Epidemiology

Cholangiocarcinoma is an uncommon cancer with an incidence of 1–2 per 100 000 in the USA and approximately 5000–8000 new cases diagnosed each year.² Overall, men are affected 1.5 times as often as women and the majority of patients are greater than 65 years of age, with the peak incidence occurring in the eighth decade of life.² Cholangiocarcinomas are classified according to their site of origin within the biliary tree, with those involving the biliary confluence, or hilar cholangiocarcinoma, being the most common and accounting for approximately 60% of all cases.^3–6 Twenty to thirty per cent of cholangiocarcinomas originate in the distal bile duct, while approximately 10% arise within the intrahepatic biliary tree.^7–9 Rarely, patients will present with multifocal or diffuse involvement of the biliary tree.¹⁰ More recent studies have documented the marked increase in incidence of intrahepatic cholangiocarcinoma, which may become the more common variant (see below).^11–14

Natural history

The vast majority of patients with unresectable bile duct cancer die within 6 months to 1 year of diagnosis, usually from liver failure or infectious complications secondary to biliary obstruction.^3,15–17 The prognosis is often worse for hilar lesions and better for lesions of the distal bile duct, which probably reflects the greater complexity and difficulty in effectively managing proximal lesions, more so than differences in tumour biology. Indeed, it has been shown that location within the biliary tree (proximal versus distal) has no impact on survival provided that complete resection is performed.⁴ That being said, patients with intrahepatic cholangiocarcinoma often present with advanced lesions due to the absence of symptoms, such as jaundice or biliary tract-related sepsis.

Aetiology

Most cases of cholangiocarcinoma in the West are sporadic and have no obvious risk factors. However, certain pathological conditions are associated with an increased incidence, the most common of which is primary sclerosing cholangitis (PSC). The majority of patients (70–80%) with PSC have associated ulcerative colitis, while a minority of patients with ulcerative colitis develop PSC.¹⁸ The natural history of PSC is variable, and the true incidence of cholangiocarcinoma is unknown. In a Swedish series of 305 patients followed over several years, 8% of patients eventually developed cancer. On the other hand, occult cholangiocarcinoma has been reported in up to 40% of autopsy specimens and in up to 36% of liver explants from patients with PSC.^18,19 Patients with cholangiocarcinoma associated with PSC are often not candidates for resection because of multifocal disease or severe underlying hepatic dysfunction.

Congenital biliary cystic disease (i.e. choledochal cysts) is also associated with an increased risk for the development of biliary tract cancer.^20,21 This appears to be related to the finding that these patients have an abnormal choledochopancreatic duct junction, which predisposes to reflux of pancreatic secretions into the biliary tree, chronic inflammation and bacterial contamination.^21–24 A similar mechanism may also explain the increased incidence of cholangiocarcinoma reported in patients subjected to transduodenal sphincteroplasty or endoscopic sphincterotomy. In a series of 119 patients subjected to this procedure for benign conditions, Hakamada et al. found a 7.4% incidence of cholangiocarcinoma over a period of 18 years.²⁵

Hepatolithiasis is a well-known risk factor for the development of cholangiocarcinoma in Japan and parts of southeast Asia, arising in 10% of those affected. Chronic portal bacteraemia and portal phlebitis lead to intrahepatic pigment stone formation, obstruction of intrahepatic ducts, and recurrent episodes of cholangitis and stricture formation.^26,27 This recurrent inflammatory state is likely the main contributing factor to cholangiocarcinogenesis. Biliary parasites (Clonorchis sinensis, Opisthorchis viverrini) are also prevalent and endemic in parts of Asia, such as Thailand, and are similarly associated with an increased risk of cholangiocarcinoma.¹⁹ Finally, exposure to several radionuclides and chemical carcinogens, such as thorium, radon, nitrosamines, dioxin and asbestos, may also increase the risk of cholangiocarcinoma.

Histopathology

Three macroscopic subtypes of extrahepatic cholangiocarcinoma are described: sclerosing, nodular and papillary, of which the first two are often combined into one (i.e. nodular-sclerosing) since features of both types are often seen together.²⁸ The histopathology is distinct between cholangiocarcinomas arising from the extrahepatic and intrahepatic biliary system. For extrahepatic cholangiocarcinoma, the overwhelming majority is adenocarcinoma, and most are firm, sclerotic tumours with a paucity of cellular components within a dense fibrotic, desmoplastic background. As a consequence, a non-diagnostic preoperative biopsy is not uncommon.^2,28,29 Papillary tumours represent a less common morphological variant, accounting for approximately 10% of tumours arising from the extrahepatic biliary tree.²⁸ Papillary tumours are soft and friable, may be associated with little transmural invasion, and are characterised by a mass that expands rather than contracts the duct (Fig. 12.1). Although papillary tumours may grow to significant size, they often arise from a well-defined stalk, with the bulk of the tumour mobile within the ductal lumen. Despite this histological variant being the minority of cases, recognition of this entity is important since they are more often resectable and have a more favourable prognosis than the other types.^19,30

