Extrahepatic bile duct cancer is further classified as hilar or distal. Hilar cholangiocarcinoma, also called a Klatskin tumor[6], is located within 2 cm of the bifurcation of the common bile duct and is further divided into four types by Bismuth and Corlette based on the anatomic location of the tumor[1,7]. Approximately 60% to 70% of cholangiocarcinoma is reported to be located at the hilum, 20% to 30% at the distal bile duct, and 5% to 10% intrahepatic bile duct[1,8,9]. The anatomic schema is presented in Figure 1. This classification, based on the longitudinal location of the bile duct cancer, defines the surgical strategy, including the operability and curability.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Anatomic classification of cholangiocarcinoma. A: The majority of cholangiocarcinoma (60%-70%) develop in the hilar bile duct and are called Klatskin tumors. The distal bile duct is involved in 20% to 30% of cases, while intrahepatic cholangiocarcinomas represent 5% to 10% of the tumors originating from the biliary tract; B: Bismuth-Corlette classification of hilar bile duct cancer. Type I, cholangiocarcinoma confined to the common bile duct; Type II, cholangiocarcinoma involves the bifurcation of the common bile duct; Type IIIa, cholangiocarcinoma involves the bifurcation and the right hepatic duct; Type IIIb, cholangiocarcinoma involves the bifurcation and the left hepatic duct; Type IV, cholangiocarcinoma involves the bifurcation and extends to both the right and left hepatic ducts.

Most cholangiocarcinomas are well-to-moderately differentiated adenocarcinomas with a tendency to develop desmoplastic reactions and early perineural invasion[1,2,10]. Macroscopically, extrahepatic cholangiocarcinoma develops sclerosing strictures, nodular lesions, or papillary growth[11]. The sclerosing type is the most common, while the papillary pattern is rare but associated with a more favorable prognosis[12,13].

The longitudinal spread determines the type of radical operation, including pancreaticoduodenectomy (PD), extended hepatectomy and hepatopancreaticoduodenectomy, and apparent infiltration beyond the secondary branches on both sides of the biliary tree is generally considered unresectable.

Microscopic extension of bile duct cancer beyond the tumor margin visualized by present diagnostic modalities or the margin observed macroscopically is often encountered. Longitudinal extension consists of superficial and submucosal infiltration depending on the tumor growth pattern, and sometimes includes direct infiltration along lymphatic and perineural tissues[13]. Sakamoto et al[17] reported a correlation between the gross tumor type and the pattern of infiltration, and demonstrated that mucosal extension predominantly occurs with papillary (intraductal) and nodular (mass-forming) tumors, while submucosal extension mainly occurs with sclerosing (infiltrating) tumors. Tumor spread beyond the macroscopic margin is determined by the type of invasion, with a mean length of 6-10 mm for submucosal spread and 10-20 mm for mucosal spread[18]. Therefore, a gross surgical margin of more than 1 cm in the infiltrating type and more than 2 cm in the papillary and nodular types is recommended to obtain microscopically negative margins.

The evaluation of vertical infiltration is critical to the patient’s prognosis, because it usually defines resectability and curability. Apparent distant metastasis and para-aortic lymph node metastasis are absolute contraindications for radical surgery[19-21], while direct invasion of major vessels, regardless of the prognosis, present surgical challenges whose indication and operative procedures continue to be debated[22-28].

Patients with cancer invasion confined to the fibromuscular layer of the extrahepatic bile duct have a better postoperative survival rate (80%-100% 5-year survival), while those with cancer that extends beyond the fibromuscular layer have a poor prognosis[29-31].

Perineural invasion beyond the bile duct wall is a unique characteristic of cholangiocarcinoma that is observed in 75% of cases and has proved to be a significant prognostic factor for poor outcome[14,16,32].

Hilar cholangiocarcinoma easily infiltrates into the hepatic parenchyma and the hepatoduodenal ligament, in which the hepatic artery and portal vein are located adjacent to the bile duct, while distal bile duct cancer directly invades the pancreas or duodenum[33,34]. Up to 80% of hilar cholangiocarcinoma directly infiltrates the liver parenchyma and surrounding connective tissues of the hepatoduodenal ligament[34,35], thus necessitating meticulous three-dimensional knowledge of the hepatic hilum and challenging operations for this disease entity[36]. One of the most important points of anatomic consideration for the vertical invasion of hilar cholangiocarcinoma is the need for hepatectomy with complete caudal lobectomy. Hilar cholangiocarcinoma spreads not only longitudinally along the right and left hepatic ducts, but also in the cranial and dorsal directions along the thin bile ducts. Hence, it is necessary to remove the liver parenchyma adjacent to the hepatic hilum together with the hilar plate to achieve a complete curative resection. In this sense, complete caudate lobectomy and resection of the inferior part of Couinaud’s segment IV coupled with right or left hemihepatectomy (according to the predominant tumor location) is the main goal of surgical resection of hilar cholangiocarcinoma[37-41]. Major hepatectomy with caudate lobectomy for hilar cholangiocarcinoma is associated with improved outcome[26,41-45].

