- Email: [email protected]
Preoperative vascular imaging in pediatric liver transplantation
Journal of Pediatric Surgery (2005) 40, 643 – 647
www.elsevier.com/locate/jpedsurg
Preoperative vascular imaging in pediatric liver transplantation Kadiyala V. Ravindraa, J. Ashley Guthrieb, Helen Woodley b, Suzanne Davisona, Patricia McCleana, K. Raj Prasada, Mark D. Stringer a,* a
Children’s Liver and GI Unit, St James’s University Hospital, Beckett St, Leeds LS9 7TF, UK Department of Clinical Radiology, St James’s University Hospital, Beckett St, Leeds LS9 7TF, UK
b
Index words: Liver transplant; Magnetic resonance angiography; Doppler ultrasound; Biliary atresia
Abstract Background/Purpose: There is a considerable variation in the use of vascular imaging techniques in the preoperative assessment of children scheduled for liver transplantation. Duplex Doppler ultrasound scan (US), magnetic resonance angiography (MRA), and conventional angiography are used to varying extents. The authors compared the results of preoperative vascular imaging studies with operative findings to determine their accuracy and usefulness. Methods: Results of preoperative vascular imaging in 37 consecutive children undergoing cadaveric liver transplantation were compared with operative findings. Those undergoing relatively elective transplantations were investigated by US and MRA (group 1), whereas those requiring urgent transplants were assessed only by US (group 2). Results: The median age of the cohort (15 boys; 22 girls) was 4 years (19 days to 16 years) and the median weight was 17 kg (2.9 to 82 kg). In group 1 (n = 26), 20 children had a normal-caliber, patent portal vein at transplant and 6 had a narrow but patent portal vein requiring venous reconstruction in 4. The sensitivity and specificity of MRA in the detection of an abnormally narrow portal vein were 100% (6/6) and 95% (19/20), respectively. If reversed or absent flow in the portal vein on US was taken as an indication of a potentially abnormal vein, the sensitivity and specificity of Doppler US were 83% (5/6) and 95% (19/20), respectively. Magnetic resonance angiography revealed arterial anomalies in 4 children but failed to detect small accessory hepatic arteries in 5. The single patient with an aberrant vena cava was identified by MRA. In group 2 (n = 11), venous findings at operation and on US were concordant in 10 (91%) cases; one infant with reversed flow in the portal vein on US had a thrombosed vein at surgery. Magnetic resonance angiography was useful in 2 patient groups: those with reversed flow on Doppler US or suspected portal vein thrombosis in whom an abnormal portal vein was present in 86% (6/7) and infants with the biliary atresia splenic malformation syndrome who had multiple venous and arterial anomalies. Conclusions: A detailed Doppler examination of the hepatic vasculature by an experienced sonographer/radiologist provides sufficient vascular imaging for most children scheduled for cadaveric liver transplantation. Routine MRA is recommended in children with the biliary atresia splenic malformation syndrome and in those with abnormal duplex Doppler US findings. Although there are
Abstract presented at the Second Congress of the International Pediatric Transplant Association, Rio de Janeiro, Brazil, April 2003. T Corresponding author. Tel.: +44 113 206 6689; fax: +44 113 206 6691. E-mail address: [email protected] (M.D. Stringer). 0022-3468/05/4004-0005$30.00/0 D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jpedsurg.2004.12.003
644
K.V. Ravindra et al. limited data in this study, MRA is also valuable in children with Budd-Chiari syndrome, liver tumors, or a previous portosystemic shunt. D 2005 Elsevier Inc. All rights reserved.
Rapid advances in orthotopic liver transplantation (OLT) have increased the safety and success of the procedure in children. A key determinant of outcome is the patency of the vascular anastomoses. Prior knowledge of the status of the portal vein, hepatic artery, and inferior vena cava (IVC) assists surgeons in planning the procedure and anticipating the need for vascular reconstruction. This is particularly relevant in children with smaller caliber vessels and a greater tendency to anatomical variation. Portal venous anomalies include absent, small/hypoplastic, and thrombosed veins [1]. Congenital malformations of the IVC are generally rare [2] but occur frequently in the biliary atresia splenic malformation syndrome (BASM) [3]. Knowledge of the hepatic venous anatomy is critical in patients with Budd-Chiari syndrome. Variations in arterial anatomy are common, particularly the presence of accessory hepatic arteries [4]. Various imaging modalities including duplex ultrasound (US), color Doppler, conventional angiography, computerized tomography, and magnetic resonance angiography (MRA) have been used to delineate hepatic vascular anatomy before an OLT [5,6]. Although these investigations offer individual advantages and are, to some extent, complementary, MRA is now regarded by many as the investigation of choice [6]. In an attempt to determine the accuracy and value of preoperative Doppler US and MRA in children scheduled for OLT, the results of these imaging techniques were compared with operative findings in a consecutive series of pediatric liver transplant recipients.
1. Materials and methods A retrospective analysis of the role of preoperative hepatic vascular imaging in children undergoing cadaveric OLT at St James’s University Hospital was undertaken. The patients consisted of a consecutive series of children who underwent a transplant between November 2000 and June 2003. According to unit policy, all children scheduled for elective or semiurgent OLT were assessed by duplex Doppler US and MRA to define vascular anatomy. Doppler studies were repeated at 3-month intervals in those with protracted waiting periods. Children requiring urgent OLT were investigated by Doppler US alone.