Hilar cholangiocarcinoma is typically highly invasive within the hepatoduodenal ligament. Direct invasion of the liver or perihepatic structures, such as the portal vein or hepatic artery, is a common feature and has important clinical implications regarding resectability.²⁸ The liver is also a common site of metastatic disease, as are the regional lymph node basins, but spread to distant extra-abdominal sites is uncommon at initial presentation.^3,31 These tumours also have a propensity for longitudinal spread along the duct wall and periductal tissues, which is an important pathological feature of cholangiocarcinomas as it pertains to the margin of resection.²⁸ There may be substantial extension of tumour beneath an intact epithelial lining, as much as 2 cm proximally and 1 cm distally, thus predisposing to a radiographic underestimation of tumour extent.³² This predilection for submucosal extension underscores the difficulty in achieving a complete resection. Frozen section analysis of the duct margin during operation may be helpful in this regard but caution is necessary when interpreting the results. Indeed, the authors recently analysed their experience with intraoperative frozen sections and found a substantial false-negative rate. In addition, the benefits of extending the resection with a positive frozen section result were questionable.³³

Gross examination of intrahepatic cholangiocarcinoma reveals a grey scirrhous mass, often infiltrative into the liver parenchyma.³⁴ These tumours are adenocarcinomas and the diagnosis of intrahepatic or peripheral cholangiocarcinoma should be considered in all patients presenting with a presumptive diagnosis of metastatic adenocarcinoma with an unknown primary, particularly in the setting of a large, solitary hepatic mass. A small number show different patterns with focal areas of papillary carcinoma with mucous production, signet-ring cells, squamous cell, mucoepidermoid and spindle cell variants.³⁵ The Liver Cancer Study group of Japan established a subclassification of these tumours based on morphology: (1) mass forming type; (2) periductal infiltrating type; (3) intraductal growth type.³⁶ Although some studies have suggested a correlation with outcome based on morphological subtype, this classification scheme has not gained wide acceptance. Positive immunohistochemical staining usually includes carcinoembryonic antigen (CEA), and tumour markers CA50 and CA19-9. K-ras mutations have also been detected in up to 70% of intrahepatic cholangiocarcinomas, although the frequency of this mutation is quite variable.^37,38 Metastatic disease at the time of exploration is not an infrequent finding. Tumours with both hepatocellular and cholangiocellular differentiation (combined tumours) are rare but well described. Their clinical behaviour more closely approximates that of cholangiocarcinoma than hepatocellular carcinoma, and they tend to display aggressive biology.³⁹

Clinical presentation and diagnosis

The early symptoms of hilar cholangiocarcinoma are often non-specific, with abdominal pain, discomfort, anorexia, weight loss and/or pruritus seen in about one-third of patients.^6,19,40,41 Most patients present with jaundice or incidentally discovered abnormal liver function tests. Pruritus may precede jaundice by some weeks, and this symptom should prompt an evaluation, especially if associated with abnormal liver function tests. Patients with papillary tumours of the hilus may give a history of intermittent jaundice, perhaps due to the ball-valve effect of a pedunculated mass within the lumen or, more likely, small fragments of tumour having passed into the common bile duct. Clinical findings are often non-specific but may provide some useful information. Jaundice is usually obvious, and patients with pruritus often have multiple excoriations of the skin. The liver may be enlarged and firm as a result of biliary tract obstruction. The gallbladder is usually decompressed and non-palpable with hilar obstruction. Thus, a palpable gallbladder suggests a more distal obstruction or an alternative diagnosis. Rarely, patients with long-standing biliary obstruction and/or portal vein involvement may have findings consistent with portal hypertension.

In patients with no previous biliary intervention, cholangitis is rare at initial presentation, despite a 30% incidence of bacterial contamination.^42,43 Endoscopic or percutaneous instrumentation significantly increases the incidence of bacterial contamination and the subsequent risk of clinical infection. In fact, the incidence of bacterobilia approaches 100% after endoscopic biliary intubation, thus making cholangitis more common.⁴³ It should be noted that bacterial contamination of the biliary tract in partial obstruction is not always clinically apparent. The presence of overt or subclinical infection at the time of surgery is a major source of postoperative morbidity and mortality. Thus, endoscopic and percutaneous intubations are both associated with greater morbidity and mortality following surgical resection or palliative bypass for hilar cholangiocarcinoma. In an analysis of 71 patients who underwent either resection or palliative biliary bypass for proximal cholangiocarcinoma, all patients stented endoscopically and 62% of those stented percutaneously had bacterobilia. Postoperative infectious complications were doubly increased in patients stented before operation compared to non-stented patients, while non-infectious complications were equal in both groups.⁴³ Enterococcus, Klebsiella, Escherichia coli, Streptococcus viridans and Enterobacter aerogenes are the most common organisms, and this spectrum of bacteria must be considered when administering perioperative antibiotics; it is imperative to take intraoperative bile specimens for culture in order to guide selection of postoperative antibiotic therapy.

While gallstones or even common bile duct stones may coexist with bile duct cancer, in the absence of certain predisposing conditions (e.g. PSC, recurrent pyogenic cholangitis (previously referred to as Oriental cholangiohepatitis)), it is uncommon for choledocholithiasis to cause obstruction at the biliary confluence. Furthermore, the degree of bilirubin elevation tends to be higher (e.g. 10–18 mg/dL) for malignant obstruction compared to benign stone disease (e.g. 2–10 mg/dL). That being said, other conditions may mimic hilar cholangiocarcinoma on imaging studies, such as benign idiopathic focal stenosis of the hepatic ducts (malignant masquerade), Mirizzi’s syndrome resulting from a large stone impacted in the neck of the gallbladder, and gallbladder cancer.⁴⁴ Nevertheless, it is imperative to fully investigate and delineate the level and nature of any obstructing lesion causing jaundice to avoid missing the diagnosis of carcinoma.