The efficacy of preoperative biliary drainage for patients with obstructive jaundice remains controversial. Based on previous reports emphasizing the adverse effects of biliary drainage, such as pancreatitis, cholangitis, and tract seeding[76,86-91], and the recent meta-analysis comparing surgery with preoperative biliary drainage to that without drainage, which showed no beneficial effect of preoperative drainage[77], routine biliary drainage is not recommended for all jaundiced patients, especially for distal bile duct cancer without any complications[92]. Furthermore, preoperative biliary drainage seems to increase the risk of perioperative infections and a longer postoperative hospital stay[76,77]. Preoperative biliary drainage has proved to be beneficial, however, in the presence of cholangitis, severe malnutrition, and coagulation abnormalities[93,94], and it is absolutely indicated for patients requiring major hepatic resection for curative surgery[36,71]. Prolonged preoperative jaundice is associated with increased postoperative morbidity and mortality after hepatic resection due to severe cholestatic liver dysfunction[95,96]. When biliary drainage is performed for patients awaiting major hepatectomy, radical surgery should be postponed for several weeks until the serum total bilirubin is less than 2.0 mg/dL to allow for sufficient restoration of hepatic function[36]. The use of sequential liver volumetric analyses and hepatic functional studies is warranted when anticipating extended hepatectomy to secure a sufficient volume and function of the future remnant liver, thereby minimizing the risk of postoperative liver failure[97].

As discussed above, hilar cholangiocarcinoma usually requires extended hepatectomy (i.e. extended right hepatectomy, and right or left trisegmentectomy), which is related with a rather high rate of perioperative mortality (0% to 19%)[98]. This is partly due to the increased rate of postoperative liver failure with major hepatic resection[99-101]. Portal vein embolization, which was first indicated for hilar cholangiocarcinoma by Makuuchi et al[102], is now widely accepted as a valuable preoperative measure in anticipation of extensive liver resection with a subsequent small liver remnant volume[103-108]. Compensatory hypertrophy of the remnant liver parenchyma, usually an increase of 8% to 20% within 2 to 6 wk, is induced in association with atrophy of the future resected liver by selectively occluding the main portal branch to the liver parenchyma to be removed. In general, portal vein embolization can benefit patients requiring a future liver remnant volume of less than 25% to 35% of the original liver volume, yet the indication is still controversial, especially for patients with normal liver function. Currently, there seems to be no objection to portal vein embolization for potentially resectable patients with normal liver function when the anticipated future liver remnant volume is less than 20% of the total liver volume, or in patients with compromised liver function when the anticipated future liver remnant volume is less than 40% of the total liver volume[98]. Once portal vein embolization is performed, sequential evaluation of liver function and volumetry is mandatory not to miss the optimal timing for radical surgery[109].

Pancreaticoduodenectomy (PD), including pylorus preserving PD (PPPD) coupled with porta hepatis lymphadenectomy, is the standard treatment of choice for the complete removal of distal bile duct cancer, but extended lymphadenectomy including the para-aortic nodes is not justified because it does not provide a survival advantage and is associated with increased perioperative morbidity[20,21,110,111]. Both PD and PPPD provide equal outcome for distal bile duct cancer[20,21,112-114], while segmental bile duct excision is rarely an option because only 10% of patients undergoing bile duct excision alone obtain curative resection margins on final pathology[110,115]. Based on the equivalent outcome between segmental bile duct resection and PD recently reported by Lee et al[116], segmental bile duct resection with excision of surrounding lymph nodes and connective tissues seems to be a possible strategy[117], yet it is not accepted as a standard operation for distal bile duct cancer for obtaining a negative surgical margin[118].

