Chapter 51. Cancer of the Gallbladder and Bile Ducts

This chapter focuses on biliary tract cancers, including those of the gallbladder and intrahepatic and extrahepatic bile ducts. Because the epidemiology, clinical presentation, and surgical approach associated with gallbladder cancer and bile duct cancer are distinct, these two cancers are discussed separately.

Epidemiology

With an incidence of 6500 cases annually in the United States, gallbladder cancer is the fifth most common gastrointestinal tract malignancy in this country.1 Incidence increases with age and is two to six times higher in women than in men. Worldwide, the highest incidence rates (up to 7.5 per 100,000 in men and 23 per 100,000 in women) occur among populations in the Western part of South America (eg, Chile and Peru), in North American Indians, in Mexican Americans, and in northern India.2 The best characterized risk factor for the development of gallbladder cancer is chronic inflammation associated with gallstones (Table 51-1). Although only 0.5–3% of patients with cholelithiasis will develop gallbladder cancer, gallstones are present in 70–90% of patients diagnosed with gallbladder cancer.2–4 Further, the geographic pattern of gallbladder cancer incidence correlates with that of cholelithiasis.

Other factors implicated to increase the risk of developing gallbladder cancer include porcelain gallbladder (the incidence of gallbladder cancer is reported to range from 12.5 to 60% in patients with this condition),2–4 adenomatous polyps of the gallbladder (in contrast, cholesterol and inflammatory polyps and adenomyomas are not believed to be the risk factors), chronic infection with Salmonella typhi, carcinogen exposure (eg, increased risk has been reported for miners exposed to radon), and abnormal pancreaticobiliary duct junction (APBDJ). In this latter condition, a long common channel, formed by an abnormally proximal junction between the pancreatic and common bile ducts (CBDs), and elevated sphincter of Oddi pressures create a predisposition to reflux pancreatic exocrine secretions into the bile ducts. APBDJ is most prevalent in Asian countries and appears to increase the risk of development of biliary cancers, especially gallbladder cancer.5 Gallbladder cancers arising in patients with APBDJ tends to occur at a younger age, to have a lesser degree of female predominance, and to be less often associated with cholelithiasis than those arising in patients without APBDJ.

Pathogenesis and Pathology

Chronic inflammation of the gallbladder mucosa related to gallstones is hypothesized to be the major factor leading to malignant transformation in most cases of gallbladder cancer. The progression from dysplasia, to carcinoma in situ (CIS), then to invasive cancer has been described for gallbladder cancer. The molecular changes associated with this progression are under investigation: K-ras mutations appear to be relatively uncommon, whereas p53 mutations are prevalent and tend to arise early during this progression.2

Gallbladder cancers arising in patients with APBDJ may be associated with a distinct pathogenetic mechanism. These cancers are associated with a high prevalence of K-ras mutations and a late onset of p53 mutations.2 In addition, there is a high prevalence of premalignant epithelial hyperplasia with a papillary or villous histology in the gallbladder mucosa of patients with APBDJ.

Eighty percent of primary gallbladder cancers are adenocarcinomas. Other histological types include small cell cancer, squamous cell carcinoma, lymphoma, and sarcoma. Gallbladder cancers are also classified according to morphology as infiltrative, nodular, papillary, or a combination of these types. Papillary cancers tend to grow within the gallbladder lumen and are less likely to invade the liver or to metastasize to lymph nodes; it is associated with the best prognosis. Infiltrative or nodular cancers have a more diffuse pattern of growth that is difficult to recognize on imaging studies. These lesions are more likely to have invaded the liver and to have metastasized to lymph nodes by the time of diagnosis.

Several staging systems for gallbladder cancer have been described. The Nevin staging system, originally put forth in 1976, is of historical interest; the tumor, node, metastasis (TNM) system is used today (Table 51-2). The seventh edition of the American Joint Committee on Cancer (AJCC) staging system, published in 2010, contains important modifications to the staging of gallbladder cancer contained in the sixth edition.6 N stage now includes N1 (metastasis to cystic duct, CBD, hepatic artery, and/or portal vein lymph nodes) and N2 (metastasis to periaortic, pericaval, superior mesenteric artery, and/or celiac artery lymph nodes) designations. Stage classifications have been revised to better reflect patient outcomes. For example, locally unresectable T4 cancers are now classified as stage IV, whereas T4N0 cancers were classified as stage III in the sixth edition. M0 cancers associated with lymph node metastasis are now classified as stage IIIB (with N1 disease) or stage IVB (with N2 disease), whereas these cancers were classified as stage IIB or III (depending on T stage) in the sixth edition.

Clinical Presentation and Diagnosis

In the absence of advanced disease, patients with gallbladder cancer are asymptomatic or have symptoms, such as abdominal pain, anorexia, nausea, and vomiting, that may be indistinguishable from those of cholelithiasis or cholecystitis. With advanced disease, patients can present with weight loss, obstructive jaundice (due to tumor invasion into the biliary tree or to liver metastases), and duodenal obstruction. Signs associated with advanced disease include palpable abdominal masses, hepatomegaly, and ascites.

