The perspective of liver transplantation for cholangiocarcinoma^†

In the early 1990s, the disappointment about the poor outcome of patients undergoing liver transplantation for cholangiocarcinoma had prompted most transplant centers to consider this type of malignancy as a contraindication for transplantation. A few centers have tried to find innovative surgical strategies for cholangiocarcinoma, including extended and new types of hepatic resections as well as extended types of transplantations, such as upper abdominal cluster transplantation.1-6 Other attempts have included adjuvant and neoadjuvant treatment protocols applying radiotherapy with or without chemotherapy.7-11 Overlooking more than a decade of experience with these different attempts, the Mayo Clinic group has emerged as the most successful among these pioneers. Their results with liver transplantation and neoadjuvant radiochemotherapy for patients suffering from cholangiocarcinoma are comparable to results of liver transplantation for early hepatocellular carcinoma as well as other chronic liver diseases.7, 12, 13 A 5-year-survival rate of 83% is an outstanding and unprecedented figure that a few years ago would not have been considered to be a realistic goal in any type of treatment for cholangiocarcinoma.

Abbreviations

DDLT, deceased donor liver transplantation; LDLT, living donor liver transplantation; PSC, primary sclerosing cholangitis.

Although the patient numbers enrolled in the Mayo Clinic protocol are constantly increasing, with 68 patients in the present report,14 it may be too early to consider neoadjuvant radiochemotherapy as a standard treatment. This caveat is related to both the efficacy and safety of this innovative therapy. One point is a favorable survival figure of 58% at 5 years even in the intention-to-treat group, which contained 71 patients at a time when only 38 patients had completed the protocol including liver transplantation.12 This surprisingly fair survival rate also among patients who turned out to be ineligible for transplantation requires a more detailed look at the reliability of the cancer diagnosis. In fact, the inclusion criteria allowed not only biopsy-proven or cytology-proven cholangiocarcinoma but also a cancer antigen 19-9 serum level exceeding 100 ng/mL in the setting of a radiographic malignant stricture.13 We consider this policy to be fully justified and closely related to the clinical decision-making but have to admit that it may also yield diagnostic inaccuracies.

From an oncological point of view, it appears that patients younger than 45 years of age suffering from a cholangiocarcinoma arising within primary sclerosing cholangitis (PSC) are likely to have the most pronounced benefit.12 In a recent report from the Mayo Clinic, Heimbach et al.13 tried to identify predictors of disease recurrence following neoadjuvant radiochemotherapy and liver transplantation. In the discussion of this article, the authors qualified their findings by stating that a multivariate analysis would be required to determine, for example, the true impact of patient age on outcome. However, the number of recurrences has been favorably low, that is, too small to allow for a multivariate analysis as yet. Of course, this protocol is the most promising strategy to date and must be continued with careful revision of the upcoming results with larger patient numbers and longer follow-up periods.

The problem of an as yet limited experience impairing the possibility of multivariate analyses pertains to safety issues as well. The Mayo Clinic group has considered the toxicity of the protocol of liver transplantation and neoadjuvant radiochemotherapy for patients suffering from cholangiocarcinoma to be significant.13 Safety issues are related to pretransplant complications such as cholangitis and sepsis and are largely due to radiation damage. Early stages of radiation damage are characterized by intravascular coagulation, with deposits of fibrin selectively obstructing the centrilobular outflow venules and larger hepatic veins.15 The vascular lesions lead to congestion and secondary degeneration and atrophy of hepatocytes, a condition termed hepatic veno-occlusive disease. With time, the fibrin is substituted by collagen and reticulin fibers, which also replace the missing parenchymal cells. Fortunately, these earlier changes can be reversed by a total hepatectomy and subsequent liver transplantation at least in those patients who proceed to this point of the protocol.

