Chapter 97E Liver transplantation for cholangiocarcinoma and other neoplastic diseases

Justin M. Burns, Charles B. Rosen, Julie K. Heimbach, Gregory J. Gores

Overview

Liver transplantation for malignancies other than hepatocellular carcinoma (HCC) remains highly controversial. Nonhepatocellular carcinomas most commonly treated with liver transplantation include cholangiocarcinoma (CCA), metastatic neuroendocrine cancer (NEC), and hepatic epithelioid hemangioendothelioma (HEHE). Single-center and multicenter results have been equivocal with discrepancies in survival and efficacy. Herein we review the results of liver transplantation for these malignancies—specifically CCA, metastatic NEC, and HEHE—with an aim to suggest guidelines for the application of liver transplantation in the treatment of these diseases.

Hilar Cholangiocarcinoma (See Chapter 50B)

CCA, the second most common primary malignant tumor of the liver, arises from the cholangiocytes of the intrahepatic and extrahepatic bile ducts. The incidence of this tumor has been estimated at 3000 to 5000 cases per year (de Groen et al, 1999; Olnes & Erlich, 2004), and the prevalence of intrahepatic disease appears to be on the rise (Shaib et al, 2004). CCA can arise within the liver, in the perihilar location, or along the extrahepatic bile duct (see Chapters 50A and 50B). CCA has three growth patterns: 1) mass-forming tumors are usually intrahepatic, 2) sclerosing tumors arise in the perihilar and extrahepatic bile ducts, and 3) polypoid tumors grow within the major intrahepatic and extrahepatic ducts. Surgical extirpation has been the standard treatment for all three tumor types.

The treatment of hilar CCA has been most troublesome because it is impossible to achieve a tumor-free (R0) margin of resection. Radical resection with partial hepatectomy was shown to improve survival (Launois et al, 1979) for patients with hilar CCA, but few patients come to medical attention with disease amenable to complete resection. Indeed, fewer than 30% of patients are candidates for resection at diagnosis because of either bilateral liver involvement, encasement of hilar vascular structures, involvement of sectoral bile ducts, and/or underlying liver disease, such as primary sclerosing cholangitis (PSC; see Chapter 41). Liver transplantation appears promising for the treatment of intrahepatic and hilar CCA, because the procedure affords a radical resection, is not limited by bilateral ductal or vascular involvement, and treats underlying liver disease.

Early Experience with Liver Transplantation

Unfortunately, early experiences with liver transplantation for the treatment of CCA were uniformly poor. Liver transplantation for both intrahepatic and hilar CCA was fraught with high recurrence rates and poor patient survival. The Cincinnati Tumor Registry reported a large multicenter analysis for patients transplanted from 1968 to 1997, wherein 1-, 3-, and 5-year patient survival rates were only 72%, 48%, and 23%, respectively (Meyer et al, 2000). The recurrence rate was 51% with a median time to recurrence of only 9.7 months. Local recurrence within the allograft was the most common initial site of recurrence (47%) followed by distant metastases to the lung (30%). Recurrence of tumor portended an extremely poor prognosis, with a median survival of only 2 months. Adjuvant therapy was not found to be beneficial, and no difference was reported in the survival rate of known tumors versus incidental tumors found at the time of orthotopic liver transplantation (OLT); results were poor for both intrahepatic and hilar tumors.

A multitude of retrospective studies have confirmed these findings. A Spanish multicenter study reported similar results for 59 patients who underwent OLT for CCA from 1988 to 2001 (Robles et al, 2004) and found 5-year survival was 30% with a 53% recurrence rate for 39 patients with hilar CCA. Results were equally poor for 23 patients with intrahepatic CCA, for which 5-year survival was 42%, and the recurrence rate was 35%. Similarly, a Scandinavian study reported a 5-year survival of 30% following OLT in a primary sclerosing cholangitis (PSC) population with early-stage CCA (Brandsaeter et al, 2004).

Several centers have reported their outcomes with incidental tumors discovered in patients undergoing transplantation for chronic liver disease. Ghali and colleagues (2005) reviewed the Canadian experience from 1996 through June 2003 and identified 10 cases, 8 arising in patients with PSC. Most of these tumors had favorable characteristics that included small size (<1 cm) and absence of perihepatic lymph node involvement; 90% were well differentiated, and 60% arose in the extrahepatic or hilar ducts. Despite the favorable characteristics, 3-year survival was only 30%. Only the University of California–Los Angeles has reported reasonable survival outcomes in incidental CCA detected in the explant after OLT (Goss et al, 1997). Ten patients with incidental CCA had a 5-year survival of 87%, which was comparable to PSC patients without CCA, although pathologic characteristics were not included in the paper. As with all other experiences, the 4 patients transplanted with known CCA had poor outcomes, and none were alive at 5 years.

