Liver transplantation for cholangiocellular carcinoma: Analysis of a single-center experience and review of the literature

ORIGINAL ARTICLES

Liver Transplantation for Cholangiocellular Carcinoma: Analysis of a Single-Center Experience and Review of the Literature Mitsugi Shimoda,* Douglas G. Farmer,† Steven D. Colquhoun,‡ Michael Rosove,§ R. Mark Ghobrial,† Hasan Yersiz,† Pauline Chen,† and Ronald W. Busuttil† Cholangiocellular carcinoma (CCC) is a biliary malignancy that frequently presents in advanced unresectable stages. The role of liver transplantation (LT) as a surgical modality is unclear. The goal of this study is to evaluate outcomes of patients with CCC undergoing LT. A retrospective analysis of all patients undergoing LT was undertaken. Only those patients with the pathological diagnosis of CCC were included on the study. Patients were divided into two groups based on primary tumor location: extrahepatic (EH)-CCC and intrahepatic (IH)-CCC. The Kaplan-Meier method was used to calculate overall and recurrence-free survival. Log-rank analysis was used to determine the significance of prognostic variables. Twenty-five patients were identified: 9 patients with EH-CCC (5 patients, Klatskin-type; 2 patients, the middle third; and 2 patients, the distal third) and 16 patients with IHCCC. Mean age was 47.1 ⴞ 10.6 years. There were 14 men and 11 women. Tumor stage was local (stages I and II; n ⴝ 9) or advanced (stages III and IV; n ⴝ 16). Overall and disease-free survival rates were 71% and 67% at 1 year and 35% and 32% at 3 years, respectively. Analysis of variables showed statistically significant improved outcomes (P < .05) for the absence of contiguous organ invasion at LT, small tumor size, and single tumor foci. This study indicates that early survival after LT for CCC is acceptable. Three-year disease-free survival is achieved in approximately 30% of patients. These outcomes can be improved by applying strict selection criteria based on prognostic variables identified in this study. (Liver Transpl 2001;7:1023-1033.)

Although surgical resection is the treatment of choice affording the best outcome,7-16 many patients have CCC that is unresectable because of associated advanced liver disease, centrally located lesions, or lesions involving portal structures. For these patients, liver transplantation (LT) theoretically provides the opportunity for wide resection margins, thus potentially expanding the number of patients eligible for surgical intervention. However, the allocation of limited donor organs to this patient population has been extremely controversial. In the literature, several centers reported the use of LT for patients with CCC10,17-28 with limited results because of high recurrence rates. The purpose of this study is to retrospectively review our experience with LT for patients with IH-CCC and EH-CCC to determine prognostic variables that may impact on patient selection and outcome.

C

From the *Department of Gastroenterological and Hepatobiliary Surgery, Dokkyo University School of Medicine, Mibu, Tochigi, Japan; Departments of †Surgery and §Medicine, The Dumont-University of California at Los Angeles Transplant Center, Division of Liver and Pancreas Transplantation, University of California at Los Angeles School of Medicine; and the ‡Department of Surgery, Cedars Sinai Medical Center, Los Angeles, CA. Supported in part by The Dumont Foundation, The Joanne Barr Foundation, The Hearst Foundation, The Herbalife Foundation, and The Joan S. and Ralph N. Goldwyn Trust. Address reprint requests to Douglas G. Farmer, MD, Division of Liver and Pancreas Transplantation, Rm 77-120 CHS, Box 957054, Los Angeles, CA 90095-7054. Telephone: 310-794-6487; FAX: 310267-0392; E-mail: [email protected] Copyright © 2001 by the American Association for the Study of Liver Diseases 1527-6465/01/0712-0010$35.00/0 doi:10.1053/jlts.2001.29419

holangiocellular carcinoma (CCC) is an infrequent tumor of the biliary tree with an estimated incidence of 0.01% to 0.8%.1-3 Histologically, most CCCs are adenocarcinomas,1 and classification based on location classically has been applied.4,5 Intrahepatic (IH) or peripheral CCC arises from IH bile ducts and is the second most common primary liver cancer behind hepatocellular carcinoma (HCC).3 Extrahepatic (EH)CCC arises from the biliary tree outside the liver parenchyma and also has been referred to as bile duct carcinoma, hilar, or central CCC. EH-CCCs classically have been described according to the classification of Longmire et al,6 which separates tumors into the proximal one third (also known as Klatskin-type5), middle one third, and distal one third lesions.

Patients and Methods General From 2,727 LTs performed at the Dumont-University of California at Los Angeles Transplant Center between February 1984 and January 2000 and 198 LTs performed at Cedars Sinai Hepatobiliary Institute (Los Angeles, CA) between Sep-

Liver Transplantation, Vol 7, No 12 (December), 2001: pp 1023-1033

1023

1024

Shimoda et al

tember 1995 and January 2000, all patients with CCC were identified and included on this study. No patients were lost to follow-up. Data were collected retrospectively from all hospital medical records and transplant databases. The study concluded in January 2000, affording a median follow-up of 22.3 months.

Table 1. pTNM Staging for IH-CCC Stage I Stage II Stage III

Preoperative Assessment All patients were evaluated by a multidisciplinary transplant selection committee. Routine pre-LT imaging included hepatic ultrasonography, abdominal computed tomography (CT), and chest radiography. Nonroutine imaging applied only to patients with suspected malignancy or high-risk patients included chest CT, magnetic resonance imaging of the abdomen, technetium 99 –labeled bone scanning, and positron emission tomographic scanning. Serum tumor markers, such as alphafetaprotein (AFP), carcinoembryonic antigen (CEA), and cancer antigen (CA19-9), were measured selectively. For patients with primary sclerosing cholangitis (PSC), cholangiography (either transhepatic or endoscopic) was used to establish the diagnosis, as well as for periodic screening for disease progression or development of malignancy. Dominant strictures were routinely biopsied or brushed for histopathologic and/or cytopathologic evaluation. In the more recent transplant experience, candidates awaiting LT for more than 1 year were reevaluated and imaged appropriately.