However, the histopathological diagnosis of hilar cholangiocarcinoma is often not made until the specimen is removed at operation. As mentioned previously, due to the dense desmoplastic reaction associated with the sclerosing variant of cholangiocarcinoma, non-diagnostic preoperative biopsies or brushings are the usual clinical scenario. In the authors’ view, histological confirmation of malignancy is not mandatory prior to exploration. With no prior suggestive history (i.e. prior biliary tract operation, PSC, hepatolithiasis), the finding of a focal stenotic lesion combined with the appropriate clinical presentation is sufficient for a presumptive diagnosis of hilar cholangiocarcinoma, which is correct in most instances.⁴⁵ Furthermore, the alternative conditions that one may encounter are often best assessed and treated at operation. It is dangerous to rely entirely on a negative result from a needle biopsy or biliary brush cytology, since they are often misleading, particularly in the face of compelling radiographic evidence of malignant disease.⁴⁶ The use of spy glass technology via endoscopic guidance has facilitated direct visualisation of the bile duct lumen and allows for targeted biopsies of the affected area.

Once a diagnosis of cholangiocarcinoma is suggested, radiological studies are crucial to determine the extent of the tumour to appropriately design a therapeutic plan.

Radiological investigation

High-quality radiological studies are necessary to accurately select patients for resection. Until recently, computed tomography (CT), percutaneous transhepatic cholangiography (PTC) and angiography were considered standard investigations. With improvements in the quality of non-invasive modalities, the authors’ current practice relies almost exclusively on magnetic resonance cholangiopancreatography (MRCP) and duplex ultrasonography (US) for preoperative assessment, which provide similar information with less risk to the patient.

Preoperative evaluation and assessment of resectability

Evaluation of patients with hilar cholangiocarcinoma is principally an assessment of resectability, since resection is the only effective therapy. First and foremost, the surgeon must assess the patient’s general condition, fitness for operation and liver function, since a complete resection usually includes a partial hepatectomy. The presence of significant comorbid conditions, chronic liver disease and/or portal hypertension generally precludes resection. In these patients, biliary drainage is the most appropriate intervention, and the diagnosis should be confirmed histologically if chemotherapy or radiation therapy is planned.

The preoperative evaluation must address four critical determinants of resectability: extent of tumour within the biliary tree, vascular invasion, hepatic lobar atrophy and the presence of metastatic disease.³ The presence of lobar atrophy is often overlooked; however, its importance in determining resectability cannot be overemphasised, since it implies portal venous involvement, suggests a more locally advanced lesion, and compels the surgeon to perform a partial hepatectomy, if the tumour is indeed resectable.⁴⁷ While long-standing biliary obstruction may cause moderate atrophy, concomitant portal venous compromise results in rapid and severe atrophy of the involved segments.

Appreciation of gross atrophy on preoperative imaging is important since it often influences both operative and non-operative therapy.⁴⁷ If the tumour is not resectable, percutaneous biliary drainage through an atrophic lobe, unless necessary to control sepsis, should be avoided since it will not effect a reduction in bilirubin level. Atrophy is apparent on cross-sectional imaging as a small, often hypoperfused lobe with crowding of the dilated intrahepatic ducts (Fig. 12.3). Tumour involvement of the portal vein is usually present if there is compression/narrowing, encasement or occlusion seen on imaging studies.^3,57

The staging systems currently used for hilar cholangiocarcinoma do not account fully for all of the tumour-related variables that influence resectability, namely biliary tumour extent, lobar atrophy and vascular involvement. The modified Bismuth–Corlette classification stratifies patients solely based on the extent of biliary duct involvement by tumour.⁵⁸ Although useful to some extent, it is not indicative of resectability or survival. Similarly, the previous AJCC T-stage system (sixth edition) was based largely on pathological criteria and had little applicability for preoperative staging. The ideal staging system should accurately predict resectability and the likelihood of associated metastatic disease, and also correlate with survival. The authors have proposed such a preoperative staging system.^3,57 This staging system places the finding of portal venous involvement and lobar atrophy into the proper context for determining resectability, especially when partial hepatectomy is an important component of the operative approach (Table 12.1). For example, a tumour with unilateral extension into second-order bile ducts that is associated with ipsilateral portal vein involvement and/or lobar atrophy would still be considered potentially resectable, while such involvement on the contralateral side would preclude a resection. The authors have found that this staging system correlated well with resectability, the likelihood of associated distant metastatic disease, and median survival (Table 12.2).⁵⁷ Independent confirmation of the utility of this classification scheme (the Blumgart clinical staging system) was recently reported in a series of 85 patients from China.⁵⁹ The authors’ criteria for unresectability are detailed in Box 12.1. This staging scheme is now incorporated in the seventh edition of the AJCC staging system for hilar cholangiocarcinoma.