A more radical approach with cluster abdominal transplantation reported by the University of Pittsburgh had equally poor results: a 3-year survival of 20% and a 57% recurrence rate (Alessiani et al, 1995). A similar experience was recently reported by Neuhaus’ team in Berlin (Seehofer et al, 2009). Sixteen patients with CCA were treated by combined liver transplantation and pancreatoduodenectomy (PD) between 1992 and 1998, and results were compared to those achieved for 8 patients who did not undergo PD, which at the time of liver transplantation was associated with significantly higher morbidity than transplantation alone. Long-term survival (>4 years) was achieved in only 3 lymph node–negative patients of 20 patients who survived the perioperative period. Neuhaus and colleagues concluded that “there is no good evidence that more radical resections alone are able to markedly improve long-term results.”

With uniformly poor results of liver transplantation for these tumors, intrahepatic and hilar CCA became widely recognized as absolute contraindications for liver transplantation. Both intrahepatic and hilar CCA are best treated by resection; unresectable disease has a prohibitively high recurrence rate after transplantation and warrants additional or palliative therapy.

Neoadjuvant Therapy and Liver Transplantation

Despite the poor results with liver transplantation alone, some patients with favorable hilar CCA—that is, those with negative margins and no regional lymph node metastases—did benefit from transplantation (Shimoda et al, 2001). In addition, a small group of patients at the Mayo Clinic treated with primary radiotherapy and chemosensitization alone, without resection, had a 5-year survival of 22% (Foo et al, 1997). Based on the known palliative efficacy of radiotherapy for CCA—and knowledge that CCA resection failures are usually due to locoregional recurrence, rather than distant metastases (Jarnagin et al, 2003)—the transplant team at the University of Nebraska pioneered a strategy of high-dose neoadjuvant brachytherapy and chemotherapy followed by liver transplantation (Sudan et al, 2002) for patients with unresectable hilar CCA.

The initial Nebraska protocol utilized high-dose intrabiliary brachytherapy, 6000 cGy, followed by daily intravenous 5-fluorouracil (5-FU) until OLT. Patients underwent operative staging when a donor liver became available for transplantation. At operation, the patients were assessed for extrahepatic metastases or regional lymph node involvement. Either finding precluded transplantation, and the donor liver was reallocated to another patient. Seventeen patients received neoadjuvant brachytherapy: 2 patients died from disease progression, and 4 were found to have extrahepatic disease at exploration; 11 patients underwent transplantation. Median survival after transplantation was 25 months; 5 (45%) were alive and disease-free at a median of 7.5 years (range, 2.8 to 14.5 years) after transplantation, 2 patients died from recurrent disease, and 4 patients died from perioperative complications. Overall survival was 30% for the 17 patients 5 years after the start of neoadjuvant therapy.

Mayo Clinic Experience

The transplant team at the Mayo Clinic embraced the concept pioneered by the team at the University of Nebraska and implemented a protocol in 1993 through a collaborative effort of oncologists, hepatologists, and surgeons. The general theory is that neoadjuvant therapy and liver transplantation should provide the best possible control of local disease. The rationale for the protocol was based on several factors: 1) the known CCA tumor response to high-dose radiotherapy; 2) hepatotoxicity of radiotherapy is obviated by liver transplantation; 3) liver transplantation achieves radical resection, including removal of residual disease following neoadjuvant therapy; 4) liver transplantation is not limited by underlying liver disease (PSC), vascular involvement, or concern about intrahepatic extension of disease; 5) neoadjuvant therapy prior to operative staging and transplantation might avoid tumor dissemination during operation; and 6) careful patient selection with operative staging prior to liver transplantation could exclude patients with advanced disease and regional lymph node metastases that are destined to develop distant metastatic disease.

Inclusion and Exclusion Criteria

Criteria for protocol enrollment are designed to select those patients least likely to develop metastatic disease, most likely to respond to neoadjuvant therapy, and who have a high probability for survival after transplantation. Appropriate patients include those with early-stage hilar CCA determined to be unresectable or those who have underlying PSC, because CCA arising in the setting of PSC has a very poor natural history following standard resection (Rosen et al, 1991).