Surgical Conduct LT was performed using standard techniques29,30 after abdominal exploration was undertaken to identify the presence of other intra-abdominal pathological states. Perihilar lymphatics were not routinely excised with the explant specimen, except in cases of suspected malignancy. For patients with known malignancy, surgical resection always was considered the first option, and LT was undertaken only for patients not eligible for resection. The application of other radical procedures was dictated by the nature of the malignancy, intraoperative findings, and adjacent organ involvement.

Postoperative Care and Immunosuppression Standard postoperative care was applied in all cases. Before July 1994, immunosuppression was cyclosporine-based (Sandimmune; Novartis Pharmaceuticals Corp, East Hannover, NJ; Neoral, Novartis Pharm AG, Basel, Switzerland). After July 1994, tacrolimus-based (FK506; Prograf; Fujisawa USA, Deerfield, IL) immunosuppression was routinely used. Corticosteroids were administered in a pulse/taper manner. The use of other agents, such as azathioprine (Imuran; Faro Pharmaceuticals Inc, Bedminister, NJ) or mycophenolate mofetil (CellCept; Roche Pharmaceuticals, Nutley, NJ) was avoided as much as possible. Acute cellular rejection was defined only by standard histopathologic criteria, regardless of the clinical scenario.31 Standard rejection therapy included pulse steroids followed by a taper, unless otherwise stated.

Stage IVA Stage IVB

T1 T2 T1 T2 T3 T3 T4 Any T

N0 N0 N1 N1 N0 N1 Any N Any N

M0 M0 M0 M0 M0 M0 M0 M1

Abbreviations: T1, solitary tumor 2 cm or less, no vascular invasion; T2, solitary tumor 2 cm or less with vascular invasion; or multiple tumors in one lobe, all less than 2 cm, no vascular invasion; or solitary tumor greater than 2 cm without vascular invasion; T3, solitary tumor greater than 2 cm with vascular invasion; or multiple tumors in one lobe all less than 2 cm, with vascular invasion; or multiple tumors greater than 2 cm in one lobe with and without vascular invasion; T4, multiple tumors in more than one lobe or tumor invasion of major portal or hepatic venous branch; N0, no regional lymph node metastasis; N1, regional lymph node metastasis; M0, no distant metastasis; M1, distant metastasis.

Screening for tumor recurrence was undertaken periodically using various imaging modalities, and appropriate serum tumor markers were identified before LT. Recurrence was defined as an elevation in serum tumor marker levels or identification of metastatic lesions on imaging. Biopsy was used to assist with the diagnosis, as needed. The use of chemotherapy was evaluated on an individual basis. Its use in an adjuvant and palliative manner was selective.

Statistical Analysis The following variables were collected for analysis: (1) sex; (2) primary site (location of the primary lesion [IH-CCC v EHCCC] judged on explant pathological characteristics; tumor foci originating in EH bile ducts were classified as such, whereas all other proximal lesions were classified as IH); (3) other organ invasion (microscopic or macroscopic invasion into contiguous organs [present v absent]; note that invasion of the liver by hilar CCC was not regarded as positive); (4) concomitant cirrhosis (at minimum, bridging fibrosis on explant pathological examination [present v absent]); (5) pathological tumor, node, metastasis (pTNM) stage (defined by the American Joint Council on Cancer32,33 [stages I ⫹ II v III ⫹ IV; Tables 1 and 2); (6) tumor size (diameter of the largest lesion on explant pathological examination [ⱕ3 v ⬎3 cm v unknown); (7) number of nodules (total number of tumor nodules described on explant pathological examination [single v multiple; more than one]); (8) vascular invasion (extension of tumor into any vascular space on explant pathological examination [present v absent v unknown]); (9) lymph node (presence of tumor in lymph nodes on explant pathological examination; failure to report lymph node analysis on explant pathological examination was classified as unknown [present v absent v unknown]); (10) neural invasion (invasion

Liver Transplantation for Cholangiocarcinoma

Table 2. pTNM Staging for EH-CCC Stage 0 Stage I Stage II Stage III Stage IVA Stage IVB

Tis T1 T2 T1 T2 T3 Any T

N0 N0 N0 N1, N2 N1, N2 Any N Any N

M0 M0 M0 M0 M0 M0 M1

Abbreviations: Tis, carcinoma in situ; T1, tumor invades mucosa or muscular layer; T2, tumor invades perimuscular connective tissue; T3, tumor invades adjacent structures and/or organs; N0, no regional lymph node metastasis; N1, metastasis in the cystic duct, pericholedochal, and/or hilar lymph nodes; N2, metastasis in the peripancreatic, periduodenal, periportal, celiac, superior mesenteric, and/or posterior pancreaticoduodenal lymph nodes; M0, no distant metastasis; M1, distant metastasis.

of tumor into any neural structure or sheath on explant pathological examination; failure to report neural analysis on explant pathological examination was classified as unknown [present v absent v unknown]); (11) incidental tumor (strictly defined as lesion[s] too small to be detected on routine preoperative imaging or intraoperative evaluation and therefore found only on explant pathological analysis [incidental v nonincidental]); (12) adjuvant chemotherapy (administration of any chemotherapy regimen after LT with the goal to prevent tumor recurrence; chemotherapy administered after tumor recurrence was diagnosed or suspected was not included [yes v no]); (13) CEA (greater than normal range [0 to 5.1 ng/mL] at any time before LT [elevated v normal]); (14) AFP (greater than normal range [0 to 6.6 ng/mL] at any time before LT [elevated v normal]); and (15) CA19-9 (greater than normal range [0 to 37 U/mL] at any time before LT [elevated v normal]). Overall and disease-free survival rates were calculated according to the Kaplan-Meier method. Disease-free survival was death censored for non– cancer-related mortality. Follow-up was completed January 2000. The log-rank test was used to compare survival and study variables. P less than .05 is considered statistically significant.