Treatment options

In patients with operable disease, the principal objective is a complete resection, obtaining negative histological margins with subsequent restoration of biliary-enteric continuity. Complete resection is associated with 5-year survival rates of approximately 25–40%, which is far superior to that obtainable with non-operative therapies. Clearly, patient selection contributes largely to this finding, as patients treated non-operatively typically have more advanced disease, and no comparative trials have been performed equating stage for stage. Nevertheless, given the relatively poor response rates with chemotherapy and chemoradiation therapy, resection has emerged as the most effective treatment.

Orthotopic liver transplantation has been attempted for unresectable hilar tumours. Klempnauer et al. reported four long-term survivors out of 32 patients who underwent transplantation for hilar cholangiocarcinoma.⁶⁰ The same group also reported a 17.1% 5-year survival for their overall transplant group.⁶¹ Comparable results were reported by Iwatsuki et al.⁶² The results of transplantation have previously not been sufficiently adequate to justify its use, and most centres now do not perform liver transplantation for cholangiocarcinoma. More recently, data from the Mayo Clinic have emerged suggesting good results with transplantation in highly selected patients with low-volume unresectable disease and combined with an intensive pre-transplant treatment regimen.^63,64 Although the data are compelling, routine use of vascular resection, even when there is no obvious tumour infiltration, will likely lead to higher perioperative morbidity; this approach would therefore seem applicable to a very small proportion of patients.

Resection

Resection is the most effective therapy for patients with potentially resectable tumours, with the primary objective being complete removal of all gross disease with clear histological margins (R0 resection). The importance of an R0 resection is clear from previous studies showing that incomplete (R1 or R2) resections do not improve survival beyond that of patients with unresectable tumours (Fig. 12.4).^3,57 There is now overwhelming evidence to support the observation that partial hepatectomy, combined with excision of the extrahepatic biliary system, is usually required to achieve this goal (Table 12.3). A review of several series in the literature shows a close correlation between the proportion of patients who underwent concomitant partial hepatectomy and the proportion of R0 resections achieved. For tumours extending into the left hepatic duct, en bloc caudate lobectomy is usually necessary to obtain a complete resection, since the principal biliary drainage of the caudate lobe is via the left hepatic duct.^65,66 A dilated caudate duct, suggesting tumour involvement, may occasionally be visualised on preoperative imaging (Fig. 12.3).

Despite improvements in preoperative imaging, a considerable number of patients are still found to have unresectable disease at the time of exploration. In a recent report from MSKCC, this number approached 50% of patients with cholangiocarcinoma explored with curative intent.³⁰ In an effort to minimise the number of non-curative laparotomies performed, staging laparoscopy has been utilised. Two recent studies specifically analysing patients with biliary cancer have shown that laparoscopy can identify a large proportion of patients with unresectable disease primarily in the form of radiographically occult metastases, the yield of which is greatest in locally advanced tumours.^67,68 Weber et al. evaluated 56 patients with potentially resectable hilar cholangiocarcinomas; 33 were ultimately determined to have unresectable disease, of which 14 (42%) were identified at laparoscopy and spared an unnecessary laparotomy. Additionally, a number of recent reports have suggested a potential role for fluoro-2-D-glucose positron emission tomography (FDG-PET) scanning as a means of identifying occult metastatic disease. However, most of these studies include small numbers of patients, and further evaluation is needed before PET can be recommended as a routine screening study for this disease.^69–71 In the authors’ experience with FDG-PET for all biliary tract cancer, the information provided influenced management in 24% of patients.⁷²

Technical aspects of intraoperative tumour assessment, exposure and resection are outside the scope of this chapter. The reader is referred to specialty texts for a detailed description of surgical techniques.⁷³ The authors’ general approach involves the use of staging laparoscopy, followed by a full exploration of the abdomen and pelvis, including intraoperative ultrasonography. Resection of the tumour involves, at a minimum, removal of the entire extrahepatic biliary tract from just above the pancreas distally to beyond the biliary confluence with a complete porta hepatis lymphadenectomy. Also, for the reasons cited above, en bloc partial hepatectomy is required in nearly every case in order to achieve complete tumour clearance. Tumour involvement of the main portal vein proximal to its bifurcation additionally requires a vascular resection and reconstruction if technically feasible. Some authors advocate a ‘no-touch’ technique where a hilar en bloc resection is performed that entails resection of the portal vein bifurcation with reconstruction.⁷⁴ The authors report a 5-year survival advantage of 58% versus 29% (P = 0.02) associated with this approach compared to a conventional hepatectomy. This is clearly an aggressive surgical approach that is likely best applied to a select population.

The extent of lymphadenectomy that should be performed remains an area of controversy. Some surgeons advocate an extended nodal dissection as some studies have demonstrated measurable 5-year actuarial survival in the presence of metastatic disease to distant nodal basins (e.g. para-aortic).^75,76 However, an analysis of studies specifically reporting 5-year survival in patients would suggest that any nodal involvement is a powerful adverse factor and that very few patients benefit from such an aggressive approach (Table 12.4). Thus, while a complete porta hepatis lymphadenectomy should be routinely performed when attempting complete resection, the authors do not advocate an extended lymph node dissection. As is the case in other tumours, the clinical implication of a negative lymph node on histopathological analysis is likely dependent on the total number of lymph nodes sampled. A study from MSKCC reported that seven lymph nodes appears to be the target sampling number in order to accurately stage hilar cholangiocarcinoma.⁷⁷ This must be weighed against the reality that, in most series, the median number of nodes sampled from a porta hepatis lymphadenectomy is usually around three.