1.1. Magnetic resonance protocol All imagings were performed with a high-performance gradient system at 1.5 T (Symphony, Siemens, Erlangen, Germany) using a head coil for small infants or combined spinal and phased array body coils for bigger children for
transmission and reception of signal. All precontrast images were acquired in the axial plane during breathholding or with suspended respiration if under general anesthesia. Sequences obtained included true free-induction steady-state precession (TR, 5.7 milliseconds; TE, 2.8 milliseconds), HASTE (TR, 1100 milliseconds; TE, 95 milliseconds; echo train length, 148), and T1-weighted in-phase (TR, 166 milliseconds; TE, 4.5 milliseconds; flip angle, 808) and opposed-phase (TR, 144 milliseconds; TE, 2.4 milliseconds; flip angle, 808) gradient echo sequences. A rectangular field of view and 6-mm slice thickness, optimized to each patient’s body habitus, were chosen and remained constant for each sequence. Gadopentetate dimeglumine (Magnevist, Schering, Berlin, Germany) or gadodiamide (Omniscan, Nycomed Amersham, Amersham, UK) was administered as a rapid bolus at a dose of 0.1 mmol/kg followed by a saline flush, and 3D gradient recalled echo T1-weighted images with fat suppression (TR, 3.4 milliseconds; TE, 1.6 milliseconds; flip angle, 158) and an effective slice thickness of 1 to 2.5 mm (depending on an infant’s or child’s size) were acquired 10, 40 and 120 seconds later. The first 2 acquisitions were obtained in an oblique coronal plane; the last, in an axial plane.
1.2. Ultrasound protocol Duplex US examinations (real time imaging with color Doppler analysis of blood flow) were undertaken by experienced sonographers and/or radiologists using a 5-MHz curvilinear multifrequency transducer (Elegra, Siemens). The portal vein characteristics —size, patency, and direction of flow — were recorded on Doppler; size and patency were documented on MRA. Regarding the hepatic artery, the Doppler study only documented the presence of, whereas the MRA could provide details on the origin (whether from the celiac or superior mesenteric arteries) and the presence of, accessory right and left hepatic arteries from the superior mesenteric and left gastric arteries, respectively. Patency and anomalies of the IVC were noted — particularly in the subgroup of patients with BASM. Comparison was made between the findings of portal vein, hepatic artery, and the IVC on Doppler, MRA, and operative findings.
2. Results During the study period, 38 children underwent liver transplantation. One child who underwent transplantation for fulminant liver failure did not undergo any preoperative imaging and was excluded from further analysis. There were 15 boys and 22 girls with a median age of 4 years (19 days16 years) and median weight of 17 kg (2.9-82 kg). Ten were
Preoperative vascular imaging in pediatric liver transplantation infants. Children undergoing elective or semiurgent transplantation (median waiting time, 32 days) were investigated by both Doppler US and MRA (group 1, n = 26). Those undergoing relatively urgent OLT (median waiting time, 2 days) were assessed preoperatively by Doppler US alone (group 2, n = 11). The indications for OLT in both groups are shown in Table 1.
Table 2 Concordance of preoperative Doppler and MRA portal vein imaging with operative findings in 26 pediatric liver transplant recipients (group 1) Portal vein at surgery
Doppler
MRA
Normal caliber and patent = 20 Narrow but patent = 6
Normal = 20 (reversed flow = 1) Normal = 1, Thrombosed = 1 (reversed flow = 4)
Normal = 19, Thrombosed = 1 Narrow but patent = 6
2.1. Group 1 Doppler imaging showed a normal-caliber, patent portal vein with forward flow in 20 patients; in 19, these findings were concordant with those at MRA and were confirmed at surgery. One patient with BASM was found to have a narrow-caliber (3 mm), preduodenal portal vein at MRA and at surgery and required an interposition iliac vein graft at transplant. In the 6 remaining patients, Doppler studies revealed reversed flow in a patent portal vein on at least one occasion in 5 patients (2 with biliary atresia [BA] and 1 each with cystic fibrosis, Alagille syndrome with hepatopulmonary syndrome, and neonatal liver failure) and a thrombosed portal vein in 1 infant with BASM. In the 5 patients with reversed flow on Doppler, MRA showed a narrow but patent portal vein in 4 and a thrombosed portal vein in 1. At surgery, MRA findings were confirmed in the 4 patients with a narrow portal vein, but the patient with a presumed thrombosed portal vein on MRA (associated with cystic fibrosis liver disease and massive mesenteric collaterals) had a patent 1-cm-diameter healthy portal vein at transplant. The single patient in whom Doppler studies identified a thrombosed portal vein was found to have a narrow but patent vein both at MRA and at subsequent transplant. Thus, the portal vein was abnormal at surgery in 5 of the 6 patients in whom Doppler studies demonstrated reversed flow or possible thrombosis. Four of these patients required anastomosis of a donor venous conduit to the confluence of splenic and superior mesenteric veins. Table 1
Indications for liver transplantation
Group 1 (n = 26) a
BA Cystic fibrosis a-1 antitrypsin deficiency Hepatopulmonary syndrome Autoimmune hepatitis Neonatal liver failure Langerhans cell histiocytosis Secondary biliary cirrhosis Liver tumor Budd-Chiari syndrome PFIC
n Group 2 (n = 11)
n
10 Non–A-E fulminant hepatitis 3 Autoimmune hepatitis 2 Fulminant hepatitis A
3 2 1
2 Wilson’s disease
1
1 Drug-induced acute liver failure 1 2 TPN/Short gut liver failure 1 1 BASM 1 1 PFIC (acute decompensation) 2 1 1
PFIC indicates progressive familial intrahepatic cholestasis. a Two patients with BASM syndrome.