Literature Review A search of all English-language articles related to the topic was undertaken using MEDLINE 1985, 1990, 1995, and 2000 (current/default database). A keyword search was used for each database using the following words or combination of words: cholangiocarcinoma, hepatobiliary carcinoma, liver transplantation, liver resection, cholangiocarcinoma and liver transplantation, cholangiocarcinoma and liver resection, hepatobiliary carcinoma and liver transplantation, and hepatobiliary carcinoma and liver resection. All pertinent articles were obtained and assessed. A secondary search of the bibliography of each article also was undertaken. Results were

1025

compiled according to medical center, with the latest publication referenced. Reports citing fewer than 10 patients with CCC undergoing LT were not included in the analysis. Single-center data were preferred, but pooled data from several centers or registries were included. The most recent publication from each center was used, except when unique information was obtained from an earlier reference.

Results General Twenty-five patients with CCC were identified and included in the analysis. Mean age of patients was 47.1 ⫾ 10.6 years. There were 14 men and 11 women. Clinical findings consistent with end-stage liver disease were present in 13 patients, with underlying causes of PSC (n ⫽ 8), alcohol alone (n ⫽ 1), alcohol plus hepatitis C (n ⫽ 2), hepatitis B alone (n ⫽ 1), and hepatitis B plus hepatitis C (n ⫽ 1). Indications for LT in the 25 patients were as follows: advanced liver disease without known malignancy (n ⫽ 10), recurrent and/or intractable biliary complications without known malignancy (n ⫽ 4), advanced liver disease with known malignancy (n ⫽ 3), and unresectable hepatobiliary malignancy without advanced liver disease (n ⫽ 8). In 14 patients (5 patients, IH-CCC; 9 patients, EHCCC), preoperative imaging failed to detect evidence of malignancy. Preoperative imaging (obtained during evaluation and follow-up) in this group included ultrasonography (n ⫽ 22; 1.6 examinations/patient), abdominal CT (n ⫽ 9; 0.6 examinations/patient), abdominal magnetic resonance imaging (n ⫽ 3; 0.2 examinations/patient), cholangiography (n ⫽ 29; 2.1 examinations/patient), and angiography (n ⫽ 2; 0.1 examinations/patient). In the remaining 11 patients, malignancy was diagnosed by preoperative imaging that included ultrasonography (n ⫽ 15; 1.4 examinations/patient), abdominal CT (n ⫽ 21; 1.9 examination/patient), abdominal magnetic resonance imaging (n ⫽ 3; 0.3 examinations/patient), and visceral angiography (n ⫽ 1; 0.1 examinations/patient). In the latter group, lesions were deemed unresectable because of size, location, or severity of underlying liver disease. Metastatic workup in the latter group also included chest CT (n ⫽ 11) and positron emission tomographic scan (n ⫽ 2). CCC was correctly diagnosed pre-LT in only three patients (12% of the total group; 27% of the known malignancy group); the other eight patients were given an incorrect diagnosis of HCC (n ⫽ 7) or renal cell carcinoma with inferior vena cava (IVC) involvement (n ⫽ 1). Four patients, all with IH lesions, were admin-

1026

Shimoda et al

istered preoperative chemotherapy for pre-LT malignant diagnoses of mixed cholangiohepatoma (n ⫽ 1), CCC (n ⫽ 1), and HCC (n ⫽ 2). Regimens administered were 5-fluorouracil plus carboplatin, 5-fluorouracil alone, 5-fluorouracil plus doxorubicin, or carboplatin plus gemcitabine. One patient with a mixed cholangiohepatoma had complete response in an EH pancreatic mass and partial response in an IH mass. The other three patients had either stable disease or partial response to pre-LT chemotherapy. The tumor marker CEA was analyzed pre-LT in 10 patients with IH-CCC, and levels were elevated in half those patients. CEA was analyzed pre-LT in all patients with EH-CCC, but only 1 patient had an elevated level. The tumor marker AFP was analyzed pre-LT in all patients with IH-CCC, and levels were elevated in 9 patients. Eight patients with EH-CCC had pre-LT AFP measured; only 1 patient had an elevated level. Finally, the tumor marker CA19-9 was analyzed pre-LT in only 6 patients with IH-CCC and 3 patients with EH-CCC, with abnormal levels in 4 and 1 patients, respectively. Surgical Procedures The surgical procedure was altered in some patients because of the presence of known or suspected malignancy. Intraoperative frozen sections were obtained in 13 patients, with analysis of the bile duct (n ⫽ 9), portal lymph nodes (n ⫽ 4), pancreatic duct (n ⫽ 2), and diaphragm (n ⫽ 1). In 2 patients (IH-CCC and EHCCC) with previously unknown malignancy and PSC, CCC was diagnosed intraoperatively using frozen section. In a third patient with PSC, EH-CCC was suspected during LT because of portal inflammation, but intraoperative biopsies were nondiagnostic. Thus, surgery identified CCC in another 3 patients, leaving 9 patients (36%) with CCC undiagnosed through preoperative workup or surgical exploration. Interestingly, eight of these incidental lesions (88.9%) were EH-CCC. Contiguous organ involvement in the malignant process was noted intraoperatively in five patients and included invasion of the duodenum and right adrenal gland (n ⫽ 1), tumor thrombus in the IVC (n ⫽ 1), periportal and pancreatic lymph node metastases (n ⫽ 1), and invasion into the right hemidiaphragm (n ⫽ 2). To obtain disease-free surgical margins, extended resections were performed that included an en bloc hepatectomy plus adrenalectomy and gastric antrectomy (n ⫽ 1), IVC and right atrial thrombectomy (n ⫽ 1), radical pancreaticoduodenectomy (n ⫽ 1), and diaphragmatic resection (n ⫽ 2). In three patients, radical portal