Results of resection

Long-term survival after resection of hilar cholangiocarcinoma can be achieved and has improved over recent years.^{3,4,6,65,78,79} It is clear, however, that the results of resection depend critically on the status of the resection margins. The authors firmly believe that an increase in the use of hepatic resection is responsible for the increase in the percentage of R0 resections (negative histological margins) and the observed improvement in survival after resection. This point is emphasised by a reported series of 269 patients accumulated over a 20-year period demonstrating a progressive increase in the proportion of patients subjected to partial hepatectomy, with a corresponding increase in the incidence of negative histological margins and in survival.⁸⁰ A more recent study from MSKCC reported results of resection in 106 consecutive patients and showed a median survival of 43 months in patients who underwent an R0 resection compared to 24 months in those with involved resection margins.³⁰ Multivariate analysis showed that an R0 resection, a concomitant hepatic resection, well-differentiated histology and papillary tumour phenotype were independent predictors of long-term survival.

Adjuvant therapy

The rarity of cholangiocarcinoma has prevented any meaningful clinical trials evaluating the use of adjuvant therapy. Several small, single-centre studies have attempted to investigate the benefit of postoperative adjuvant chemoradiation therapy in patients with hilar cholangiocarcinoma. Cameron et al. and Pitt et al. from Johns Hopkins, in two separate reports, provided data suggesting no benefit of adjuvant external beam or intraluminal radiation therapy.^81,82 In contrast, Kamada et al. suggested that radiation may improve survival in patients with histologically positive hepatic duct margins.⁸³ Additionally, in a small series of patients, five with hilar cholangiocarcinoma, resectability was reportedly greater in patients given neoadjuvant radiation therapy prior to exploration.⁸⁴ It must be noted, however, that none of these studies were randomised and most consisted of a small, heterogeneous group of patients. At the present time, there are no data to support the routine use of adjuvant or neoadjuvant radiation therapy, except in the context of a controlled trial.

The only phase III trial investigating adjuvant chemotherapy, which used mitomycin/5-fluorouracil (5-FU), included 508 patients with resected bile duct tumours (n = 139), gallbladder cancers (n = 140), pancreatic cancers (n = 173) and ampullary tumours (n = 56).⁸⁵ On subset analysis, there were no significant differences in overall or disease-free survival for bile duct tumours. As with radiation therapy, there are no data to support the routine use of chemotherapy in the adjuvant setting, until newer agents, such as oxaliplatin, are tested in a randomised controlled fashion.

Palliation

All patients should be properly assessed for possible resection; however, unfortunately the majority of patients with hilar cholangiocarcinoma are not candidates for resection. In this setting, the management goals include biliary decompression and/or supportive care. Jaundice alone is not necessarily an indication for biliary decompression, given the associated morbidity and mortality. The indications for biliary decompression include intractable pruritus, recurrent cholangitis, the need for access for intraluminal radiotherapy and finally to allow recovery of hepatic parenchymal function in patients receiving chemotherapeutic agents. In fact, supportive care alone is probably the best approach for elderly patients with significant comorbid conditions, provided that pruritus is not a major feature. In patients who are found to be unresectable at operation, an operative biliary decompression can be performed and can be so constructed as to provide access to the biliary tree for postoperative irradiation.^3,86

If the patient is deemed unresectable, the diagnosis should be confirmed with a biopsy. Biliary decompression can be obtained either by a percutaneous transhepatic route or by endoscopic stent placement, although hilar tumours are more difficult to transverse endoscopically. Moreover, the failure rates and incidence of subsequent cholangitis associated with endoscopic decompression are high.⁸⁷ Thus, most are probably better palliated via a percutaneous approach.

Clinical presentation and diagnosis

The clinical presentation of distal bile duct cancer is generally indistinguishable from that of hilar cholangiocarcinoma or other periampullary malignancies. Progressive jaundice is seen in 75–90% of patients, with serum bilirubin levels often exceeding 10 mg/dL.¹⁰⁷ Abdominal pain, weight loss, fever or pruritus occurs in one-third or fewer.^6,104 Distal bile duct tumours are frequently mistaken for adenocarcinoma of the pancreas, the most common periampullary malignancy. Endoscopic retrograde cholangiopancreatography (ERCP) can provide valuable information regarding the level of obstruction, may show that the obstruction is arising from the bile duct without involvement of the pancreatic duct, and can be both diagnostic and therapeutic in cases of choledocholithiasis. Percutaneous transhepatic cholangiography is generally less useful for tumours of the distal bile duct. A good-quality cross-sectional imaging study is also required, usually a CT with angiography to assess for vascular involvement and/or metastatic disease. It is not uncommon that CT does not reveal a mass given the frequent small tumour size at presentation. Increasingly, magnetic resonance cholangiopancreatography (MRCP) is being used to evaluate periampullary tumours. As is true for hilar lesions, MRCP can provide information of the distal bile duct previously obtainable only with the combination of ERCP and CT.¹⁰⁸

In patients with a stricture of the distal bile duct and a clinical presentation consistent with cholangiocarcinoma (or any other periampullary malignancy), histological confirmation of malignancy is generally unnecessary, unless non-operative therapy is planned. Benign strictures do occur in the lower bile duct, but these are difficult to differentiate definitively from malignant strictures without resection. In addition, endoscopic brushings of the bile duct have an unacceptably low sensitivity, making a negative result virtually useless.¹⁰⁹ Excessive reliance on the results of percutaneous or brush biopsies serves only to delay therapy.