1
645
The overall comparison of portal vein imaging with Doppler and MRA is shown in Table 2. The sensitivity and specificity of MRA in the detection of an abnormally narrow portal vein were 100% (6/6) and 95% (19/20), respectively. If reversed or absent flow in the portal vein was taken as an index of a potentially abnormally narrow vein in this pediatric population, the sensitivity and specificity of Doppler US were 83% (5/6) and 95% (19/20), respectively. With respect to hepatic arterial anatomy, Doppler studies simply confirmed a patent hepatic artery at the liver hilum in all cases. Magnetic resonance angiography additionally revealed the presence of 1 replaced and 2 accessory right hepatic arteries arising from the superior mesenteric artery and 1 left accessory artery. These findings were confirmed at surgery, but there were 4 additional patients with an accessory left hepatic artery and 1 with an accessory right hepatic artery. These vessels did not measure more than 2 to 3 mm in diameter. A detailed retrospective review of the magnetic resonance (MR) angiograms identified 2 of the 4 bmissedQ accessory left hepatic arteries and the accessory right hepatic artery although the arterial acquisition phase was suboptimal in the latter study. The IVC was reported to be normal at preoperative Doppler and MRA in 25 patients. One child with BASM and normal US appearances of the IVC was reported to have azygos continuation of the IVC on MRA, which was consistent with the findings at surgery. In the patient with Budd-Chiari syndrome, both Doppler US and MRA correctly identified patency of the IVC with occlusion of the hepatic veins and right portal vein.
2.2. Group 2 This group comprised 11 patients who underwent relatively urgent OLTs and who were assessed preoperatively by Doppler US alone. A patent, normal-caliber portal vein was demonstrated in 10 patients and confirmed at surgery. One sick infant with BA, polysplenia and situs inversus (BASM), and marked ascites in whom an MRA under general anesthesia would have been detrimental was found to have reversed flow in the portal vein on Doppler. At transplant 3 days later, he was noted to have a narrow thrombosed portal vein and required an interposition iliac vein graft reconstruction. Thus, Doppler US proved reliable in imaging the portal
646 vein in 91% (10/11) of the patients. Reversed flow in the portal vein was again indicative of an abnormality. Two patients had an accessory right hepatic artery at surgery that was not detected during preoperative Doppler studies. The IVC was reported to be normal in 10 patients and the infant with BASM (see above) was judged to have an interrupted IVC. These findings were verified at surgery.
3. Discussion Simultaneous with advances in liver transplantation, there have been significant developments in hepatobiliary imaging. In the 1980s, radiological assessment of liver transplant candidates included US imaging, computerized tomography (with rapid sequence enhancement) to delineate abnormalities of the portal and hepatic veins, and angiography to establish arterial anatomy [5]. Pediatric units relied predominantly on Doppler US for imaging the vessels, even in BASM [7]. Although MR imaging has been recognized to be useful in evaluating venous anatomy before a liver transplant, as recently as 1997, selective arteriography of the celiac trunk and superior mesenteric artery has been considered by some authors to be the definitive test for assessing hepatic arterial and portal venous anomalies [8]. Subsequent studies in adults and children have shown MR to be equal [9,10] or superior [11] to digital subtraction angiography in portal vein evaluation. Few studies have directly compared MRA with duplex Doppler sonography in transplant candidates. In a study from Boston, USA, Finn et al [12] demonstrated MRA to be superior to duplex Doppler imaging in evaluating the portal venous system of adults with cirrhosis awaiting transplant. In our own unit in Leeds, Naik et al [13] concluded that MR provided superior imaging of the intrahepatic portal branches, varices, and splenorenal shunts in adults and recommended MRA in all patients with abnormal portal venous anatomy or with equivocal findings on Doppler US. Portal vein imaging is a critical part of the preoperative evaluation of potential transplant candidates. The prevalence of preoperative portal vein thrombosis is between 2% and 13% in large transplant series [14,15]. Once considered to be a contraindication to transplant, it no longer precludes a successful outcome but does have an impact on the technical complexity of the operation and the perioperative morbidity. Prior knowledge of the presence and extent of portal vein thrombosis allows surgeons to anticipate difficulties and plan technical solutions [16]. Other anomalies of the portal vein such as a preduodenal or hypoplastic vein are associated with BA, especially its syndromic variety (BASM) [1]. When the portal vein is less than 4 mm in diameter, dissection back to the confluence of the splenic and superior mesenteric veins, with or without anastomosis of a venous conduit, is often necessary to achieve a wide anastomosis. An unrecognized preduodenal portal vein is at risk of iatrogenic injury early in the dissection of the portal triad [17]. It is advisable for
K.V. Ravindra et al. transplant surgeons to know the exact anatomy of the portal vein in these patients. In our study, Doppler studies were inaccurate in all 3 infants with BASM but MRA correctly identified the portal vein characteristics in the 2 elective cases. On the basis of this limited study, we continue to recommend routine MRA imaging of the portal vein in all pediatric transplant candidates with BASM. Similar imaging is also beneficial in the evaluation of transplant candidates with liver tumors or Budd-Chiari syndrome, but for most other transplant indications in children, if the portal vein appears normal on Doppler, further imaging is unlikely to be helpful. In our study, suspected portal vein thrombosis or reversed flow on Doppler US was associated with an abnormal portal vein in 6 of 7 (86%) children, 4 of whom were infants. Therefore, children with abnormal US findings may be an additional group in which preoperative MRA imaging of the portal vein is valuable. Inferior vena caval anomalies are rare. Lerut et al [2] reported an incidence of 2.9% in a large series of predominantly adult liver transplant patients. Most IVC anomalies occur in the pediatric population with