lymphadenectomy and LT were undertaken because of suspected IH or proximal hilar malignancy. A radical EH bile duct resection was combined with LT in one patient with intraoperatively diagnosed EH-CCC. Two patients underwent pancreaticoduodenectomy at the time of LT for suspected benign conditions (pancreatitis and biliary papillomatosis). Two patients returned to the operating room for radical pancreaticoduodenectomy after LT when explant pathological examination showed EH-CCC with positive margins. The remaining 12 patients required no additional procedures or modifications of LT to excise the tumors. Complications There were no operative deaths in this series. Re-LT was required in four patients (16%) because of bile duct necrosis (n ⫽ 1), primary nonfunction (n ⫽ 1), and hepatic artery thrombosis (n ⫽ 2). The average time to re-LT was 48 days (range, 2 to 139 days). The 30-day mortality rate was 12% as a result of poor graft function and/or sepsis. Other major technical complications included bile leak (n ⫽ 5) and pancreaticojejunostomy leak (n ⫽ 5). Histopathologic Examination and Staging Analysis of explant pathological characteristics showed CCC (n ⫽ 17) or mixed cholangiohepatoma (n ⫽ 8).34 One patient had a single focus of CCC, as well as two separate and distinct foci of HCC. Sixteen patients had CCC arising within the hepatic parenchyma (IH-CCC). Associated hepatic pathological states included cirrhosis secondary to hepatitis B (n ⫽ 1), hepatitis C (n ⫽ 1), combined hepatitis B and hepatitis C (n ⫽ 1), PSC (n ⫽ 2), alcohol (n ⫽ 2), and chronic biliary obstruction (n ⫽ 1). Explant histopathologic examination in patients with IH-CCC without cirrhosis showed either normal underlying hepatic parenchyma (n ⫽ 4), hepatic venous outflow obstruction (n ⫽ 1), biliary outflow obstruction (n ⫽ 1), or steatosis (n ⫽ 2). The average size of the largest nodule was 5.5 ⫾ 3.1 cm (n ⫽ 14). Twelve patients had multifocal (n ⬎ 2) lesions. Vascular invasion was present in three patients, neural invasion was present in three patients, and two patients had lymph node involvement on explant histopathologic analysis. pTNM stages were I (n ⫽ 2), II (n ⫽ 2), III (n ⫽ 3), IVa (n ⫽ 7), and IVb (n ⫽ 2). Nine patients had CCC localized to the EH bile ducts (EH-CCC): five patients had proximal (Klatskintype), two patients had the middle one third, and two patients had the distal one third lesions. Associated biliary pathological states included PSC (n ⫽ 5) and

Liver Transplantation for Cholangiocarcinoma

biliary papillomatosis (n ⫽ 1). Secondary biliary cirrhosis was present in four patients. The average size of the largest lesion was 2.3 ⫾ 0.8 cm (n ⫽ 5). A single focus of EH-CCC was seen in eight patients. Two patients had vascular invasion, four patients had neural invasion, and four patients had hepatic invasion/metastasis. pTNM stages were I (n ⫽ 1), II (n ⫽ 4), III (n ⫽ 0), and IVa (n ⫽ 4). Adjuvant Postoperative Chemotherapy Patients were selectively considered for adjuvant chemotherapy based on the final histopathologic tumor characteristics and clinical recovery after LT. Adjuvant chemotherapy was administered to nine patients. The diagnostic staging treated in this manner was IH-CCC stages II (n ⫽ 1), III (n ⫽ 2), and IVa (n ⫽ 4) and EH-CCC stage II (n ⫽ 2). Adjuvant therapy was administered primarily because of the advanced nature of the tumor (n ⫽ 4), large size of the primary tumor (n ⫽ 1), presence of vascular invasion (n ⫽ 2), or responsiveness to preoperative therapy (n ⫽ 1). The protocols administered included 5-fluorouracil alone (n ⫽ 1), 5-fluorouracil plus mitomycin C (n ⫽ 3), 5-fluorouracil plus doxorubicin and carboplatin/cisplatin (n ⫽ 2), carboplatin plus gemcitabine with or without 5-fluorouracil (n ⫽ 2), and unspecified regimen (n ⫽ 1). Major side effects from the protocols included nausea, cytopenia, or stomatitis. Patient Survival Actuarial patient survival at 1 and 3 years after LT for all patients with CCC (n ⫽ 25) was 71% and 35%, respec-

Figure 1. Actuarial patient survival calculated using the Kaplan-Meier method at 1 and 3 years after LT for CCC. Survival is shown for the total group.

1027

tively (Fig. 1). Actuarial patient survival at 1 and 3 years after LT for those patients with IH-CCC (n ⫽ 16) was 62% and 39%, respectively (Fig. 1). Actuarial patient survival at 1 and 3 years after LT for those patients with EH-CCC (n ⫽ 9) was 86% and 31%, respectively (Fig. 1). Recurrent CCC was the major cause of death after LT (40%). Sepsis was the primary nonmalignant cause of death (16%). Three patients died within 30 days of LT and were eliminated from further tumor-related analysis because follow-up was too short for meaningful oncological conclusions. Disease-Free Survival Death-censored actuarial disease-free survival at 1 and 3 years after LT for all patients with CCC (n ⫽ 22) was 67% and 42%, respectively (Fig. 2). Death-censored actuarial disease-free survival at 1 and 3 years after LT for patients with IH-CCC (n ⫽ 13) was 70% and 35%, respectively (Fig. 2). Death-censored actuarial diseasefree survival at 1 and 3 years after LT for patients with EH-CCC (n ⫽ 9) was 57% and 57%, respectively (Fig. 2). Recurrence rates for patients with CCC, IH-CCC only, or EH-CCC only were 41%, 53.8%, and 22%, respectively. The median time to recurrence was 10.6 months (range, 4.6 to 32.8 months). Tumor recurrence was diagnosed within 2 years after LT in 77.8% of patients. The liver was the most common site of first recurrence (55.5%). Recurrence also was seen in bone, lymph nodes, lung, and abdominal wall. Palliative therapies for recurrent tumors were no therapy (n ⫽ 1),

1028

Shimoda et al

Figure 2. Actuarial diseasefree survival patient survival calculated using the KaplanMeier method at 1 and 3 years after LT for CCC. Patients who died within 30 days of LT were eliminated. Survival is shown for the total group.

surgical resection (n ⫽ 1), chemotherapy (n ⫽ 3), and radiation therapy (n ⫽ 1). Median survival after the diagnosis of recurrence was 9.2 months (range, 1.8 to 20.7 months). Univariate Statistical Analysis Results of univariate analysis of variables are shown for patients with CCC separated into 1- and 3-year patient and disease-free survival (Table 3). Only variables that significantly impacted on survival are shown, and because of small numbers, only the analysis for the total group is shown. When present, the variable “other organ invasion” negatively influenced both patient and disease-free survival at 1 and 3 years (P ⬍ .05). Using 3 cm as the cutoff, the variable “tumor size” also influenced survival. One- and 3-year disease-free survival was significantly improved for all patients with small lesions compared with those with tumors greater than 3 cm or unknown size. The unknown group represents six patients with either diffuse IH-CCC or EH-CCC with extensive periportal inflammation; both conditions rendered accurate measurement of tumor diameter impossible. In this case, the unknown group behaved more like the larger-sized tumor group. The variable “number of nodules” also affected outcome. Single lesions showed better 1-year survival than multifocal lesions. This difference reached statistical significance at 1-year patient survival only. No significant difference in patient survival at 3 years or disease-free survival at 1 and 3 years was seen, indicating the unreliable nature of this variable alone to predict outcome.