Staging and assessment of resectability

Carcinomas of the distal common bile duct are staged according to the AJCC system (seventh edition) for tumours of the extrahepatic bile ducts. This system is of limited clinical use, as it is based on pathological information and does not provide any information pertaining to factors that define resectability. The most important of these is the presence of tumour involvement of the portal vein, superior mesenteric artery or common hepatic artery. Tumours involving a short segment of the portal vein (< 2 cm) may be resected with reconstruction of the vein. Metastatic disease to distant sites, such as the liver or peritoneum, represents an absolute contraindication to proceeding with resection; the involvement of regional nodal basins should perhaps be viewed as a relative contraindication, given the poor survival in patients with node-positive disease. Along with good-quality preoperative imaging, staging laparoscopy may help to reduce the number of non-curative laparotomies performed. Endoscopic ultrasonography (EUS) can provide additional staging information (nodal or vascular involvement) and allows an opportunity to biopsy the lesion, if necessary; however, EUS is generally not required if resection is planned based on the results of high-quality cross-sectional imaging studies.

Treatment options

Complete resection is the only effective therapy for cancers of the distal bile duct.4–6^,104–106 Reported 5-year survival rates range from 14% to 40% after complete resection, and in most studies survival beyond 1 year was uncommon in patients with tumours not amenable to resection.^5,38,104,105 Nearly all distal bile duct cancers require a pancreaticoduodenectomy for complete excision. In a series from MSKCC, 13% of patients (6 of 45) underwent bile duct excision alone, while in the Veterans Hospital study this figure was only 9% (3 of 34).^104,105 In addition to resection margin status (i.e. an R0 resection), metastatic disease to regional lymph nodes is a critical determinant of outcome. Fong et al. found that lymph node status was the only independent predictor of long-term survival after complete resection, with positive nodes conferring a 6.7 times greater likelihood of recurrence and death.¹⁰⁴ Ito et al. reported that 11 lymph nodes needed to be evaluated to accurately assess lymph node involvement for distal cholangiocarcinoma.⁷⁷

Survival after resection of distal bile duct tumours is comparable to, and maybe better than, that for pancreatic cancer.^6,104,105 Furthermore, it has been erroneously assumed that survival is greater than that after resection of hilar cholangiocarcinomas as well.⁵ However, if adjusted for stage and completeness of resection, the survival rates between the two are similar.⁴ Adjuvant therapy after resection (chemotherapy and radiation therapy) has not been proved to improve survival, although this issue has not been evaluated in a prospective fashion.⁶

Surgical bypass (hepaticojejunostomy or choledochojejunostomy) or biliary endoprostheses can be used for palliation of symptomatic biliary obstruction. Endoprostheses for distal biliary obstruction are easier to place and have a greater long-term patency than those placed for hilar obstruction.⁸⁸ Surgical bypass provides excellent relief of jaundice, but is typically used when unresectability is found at laparotomy. The authors generally use biliary endoprostheses in patients with clear-cut unresectable disease, discovered preoperatively or at staging laparoscopy, and in those unfit for operation. Laparoscopic biliary enteric bypass is also an option, but the expertise needed to perform this procedure is not widely available.

Clinical presentation

Intrahepatic cholangiocarcinoma (IHC), also referred to as peripheral cholangiocarcinoma, originates from the intrahepatic biliary radicles. IHC is rare in Western countries, accounts only for approximately 10% of all cholangiocarcinomas and is less frequently associated with underlying liver parenchymal disease than hepatocellular carcinoma, although an association appears to exist. Recently, a marked increase in the incidence and age-adjusted mortality has been identified, the reasons for which are unclear but may be related to the rising incidence of obesity-related, non-alcoholic fatty liver disease or chronic hepatitis C infection.^12,13 The presenting symptoms are subtle and often only include pain either directly or indirectly related to a large lesion. Malaise, weight loss and fever are uncommon, but jaundice and pruritus may be seen in up to one-third of cases, which is generally indicative of compression or invasion of the biliary confluence. Small lesions often present as incidental findings on imaging studies undertaken for unrelated symptoms.

Diagnosis

A solitary, intrahepatic tissue mass at first raises concern for hepatocellular carcinoma, a more common disease than IHC. However, in the absence of chronic hepatic parenchymal disease, chronic hepatitis or an elevated serum α-fetoprotein level, IHC must be considered. Percutaneous needle biopsy is often performed and will demonstrate adenocarcinoma; however, a definitive diagnosis of intrahepatic cholangiocarcinoma often cannot be made based on a needle biopsy alone. Patients should be investigated for evidence of a primary tumour elsewhere (gastrointestinal tract, lung, breast), since the most common diagnosis for adenocarcinoma in the liver is metastatic disease. In the absence of an extrahepatic primary site, patients with biopsy-proven adenocarcinoma in the liver should be considered to have an intrahepatic cholangiocarcinoma. Immunohistochemical staining of the biopsy specimen may further support the diagnosis by demonstrating a lesion of pancreaticobiliary origin.