BASM. Boillot et al [3] reported a 3.7% incidence of IVC anomalies in BA. In all cases, the IVC was interrupted between the renal veins and the hepatic veins. Two patients in our series, both with BASM, had an interrupted cava. This was correctly identified by MRA in the group 1 patient and by duplex Doppler US in the group 2 patient. However, given the spectrum of congenital IVC anomalies, prior knowledge of IVC anatomy is not critical except in cases of tumor and Budd-Chiari syndrome. Conventional angiography is now seldom used in pretransplant evaluation of arterial anatomy. However, MRA may not identify an accessory artery if it is small (V1-2 mm) [11,18], which is more likely in children, or in the case of a left accessory artery if it is obscured by left gastric varices. In our study, 4 accessory left hepatic arteries and 1 accessory right hepatic artery were not initially reported on MRA, 3 of which were present on retrospective review of the imaging. Detection of an accessory left hepatic artery could be improved by performing a left oblique coronal acquisition with a second injection of gadolinium, but the value of this is questionable because arterial anomalies in the recipient only have a limited impact on the technique of cadaveric liver transplantation in children. More important anomalies such as celiac artery stenosis [19] have not been reported in children undergoing OLT. With the availability of vascular allografts, the arterial inflow can
Table 3 Indications for MRA in pediatric cadaveric liver transplant evaluation !BASM syndrome !Abnormal portal vein on duplex Doppler US !Budd-Chiari syndrome !Tumors (eg, hepatoblastoma) !Prior portosystemic shunt
Preoperative vascular imaging in pediatric liver transplantation be reconstructed if the recipient hepatic artery is inadequate or if there are donor arterial anomalies [20]. Thus, routine MR imaging of arterial anatomy is not strictly necessary in pediatric cadaveric liver transplantation. However, because the incidence of hepatic arterial anomalies is particularly high in infants with BASM [17], MRA may still be valuable in helping surgeons anticipate the need for arterial reconstruction in these patients. Few studies have addressed the role of preoperative vascular imaging in pediatric liver transplant candidates and there is little evidence base for the selection of appropriate investigations. The present study suggests that a detailed Doppler examination of the hepatic vasculature by an experienced sonographer/radiologist is sufficient for most children scheduled for cadaveric liver transplantation. Routine MRA is recommended in children with the BASM (which is almost always apparent from a previous laparotomy or US scan), and in those with abnormal duplex Doppler US findings (Table 3). In addition, although there is insufficient evidence from the present study, we consider that pretransplant MRA (particularly to assess venous anatomy) is also indicated in children with tumors, those who are known to have a portosystemic shunt, and in patients with Budd-Chiari syndrome.
References [1] Whitington PF, Alonso EM. General criteria for pediatric transplantation. In: Busuttil RW, Klintmalm GB, editors. Transplantation of the liver. Philadelphia7 WB Saunders; 1996. p. 159 - 75. [2] Lerut J, Tzakis AG, Bron K, et al. Complications of venous reconstruction in human orthotopic liver transplantation. Ann Surg 1987;205:404 - 14. [3] Boillot O, Sarfati PO, Bringler J, et al. Orthotopic liver transplantation and pathology of inferior vena cava. Transplant Proc 1990;22:1567 - 8. [4] Michels NA. Blood supply and anatomy of the upper abdominal organs with a descriptive atlas. Philadelphia (Pa)7 Lippincott; 1955. [5] Dixon A, Sherwood T. Radiological assessment in liver transplantation. In: Calne SR, editor. Liver transplantation. 2nd ed. London7 Grune & Stratton Inc; 1987. p. 91 - 5.
647
[6] Boeve WJ, Kok T, Haagsma EB, et al. Superior diagnostic strength of combined MR-angiography and MR-imaging compared to intraarterial DSA in liver transplantation candidates. Magn Reson Imaging 2001;19:609 - 22. [7] Kelly DA. Paediatric liver transplantation. In: Neuberger J, Lucey MR, editors. Liver transplantation—practice and management. Plymouth7 BMJ Publishing Group; 1994. p. 97. [8] Suhocki P, Chari R, McCann R, et al. Vascular aspects of liver transplantation. In: Killenberg PG, Clavien PA, editors. Medical care of the liver transplant patient. Massachusetts7 Blackwell Science Inc; 1997. p. 180 - 2. [9] Kreft B, Strunk H, Flacke S, et al. Detection of thrombosis in the portal venous system: comparison of contrast-enhanced MR angiography with intraarterial digital subtraction angiography. Radiology 2000;216:86 - 92. [10] Cheng YF, Huang TL, Lui CC, et al. Magnetic resonance venography in potential pediatric liver transplant recipients. Clin Transplant 1997;11:121 - 6. [11] Kopka L, Rodenwaldt J, Vosshenrich R, et al. Hepatic blood supply: comparison of optimised dual phase contrast-enhanced threedimensional MR angiography and digital subtraction angiography. Radiology 1999;211:51 - 8. [12] Finn JP, Kane RA, Edelman RR, et al. Imaging of the portal venous system in patients with cirrhosis: MR angiography vs duplex Doppler sonography. AJR Am J Roentgenol 1993;161:989 - 94. [13] Naik KS, Ward J, Irving HC, et al. Comparison of dynamic contrast enhanced MRI and Doppler ultrasound in the pre-operative assessment of the portal venous system. Br J Radiol 1997;70:43 - 9. [14] Stieber AC, Zetti G, Todo S, et al. The spectrum of portal vein thrombosis in liver transplantation. Ann Surg 1991;213:199 - 206. [15] Busuttil RW, Shaked A, Millis JM, et al. One thousand liver transplants: the lessons learned. Ann Surg 1994;219:490 - 9. [16] Olthoff KM, Busuttil RW. Venous anomalies including portal vein thrombosis and prior portosystemic shunts. In: Busuttil RW, Klintmalm GB, editors. Transplantation of the liver. Philadelphia7 WB Saunders; 1996. p. 465 - 73. [17] Varela-Fascinetto G, Castaldo P, Fox IJ, et al. Biliary atresia– polysplenia syndrome: surgical and clinical relevance in liver transplantation. Ann Surg 1998;227:583 - 9. [18] Ward J, Guthrie AJ, Hughes T, et al. Anatomy of the arterial supply to the liver demonstrated by MRI. Eur Radiol 1997;7:893 - 9. [19] Garcia-Criado A, Gilabert R, Bargallo X, et al. Radiology in liver transplantation. Semin Ultrasound, CT MR 2002;23:114 - 29. [20] Bowen A, Hungate RG, Kaye RD, et al. Imaging in liver transplantation. Radiol Clin North Am 1996;34:757 - 77.