Literature Review Results from the literature search are shown separated into large single-center series (Table 4)10,17-28 and pooled registry data (Table 5).35-37 Data from this study are included in Table 4. Additional experience with LT for CCC was obtained by reviewing the literature related to LT for PSC.27,38-42 This series represents the fourth largest single-center series published on the use of LT for the treatment of unresectable CCC. In the combined single-center experience, the 3-year patient survival rate is between 10% and 53%. Improved 3-year patient survival rates are seen in the two smaller (⬎10 and ⬍20 patients) series (53% and 41%) compared with the three larger (⬎20 patients) series (10%, 21.4%, and 29%). Only three published series calculated disease-free survival after LT for CCC. Three-year disease-free survival rates ranged from 13% to 53%. Recurrence rates were reported to be between 27% and 87%. Results from two large studies representing pooled registry data for 189 patients undergoing LT for unresectable CCC are listed in Table 5. The 3-year patient survival rate was 25% or less, whereas the censored disease-free survival rate in one study was 50%.

Discussion In the era of organ shortages, LT for hepatobiliary malignancies remains controversial, mostly because of reports citing poor outcome and high recurrence rates. In this retrospective study, we analyzed outcomes of 25

1029

Liver Transplantation for Cholangiocarcinoma

Table 3. Univariate Statistical Analysis

Total CCC Other organ invasion Present Absent Tumor size ⬎3 cm Unknown ⱕ3 cm Nodule Multiple Single

Survival (%)

No. of Patients

1 Yr

P

3 Yr

P

5 20

20 84

.0005

0 46

11 6 8

61 67 88

.498 .396

13 12

44 91

.009

Disease-Free (%)

No. of Patients

1 Yr

P

3 Yr

P

.0002

2 20

0 75

⬍.0001

0 47

⬍.0001

34 17 58

.302 .072

10 5 7

63 75 100

.056 .019

34 0 100

.031 .006

35 33

.168

10 12

58 73

.738

39 48

.786

patients with unresectable IH-CCC or EH-CCC undergoing LT to determine survival, recurrence, and prognostic indicators. This study represents one of the largest single-center series of its kind. From results of this study, several important conclusions can be drawn that potentially impact on the decision for LT in a patient with CCC. The first point is the definition and outcome of incidental CCC. This study does not and was not

intended to evaluate the incidence of incidental CCC. To do so would require that the total number of patients undergoing LT for diagnoses of PSC, hepatitis C virus, hepatitis B virus, and alcohol be evaluated. Therefore, 9 of 25 patients with tumors not diagnosed before LT represent only an observation in this group of patients with CCC, not a true prevalence of incidental CCC. Compounding this problem, the definition of this term varies widely in the CCC literature. Because

Table 4. Literature Review: Single-Center Data

Institution Kings College Hospital, London, UK Niguarda Hospital, Milan, Italy Hochschule Hannover, Hannover, Germany University of Pittsburgh

Baylor Medical Center, Dallas, TX Mayo Clinic, Rochester, MN University of California at Los Angeles, CA

Patient Survival (%)

Tumor-Free Survival (%)

Location

1-Yr

3-Yr

5-Yr

1-Yr

3-Yr

5-Yr

38 30 89 — — 13.9 20.8 60 24 50 72 70 59.3 54.6 53

10 10 53 — — — 0 21.4 24 0 — 29 36.2 9.1 41

10 10 53 — — — — 17.1 0 — 26.7 18 36.2 9.1 —

— — 67 — — — — —

— — 53 — — — — —

— — 53 — — — — —

— — 67 — — 40

— — 31 — — 13

— — 31 — — —

— — —

67 70 57

42 35 57

— — —

Reference

Year

No. of Patients

17

1987

18

1994

19 20 10 26

1995 1995 1996 1990

21 22 23

1993 1997 1998

24, 25

1998

13 13 11 5 6 18 22 25 19 2 18 20 27 11 14

IH EH Total IH EH IH IH EH EH IH Total IH EH EH Total

28

2000

11

Total

100

Present study

2000

25 16 9

Total IH EH

71 62 86

35 39 31

1030

Shimoda et al

Table 5. Literature Review: Registry Data Tumor-Free Survival (%)

Patient Survival (%) Institution European Transplant Registry

Transplant Tumor Registry

Reference

Year

No. of Patients

35

1988

80 42 38 109 207

36 37

1991 2000

several centers reported that patients with incidental CCC fare better than those with known CCC,21,43 it is important to evaluate the definition of incidental tumor in each study. Because most incidental lesions are solitary and small, the more accurate statement is that patients with early-stage CCC lesions do better than those with more advanced lesions. Our data corroborate these findings. Furthermore, this investigation highlights the difficulty obtaining a definitive preoperative diagnosis of CCC. The majority of our patients with undiagnosed CCC had EH-CCC associated with PSC, and all had undergone extensive preoperative screening, including cholangiograms, biopsies, or brushings. Although the diagnostic accuracy of such modalities as ultrasound, CT, and magnetic resonance imaging is high for IH lesions, these modalities are far less accurate for EH ductal lesions.44,45 Consequently, most centers, including our own, rely on cholangiograms, endoscopic duct biopsies and brushings, and intraoperative frozen sections to diagnose CCC. These latter modalities are notoriously inaccurate in the diagnosis of tumors of bile duct origin.21 Although the use of such tumor markers as CA19-9 can potentially increase the accuracy of the diagnostic evaluation, we did not systematically use these markers in many of the earlier patients and thus cannot make definitive conclusions. Therefore, the limitations associated with state-of-the-art imaging and evaluation will inevitably lead to continued transplantation in patients with undiagnosed CCC. Despite these limitations, we still advocate aggressive and repetitive monitoring of high-risk patients for the development of occult CCC. The association between PSC and the development of CCC is well known. Most series showed excellent outcomes for patients with uncomplicated PSC or those with PSC complicated by incidental CCC.21,27,38,39,41,42 However, close scrutiny of results from the larger series shows 5-year survival rates of only 0% to 47% and