Criteria for anatomic unresectability include bilateral segmental ductal extension, encasement of the main portal vein, unilateral segmental ductal extension with contralateral vascular encasement, and unilateral atrophy with either contralateral segmental ductal or vascular involvement. Due to the difficulty of assessing extent of disease along the bile duct, especially in the setting of PSC, no longitudinal limits exist for bile duct involvement.

Original criteria required that hilar CCA not extend lower than the cystic duct, but it was subsequently found that early CCA arising in PSC with unsuspected common bile duct involvement found at transplantation was amenable to transplantation with PD; however, patients with CCA extending below the cystic duct on cholangiography are excluded, because they have larger tumors that are more likely to abut the portal vein and be less amenable to complete extirpation during transplantation.

Vascular encasement of the hilar vessels is not a contraindication to transplantation. The upper limit of tumor size is 3 cm when a mass is visible on cross-sectional imaging studies, and there must be no evidence of intrahepatic or extrahepatic metastases by chest computed tomography (CT), abdominal CT or magnetic resonance imaging (MRI), ultrasonography, or bone scan. Endoscopic ultrasound (EUS) is performed prior to neoadjuvant therapy to exclude patients with regional lymph node metastases.

The Mayo Clinic protocol specifically excludes patients with evidence of intrahepatic or extrahepatic metastases or gallbladder involvement. Surgical intervention and any type of transperitoneal biopsy or fine needle aspiration have emerged as absolute contraindications to enrollment. These procedures result in an unacceptable rate of peritoneal seeding, which has been discovered during operative staging (unpublished data). Candidates must have no active infections or medical conditions that preclude either neoadjuvant therapy or OLT.

Tumor Diagnosis

Diagnostic criteria for neoadjuvant therapy and liver transplantation require a malignant-appearing stricture on cholangiography and either visualization of a mass on cross-sectional imaging, cytologic or histologic confirmation of CCA by transluminal brushing or biopsy, elevation of carboydrate antigen (CA)19-9 greater than 100 U/mL, or polysomy by fluorescence in situ hybridization (FISH). Patients with indeterminate diagnostic criteria—such as FISH trisomy (7 or 3), dysplasia from brushings, FISH polysomy in the absence of a malignant-appearing stricture, or a malignant-appearing stricture in the absence of a mass lesion are followed closely with repeat endoscopic retrograde cholangiography (ERCP) with brushings, cross-sectional imaging, and laboratory testing.

Neoadjuvant Therapy

Neoadjuvant therapy is administered by external beam radiation (40 to 45 Gy), followed by transcatheter radiation (20 to 30 Gy) with iridium wires. Wires are placed preferentially by ERCP and percutaneous transhepatic cholangiography (PTC) when ERCP is technically not possible. Intravenous 5-fluorouracil (5-FU) is administered for chemosensitization during radiation therapy, and capecitabine is administered afterward, while patients await transplantation.

Staging Operation

All patients undergo operative staging prior to OLT. Operative staging includes a thorough abdominal exploration with careful palpation of the liver to identify small, previously undetected intrahepatic metastases, biopsy of any suspicious nodules, and excision of a proximal proper hepatic artery lymph node (at the take-off of the gastroduodenal artery) and a pericholedochal lymph node posterior to the common bile duct just superior to the pancreas. The caudate process and retrohepatic vena cava are assessed for suitability of a caval-sparing hepatectomy, which is necessary for recipients of living-donor liver grafts. Extrahepatic or intrahepatic metastases, lymph node involvement, or locally extensive disease preclude transplantation. Survival for patients with these findings is low, and survival beyond a year is very rare. The staging operation was initially performed through a right subcostal incision with extension along the future liver transplant incision as necessary; however, during the past few years, most procedures have been accomplished by hand-assisted laparoscopy utilizing a smaller, right subcostal incision.

Timing of the staging operation depends on the possibility of a living-donor liver transplantation (LDLT) or the anticipated waiting time for a deceased-donor liver transplantation (DDLT). The staging procedure is performed 1 to 2 days prior to LDLT or as the patient nears the time for DDLT. Patients with underlying cirrhosis as a result of PSC or liver dysfunction as a result of cholestasis are somewhat prone to decompensation following the staging procedure, which may lead to an increase in a patient’s MELD score and advance their position on the deceased-donor waiting list, but it also leads to an increase in perioperative morbidity and mortality. Unfortunately, only supportive care is possible for those patients who decompensate after a positive staging operation, and few survive the perioperative period.