www.elsevier.com/locate/jpedsurg
Preoperative vascular imaging in pediatric liver transplantation Kadiyala V. Ravindraa, J. Ashley Guthrieb, Helen Woodley b, Suzanne Davisona, Patricia McCleana, K. Raj Prasada, Mark D. Stringer a,* a
Children’s Liver and GI Unit, St James’s University Hospital, Beckett St, Leeds LS9 7TF, UK Department of Clinical Radiology, St James’s University Hospital, Beckett St, Leeds LS9 7TF, UK
b
Index words: Liver transplant; Magnetic resonance angiography; Doppler ultrasound; Biliary atresia
Abstract Background/Purpose: There is a considerable variation in the use of vascular imaging techniques in the preoperative assessment of children scheduled for liver transplantation. Duplex Doppler ultrasound scan (US), magnetic resonance angiography (MRA), and conventional angiography are used to varying extents. The authors compared the results of preoperative vascular imaging studies with operative findings to determine their accuracy and usefulness. Methods: Results of preoperative vascular imaging in 37 consecutive children undergoing cadaveric liver transplantation were compared with operative findings. Those undergoing relatively elective transplantations were investigated by US and MRA (group 1), whereas those requiring urgent transplants were assessed only by US (group 2). Results: The median age of the cohort (15 boys; 22 girls) was 4 years (19 days to 16 years) and the median weight was 17 kg (2.9 to 82 kg). In group 1 (n = 26), 20 children had a normal-caliber, patent portal vein at transplant and 6 had a narrow but patent portal vein requiring venous reconstruction in 4. The sensitivity and specificity of MRA in the detection of an abnormally narrow portal vein were 100% (6/6) and 95% (19/20), respectively. If reversed or absent flow in the portal vein on US was taken as an indication of a potentially abnormal vein, the sensitivity and specificity of Doppler US were 83% (5/6) and 95% (19/20), respectively. Magnetic resonance angiography revealed arterial anomalies in 4 children but failed to detect small accessory hepatic arteries in 5. The single patient with an aberrant vena cava was identified by MRA. In group 2 (n = 11), venous findings at operation and on US were concordant in 10 (91%) cases; one infant with reversed flow in the portal vein on US had a thrombosed vein at surgery. Magnetic resonance angiography was useful in 2 patient groups: those with reversed flow on Doppler US or suspected portal vein thrombosis in whom an abnormal portal vein was present in 86% (6/7) and infants with the biliary atresia splenic malformation syndrome who had multiple venous and arterial anomalies. Conclusions: A detailed Doppler examination of the hepatic vasculature by an experienced sonographer/radiologist provides sufficient vascular imaging for most children scheduled for cadaveric liver transplantation. Routine MRA is recommended in children with the biliary atresia splenic malformation syndrome and in those with abnormal duplex Doppler US findings. Although there are
Abstract presented at the Second Congress of the International Pediatric Transplant Association, Rio de Janeiro, Brazil, April 2003. T Corresponding author. Tel.: +44 113 206 6689; fax: +44 113 206 6691. E-mail address: [email protected] (M.D. Stringer). 0022-3468/05/4004-0005$30.00/0 D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jpedsurg.2004.12.003
644
K.V. Ravindra et al. limited data in this study, MRA is also valuable in children with Budd-Chiari syndrome, liver tumors, or a previous portosystemic shunt. D 2005 Elsevier Inc. All rights reserved.
Rapid advances in orthotopic liver transplantation (OLT) have increased the safety and success of the procedure in children. A key determinant of outcome is the patency of the vascular anastomoses. Prior knowledge of the status of the portal vein, hepatic artery, and inferior vena cava (IVC) assists surgeons in planning the procedure and anticipating the need for vascular reconstruction. This is particularly relevant in children with smaller caliber vessels and a greater tendency to anatomical variation. Portal venous anomalies include absent, small/hypoplastic, and thrombosed veins [1]. Congenital malformations of the IVC are generally rare [2] but occur frequently in the biliary atresia splenic malformation syndrome (BASM) [3]. Knowledge of the hepatic venous anatomy is critical in patients with Budd-Chiari syndrome. Variations in arterial anatomy are common, particularly the presence of accessory hepatic arteries [4]. Various imaging modalities including duplex ultrasound (US), color Doppler, conventional angiography, computerized tomography, and magnetic resonance angiography (MRA) have been used to delineate hepatic vascular anatomy before an OLT [5,6]. Although these investigations offer individual advantages and are, to some extent, complementary, MRA is now regarded by many as the investigation of choice [6]. In an attempt to determine the accuracy and value of preoperative Doppler US and MRA in children scheduled for OLT, the results of these imaging techniques were compared with operative findings in a consecutive series of pediatric liver transplant recipients.
1. Materials and methods A retrospective analysis of the role of preoperative hepatic vascular imaging in children undergoing cadaveric OLT at St James’s University Hospital was undertaken. The patients consisted of a consecutive series of children who underwent a transplant between November 2000 and June 2003. According to unit policy, all children scheduled for elective or semiurgent OLT were assessed by duplex Doppler US and MRA to define vascular anatomy. Doppler studies were repeated at 3-month intervals in those with protracted waiting periods. Children requiring urgent OLT were investigated by Doppler US alone.