Laboratory tests may suggest obstructive jaundice if this condition is present; otherwise, they are not helpful in the diagnosis of gallbladder cancer. Tumor markers carcinoembryonic antigen (CEA) or CA 19-9 may be elevated; however, they lack sufficient sensitivity or specificity to be useful in clinical decision making for individual patients.

Patients with suspected gallstone- or gallbladder-related conditions typically undergo transabdominal ultrasonography (US). Findings suggestive of gallbladder cancer on ultrasonography include mural thickening or calcification, a gallbladder mass greater than 1 cm in diameter, and loss of the normal gallbladder wall–liver interface (Fig. 51-1). Relative to transabdominal ultrasonography, endoscopic ultrasonography (EUS) offers greater accuracy in assessing depth of gallbladder wall penetration by masses and regional lymph node enlargement. Selective application of EUS in patients with a gallbladder mass can help in the determination of whether the mass is non-neoplastic (eg, cholesterol pseudopolyp) or neoplastic. In addition, EUS-guided biopsy is an effective technique in cases in which a tissue diagnosis is required.

Computed tomography (CT) scanning should be performed on patients suspected of having gallbladder cancer. Findings of gallbladder cancer include a mass protruding into the gallbladder lumen or completely replacing the gallbladder and focal or diffuse thickening of the gallbladder wall (Fig. 51-2). CT scanning also offers information on the presence or absence of distant metastases, regional lymph node involvement, and local invasion into the liver and porta hepatis.

Magnetic resonance imaging (MRI) and magnetic resonance cholangiopancreatography (MRCP) can offer additional information on local invasion, particularly into the porta hepatis. These tests are used selectively, if CT findings are equivocal. Similarly, endoscopic or percutaneous cholangiography is not routinely indicated; they are used primarily for palliation or preoperative management of obstructive jaundice.

Surgical Therapy

Surgical resection is the only known curative form of therapy for gallbladder cancer. For patients in whom surgical exploration is contraindicated because of medical comorbidities or evidence of unresectable disease on imaging studies (eg, metastatic disease), a percutaneous or endoscopic needle biopsy can be obtained to confirm the diagnosis. For patients in whom surgery is planned, a preoperative biopsy is contraindicated, as gallbladder cancer has a propensity for dissemination along needle tracts.

Recommendations for extent of surgical resection according to disease stage are given below. Specific technical issues are discussed subsequently.

Stages 0 and I

For Tis (carcinoma in situ) and T1a (cancer that invades the lamina propria but does not extend into the muscularis) lesions, the available retrospective data suggest that simple cholecystectomy is sufficient therapy in most cases. These lesions are most frequently detected on pathological examination of gallbladders removed for presumed benign disease. Patients diagnosed with gallbladder cancer in this manner should undergo formal imaging-based staging, and the cholecystectomy specimen should be carefully examined to ensure that all margins are negative for cancer. Patients with imaging studies that reveal no evidence of residual or metastatic gallbladder cancer and are found to have a cystic duct margin that is positive for cancer should undergo re-exploration with common duct excision, regional lymphadenectomy, and hepaticojejunostomy. In contrast, patients with negative margins and negative imaging studies who undergo no additional treatment for their gallbladder cancer have excellent outcomes that are unlikely to be improved by radical surgery.7

The management of T1b (cancer that invades the muscularis but does not extend into the perimuscular connective tissues) lesions has been controversial. In published series, the 5-year survival rate for patients with T1b gallbladder cancer having undergone radical resection averages 87.5%, whereas it averages only 61.3% in patients having undergone simple cholecystectomy alone.8 Further, a recently published decision analysis suggests that radical surgery (described later for stage II cancers) is associated with improved survival compared to that associated with simple cholecystectomy alone in most patients with T1b gallbladder cancer.8 Therefore we treat patients with T1b gallbladder cancer in the same way we treat patients with T2 gallbladder cancer.

Stage II

Patients found to have a T2 (cancer invasion into the perimuscular connective tissues of the gallbladder) lesion in their cholecystectomy specimen following surgery for presumed benign disease should undergo staging (as described earlier), and in the absence of contraindications, radical resection. Simple cholecystectomy is usually performed using a subserosal dissection plane, and, hence, may leave positive margins in the gallbladder fossa. Indeed, re-exploration reveals residual tumor in 40–76 % of these cases.9–12 In addition, the probability of regional lymph node metastasis in patients with T2 gallbladder cancer has been reported to range from 28 to 63%.9–12 These findings provide rationale for performing re-exploration with liver resection and regional lymphadenectomy of the hepatoduodenal ligament. There is convincing, albeit retrospective, evidence that such radical surgery is associated with improved survival for patients with T2 gallbladder cancer.9–12 Given the propensity of gallbladder cancer to seed wound sites, re-excision of all surgical wounds, including laparoscopic port sites, during re-exploration has traditionally been recommended. However, re-excising port sites can be difficult (the trajectory through which ports had traversed the abdominal wall during the initial operation may be impossible to determine at the time of definitive surgery), and the value of this practice is unproven.