Location

1-Yr

3-Yr

5-Yr

IH EH

26 40 50 72

9 (2-yr) 16 (2-yr) 25 —

2 0 17 23

1-Yr

3-Yr

5-Yr

55

50

30

recurrence rates between 28% and 100%. The only two series with particularly favorable results after LT for incidental CCC are our own and that recently published from the Mayo Clinic. In the former, patients with incidental CCC found after LT for PSC had an 83% 5-year survival rate. Similarly, the Mayo Clinic study showed excellent outcome, but stringent preoperative selection criteria, as well as adjuvant chemoirradiation, were applied. Our current data with longer follow-up do not support our earlier findings, and we believe tumor stage is a more accurate predictor of outcome. Several investigators19,36 reported that patients with IH-CCC fared worse than those with EH-CCC. This remains controversial because others reported equivalent outcomes between patients with IH-CCC and EHCCC.16,17,22 This investigation found no significant difference in patient or disease-free survival between lesions arising from IH or EH bile ducts despite relative equivalency of tumor stage in both groups. Thus, we do not use tumor location as a selection determinant when considering LT for patients with CCC. The use of extended resections combined with LT to achieve negative tumor margins is addressed in this investigation and must be clarified. Two surgical scenarios are involved when describing the extended resection. Tumors may be confined to the bile duct, but require resection of another organ to obtain negative proximal or distal margins. A common example of this situation is resection of the head of the pancreas and duodenum for a mid or distal bile duct tumor. A second scenario is one in which the bile duct tumor egresses from the duct to invade contiguous organs, such as the stomach, colon, liver, or duodenum. We believe these two situations portend different prognoses and therefore should be discussed separately. In the former (tumor confined to the bile duct, but requiring wider resection for proximal or distal margins), our data predict a more favorable prognosis. Four

Liver Transplantation for Cholangiocarcinoma

of our patients with EH-CCC underwent more extended resections to achieve negative margins (pancreaticoduodenectomy or radical bile duct resection). To date, there has been no tumor recurrence in this subset with a median follow-up of 39.5 months. Several other groups discussed the use of more radical procedures with mixed outcomes.10,46-49 Studies assessing the impact of positive surgical margins on outcomes after LT for CCC indirectly support our observations in that outcomes are uniformly poor for patients with CCC undergoing LT with positive resection margins.22,23 Obtaining disease-free surgical margins appears to be of paramount importance when evaluating potential candidates for LT. However, the use of extended resections in immunosuppressed patients can be associated with significant complication rates, as seen in this series. Conversely, patients with tumor invading beyond the bile duct or liver and into contiguous organs did not fare well in this series. This group represents patients statistically analyzed as “other organ invasion.” This scenario consistently portended poor patient survival despite the use of radical intraoperative extirpations. Other groups have applied more radical resections and cluster transplantation to patients with extensive CCC undergoing LT.23,50,51 Outcomes are not good; therefore, patients with CCC extending into other organs probably should not be offered LT. Other variables, such as tumor size, stage, number of nodules, presence of vascular invasion, neural invasion, or lymph node involvement, variably have been reported to predict poor outcome.10,18,22,23 In this study, analysis of these variables corroborated some of these findings (Table 3); specifically, tumor size greater than 3 cm and presence of multiple tumor foci. However, these variables, as well as others analyzed, were not consistent. This inconsistency is largely caused by the relatively small number of patients on this study and high incidence of “unknown” for each variable analyzed. These findings highlight limitations built into retrospective studies. Furthermore, because so few patients undergo LT for CCC at each individual transplant center, the analysis is potentially restricted by patient numbers. Although limited for other reasons, pooled or registry data can help identify high-risk variables by using large numbers of patients. The largest and most recently analyzed registry study37 also did not support tumor size or number of nodules as variables predictive of outcome. Mixed hepato-CCC is a rare lesion that contains features characteristic of both lesions.34 In this investigation, eight tumors fulfilled these histopathologic criteria. Outcomes of this subgroup did not differ signifi-

1031

cantly from patients with pure CCC (data not shown). Pichlmayer et al19 described 12 patients with a “mixedtype hepato-CCC” who underwent either resection or LT with similar outcomes. The Cincinnati Transplant Tumor Registry database identified 12 patients with mixed hepato-CCC and, again, found no difference in survival. Therefore, this lesion does not appear to alter prognosis after LT. The role of adjuvant therapy in the treatment of patients with CCC undergoing LT is controversial, but potentially promising. In this study, application of these treatment modalities was not uniform and was dictated by such variables as tumor histological characteristics, as well as patient response to pre-LT therapy and performance status after LT. Mainly high-risk patients with good recovery were administered adjuvant chemotherapy. Although statistical analysis did not show a survival advantage in this group of patients, there was a trend toward better patient and disease-free survival. Studies designed with strict inclusion criteria shed more light on the efficacy of adjuvant therapy in this setting. Two such investigations using different regimens of adjuvant and neoadjuvant chemoirradiation have been published to date.24,28 In the earlier study, Goldstein et al24 from Baylor University Medical Center administered an adjuvant chemoirradiation protocol consisting of regional radiotherapy and systemic 5-fluorouracil to 14 patients with CCC undergoing LT. Although toxicities were limited, poor disease-free survival led the investigators to conclude that this regimen offered little benefit. Conversely, De Vreede et al28 from the Mayo Clinic developed a protocol in which patients with CCC were selected to undergo sequential pre-LT external-beam irradiation, transcatheter irradiation, and systemic 5-fluorouracil infusion followed by LT. Eight patients, mostly with early-stage CCC, were studied, with more than 1 year of follow-up. Excellent survival with only one recurrence was reported, indicating the potential use of this regimen in highly selected patients with CCC. In conclusion, this study shows that 3-year diseasefree survival (death censored) can be obtained in approximately 40% of patients with CCC undergoing LT. Although these outcomes represent significant improvements over palliative cancer therapies, they clearly are inferior to outcomes after LT for most nonmalignant indications. Unfortunately, the question about application of a limited donor-organ pool to the treatment of patients with CCC cannot be answered fully by the current study or those reviewed. However, results clearly point to difficulties encountered with the