1.1. Magnetic resonance protocol All imagings were performed with a high-performance gradient system at 1.5 T (Symphony, Siemens, Erlangen, Germany) using a head coil for small infants or combined spinal and phased array body coils for bigger children for
transmission and reception of signal. All precontrast images were acquired in the axial plane during breathholding or with suspended respiration if under general anesthesia. Sequences obtained included true free-induction steady-state precession (TR, 5.7 milliseconds; TE, 2.8 milliseconds), HASTE (TR, 1100 milliseconds; TE, 95 milliseconds; echo train length, 148), and T1-weighted in-phase (TR, 166 milliseconds; TE, 4.5 milliseconds; flip angle, 808) and opposed-phase (TR, 144 milliseconds; TE, 2.4 milliseconds; flip angle, 808) gradient echo sequences. A rectangular field of view and 6-mm slice thickness, optimized to each patient’s body habitus, were chosen and remained constant for each sequence. Gadopentetate dimeglumine (Magnevist, Schering, Berlin, Germany) or gadodiamide (Omniscan, Nycomed Amersham, Amersham, UK) was administered as a rapid bolus at a dose of 0.1 mmol/kg followed by a saline flush, and 3D gradient recalled echo T1-weighted images with fat suppression (TR, 3.4 milliseconds; TE, 1.6 milliseconds; flip angle, 158) and an effective slice thickness of 1 to 2.5 mm (depending on an infant’s or child’s size) were acquired 10, 40 and 120 seconds later. The first 2 acquisitions were obtained in an oblique coronal plane; the last, in an axial plane.
1.2. Ultrasound protocol Duplex US examinations (real time imaging with color Doppler analysis of blood flow) were undertaken by experienced sonographers and/or radiologists using a 5-MHz curvilinear multifrequency transducer (Elegra, Siemens). The portal vein characteristics —size, patency, and direction of flow — were recorded on Doppler; size and patency were documented on MRA. Regarding the hepatic artery, the Doppler study only documented the presence of, whereas the MRA could provide details on the origin (whether from the celiac or superior mesenteric arteries) and the presence of, accessory right and left hepatic arteries from the superior mesenteric and left gastric arteries, respectively. Patency and anomalies of the IVC were noted — particularly in the subgroup of patients with BASM. Comparison was made between the findings of portal vein, hepatic artery, and the IVC on Doppler, MRA, and operative findings.
2. Results During the study period, 38 children underwent liver transplantation. One child who underwent transplantation for fulminant liver failure did not undergo any preoperative imaging and was excluded from further analysis. There were 15 boys and 22 girls with a median age of 4 years (19 days16 years) and median weight of 17 kg (2.9-82 kg). Ten were
Preoperative vascular imaging in pediatric liver transplantation infants. Children undergoing elective or semiurgent transplantation (median waiting time, 32 days) were investigated by both Doppler US and MRA (group 1, n = 26). Those undergoing relatively urgent OLT (median waiting time, 2 days) were assessed preoperatively by Doppler US alone (group 2, n = 11). The indications for OLT in both groups are shown in Table 1.
Table 2 Concordance of preoperative Doppler and MRA portal vein imaging with operative findings in 26 pediatric liver transplant recipients (group 1) Portal vein at surgery
Doppler
MRA
Normal caliber and patent = 20 Narrow but patent = 6
Normal = 20 (reversed flow = 1) Normal = 1, Thrombosed = 1 (reversed flow = 4)
Normal = 19, Thrombosed = 1 Narrow but patent = 6
2.1. Group 1 Doppler imaging showed a normal-caliber, patent portal vein with forward flow in 20 patients; in 19, these findings were concordant with those at MRA and were confirmed at surgery. One patient with BASM was found to have a narrow-caliber (3 mm), preduodenal portal vein at MRA and at surgery and required an interposition iliac vein graft at transplant. In the 6 remaining patients, Doppler studies revealed reversed flow in a patent portal vein on at least one occasion in 5 patients (2 with biliary atresia [BA] and 1 each with cystic fibrosis, Alagille syndrome with hepatopulmonary syndrome, and neonatal liver failure) and a thrombosed portal vein in 1 infant with BASM. In the 5 patients with reversed flow on Doppler, MRA showed a narrow but patent portal vein in 4 and a thrombosed portal vein in 1. At surgery, MRA findings were confirmed in the 4 patients with a narrow portal vein, but the patient with a presumed thrombosed portal vein on MRA (associated with cystic fibrosis liver disease and massive mesenteric collaterals) had a patent 1-cm-diameter healthy portal vein at transplant. The single patient in whom Doppler studies identified a thrombosed portal vein was found to have a narrow but patent vein both at MRA and at subsequent transplant. Thus, the portal vein was abnormal at surgery in 5 of the 6 patients in whom Doppler studies demonstrated reversed flow or possible thrombosis. Four of these patients required anastomosis of a donor venous conduit to the confluence of splenic and superior mesenteric veins. Table 1
Indications for liver transplantation
Group 1 (n = 26) a
BA Cystic fibrosis a-1 antitrypsin deficiency Hepatopulmonary syndrome Autoimmune hepatitis Neonatal liver failure Langerhans cell histiocytosis Secondary biliary cirrhosis Liver tumor Budd-Chiari syndrome PFIC
n Group 2 (n = 11)
n
10 Non–A-E fulminant hepatitis 3 Autoimmune hepatitis 2 Fulminant hepatitis A
3 2 1
2 Wilson’s disease
1
1 Drug-induced acute liver failure 1 2 TPN/Short gut liver failure 1 1 BASM 1 1 PFIC (acute decompensation) 2 1 1
PFIC indicates progressive familial intrahepatic cholestasis. a Two patients with BASM syndrome.