Radiological investigations

The radiological features of IHC on cross-sectional imaging are well described, and when combined with histological findings from a needle biopsy, can be virtually diagnostic. On MRI, these tumours are generally hypointense on T1-weighted images and heterogeneously hyperintense on T2-weighted images. These lesions demonstrate initial rim enhancement characterised by progressive and concentric enhancement post-administration of contrast material. Generally the lesions do not completely enhance post-contrast. In the absence of a separate primary source of disease, a lesion in the liver with this morphology on MRI evaluation can be considered virtually diagnostic of cholangiocarcinoma without a tissue diagnosis. On contrast-enhanced CT, variable rim-like enhancement is also seen, predominantly on the arterial phase images with gradual centripetal enhancement on delayed imaging (Fig. 12.5). Intrahepatic cholangiocarcinomas may only enhance completely on delayed imaging obtained hours after contrast administration, a finding related to the desmoplastic nature of the tumour. Capsular retraction may also be seen.^110,111

Staging and assessment of resectability

Currently, there is no useful clinical staging system for intrahepatic cholangiocarcinoma. The AJCC TNM classification for primary liver cancers is applied both to hepatocellular carcinoma and IHC, but is of little clinical value. Because IHCs tend to be relatively silent lesions, they are often large at presentation. Thirty per cent of patients will have peritoneal or hepatic metastases at presentation and many of these will not be detected until staging laparoscopy or exploratory laparotomy is performed. In a review of 53 peripheral cholangiocarcinomas treated at MSKCC over an 8-year period, the median tumour diameter was 7.1 cm at presentation.¹¹² Twenty patients were found to be unresectable at exploration for a 62% overall resectability rate. Operative findings precluding resection were intrahepatic metastases (35%), peritoneal metastases (30%), coeliac lymph node metastases (25%) and portal vein involvement (10%). Staging laparoscopy was conducted in 22 patients, of whom six were spared laparotomy secondary to findings of peritoneal and intrahepatic metastases. In a more recent review at the authors’ institution, a total of 270 IHC patients were seen over a 16-year period, representing an average annual increase of 14% in patients with this diagnosis over the study period. Of the patients treated at MSKCC, 54% had unresectable disease at presentation; ultimately, 34% of the entire cohort underwent a potentially curative resection (70% of those explored with curative intent).¹⁴

Treatment options

Hepatic resection with negative histological margins remains the only potentially curative treatment for this disease. Unfortunately, only about one-third to one-half of patients have potentially resectable lesions at the time of presentation. Additionally, a significant proportion of these patients will have findings at operation that preclude resection.^14,112 Median survival after resection was approximately 36 months in a recent study by Endo et al., compared to 9 months for patients with advanced disease.¹⁴ Unfortunately, even after a complete resection, recurrence was common and was predicted by tumour size > 5 cm, the presence of multiple liver tumours or metastatic disease to regional lymph nodes; the liver was the single most common site of recurrence.

An international study group for IHC has recently advocated routine portal lymph node dissection at the time of resection, as approximately 30% of patients who underwent evaluation were found to have lymph node involvement.¹¹³ Although this does not seem to provide any therapeutic benefit, it may allow for better prognostication and patient selection for adjuvant therapy. The idea of adjuvant therapy needs to be revisited, given the promising results from the ABC-02 trial that reported improved survival with gemcitabine plus cisplatin.¹⁰³ Appropriate patient selection will likely play a paramount role in the application of these data and practice to the adjuvant setting. Given the low yield of lymph nodes from a portal lymphadenectomy (median 3), the accuracy of lymph node evaluation is questionable, and thus its use as a selection criteria for adjuvant therapy is controversial. Perhaps pathological criteria from evaluation of the primary tumour, such as lymphovascular invasion and perineural invasion, should be used instead as selection criteria for adjuvant therapy, as the presence of these factors has been associated with poor survival that is similar to lymph node-positive disease.¹¹⁴

Orthotopic liver transplantation has been utilised in the management of some patients.^115,116 However, many of these lesions are suitable for resection, which would likely produce similar results. Given the critical shortage of liver grafts, transplantation for IHC is not performed in most centres, unless it is done in the context of a clinical trial. The use of chemotherapy has not been shown to improve survival, either as adjuvant therapy following resection or in patients with unresectable lesions.¹¹⁷ However, as mentioned above, given the results of the ABC-02 trial for advanced disease, administration of the double regimen in the adjuvant setting needs to be evaluated. External beam radiation therapy, intraoperative radiation and intraluminal radiation therapy have all been evaluated as well, albeit in small, not well controlled, primarily retrospective studies. Similar to chemotherapy, none have shown a significant survival benefit in patients with unresectable disease. However, Ibrahim et al. reported a median survival of 31.8 months in patients with unresectable IHC treated with yttrium-90 (Y-90) who had a performance status of ECOG 0. This study included only 24 patients and a positive effect of Y-90 was not observed in patients with ECOG performance status of 1 or 2.¹¹⁸ The senior author has also reported the experience using hepatic arterial infusion pump therapy (FUDR) with and without bevacizumab for advanced disease. The median survival of patients treated with intra-arterial therapy was 29.5 months, better than that usually achieved with systemic chemotherapy.¹¹⁹ It also appeared that the addition of bevacizumab increased the incidence of biliary toxicity without any improvement in survival (31.1 vs. 29.5 months; P = NS).¹²⁰