Late effects are really the dose-limiting factor in radiation therapy. These include necrosis, fibrosis, fistula formation, nonhealing ulceration, and organ damage. One hypothesis is that late effects result from damage to vasculoconnective stroma.16 In the setting of liver transplantation and neoadjuvant hilar radiotherapy, late complications are likely to occur in the immediate confines of the radiation field, that is, within the remnant structures of the lower margin of the hepatoduodenal ligament, including the common hepatic artery and the portal confluens. Generally, radiation damages lymphatic vessels, veins, and arteries with decreasing frequency, and major blood vessels do not ordinarily show signs of therapeutic radiation damage.15, 17 However, arterial damage may appear with a latency of decades, as has been shown in former reports on carotid artery occlusion after 16, 30, and 52 years.18-20

In the present article, Mantel et al.14 report an unusually high incidence of vascular complications. Vascular complications developed in 40% of the 68 patients, with arterial and portal venous complication rates of 21% and 22%, respectively. Seven patients died during the study period, and 2 of these deaths were attributable to arterial complications. Five patients required retransplantation for arterial complications. In contrast, portal venous complications were equally common but did not lead to graft loss or patient death. Two groups of patients are distinctly examined in this report: patients undergoing deceased donor liver transplantation (DDLT) and patients undergoing living donor liver transplantation (LDLT).

The DDLT group (n = 51) was further divided into an early cohort undergoing transplantation without an arterial interposition graft (n = 7) and a cohort reflecting the current policy of using deceased donor iliac arteries for reconstruction with an interposition graft to the infrarenal aorta in the recipient (n = 45). Formerly, DDLT without an arterial interposition graft had been abandoned by the Mayo Clinic group when applying neoadjuvant radiotherapy because of arterial complications in 2 of 7 patients necessitating retransplantation in 1 patient. This change of policy appears worthwhile in the light of former results emphasizing the importance of extra-anatomic approaches in the treatment of radiation-induced arterial occlusions.21 Indeed, the authors did not find any difference in early or late arterial complications between the study group and a matched patient population that had undergone DDLT using an arterial interposition graft to the infrarenal aorta but for an indication different from cholangiocarcinoma. On the other hand, the risk of arterial complications was 11% in the DDLT study group with an arterial interposition graft versus 15% in the matched group. These rates represent arterial stenoses and thromboses and are difficult to compare to previous reports. The reported overall rates of early and late posttransplant hepatic artery thromboses in adults range from 2% to 9% (reviewed in Stange et al.22). Although it is certainly justified to select patients with an arterial interposition graft for the matched control group, the question arises whether an arterial interposition graft is a risk factor for arterial complications itself. In an analysis of our experience after 1200 liver transplantations, arterial interposition grafts to the supraceliac aorta, which was our standard approach, or to the infrarenal aorta carried a risk of hepatic artery thrombosis of 10% or 6%, respectively.22 The risk in all other types of arterial reconstruction was only 2%. If the Mayo Clinic group had similar results, it would be rather obvious that an increased rate of hepatic artery complications that is due to a specific type of arterial reconstruction is an inherent risk of the protocol requiring this specific type of reconstruction.

It remains speculative whether such an inherent risk exists. Moreover, it is speculation whether hypercoagulability in patients with PSC adds to this risk by translating into an increased rate of posttransplant vascular complications.23 Indeed, it has been shown that platelet function differs between patients suffering from cholestatic and noncholestatic liver diseases and may be hyperactive in patients with PSC.24 One molecular mechanism is triggered by a marked systemic inflammatory activity in PSC patients. A variety of chemokines and leukocyte interactions via surface receptors have been reported to modulate platelet function and contribute to activation of platelets.25-27 Theoretically, such a marked systemic inflammatory activity might be further increased by bile duct stenosis, malignancy, and radiotherapy. However, there is no convincing evidence that these mechanisms really have a clinical impact post transplant. In our experience, the rate of posttransplant hepatic artery thromboses was 1.5% in PSC patients and 0.9% in those suffering from malignant diseases versus 3.2% in all other patients.22