1
645
The overall comparison of portal vein imaging with Doppler and MRA is shown in Table 2. The sensitivity and specificity of MRA in the detection of an abnormally narrow portal vein were 100% (6/6) and 95% (19/20), respectively. If reversed or absent flow in the portal vein was taken as an index of a potentially abnormally narrow vein in this pediatric population, the sensitivity and specificity of Doppler US were 83% (5/6) and 95% (19/20), respectively. With respect to hepatic arterial anatomy, Doppler studies simply confirmed a patent hepatic artery at the liver hilum in all cases. Magnetic resonance angiography additionally revealed the presence of 1 replaced and 2 accessory right hepatic arteries arising from the superior mesenteric artery and 1 left accessory artery. These findings were confirmed at surgery, but there were 4 additional patients with an accessory left hepatic artery and 1 with an accessory right hepatic artery. These vessels did not measure more than 2 to 3 mm in diameter. A detailed retrospective review of the magnetic resonance (MR) angiograms identified 2 of the 4 bmissedQ accessory left hepatic arteries and the accessory right hepatic artery although the arterial acquisition phase was suboptimal in the latter study. The IVC was reported to be normal at preoperative Doppler and MRA in 25 patients. One child with BASM and normal US appearances of the IVC was reported to have azygos continuation of the IVC on MRA, which was consistent with the findings at surgery. In the patient with Budd-Chiari syndrome, both Doppler US and MRA correctly identified patency of the IVC with occlusion of the hepatic veins and right portal vein.
2.2. Group 2 This group comprised 11 patients who underwent relatively urgent OLTs and who were assessed preoperatively by Doppler US alone. A patent, normal-caliber portal vein was demonstrated in 10 patients and confirmed at surgery. One sick infant with BA, polysplenia and situs inversus (BASM), and marked ascites in whom an MRA under general anesthesia would have been detrimental was found to have reversed flow in the portal vein on Doppler. At transplant 3 days later, he was noted to have a narrow thrombosed portal vein and required an interposition iliac vein graft reconstruction. Thus, Doppler US proved reliable in imaging the portal
646 vein in 91% (10/11) of the patients. Reversed flow in the portal vein was again indicative of an abnormality. Two patients had an accessory right hepatic artery at surgery that was not detected during preoperative Doppler studies. The IVC was reported to be normal in 10 patients and the infant with BASM (see above) was judged to have an interrupted IVC. These findings were verified at surgery.
3. Discussion Simultaneous with advances in liver transplantation, there have been significant developments in hepatobiliary imaging. In the 1980s, radiological assessment of liver transplant candidates included US imaging, computerized tomography (with rapid sequence enhancement) to delineate abnormalities of the portal and hepatic veins, and angiography to establish arterial anatomy [5]. Pediatric units relied predominantly on Doppler US for imaging the vessels, even in BASM [7]. Although MR imaging has been recognized to be useful in evaluating venous anatomy before a liver transplant, as recently as 1997, selective arteriography of the celiac trunk and superior mesenteric artery has been considered by some authors to be the definitive test for assessing hepatic arterial and portal venous anomalies [8]. Subsequent studies in adults and children have shown MR to be equal [9,10] or superior [11] to digital subtraction angiography in portal vein evaluation. Few studies have directly compared MRA with duplex Doppler sonography in transplant candidates. In a study from Boston, USA, Finn et al [12] demonstrated MRA to be superior to duplex Doppler imaging in evaluating the portal venous system of adults with cirrhosis awaiting transplant. In our own unit in Leeds, Naik et al [13] concluded that MR provided superior imaging of the intrahepatic portal branches, varices, and splenorenal shunts in adults and recommended MRA in all patients with abnormal portal venous anatomy or with equivocal findings on Doppler US. Portal vein imaging is a critical part of the preoperative evaluation of potential transplant candidates. The prevalence of preoperative portal vein thrombosis is between 2% and 13% in large transplant series [14,15]. Once considered to be a contraindication to transplant, it no longer precludes a successful outcome but does have an impact on the technical complexity of the operation and the perioperative morbidity. Prior knowledge of the presence and extent of portal vein thrombosis allows surgeons to anticipate difficulties and plan technical solutions [16]. Other anomalies of the portal vein such as a preduodenal or hypoplastic vein are associated with BA, especially its syndromic variety (BASM) [1]. When the portal vein is less than 4 mm in diameter, dissection back to the confluence of the splenic and superior mesenteric veins, with or without anastomosis of a venous conduit, is often necessary to achieve a wide anastomosis. An unrecognized preduodenal portal vein is at risk of iatrogenic injury early in the dissection of the portal triad [17]. It is advisable for
K.V. Ravindra et al. transplant surgeons to know the exact anatomy of the portal vein in these patients. In our study, Doppler studies were inaccurate in all 3 infants with BASM but MRA correctly identified the portal vein characteristics in the 2 elective cases. On the basis of this limited study, we continue to recommend routine MRA imaging of the portal vein in all pediatric transplant candidates with BASM. Similar imaging is also beneficial in the evaluation of transplant candidates with liver tumors or Budd-Chiari syndrome, but for most other transplant indications in children, if the portal vein appears normal on Doppler, further imaging is unlikely to be helpful. In our study, suspected portal vein thrombosis or reversed flow on Doppler US was associated with an abnormal portal vein in 6 of 7 (86%) children, 4 of whom were infants. Therefore, children with abnormal US findings may be an additional group in which preoperative MRA imaging of the portal vein is valuable. Inferior vena caval anomalies are rare. Lerut et al [2] reported an incidence of 2.9% in a large series of predominantly adult liver transplant patients. Most IVC anomalies occur in the pediatric population with