Patients suspected of having a T2 gallbladder cancer preoperatively (prior to cholecystectomy) should undergo staging, and in the absence of contraindications, exploration with en bloc resection of the gallbladder and adjacent liver to a depth of at least 2 cm, in addition to regional lymphadenectomy of the hepatoduodenal ligament. Although a nonanatomic liver resection encompassing the gallbladder fossa can be applied at the time of re-exploration or en bloc with the gallbladder during the initial procedure, anatomical resection of liver segments 4b and 5 may be associated with less intraoperative bleeding.

Stage III

A role for aggressive surgical resection for some stage III gallbladder cancers has been receiving increasing recognition. This stage includes T3 lesions (locally advanced cancers that perforate the gallbladder serosa or directly invade the liver and/or one adjacent organ) and T1–3 lesions associated with regional lymph node metastasis.

Surgery for patients with T3 lesions requires careful planning and must be tailored to individual patients. For some patients with liver invasion, hepatic resections encompassing segments 4b and 5 may be sufficient. However, because the gallbladder fossa bridges both right and left hepatic lobes, trisegmentectomy is often required. Adjacent involved structures, such as the hepatic flexure of the colon, should be resected en bloc. Long-term survival rates ranging from 15 to 63% have been reported from some centers to be associated with these extended procedures for T3 lesions.9–12

Stage IV

Stage IVA (invasion of the main portal vein, common hepatic artery, multiple extrahepatic organs) and stage IVB (N2 and/or distant metastasis) disease meet criteria for unresectability. Anecdotal reports of super-radical procedures involving resection of the main portal vein and/or common hepatic artery exist, but these procedures are associated with substantial morbidity and mortality rates and are unlikely to confer any survival benefits.

There is no evidence to support the application of debulking cholecystectomy to prevent subsequent episodes of cholecystitis; we do not recommend it.

Surgical Technique

For patients suspected of having resectable gallbladder cancer, we begin surgical exploration with laparoscopy. In the absence of disseminated disease, we proceed with open laparotomy. Because of the risk for gallbladder perforation and tumor spillage, we recommend against laparoscopic cholecystectomy in patients suspected of having gallbladder cancer. We also recommend early conversion to open laparotomy in patients undergoing laparoscopic cholecystectomy for presumed benign disease in whom the suspicion of gallbladder cancer arises intraoperatively.

We use a right subcostal incision, as it easily can be extended to a chevron incision if necessary. We then conduct a thorough examination for metastases, especially in the liver and on the peritoneal surfaces. For patients in whom the suspicion of gallbladder cancer is low at this point a simple cholecystectomy is done, and the gallbladder is examined using frozen-section analysis. Confirmation of T1b, T2, or T3 disease should prompt radical resection, as described later. If the diagnosis based on frozen-section analysis is ambiguous (ie, the presence of gallbladder cancer cannot be confirmed), radical surgery should be deferred. For patients in whom the suspicion of gallbladder cancer is high because of the presence of a firm mass, we obtain a small biopsy of the lesion. If the diagnosis of gallbladder cancer is confirmed on frozen-section analysis, the gallbladder is resected en bloc with the adjacent liver, as described later. Although determining depth of cancer invasion can be difficult on frozen sections, these grossly apparent cancers are likely to be T2 or more advanced lesions.

If radical resection is indicated, we then perform a Kocher maneuver to mobilize the duodenum and the head of the pancreas. Enlarged retropancreatic, celiac, superior mesenteric, or para-aortic lymph nodes are sampled and subjected to frozen-section analysis. If these lymph nodes are positive for metastases, N2 disease is present, and radical resection is aborted.

In the absence of N2 disease, we then perform regional lymphadenectomy. We skeletonize CBD and common hepatic duct, hepatic artery, and portal vein from the superior border of the duodenum to the liver hilum. During this dissection, lymph node–bearing fibrofatty tissues are swept toward the gallbladder and removed as a specimen. Tumor invasion of the portal vasculature is assessed during this dissection. We do not perform major vascular resection for advanced gallbladder cancer at our institution.

In contrast, we do perform common duct resection if the gallbladder cancer has invaded this structure. Common duct resection may also facilitate resection of bulky nodal disease in the hepatoduodenal ligament. The CBD is clamped and transected at the superior border of the duodenum, and its stump is oversewn with a nonabsorbable monofilament suture. Similarly, the common hepatic duct is transected near its bifurcation. We take care to minimize spillage of bile that may contain cancer cells.

We then perform en bloc resection of the gallbladder and the adjacent liver (or the liver resection alone if the patient has already undergone cholecystectomy). If the CBD has not been transected, the cystic duct is divided near its junction with the CBD. Similarly, the cystic artery is ligated and divided near its origin. For T2 cancers, either a nonanatomic wedge resection of the liver that encompasses the gallbladder fossa to a depth of 2 cm or anatomical resection of liver segments 4b and 5 is acceptable (Fig. 51-3). The capsule of the liver is scored with electrocautery to mark the resection plane. Overlapping chromic liver sutures are then placed around the periphery of the resection plane for hemostasis and retraction. The liver parenchyma is then transected using one of the standard methods (we usually use a combination of electrocautery and argon-beam coagulation). Care should be taken near the base of the liver resection margins to avoid injuring the right hepatic artery as it traverses inferiorly in the gallbladder fossa.