Reports from high-volume centers during the last decade are summarized in Table 1[21,33,116,119-131]. The overall 3- and 5-year survival rates after radical surgery ranged from 33% to 63% and 16% to 52%, respectively. A recent meta-analysis by Japanese Biliary Tract Cancer Statistics Registry revealed that the overall 3- and 5-year survival rates of distal bile duct cancer after radical resection were 58% and 44%, respectively, among 779 patients who received PD or PPPD between 1998 and 2004 in Japan[132,133].

Based on the mode of tumor extension and the radicality and simplicity of the procedure, extended right- or left-hemihepatectomy is regarded as the standard radical operation for hilar cholangiocarcinoma[37-39,134], even for patients with Bismuth and Corlette type I or II[118]. Extended right hemihepatectomy consists of the resection of the right liver, the inferior part of segment IV, the hilar plate, and the entire caudate lobe, while extended left hemihepatectomy consists of resection of the left liver, the hilar plate of the right paramedian sector, and most of the caudate lobe, both coupled with complete resection of the extrahepatic bile duct and porta hepatis lymphadenectomy. “Right” or “left” is dependent on the predominance of the tumor, but an extended right-hemihepatectomy is preferentially indicated for even centrally located tumors, because of the length of each hepatic duct, location of the hilar common bile duct in the hepatoduodenal ligament, facility of complete caudate lobectomy, and the ease of portal vein reconstruction[39,135]. When the tumor spreads diffusely into the intrapancreatic bile duct, PD should be added to extended hemihepatectomy simultaneously[136-138].

Right or left trisegmentectomy is one of the most extensive hepatic resection procedures because of the massive volume loss of the liver parenchyma, and hilar cholangiocarcinoma invading the hepatic hilum sometimes requires trisegmentectomy for curative resection. Trisegmentectomy is advantageous in terms of obtaining a negative margin of the bile ducts[139,140]. A negative ductal margin is achieved in 75% and 88% of cases with left and right trisegmentectomy, respectively, both of which are higher than the negative rate obtained with extended hemihepatectomy[139,140]. On the other hand, minimally invasive procedures, including parenchyma-preserving hepatectomy and bile duct segmental resection without hepatectomy, could be an option for Bismuth-Corlette type I and II or cancer of the middle bile duct, but in terms of obtaining R0, those procedures are indicated for strictly localized tumors or for patients with a poor general condition or high-risk factors[36,41,116,117]. Many authors report improved survival with complete caudate lobectomy and major hepatectomy[26,41-45].

Metastasis to regional lymph nodes is one of the most important prognostic factors influencing survival after resection for hilar cholangiocarcinoma[45,141]. Patients with nodal involvement beyond the hepatoduodenal ligament, including para-aortic nodal metastases, were shown to have dismal prognosis with a 5-year survival of 0% to 12%[141-144]. Therefore, routine lymph node dissection beyond the hepatoduodenal ligament is not recommended. On the other hand, portal vein resection and reconstruction has been performed for hilar cholangiocarcinoma with conflicting results[145]. With recent technical advances, several retrospective series have shown no difference in surgical mortality between patients undergoing major hepatectomy with portal vein resection and without it, advocating routine resection of the portal vein for en-bloc, “no-touch” resection[25,26], but the impact of portal vein resection on long-term survival seems to be less clear. Moreover, other studies have shown equivalent or worse survival in patients undergoing portal vein resection[23,27,28]. When there is severe adhesion between the tumor and portal vein, combined resection and reconstruction is necessary to obtain a possible negative surgical margin, yet routine resection of the portal vein might not be recommended unless supported by findings from a randomized clinical trial.

Reports from high-volume centers during the last decade are summarized in Table 2[26,45,71,73,128,135,142,146-160]. The overall 3- and 5-year survival rates after radical surgery ranged from 37% to 60% and 20% to 42%, respectively. A recent meta-analysis by the Japanese Biliary Tract Cancer Statistics Registry revealed that the overall 3- and 5-year survival rates of hilar bile duct cancer after radical resection were 47% and 39%, respectively, among 255 patients who received major hepatectomy during 1998 and 2004 in Japan[132,133].