BASM. Boillot et al [3] reported a 3.7% incidence of IVC anomalies in BA. In all cases, the IVC was interrupted between the renal veins and the hepatic veins. Two patients in our series, both with BASM, had an interrupted cava. This was correctly identified by MRA in the group 1 patient and by duplex Doppler US in the group 2 patient. However, given the spectrum of congenital IVC anomalies, prior knowledge of IVC anatomy is not critical except in cases of tumor and Budd-Chiari syndrome. Conventional angiography is now seldom used in pretransplant evaluation of arterial anatomy. However, MRA may not identify an accessory artery if it is small (V1-2 mm) [11,18], which is more likely in children, or in the case of a left accessory artery if it is obscured by left gastric varices. In our study, 4 accessory left hepatic arteries and 1 accessory right hepatic artery were not initially reported on MRA, 3 of which were present on retrospective review of the imaging. Detection of an accessory left hepatic artery could be improved by performing a left oblique coronal acquisition with a second injection of gadolinium, but the value of this is questionable because arterial anomalies in the recipient only have a limited impact on the technique of cadaveric liver transplantation in children. More important anomalies such as celiac artery stenosis [19] have not been reported in children undergoing OLT. With the availability of vascular allografts, the arterial inflow can
Table 3 Indications for MRA in pediatric cadaveric liver transplant evaluation !BASM syndrome !Abnormal portal vein on duplex Doppler US !Budd-Chiari syndrome !Tumors (eg, hepatoblastoma) !Prior portosystemic shunt
Preoperative vascular imaging in pediatric liver transplantation be reconstructed if the recipient hepatic artery is inadequate or if there are donor arterial anomalies [20]. Thus, routine MR imaging of arterial anatomy is not strictly necessary in pediatric cadaveric liver transplantation. However, because the incidence of hepatic arterial anomalies is particularly high in infants with BASM [17], MRA may still be valuable in helping surgeons anticipate the need for arterial reconstruction in these patients. Few studies have addressed the role of preoperative vascular imaging in pediatric liver transplant candidates and there is little evidence base for the selection of appropriate investigations. The present study suggests that a detailed Doppler examination of the hepatic vasculature by an experienced sonographer/radiologist is sufficient for most children scheduled for cadaveric liver transplantation. Routine MRA is recommended in children with the BASM (which is almost always apparent from a previous laparotomy or US scan), and in those with abnormal duplex Doppler US findings (Table 3). In addition, although there is insufficient evidence from the present study, we consider that pretransplant MRA (particularly to assess venous anatomy) is also indicated in children with tumors, those who are known to have a portosystemic shunt, and in patients with Budd-Chiari syndrome.
References [1] Whitington PF, Alonso EM. General criteria for pediatric transplantation. In: Busuttil RW, Klintmalm GB, editors. Transplantation of the liver. Philadelphia7 WB Saunders; 1996. p. 159 - 75. [2] Lerut J, Tzakis AG, Bron K, et al. Complications of venous reconstruction in human orthotopic liver transplantation. Ann Surg 1987;205:404 - 14. [3] Boillot O, Sarfati PO, Bringler J, et al. Orthotopic liver transplantation and pathology of inferior vena cava. Transplant Proc 1990;22:1567 - 8. [4] Michels NA. Blood supply and anatomy of the upper abdominal organs with a descriptive atlas. Philadelphia (Pa)7 Lippincott; 1955. [5] Dixon A, Sherwood T. Radiological assessment in liver transplantation. In: Calne SR, editor. Liver transplantation. 2nd ed. London7 Grune & Stratton Inc; 1987. p. 91 - 5.
647
[6] Boeve WJ, Kok T, Haagsma EB, et al. Superior diagnostic strength of combined MR-angiography and MR-imaging compared to intraarterial DSA in liver transplantation candidates. Magn Reson Imaging 2001;19:609 - 22. [7] Kelly DA. Paediatric liver transplantation. In: Neuberger J, Lucey MR, editors. Liver transplantation—practice and management. Plymouth7 BMJ Publishing Group; 1994. p. 97. [8] Suhocki P, Chari R, McCann R, et al. Vascular aspects of liver transplantation. In: Killenberg PG, Clavien PA, editors. Medical care of the liver transplant patient. Massachusetts7 Blackwell Science Inc; 1997. p. 180 - 2. [9] Kreft B, Strunk H, Flacke S, et al. Detection of thrombosis in the portal venous system: comparison of contrast-enhanced MR angiography with intraarterial digital subtraction angiography. Radiology 2000;216:86 - 92. [10] Cheng YF, Huang TL, Lui CC, et al. Magnetic resonance venography in potential pediatric liver transplant recipients. Clin Transplant 1997;11:121 - 6. [11] Kopka L, Rodenwaldt J, Vosshenrich R, et al. Hepatic blood supply: comparison of optimised dual phase contrast-enhanced threedimensional MR angiography and digital subtraction angiography. Radiology 1999;211:51 - 8. [12] Finn JP, Kane RA, Edelman RR, et al. Imaging of the portal venous system in patients with cirrhosis: MR angiography vs duplex Doppler sonography. AJR Am J Roentgenol 1993;161:989 - 94. [13] Naik KS, Ward J, Irving HC, et al. Comparison of dynamic contrast enhanced MRI and Doppler ultrasound in the pre-operative assessment of the portal venous system. Br J Radiol 1997;70:43 - 9. [14] Stieber AC, Zetti G, Todo S, et al. The spectrum of portal vein thrombosis in liver transplantation. Ann Surg 1991;213:199 - 206. [15] Busuttil RW, Shaked A, Millis JM, et al. One thousand liver transplants: the lessons learned. Ann Surg 1994;219:490 - 9. [16] Olthoff KM, Busuttil RW. Venous anomalies including portal vein thrombosis and prior portosystemic shunts. In: Busuttil RW, Klintmalm GB, editors. Transplantation of the liver. Philadelphia7 WB Saunders; 1996. p. 465 - 73. [17] Varela-Fascinetto G, Castaldo P, Fox IJ, et al. Biliary atresia– polysplenia syndrome: surgical and clinical relevance in liver transplantation. Ann Surg 1998;227:583 - 9. [18] Ward J, Guthrie AJ, Hughes T, et al. Anatomy of the arterial supply to the liver demonstrated by MRI. Eur Radiol 1997;7:893 - 9. [19] Garcia-Criado A, Gilabert R, Bargallo X, et al. Radiology in liver transplantation. Semin Ultrasound, CT MR 2002;23:114 - 29. [20] Bowen A, Hungate RG, Kaye RD, et al. Imaging in liver transplantation. Radiol Clin North Am 1996;34:757 - 77.