1032

Shimoda et al

identification of patients with CCC before LT, thus predicting continued use of this modality in the treatment of CCC, even if inadvertent. Furthermore, several subsets of patients can be selected that may benefit from LT for CCC. Based on these results, we offer the following recommendations: (1) aggressive and repeated attempts to accurately diagnose and stage patients awaiting LT who are at risk for CCC; (2) selective application of LT for the treatment of early-stage CCC on a case-by-case basis; (3) wholesale exclusion of patients with advanced CCC invading contiguous organs, even when radical surgical procedures enable disease-free margins; and (4) aggressive surgical resection to obtain disease-free margins for patients with CCC limited to the bile duct without invasion into extrabiliary organs.

References 1. Yeo CJ, Pitt HA, Cameron JL. Cholangiocarcinoma. Surg Clin North Am 1990;70:1429-1447. 2. Rossi RL, Heiss FW, Beckmann CF, Braasch JW. Management of cancer of the bile duct. Surg Clin North Am 1985;65:59-78. 3. Wanebo HJ, Falkson G, Order SE. Cancer of the hepatobiliary system. In: DeVita VT, Hellman S, Rosenberg SA (eds). Cancer, principles and practice of oncology (ed 3). Philadelphia, Lippincott, 1989:856-874. 4. Altemeier WA, Gall EA, Zinninger MM, Hoxworth PI. Sclerosing carcinoma of the major intrahepatic bile ducts. Arch Surg 1957;75:450-461. 5. Klatskin G. Adenocarcinoma of the hepatic duct at its bifurcation within the porta hepatis: An unusual tumor with distinctive clinical and pathologic features. Am J Med 1965;38:241-256. 6. Longmire WP Jr, McArthur MS, Bastounis EA, Hiatt J. Carcinoma of the extrahepatic biliary tract. Ann Surg 1973;178:333345. 7. Busuttil RW, Farmer DG. The surgical treatment of primary hepatobiliary malignancy. Liver Transpl Surg 1996;2(suppl 1):S1-S18. 8. Washburn WK, Lewis WD, Jenkins RL. Aggressive surgical resection for cholangiocarcinoma. Arch Surg 1995;130:270-276. 9. Burke EC, Jarnagin WR, Hochwald SN, Pisters PWT, Fong Y, Blumgart LH. Hilar cholangiocarcinoma: Patterns of spread, the importance of hepatic resection for curative operation, and a presurgical clinical staging system. Ann Surg 1998;228:385-394. 10. Pichlmayr R, Weimann A, Klempnauer J, Oldhafer KJ, Maschek H, Tusch G, Ringe B. Surgical treatment in proximal bile duct cancer: A single-center experience. Ann Surg 1996;224:628-638. 11. Tsao JI, Nimura Y, Kamiya J, Hayakawa N, Kondo S, Nagino M, et al. Management of hilar cholangiocarcinoma: Comparison of an American and a Japanese experience. Ann Surg 2000;232: 166-174. 12. Fortner JG, Vitelli CE, Maclean BJ. Proximal extrahepatic bile duct tumors: Analysis of a series of 52 consecutive patients treated over a period of 13 years. Arch Surg 1989;124:12751279. 13. Lai ECS, Tompkins RK, Roslyn JJ, Mann LL. Proximal bile duct cancer: Quality of survival. Ann Surg 1987;205:111-118.

14. Bismuth H, Nakache R, Diamond T. Management strategies in resection for hilar cholangiocarcinoma. Ann Surg 1992;215:3138. 15. Stain SC, Baer HU, Dennison AR, Blumgart LH. Current management of hilar cholangiocarcinoma. Surg Gynecol Obstetr 1992;175:579-588. 16. Nakeeb A, Pitt HA, Sohn TA, Coleman J, Abrams RA, Piantadosi S, et al. Cholangiocarcinoma: A spectrum of intrahepatic, perihilar, and distal tumors. Ann Surg 1996;224:463-475. 17. O’Grady JG, Polson RJ, Rolles K, Calne RY, Williams R. Liver transplantation for malignant disease. Ann Surg 1988;207:373379. 18. Sansalone CV, Colella G, Caccamo L, Rondinara GF, Fassati LR, Rossetti O, et al. Orthotopic liver transplantation for primary biliary tumors: Milan multicenter experience. Transplant Proc 1994;26:3561-3563. 19. Pichlmayr R, Lamesch P, Weimann A, Tusch G, Ringe B. Surgical treatment of cholangiocellular carcinoma. World J Surg 1995;19:83-88. 20. Pichlmayr R, Weimann A, Oldhafer KJ, Schlitt HJ, Klempnauer J, Bornscheuer A, et al. Role of liver transplantation in the treatment of unresectable liver cancer. World J Surg 1995;19:807813. 21. Abu-Elmagd KM, Selby R, Iwatsuki S, Fung J, Tzakis A, Todo S, et al. Cholangiocarcinoma and sclerosing cholangitis: Clinical characteristics and effect on survival after liver transplantation. Transplant Proc 1993;25:1124-1125. 22. Casavilla FA, Marsh JW, Iwatsuki S, Todo S, Lee RG, Madariaga JR, et al. Hepatic resection and transplantation for peripheral cholangiocarcinoma. J Am Coll Surg 1997;185:429-436. 23. Iwatsuki S, Todo S, Marsh JW, Madariaga JR, Lee RG, Dvorchik I, et al. Treatment of hilar cholangiocarcinoma (Klatskin tumors) with hepatic resection or transplantation. J Am Coll Surg 1998;187:358-364. 24. Goldstein RM, Stone M, Tillery GW, Senzer N, Levy M, Husberg BS, et al. Is liver transplantation indicated for cholangiocarcinoma? Am J Surg 1993;166:768-772. 25. Jeyarajah DR, Klintmalm GB. Is liver transplantation indicated for cholangiocarcinoma? J Hepatobiliary Pancreat Surg 1998;5: 48-51. 26. Yokoyama I, Todo S, Iwatsuki S, Starzl TE. Liver transplantation in the treatment of primary liver cancer. Hepatogastroenterology 1990;37:188-193. 27. Marsh JW Jr, Iwatsuki S, Makowka L, Esquivel CO, Gordon RD, Todo S, et al. Orthotopic liver transplantation for primary sclerosing cholangitis. Ann Surg 1988;207:21-25. 28. De Vreede I, Steers JL, Burch PA, Rosen CB, Gunderson LL, Haddock MG, et al. Prolonged disease-free survival after orthotopic liver transplantation plus adjuvant chemoirradiation for cholangiocarcinoma. Liver Transpl 2000;6:309-316. 29. Busuttil RW, Shaked A, Millis JM, Jurim O, Colquhoun SD, Shackleton CR, et al. One thousand liver transplants—The lessons learned. Ann Surg 1994;219:490-499. 30. Busuttil RW, Colonna JO II, Hiatt JR, Brems JJ, El Khoury G, Goldstein LI, et al. The first 100 liver transplantation at UCLA. Ann Surg 1987;206:387-402. 31. Demetris AJ, Batts KP, Dhillon AP, Ferrell L, Fung JJ, Geller SA, et al. Banff schema for grading liver allograft rejection. An international consensus document. Hepatology 1997;25:658-663. 32. Liver (including intrahepatic bile ducts). In: Bearhs OH, Henson DE, Hutter RVP, et al (eds). American Joint Committee on