If the common duct has been resected, a 60-cm Roux-en-Y limb of jejunum is used to create a hepaticojejunostomy. The anastomosis is constructed using a single layer of 5-0 absorbable sutures.

Adjuvant Therapies

Adjuvant chemoradiotherapy is commonly administered after resection of gallbladder cancers. External beam or intraoperative radiation therapy alone or in combination with 5-flourouracil (5-FU) is associated with diminished rates of local recurrence. The impact of these regimens on survival is unclear; no data derived from prospective randomized clinical trials on the efficacy of these regimens exist.

Palliation

The goals of palliative therapy are relief of pain, manifestation of biliary obstruction (eg, pruritis and cholangitis) and bowel obstruction. Given the limited expected survival duration of patients diagnosed with unresectable gallbladder cancer (weeks to months), endoscopic or percutaneous stenting, rather than surgical bypass, is generally recommended for relief of symptomatic biliary obstruction (Figs.51-4 and 51-5). Biliary stents are discussed in greater detail later in the section on palliation of bile duct cancers.

Palliative radiation therapy, regional intra-arterial chemotherapy, systemic chemotherapy, and chemoradiotherapy all have been applied in patients with advanced gallbladder cancer. Results of the ABC-02 trial were recently published.13 Data from this multicenter phase III trial of patients with locally advanced or metastatic biliary tract cancer (of whom ~36% had gallbladder cancer) demonstrated that the combination of gemcitabine plus cisplatin is associated with improved overall and progression-free survival than gemcitabine alone. As such, this gemcitabine-cisplatin combination represents the current standard treatment option for patients with advanced biliary tract cancers, including gallbladder cancer.14

Outcomes

Data derived from the National Cancer Database support the nihilistic view traditionally associated with gallbladder cancer.15 In these population-based data, 5-year survival rates for patients with T1N0, T2N0, and T3N0 (or node-positive) disease are 39, 15, and 5%, respectively.

However, contemporary surgical series suggest that substantially improved outcomes can be achieved by the application of surgical resection of gallbladder cancers.16 In these reports, 5-year survival rates following resection of T1 lesions ranges from 85 to 100%. With radical resection of T2, T3, and T4 lesions, reported 5-year postoperative survival rates range from 80 to 90%, 15 to 63%, and 2 to 25%, respectively. Radical resection of node-positive disease has been reported to be associated with 5-year survival in as high as 60% of patients, although some reported series contained no patients who survived 2 or more years among those with lymph node metastasis.9–12

Reported morbidity and mortality rates associated with resection of gallbladder cancers range from 5 to 54% and from 0 to 21%, respectively. In general the highest morbidity and mortality rates are associated with series describing more extensive resections.

The best reported outcomes among patients with unresectable biliary tract cancers are those from the ABC-02 trial.13 The median overall survival among patients treated with the combination of gemcitabine and cisplatin was 11.7 months, whereas it was 8.1 months in those treated with gemcitabine alone.13

Epidemiology

In this discussion, the term cholangiocarcinoma is used interchangeably with bile duct cancer and is used to denote cancers arising in the intrahepatic or extrahepatic biliary tree, exclusive of the ampulla of Vater and gallbladder.

Approximately 6000 new cases of cholangiocarcinoma are diagnosed annually in the United States.1 Most patients are diagnosed in the fifth through the seventh decades of life. Unlike gallbladder cancer, for which there is a clear female predominance, the incidence of bile duct cancer is slightly higher in males than in females.

Although most patients diagnosed with cholangiocarcinoma have no identifiable predisposing factors, several conditions clearly increase the risk of developing this cancer (Table 51-3). In Western countries, primary sclerosing cholangitis (PSC) is the most important risk factor; indeed, approximately 30% of cases of cholangiocarcinoma in the West are diagnosed in patients with PSC. Among patients with PSC, the estimated lifetime incidence of cholangiocarcinoma ranges from 10 to 15%, with an annual incidence of 0.6–1.5%.17 In addition, cholangiocarcinoma tends to be diagnosed at an earlier age (third through fifth decades of life) in patients with PSC than in the general population.

In Asian countries, infestation with the liver flukes Opisthorchis viverrini or Clonorchis sinensis and hepatolithiasis are important factors for cholangiocarcinoma. Other risk factors include choledochal cysts, Caroli's disease, and exposure to the radiological contrast agent Thorotrast. Increased risk has been reported for workers in the auto, rubber, chemical, and wood-finishing industries and among patients with hepatitis C viral infection. Two genetic conditions (Lynch syndrome II and multiple biliary papillomatosis) have been identified as increasing the risk of developing bile duct cancer.