T1 tumours

T1a tumours, or those that are confined to the lamina propria, are most often discovered after, and adequately treated with, a simple cholecystectomy. Because the potential for nodal involvement is small, cure rates approach 85–100% if negative margins are achieved.^140,141 T1b tumours, i.e. those tumours that have extended into, but not through, the muscle layer, in theory should be cured by a simple cholecystectomy as well. However, there have been reports in the literature documenting recurrence and death following a simple cholecystectomy for T1b tumours.¹⁴² Given the limited data regarding T1b gallbladder cancers in the literature, the decision to perform a simple cholecystectomy versus a more radical procedure should be made on a case-by-case basis.

T2 tumours

T2 lesions, or tumours that extend into the perimuscular connective tissues, should be treated with an aggressive resection, including removal of adjacent liver, lymphadenectomy of the hepatoduodenal ligament and a bile duct resection only if necessary to obtain a negative margin on the cystic duct. As discussed above, the extent of hepatic resection required depends on whether or not there is tumour involvement of the right portal pedicle (i.e. major inflow vascular structures or right hepatic duct). In the absence of such involvement, the authors prefer to perform a segmental resection of segments IVb and V, and most T2 tumours are amenable to such an approach. It should be noted that the normal plane of dissection of simple cholecystectomy, open or laparoscopic, is within the perimuscular connective tissue intimately associated with the liver. Thus, a simple cholecystectomy will not achieve tumour clearance with certainty. A lymphadenectomy is performed in the treatment of T2 tumours given that up to 50% of these lesions have associated lymph node metastases.¹³¹ The benefit of an extended resection over simple cholecystectomy is supported by data that demonstrate improved survival. This is underscored by the fact that liver involvement can be found after radical resection in up to a quarter of patients with presumed T2 disease after cholecystectomy alone, a finding that is associated with markedly reduced recurrence-free and disease-specific survival.¹³⁸ De Aretxabala et al. reported 5-year survival rates of 70% compared with only 20% after simple cholecystectomy alone.¹⁴³

T3 tumours

T3 tumours penetrate the serosa and may extend into the liver parenchyma or a single extrahepatic organ. These tumours require a hepatic resection and porta hepatis lymphadenectomy at a minimum. As with T2 tumours, if a limited partial hepatectomy can be performed to achieve the objectives, then this is preferred; however, one should not hesitate to perform a more extensive partial hepatectomy and/or bile duct resection if necessary. When a complete resection is achieved, 5-year survival rates of 30–50% can be obtained in this patient population.^131,136,139

T4 tumours

T4 lesions, as defined by the sixth edition staging system, generally reflect unresectable disease.

Preoperative suspicion of malignancy

If gallbladder cancer is suspected on preoperative imaging studies, a staging laparoscopy prior to laparotomy is helpful to assess the abdomen for evidence of peritoneal spread or discontiguous liver disease. In general, however, performing a laparoscopic cholecystectomy should be avoided.^67,68,144 One needs to be prepared to proceed with resection of an invasive malignancy, unless proven otherwise. It is not unreasonable to obtain intraoperative frozen section histology to prove malignancy before proceeding with hepatic resection.

Unsuspected malignancy at exploration

It should be routine to inspect the gallbladder mucosa after simple cholecystectomy. Suspicious lesions should be sent immediately for frozen section. If a carcinoma is diagnosed, the need to perform additional surgery is dictated by the T stage on frozen section, although the information will be limited since a full histopathological evaluation is not available at the time of operation. The authors prefer to perform an oncologically correct resection, suitable for an invasive lesion, at the time it is discovered, unless there are extenuating circumstances that mandate otherwise. However, if the surgeon is not comfortable with performing a radical cholecystectomy/hepatic resection, the patient is best served by transferring them to a centre/surgeon with experience in performing the appropriate operation. A delayed radical and appropriate resection does not negatively influence the patient’s outcome.¹³⁹

Malignancy diagnosed post-cholecystectomy

When the cancer is diagnosed by postoperative histology, the need for a more radical resection will be based on T stage, as outlined above. As mentioned above, Fong et al. demonstrated a much improved 5-year survival rate in patients undergoing a second operation compared to those who did not. Five-year survival rates of 61% were achieved in patients who were re-resected compared to 19% for patients who did not undergo a radical second operation.¹³⁹ However, prior to undertaking a second operation, high-quality cross-sectional imaging (CT/MRI) should be obtained to appropriately stage the disease. Postoperative inflammatory changes may be indistinguishable from tumour and thus may necessitate bile duct resection or a more aggressive hepatic resection to ensure complete tumour eradication.

Given that inadvertent cholecystotomy during cholecystectomy is rarely documented, it is difficult to predict who is at increased risk for peritoneal dissemination and, specifically, port site recurrence. In the past, routine resection of laparoscopic port sites was recommended, in an effort to ensure clearance of microscopic disease that may have implanted during the laparoscopic procedure. However, there is little evidence to support the efficacy of routine resection of all port sites at re-operation.¹⁴⁵ In the authors’ experience, recurrence at the port sites is a harbinger of generalised peritoneal recurrence that will not be prevented with resection of these limited areas.