Liver Transplantation for Cholangiocarcinoma

33.

34.

35.

36. 37.

38.

39.

40.

41.

42.

Cancer. Manual for staging of cancer (ed 4). Philadelphia: Lippincott, 1992:89-90. Extrahepatic bile ducts. In: Bearhs OH, Henson DE, Hutter RVP, et al (eds). American Joint Committee on Cancer. Manual for staging of cancer (ed 4). Philadelphia: Lippincott, 1992:99101. Maeda T, Adachi E, Kajiyama K, Sugimachi K, Ysuneyoshi M. Combined hepatocellular and cholangiocarcinoma: Proposed criteria according to cytokeratin expression and analysis of clinicopathologic features. Hum Pathol 1995;26:956-964. Bismuth H, Castaing D, Ericzon BG, Otte JB, Rolles K, Ringe B, et al. Hepatic transplantation in Europe. First report of the European Liver Transplant Registry. Lancet 1987;2:674-676. Penn I. Hepatic transplantation for primary and metastatic cancers of the liver. Surgery 1991;110:726-735. Meyer CG, Penn I, James L. Liver transplantation for cholangiocarcinoma: Results in 207 patients. Transplantation 2000;69: 1633-1637. Goss JA, Shackleton CR, Farmer DG, Arnaout WS, Seu P, Markowitz JS, et al. Orthotopic liver transplantation for primary sclerosing cholangitis. A 12-year, single-center experience. Ann Surg 1997;225:472-481. Abu-Elmagd KM, Malinchoc M, Dickson ER, Fung JJ, Murtaugh PA, Langworthy AL, et al. Efficacy of liver transplantation in patients with primary sclerosing cholangitis. Surg Gynecol Obstet 1993;177:335-344. Narumi S, Roberts JP, Emond JC, Lake J, Ascher NL. Liver transplantation for sclerosing cholangitis. Hepatology 1995;22: 451-457. Nashan B, Schlitt HJ, Tusch G, Oldhafer KJ, Ringe B, Wagner S, et al. Biliary malignancy in primary sclerosing cholangitis: Timing for liver transplantation. Hepatology 1996;23:11051111. Graziadei IW, Wiesner RH, Marotta PJ, Porayko MK, Hay JE,

43.

44. 45.

46.

47.

48.

49.

50.

51.

1033

Charlton MR, et al. Long-term results of patients undergoing liver transplantation for primary sclerosing cholangitis, Hepatology 1999;30:1121-1127. Loinaz C, Abradelo M, Gomez R, Colina F, Rey P, Ochando F, et al. Liver transplantation and incidental primary liver tumors. Transplant Proc 1998;30:3301-3302. Mahfouz AE, Vogl T, Hamm B. Tumor diagnosis in the adult liver transplant candidate. Eur Radiol 1999;9:841-852. Weisner RH, Porayko MK, Hay JE, LaRusso NF, Steers JL, Krom RAF, et al. Liver transplantation for primary sclerosing cholangitis: Impact of risk factors on outcome. Liver Transpl Surg 1996;2:99-108. Farmer DG, Shaked A, Colonna JO II, Olthoff KM, Jurim O, Colquhoun SD, et al. Radical resection combined with liver transplantation for Foregut tumors. Am Surg 1993;59:806-812. Neuhaus P, Blumhardt G. Extended bile duct resection—A new oncological approach to the treatment of central bile duct carcinomas? Description of method and early results. Langenbecks Arch Surg 1994;379:123-128. Cherqui D, Alon R, Piedbois P, Duvoux C, Dhumeaux D, Julien M, et al. Combined liver transplantation and pancreatoduodenectomy for irresectable hilar bile duct carcinoma. Br J Surg 1995;82:397-398. Tsukada K, Toshida K, Aono T, Koyama S, Shirai Y, Uchida K, et al. Major hepatectomy and pancreatoduodenectomy for advanced carcinoma of the biliary tract. Br J Surg 1994;81:108110. Starzl TE, Todo S, Tzakis A, Podesta L, Mieles L, Demetris A, et al. Abdominal organ cluster transplantation for the treatment of upper abdominal malignancies. Ann Surg 1989;210:374-386. Alessani M, Tzakis A, Todo S, Demetris AJ, Fung JJ, Starzl TE. Assessment of 5-year experience with abdominal organ cluster transplantation. J Am Coll Surg 1995;180:1-9.