Pathogenesis and Pathology

Malignant transformation in the bile duct epithelium, as in other regions of the gastrointestinal tract, is hypothesized to arise in association with a step-wise accumulation of genetic abnormalities. A range of mutations and other abnormalities involving oncogenes (eg, K-ras, c-myc, c-neu, c-erbB-2, and c-met) and tumor suppressor genes (eg, p53) have been reported to be prevalent in bile duct cancers; the biological and clinical significance of these abnormalities remains to be characterized.17

Greater than 90% of bile duct cancers are adenocarcinomas. Other cancer types include squamous cell carcinoma, small cell carcinoma, and sarcomas. Adenocarcinomas of the bile duct are classified as sclerosing, nodular, or papillary (analogous to the classification scheme for gallbladder adenocarcinomas). Sclerosing (scirrous) tumors, which comprise over 80% of cholangiocarcinomas, are associated with an intense desmoplastic reaction, tend to be highly invasive, and are associated with low resectability rates. Nodular tumors have the appearance of constricting annular lesions and are also associated with low resectability rates. Papillary tumors are rare and present as bulky masses that project into the bile duct lumen. Because these lesions tend to cause symptomatic obstructive jaundice relatively early in their progression, they are associated with higher resectability rates than sclerosing or nodular tumors.

Cholangiocarcinomas are also classified into three groups according to their anatomical location: (1) intrahepatic or peripheral (10% of cases), (2) perihilar (65% of cases), and (3) distal (25% of cases). The transition between perihilar and distal locations occurs where the CBD becomes a retroduodenal structure. Bile duct tumors involving the hepatic duct bifurcation are known as Klatskin tumors. An additional anatomical classification system for perihilar cholangiocarcinomas, originally proposed by Bismuth,18 is useful in surgical planning (Table 51-4).

The seventh edition of the AJCC staging system, published in 2010, contains separate staging systems for intrahepatic (Table 51-5), perihilar (Table 51-6), and distal bile duct cancers (Table 51-7).6 These systems represent a significant departure from the AJCC sixth edition, in which intrahepatic cholangiocarcinomas were staged in the same manner as hepatocellular carcinomas, and all extrahepatic cholangiocarcinomas were grouped together in a single staging system.

Clinical Presentation and Diagnosis

Intrahepatic cholangiocarcinomas typically present with nonspecific symptoms, such as abdominal pain, anorexia, weight loss, and malaise. Another mode of presentation for these cancers is the incidental detection of an intrahepatic mass on imaging studies. The most common presentation of extrahepatic cholangiocarcinomas is painless jaundice. Other manifestations of biliary obstruction, such as acholic stools, dark urine, and pruritis, are also prevalent. Abdominal pain, fatigue, malaise, and weight loss can occur with advanced disease. Signs of advanced bile duct cancer include right upper quadrant abdominal tenderness, hepatomegaly, and a palpable gallbladder. Cholangitis is unusual in the absence of prior biliary tract instrumentation.

The differential diagnosis for patients with these presentations includes primary and metastatic hepatobiliary and pancreatic neoplasms, and benign biliary strictures due to conditions such as PSC, choledocholithiasis, Mirizzi's syndrome, and postoperative strictures.

In patients with intrahepatic cholangiocarcinoma, laboratory studies usually reveal an increased alkaline phosphatase level in the setting of normal bilirubin levels. In patients with extrahepatic cholangiocarcinoma, laboratory tests are usually consistent with the presence of obstructive jaundice. Tumor markers (eg, CEA, CA 19-9, and CEA in combination with CA 19-9) may have utility in surveillance of patients with PSC; however, their sensitivities and specificities are too low for them to be applicable to screening or diagnosis in the general population.

Transabdominal ultrasonography may reveal dilation of the biliary tree, which, in the absence of choledocholithiasis, suggests a possible biliary or pancreatic malignancy and should prompt contrast-enhanced spiral CT scanning. CT scan findings of intrahepatic cholangiocarcinomas include a liver mass with or without peripherally dilated ducts (Fig. 51-6). With perihilar cholangiocarcinomas, the primary tumor may not be visualized; their presence is suggested by the detection of dilated intrahepatic bile ducts (often bilateral), a normal or collapsed gallbladder (if the site of biliary obstruction of proximal to the cystic duct–bile duct confluence), a normal caliber distal CBD, and a normal pancreas. Findings of distal cholangiocarcionomas include dilation of intra- and extrahepatic bile ducts and the gallbladder, with or without a mass in the head of the pancreas. In addition to offering information on the site of the primary lesion, the CT scan offers valuable information necessary for staging and planning of therapies, including the presence or absence of local vascular invasion, regional lymphadenopathy, distant metastasis, and liver atrophy. Unilobar bile duct obstruction typically results in atrophy of the affected liver lobe together with hypertrophy of the unaffected lobe (atrophy-hypertrophy complex). Absence of the atrophy-hypertrophy complex can suggest vascular encasement by tumor.

