography for portal anastomotic stenosis suffers from elevated PSV and ΔPSV, which can be observed in the absence of hemodynamically significant pathology on venography. Patients with graft dysfunction and velocity measurements above threshold values require portal venography to determine the significance of Doppler-suspected anastomotic stenosis. References 1. Stange BJ, Glanemann M, Nuessler NC, Settmacher U, Steinmuller T, Neuhaus P. Hepatic artery thrombosis after adult liver transplantation. Liver Transpl 2003; 9:612 620 2. Langnas AN, Marujo W, Stratta RJ, Wood RP, Shaw BW. Vascular complications after orthotopic liver transplantation. Am J Surg 1991; 161:76 83 3. Malassagne B, Soubrane O, Dousset B, Legmann P, Houssin D. Extrahepatic portal hypertension following liver transplantation: a rare but challenging problem. HPB Surg 1998; 10:357 364 4. Settmacher U, Nussler NC, Glanemann M, et al. Venous complications after orthotopic liver transplantation. Clin Transplant 2000; 14:235 241 5. Wozney P, Zajko AB, Bron KM, Point S, Starzl TE. Vascular complications after liver transplantation: a 5-year experience. AJR 1986; 147:657 663 6. Kok T, Slooff MJ, Thijn CJ, et al. Routine Doppler ultrasound for the detection of clinically unsuspected vascular complications in the early postoperative phase after orthotopic liver transplantation. Transpl Int 1998; 11:272 276 7. Stell D, Downey D, Marotta P, et al. Prospective evaluation of the role of quantitative Doppler ultrasound surveillance in liver transplantation. Liver Transpl 2004; 10:1183 1188 8. Tamsel S, Demirpolat G, Killi R, et al. Vascular complications after liver transplantation: evaluation with Doppler US. Abdom Imaging 2007; 32:339 347 9. Chong WK, Beland JC, Weeks SM. Sonographic evaluation of venous obstruction in liver transplants. AJR 2007; 188:1618; [web]w515 W521 10. Lee J, Ben-Ami T, Yousefzadeh D, et al. Extrahepatic portal vein stenosis in recipients of livingdonor allografts: Doppler sonography. AJR 1996; 167:85 90 11. Duffy JP, Hong JC, Farmer DG. Vascular complications of orthotopic liver transplantation: experience in more than 4200 patients. J Am Coll Surg 2009; 208:896 903 12. Funaki B, Rosenblum JD, Leef JA, et al. Portal vein stenosis in children with segmental liver transplants: treatment with percutaneous transhepatic venoplasty. AJR 1995; 165:161 165 13. Sieders E, Peeters PM, TenVergert EM, et al. Early vascular complications after pediatric liver transplantation. Liver Transpl 2000; 6:326 332 14. Bird RB, Stewart WE, Lightfoot EN. Transport phenomena. New York, NY: J. Wiley and Sons, 1960:46 15. Bolognesi M, Sacerdoti D, Bombonato G, et al. Change in portal flow after liver transplantation: effect on hepatic arterial resistance indices and role of splenic size. Hepatology 2002; 35:601 608 16. Henderson JM, Gilmore GT, Mackay GJ, Galloway JR, Dodson TF, Kutner MH. Hemodynamics during liver transplantation: the interactions between cardiac output and portal venous and hepatic arterial flows. Hepatology 1992; 16:715 718 17. Funaki B, Rosenblum JD, Leef JA, Hackworth CA, Szymski GX, Alonso EM. Angioplasty treatment of portal vein stenosis in children with segmental liver transplants: mid-term results. AJR 1997; 169:551 554 18. Ko GY, Sung KB, Lee S, et al. Stent placement for the treatment of portal vein stenosis or occlusion in pediatric liver transplant recipients. J Vasc Interv Radiol 2007; 18:1215 1221 19. Millis JM, Seaman DS, Piper JB, et al. Portal vein thrombosis and stenosis in pediatric liver transplantation. Transplantation 1996; 62:748 754 20. Miraglia R, Maruzzelli L, Caruso S, et al. Interventional radiology procedures in adult patients who underwent liver transplantation. World J Gastroenterol 2009; 15:684 693 21. Zanotelli ML, Vieira S, Alencastro R, et al. Management of vascular complications after pediatric liver transplantation. Transplant Proc 2004; 36: 945 946 22. Hamady M, Rela M, Sidhu PS. Spontaneous resolution of a portal vein stenosis over a 21 month period in a split-liver transplant: demonstration by colour Doppler ultrasound, catheter angiography and splenic pulp pressures. Eur Radiol 2002; 12: 2280 2283 23. Ko GY, Sung KB, Yoon HK, Lee SG. Early postransplantation portal vein stenosis following liver donor liver transplantation: percutaneous transhepatic primary stent placement. Liver Transpl 2007; 13:530 536 24. Yamakado K, Nakatsuka A, Tanaka N, Fujii A, Terada N, Takeda K. Malignant portal venous obstructions treated by stent placement: significant factors affecting patency. J Vasc Interv Radiol 2001; 12:1407 1415 25. Takebayashi S, Ueki T, Ikeda N, Fujikawa A. Diagnosis of the nutcracker syndrome with color Doppler sonography: correlation with flow patterns on retrograde left renal venography. AJR 1999; 172:39 43 AJR:195, December 2010 1443