An increased risk of late portal vein complications reached statistical significance in the DDLT group (18% versus 2% in the matched controls; P = 0.01). Although veins are more prone to radiation damage than arteries, not all of the late portal vein complications could be exclusively associated with a radiation injury. Instead, 1 late thrombosis was associated with recurrent cholangiocarcinoma, and 2 stenoses occurred in patients dying later from recurrent cholangiocarcinoma. It is noteworthy that portal complications could be successfully treated with percutaneous transhepatic portal angioplasty and did not result in graft loss or patient death. This finding is in accordance with previous reports on DDLT in adults and in children and on LDLT.28-31 It shows that angioplasty is also effective after neoadjuvant radiotherapy, with the reservation that mostly multiple interventions were required and that the future course is less predictable in view of the progressive nature of late portal vein stenoses. Our own experience in patients without radiation is also favorable, although we favor the transjugular intrahepatic technique of portal vein angioplasty also for extrahepatic stenoses because it is likely to reduce the risk of intra-abdominal bleeding in comparison with the percutaneous transhepatic approach.32

In the LDLT group, the strategies of vascular reconstruction differed profoundly from those of the DDLT group. In the first 4 patients, deceased donor iliac artery grafts were used for arterial reconstruction with poor results.33 At that time, the Mayo Clinic group had temporarily stopped performing LDLT for cholangiocarcinoma while accruing more experience with LDLT for other indications. In 2004, they resumed the LDLT program for cholangiocarcinoma and changed to arterial reconstructions using the recipient common hepatic artery or a distal branch (proper, left, or right hepatic artery) without an interposition graft in 11 patients until the end of the observation period of the present study. It has not been specified which distal branch was used in how many patients, although it might have been interesting to observe the effects of advancing the localization of the arterial anastomosis further into the external beam radiation field. In particular, the right hepatic artery lies in very close proximity to the tumor-bearing area within the hepatic hilum, that is, an area carrying the additional risk of damage by brachytherapy as well as an increased oncological risk of narrow safety margins and microscopic tumor cell dissemination. In contrast to DDLT and despite a smaller patient group and a shorter follow-up period, the risk of late hepatic arterial complications reached statistical significance in comparison with a matched LDLT group not suffering from cholangiocarcinoma (2 of 11 patients versus 0 of 38 patients; P = 0.047).

An increased rate of late portal vein complications after LDLT and neoadjuvant radiotherapy for cholangiocarcinoma was even more explicit (3 of 11 patients versus 0 of 38 patients; P = 0.009). It is especially noteworthy that all living donor transplants were done with a segment of deceased donor iliac vein as an interposition graft between the donor right portal vein branch and the recipient portal vein. It has not been specified whether these deceased donor iliac vein grafts had been cryopreserved. Generally, cryopreserved venous grafts for venous reconstruction do not achieve the good results obtained with cryopreserved arterial grafts for arterial reconstruction.34, 35 In pediatric LDLT, the risk of portal vein stenoses has been reported to be as high as 27% when cryopreserved veins are used for portal conduits.36 When conduits were excluded, the incidence of late portal vein complications was reduced to 1%.

One patient from the LDLT group (n = 11) died from recurrent cholangiocarcinoma. No patient death after LDLT was attributable to vascular complications. In contrast, 2 retransplantations were performed in patients with an early arterial stenosis and an arterial thrombosis. Because of the limited group size and the short follow-up period, no definite conclusions on the risks for graft survival after LDLT and neoadjuvant radiotherapy can be drawn as yet. New techniques and indications require a growing body of data and critical revision of different experiences, especially in the setting of living donation. LDLT still is a new treatment modality into which another innovative protocol has been introduced. For the moment, it remains an open issue whether the success story as which the Mayo Clinic protocol of neoadjuvant radiotherapy and DDLT for cholangiocarcinoma is appearing to emerge will also come true for the setting of LDLT.