For patients who are surgical candidates, an important goal of the preoperative evaluation is determining the proximal and distal tumor extent. If CT scanning fails to demonstrate the tumor itself (as is usually the case for resectable perihilar cholangiocarcinomas), additional imaging is helpful in surgical planning. In most centers, distal tumors are assessed by endoscopic retrograde cholangiopancreatography (ERCP, Fig. 51-7), whereas intrahepatic and perihilar tumors are best assessed by percutaneous transhepatic cholangiography (PTC). Recently, there has been increasing application of MRCP in this setting. Unlike conventional cholangiography, MRCP is noninvasive and does not require contrast material to be injected in the biliary ductal system. It also allows for visualization of the bile ducts both proximal and distal to a stricture. Some reports suggest that MRCP when applied to patients with cholangiocarcinoma offers information equivalent to that offered by CT scanning and traditional cholangiography combined.19 For these reasons, MRCP has been supplanting traditional cholangiography in the evaluation of patients with suspected cholangiocarcinoma in some centers.

Additional studies are not routinely indicated. The role of positron emission tomography (PET) scanning in the evaluation of patients with cholangiocarcinoma continues to be studied but is not yet established. If surgery is not planned, tissue diagnosis can be obtained through endoscopic or percutaneous biopsy. If surgical exploration is planned, a preoperative biopsy is not indicated.

A particularly challenging situation can arise in patients with PSC, 20–50% of whom will develop a benign dominant biliary stricture. These dominant strictures can be morphologically indistinguishable from cholangiocarcinomas on cholangiography. Cytological examination of brushings obtained during ERCP is the most common modality used for the diagnosis of cholangiocarcinoma in this setting; however, the sensitivity of this modality (40–80%) can be poor.20 The most accurate modality for diagnosing cholangiocarcinoma in patients with PSC who present with a dominant biliary stricture is endoscopic ultrasonography with fine-needle aspiration (EUS-FNA, sensitivity and specificity approaching 80 and 100%, respecively).21 EUS-FNA should be applied in patients with equivocal or negative brush cytological findings if clinical suspicion of cholangiocarcinoma being present in a dominant stricture is high.

Surgical Therapy

As is the case for gallbladder cancer, complete surgical resection is the only potentially curative therapy for patients with cholangiocarcinoma. Therefore, all patients suspected of having cholangiocarcinoma should be offered exploration unless contraindications to surgical resection exist. These contraindications include (1) major comorbidities precluding safe surgery, including cirrhosis, (2) metastatic disease, (3) invasion of the main portal vein or hepatic artery proximal to their bifurcations, (4) bilateral invasion of portal vein and/or hepatic artery branches, (5) bilateral hepatic duct involvement (up to secondary radicles bilaterally), and (6) unilateral duct and/or vessel involvement with contralateral liver lobe atrophy.

The utility of preoperative biliary stenting in patients with cholangiocarcinoma is controversial. Available retrospective data and one recently reported multicenter randomized controlled trial22 suggest that among patients undergoing pancreaticoduodenectomy for periampullary cancers, routine preoperative biliary stenting is associated with increased perioperative morbidity rates, especially with respect to infectious complications. Therefore, we do not recommend routine preoperative stenting for patients with distal bile duct cancers. Instead, selective application of stenting in patients with obstructive jaundice who experience significant delay until surgery is performed (eg, those undergoing neoadjuvant therapy) is appropriate. However, this experience should not be extrapolated to patients with perihilar cholangiocarcinomas, for whom the relationship between preoperative stenting and operative outcomes is less clear. Some authors believe stents placed preoperatively make intraoperative assessment of tumor extent more difficult. In our experience, bilateral Ring catheters, placed percutaneously into the left and right biliary systems shortly before the time of surgery, greatly facilitates the resection of perihilar cholangiocarcinomas. Our approach is described in detail later.

Surgical Technique

Resectable intrahepatic cholangiocarcinomas are treated using standard liver resections, and distal cholangiocarcinomas are treated by pancreaticoduodenectomy. These procedures are discussed elsewhere in this textbook. The following discussion focuses on our surgical approach to resectable perihilar cholangiocarcinomas.

We begin with exploratory laparoscopy to rule out the presence of disseminated disease that may have escaped detection on preoperative imaging. Reports suggest that 25–30% of patients undergoing laparoscopic exploration for cholangiocarcinoma are found to have unresectable disease during laparoscopy.23 If laparoscopic examination fails to reveal metastasis, we proceed with laparotomy through either an upper midline or a right subcostal incision (that can be extended to the left as necessary). We then conduct a thorough examination for the presence of distant metastases. Enlarged lymph nodes are biopsied and subject to frozen-section analysis. The presence of metastasis in N2 (periaortic, pericaval, superior mesenteric artery, and/or celiac artery) lymph nodes is a contraindication to radical resection.

Next, we lower the hilar plate by incising the liver capsule at the base of the quadrate lobe (segment 4) between the gallbladder fossa and the umbilical fissure (Fig. 51-8). This maneuver facilitates inspection of the porta hepatis. At this point we palpate the tumor in an attempt to assess its proximal and distal extent.