For the treatment of locally advanced hilar cholangiocarcinoma beyond the indication for resection, orthotopic liver transplantation (OLT) may offer the advantage of resection of all structures involved by the tumor, including vessels within the hepatoduodenal ligament, all intra- and extrahepatic bile ducts, and whole liver parenchyma. Patients requiring a total hepatectomy to achieve a negative margin and those with underlying liver failure precluding hepatic resection are possible candidates for OLT, but early experience with OLT for hilar cholangiocarcinoma was disappointing with early recurrence rates of more than 50% and a 5-year survival rate of 10% to 20%[161-164]. Recently, in highly selected patients undergoing neoadjuvant protocols, improved survivals were reported with a 5-year survival rate of 30% to 45%[165-167]. More recently, the so-called “Mayo protocol” was reported with the intent of treating a highly selected group of patients with cholangiocarcinoma with a strict regimen of preoperative staging and neoadjuvant chemoradiation followed by OLT[168-170]. The inclusion criteria of this protocol are as follows, (1) locally advanced unresectable disease with positive intraluminal cytology or biopsy, or CA19-9 >100 with radiographic features of malignancy; (2): primary sclerosing cholangitis with resectable bile duct cancer; and (3): absence of medical contraindication for OLT. Eligible patients receive neoadjuvant chemoradiation therapy comprising external radiation of 4500 cGy with concomitant fluorouracil, and transcatheter Iridium-192 brachytherapy of 2000-3000 cGy, followed by oral capecitabine as tolerated until OLT. Additionally, patients should undergo a staging laparotomy to rule out metastatic disease prior to OLT. In 38 patients who underwent this protocol, an 82% 5-year disease free survival was reported. Currently, the indication for OLT for the treatment of cholangiocarcinoma is reserved for highly selected patients in specialized centers.

A previous randomized trial revealed that chemotherapy significantly improved survival and quality of life compared to best supportive care for unresectable cholangiocarcinoma[177]. The most extensively studied agents in unresectable cholangiocarcinoma are fluorouracil (5-FU) and gemcitabine, which have been investigated as a single agent[177-179] and in combination with other drugs, such as mitomycin C[178], leucovorin[180], cisplatin[181], capecitabine[182], epirubicin[183], and oxaliplatin[184]. Eckel et al[185] conducted a pooled analysis of 104 chemotherapy studies in advanced bile duct cancers, which suggested that gemcitabine combined with cisplatin or oxaliplatin resulted in the best response without significantly improved survival. On the other hand, reports on adjuvant chemotherapy after resection are scarce (Table 3). A recent multicenter randomized trial evaluated the effect of adjuvant chemotherapy with mitomycin C and 5-FU versus surgery alone for patients with pancreato-biliary malignancies, in which no survival benefit was seen for 139 patients with R0 resection for cholangiocarcinoma[186]. Recent institutional retrospective experiences found that gemcitabine-based adjuvant chemotherapy after curative-intent surgery significantly improved patient survival[160,187]. In summary, gemcitabine in combination with cisplatin or oxaliplatin seems to be the most efficacious regimen in cholangiocarcinoma, but adjuvant chemotherapy alone cannot yet be considered standard therapy after resection.

Several studies have described adjuvant external beam radiotherapy, with or without dose escalation by intraluminal brachytherapy (Table 3). One prospective study and some retrospective studies found no survival benefit of adjuvant radiotherapy in patients who received a curative intent resection for cholangiocarcinoma[188-190]. In contrast, several large retrospective series suggested a survival benefit with adjuvant radiation. Todoroki et al[191] demonstrated a significantly higher 5-year survival of 34% in patients with R1 resections with adjuvant radiotherapy (intraoperative and extra beam) compared to 14% with surgery alone. Similarly, Cheng et al[192] and Schoenthaler et al[193] reported the efficacy of adjuvant radiation for the survival of patients with R1 or R2 resections. Other authors have corroborated the improved survival with adjuvant radiotherapy[194-197]. Although the bulk of retrospective data suggest that improved survival may be achieved with the use of adjuvant radiation, one prospective study was negative for survival benefit in a selected group of patients with hilar cholangiocarcinoma. Further prospective investigation is required to clarify the role of adjuvant radiation after resection.

The radiosensitization effect of chemotherapy has led to the investigation of concurrent chemoradiation as an adjuvant for resected cholangiocarcinoma. The most commonly utilized agent is 5-FU. Previous retrospective data revealed favorable outcomes of adjuvant chemoradiation (Table 3). Adjuvant chemoradiation has a survival benefit over adjuvant chemotherapy, adjuvant radiation, and surgery alone, especially in distal bile duct cancer[198-201]. Recently, three retrospective studies emphasized a comparable outcome of adjuvant chemoradiation for patients with non-curative surgery when compared to patients with curative surgery alone[202-204].