Prior to 2003, 30% to 40% of patients had findings during the staging operation that precluded transplantation. EUS-guided aspiration of regional nodes (not the primary tumor) was implemented to exclude patients with lymph node metastases prior to initiation of neoadjuvant therapy. Initial findings from 47 patients identified eight (17%) with metastases (Gleeson et al, 2008). No morphologic features of the lymph nodes were found at EUS that predicted microscopic disease. Since routine use of EUS was implemented in 2003, the percentage of patients with a positive staging exploration has been reduced to 15%. EUS avoids the morbidity and mortality of high-dose neoadjuvant therapy and prevents an unnecessary operation for patients destined to fall out at operative staging.

Liver Transplantation

The technical difficulty and nuances of transplantation outweigh those of standard OLT. The operation is performed via a standard bilateral subcostal incision with vertical extension in the midline. Hilar dissection is avoided to prevent tumor manipulation and to reduce the possibility of intraoperative dissemination. There is typically extensive scar tissue in the hepatoduodenal ligament because of the neoadjuvant therapy and the previous staging procedure, which can make the dissection very difficult. The irradiated native hepatic artery is avoided during transplantation with a deceased donor, and arterial inflow is established with a segment of deceased donor iliac artery sewn to the infrarenal aorta. This approach was initially applied to LDLT but unfortunately resulted in an unacceptable rate of hepatic artery thrombosis. Better results have been achieved using the native artery, which is sewn directly to the living donor artery despite prior radiation therapy.

Next, the bile duct is transected as close to the pancreas as possible, and it may be possible to enucleate a short segment of additional common bile duct from the head of the pancreas. The margin is submitted for frozen section examination. Microscopic tumor involvement at this margin occurs in 10% to 15% of patients with PSC, and the possibility of PD is discussed with all patients prior to initiation of neoadjuvant therapy. Bilioenteric continuity is restored with a standard Roux-en-Y hepaticojejunostomy (living-donor graft) or choledochojejunostomy (deceased-donor graft).

The portal vein may be somewhat brittle as a result of the neoadjuvant therapy, and an injury may be difficult to control. The portal vein is divided as low as possible and is not dissected free up into the hilus of the liver. Despite low division, the deceased-donor portal vein is almost always long enough for an end-to-end anastomosis. A segment of a deceased-donor iliac vein is used as an interposition graft between a living-donor right or left portal vein and the native vein during LDLT. A caval-sparing hepatectomy is performed in most cases, and the donor suprahepatic vena cava is sewn to the left/middle hepatic vein trunk during DDLT; the hepatic vein is reconstructed in an end-to-end fashion during LDLT. If there is concern for tumor extension into the caudate, which is usually detectable during the staging operation, the retrohepatic cava is excised, and the donor retrohepatic vena cava is sewn to the suprahepatic and infrahepatic cavae as an interposition graft, usually via portovenous and venovenous bypass.

Results

One hundred ninety-six patients were enrolled in the Mayo Clinic protocol from 1993 until April 10, 2010 (Fig. 97E.1). Sixteen patients died, became too debilitated for transplantation, or had disease progression prior to operative staging. Three patients underwent transplantation at other centers, and five were receiving neoadjuvant therapy and/or awaiting operative staging. One hundred seventy-two patients underwent operative staging, and 38 patients (22%) had findings precluding transplantation. After staging, 1 patient died from hepatic decompensation, 1 developed intrahepatic metastases, and 3 patients underwent transplantation at other centers. The remaining 126 patients underwent transplantation, 84 with deceased-donor grafts, 41 with living-donor grafts, and one with a domino familial amyloid donor graft.

FIGURE 97E.1 Mayo Clinic neoadjuvant therapy and liver transplantation for hilar cholangiocarcinoma, 1993 through 2010.

Overall patient survival from enrollment in the protocol (intention-to-treat analysis) was 56% ± 4% at 5 years after the start of neoadjuvant therapy (Fig. 97E.2), 63% ± 5% for patients with underlying PSC, and 44% ± 7% for patients with de novo CCA. Five-year survival after transplantation was 74% ± 5%; 80% ± 5% for patients with underlying PSC and 64% ± 8% for patients with de novo CCA (Fig. 97E.3). No difference in survival was reported among patients undergoing LDLT versus DDLT.