We then begin mobilization of the extrahepatic biliary tree from its surrounding structures. The gallbladder is mobilized, and the CBD is circumferentially dissected just proximal to where it assumes a retroduodenal location. We transect the CBD at this level and oversew the stump of the distal CBD with a nonabsorbable monofilament suture. We then dissect the extrahepatic biliary tree off of the underlying vascular structures, starting distally and working proximally (Fig. 51-9). During this step, cephalad and anterior traction is applied to the gallbladder, distal CBD, and the distal ends of the preoperatively placed Ring catheters (which in combination form a convenient handle that can be grasped). The bile duct and surrounding lymph node–bearing soft tissues should be cleared en bloc from the portal vein and the hepatic artery. Only after this step is accomplished the possibility of tumor vascular invasion is definitely eliminated.

The most difficult step in this dissection is usually encountered at the hepatic duct bifurcation, the site of Klatskin tumors. Dissection here is facilitated by placing vessel loops around the left and right hepatic ducts and placing them on traction as necessary. Because the left duct typically runs along the undersurface of the liver (segment 4) for a longer distance than the right duct, it is usually easier to dissect the left duct first and encircle it with a vessel loop prior to dissecting the right duct. We find that the Ring catheters are particularly helpful in the identification of the right and left hepatic ducts during this stage of the procedure. The resection is completed by transecting the biliary duct(s) proximal to the tumor (Fig. 51-10). Figure 51-11 shows the operative field following resection of a Klatskin tumor. The skeletonized hepatoduodenal ligament structures are visible. Frozen sections of the proximal and distal margins should be checked intraoperatively, with the goal of achieving negative microscopic margins (R0 resection).

Reconstruction following resection of Klatskin tumors consists of bilateral hepaticojejunostomies to a 60-cm retrocolic Roux-en-Y limb of jejunum. Small secondary or tertiary biliary branches should be incorporated into the anastomoses or ligated. Prior to performing the anastomoses, the Ring catheters are replaced with soft Silastic catheters (usually 14–18F) (Fig. 51-12). Catheter exchange is performed as follows: The Silastic catheters are placed over the distal ends of the cut Ring catheters (the portions protruding from the transected bile ducts), and the Ring and Silastic catheters are sewn together with cross sutures. The Ring catheters are then pulled proximally through the intrahepatic biliary tree and out the surface of the liver with the Silastic catheters attached. Finally, the Ring catheter is removed. If Ring catheters have not been placed preoperatively, the Silastic catheters can be placed as follows: Randall stone forceps are inserted into the intrahepatic biliary tree through the transected bile ducts and out the liver surface. The Silastic catheters are then sewn to the “eyelet” at the end of the forceps and pulled back down the duct. This maneuver is repeated so that a Silastic catheter is present in each of the right and left biliary systems.

The hepaticojejunostomies are created using a single layer of interrupted 5-0 absorbable monofilament sutures (Fig. 51-13). The posterior row of sutures are placed but not tied until the entire row can be “parachuted” closed. Using cautery, two small openings in the distal portion of the Roux limb are made, through which the distal ends of the Silastic catheters are placed. The anterior row of sutures are then placed and tied to complete the anastomosis (Fig. 51-14).

We then suture the Roux limb to the undersurface of the liver and to the mesocolon. We suture two large radiopaque clips to the surface of the liver at the sites where each of the Silastic tubes exit. These clips serve as permanent markers of the exit sites and allow for radiological visualization of the relationship between the liver surface and the last radiopaque marker on the Silastic catheters.

Recently, more aggressive approaches that include the routine application of liver resection, and portal vein resection in select cases, are being reported with increasing frequency. Addition of hepatic resection can extend the possibility of R0 resection to patients with Bismuth type III lesions (Fig. 51-15). Because Bismuth types II and III tumors may involve the caudate lobe ducts, some authors recommend routine caudate lobectomy when resecting these lesions. Although the highest 5-year postoperative survival rates have been reported from centers using such aggressive surgical approaches, these extended procedures should be done only if they can be performed with low perioperative morbidity and mortality rates. In addition, some centers have reported the application of preoperative portal vein embolization, to induce lobar hypertrophy and thereby extend the limits of liver resection in patients at risk of developing hepatic insufficiency postoperatively.

Finally, orthotopic liver transplantation has been applied to patients with intrahepatic and perihilar cholangiocarcinomas. However, cancer recurrence occurs in over 50% of cases, and 5-year survival rates average only 22%. Long-term survivors have been reported; most of these patients had small, peripheral cholangiocarcinomas discovered incidentally. For patients with known cholangiocarcinoma, liver transplantation following neoadjuvant therapy in carefully selected and staged patients is being studied, with some promising initial results.24 This form of therapy should not be offered outside the context of a study protocol.

Adjuvant Therapies

Adjuvant chemotherapy, radiotherapy, or chemoradiotherapy is commonly offered, based on results of retrospective series. However, clear efficacy data derived from prospective randomized clinical trials are lacking. Similarly, neoadjuvant therapy, associated with anecdotal reports of tumor response sufficient to permit margin-negative resection in patients with advanced cholangiocarcinoma, needs to be studied further.

Palliation

The major goal of palliation is relief of symptoms of biliary obstruction. Endoscopic or percutaneous biliary stenting is associated with less morbidity than surgical biliary bypass and is therefore recommended except in patients who are found to have unresectable disease at the time of surgical exploration or those in whom nonsurgical stenting cannot be accomplished. Endoscopic stenting is the preferred approach for distal bile duct cancers; proximal cancers are more difficult to stent endoscopically and usually require a percutaneous approach.

Patients with a Bismuth type I hilar cholangiocarcinoma are usually palliated effectively with a single biliary stent. Patients with Bismuth types II, III, and IV hilar cholangiocarcinomas may require two or more separate stents to decompress the entire biliary tree and prevent obstruction-related cholangitis. However, in a prospective, randomized controlled trial of patients with hilar cholangiocarcinoma, unilateral biliary drainage was found to provide adequate palliation of obstructive jaundice; patients randomized to receive bilateral biliary stents had higher complication rates (cholangitis) but no detectable benefits.25 The approach therefore needs to be individualized.

Metal stents tend to provide more durable palliation than plastic (polyethylene) stents (median stent patency of 8–12 vs 4.8 months) and are generally preferable in patients with malignant biliary obstruction. Plastic stents should be changed every 3–6 months to prevent episodes of cholangitis related to stent occlusion; these stents may be appropriate for patients with estimated survival durations of less than 3 months (eg, patients with diffuse metastases).

For patients who are found to have carcinomatosis at the time of exploratory laparoscopy, laparoscopic cholecystectomy traditionally has been recommended, to prevent subsequent development of acute cholecystitis related to biliary stent–induced cystic duct obstruction. The value of prophylactic cholecystectomy in this setting is unproven and should be performed only if it can be done safely. Stenting should be performed using percutaneous or endoscopic techniques postoperatively.

For patients who are found to have unresectable disease at the time of open exploration, available retrospective evidence suggests that surgical biliary bypass offers more durable palliation than percutaneous or endoscopic stenting. Patients with unresectable distal cholangiocarcinoma should undergo choledocho- or hepaticojejunostomy. The palliative options for patients with unresectable perihilar cholangiocarcinoma include (1) tumor debulking with Roux-en-Y hepaticojejunostomy and intraoperative placement of Silastic transhepatic catheters (as described earlier) and (2) Roux-en-Y hepaticojejunostomy using the segment 4 or 5 duct. Segment 3 or 5 bypass is used in patients with advanced perihilar cholangiocarcinoma with predominantly right- or left-sided disease, respectively. The segment 3 hepatic duct is approached by following the falciform ligament into the recess of the left lobe in the umbilical fissure (Fig. 51-16). Localization of the segment 5 duct is difficult, as no external anatomic landmarks exist and considerable parenchymal dissection if often necessary. Intraoperative ultrasonography (IOUS) considerably facilitates this procedure. These unilateral bypasses should be avoided in the presence of ipsilateral liver lobe atrophy, a finding that indicates limited functional hepatic parenchyma.

External beam radiation and transcatheter brachytherapy may contribute to pain relief and biliary decompression; however the data on the effects of radiation on survival duration are conflicting.

Recently published results of the ABC-02 trial (discussed earlier in the section on palliation of patients with advanced gallbladder cancer) indicate that the combination of gemcitabine plus cisplatin should be offered to patients with advanced bile duct cancer.13 Nearly 60% of patients who were enrolled in this multicenter phase III trial had locally advanced or metastatic bile duct cancer. Administration of the gemcitabine-cisplatin combination was associated with prolongation of overall and progression-free survival compared to administration of gemcitabine alone.

Finally, photodynamic therapy (PDT), in which endoscopic application of light activates a photosensitizer, leading to local cell death, has been associated with promising results. One prospective randomized trial, in which 19 patients with advanced cholangiocarcinoma were randomized to stenting alone or stenting followed by PDT, was terminated prematurely because patients randomized to the PDT arm were found to have a significantly longer survival (493 vs 98 days, median survival) in addition to improved biliary drainage and quality of life.26 PDT-associated prolongation of survival was observed in another small prospective randomized trial.27 Additional study of this modality is warranted.

Outcomes

Less than 50% of patients diagnosed with perihilar cholangiocarcinoma are able to undergo curative resection. Reported 5-year postoperative survival rates for patients with these cancers are highly variable; they range from 8 to more than 50%.19 In general, the highest survival rates are associated with series containing a high proportion of cases in which R0 resection was achieved. Series containing the highest R0 resection rates (>75% of cases in some published experiences) tend to be reported by institutions where liver resection is applied liberally to patients with cholangiocarcinoma.19 A caveat that should be remembered is that these same series also tend to be associated with the highest perioperative mortality rates (up to 10% in some cases).

For patients with intrahepatic cholangiocarcinoma, reported 3-year survival rates following curative resection with negative margins range from 22 to 66%. For patients with distal cholangiocarcinoma, 5-year survival rates following pancreaticoduodenectomy range from 15 to 25% in most reported series. Among patients with node-negative disease, 5-year postoperative survival rates as high as 54% have been reported.

The best reported outcomes among patients with unresectable biliary tract cancers are those from the ABC-02 trial.13 The median overall survival among patients treated with the combination of gemcitabine and cisplatin was 11.7 months, whereas it was 8.1 months in those treated with gemcitabine alone.13