Figure 63-1. Unresectable gallbladder cancer demonstrating palliative transhepatic percutaneous stent placed to relieve jaundice. Porcelain gallbladder is present (arrows).
Pathology and Staging
More than 80% of gallbladder cancers are adenocarcinomas; there are several histologic subtypes, including papillary, nodular, and tubular. Papillary tumors, which grow predominantly into the gallbladder lumen, have an improved prognosis compared with the other subtypes.12 Poor prognostic signs in gallbladder cancer include grade, and vascular invasion.12,13 The most important prognostic factor may be lymph node status, although 5-year survivors with nodal involvement have been documented.14,15 Less than 5% of cases are squamous cell carcinomas, with the remaining 10% being anaplastic lesions.
Gallbladder cancer spreads via the lymphatic and venous drainage. Because of drainage of the cholecystic veins directly into the adjacent liver, these tumors often involve hepatic parenchyma, most often portions of segments IV and V. Lymphatic spread is first to the cystic duct (Calot) node, then to pericholedochal and hilar nodes, and finally to peripancreatic, duodenal, periportal, celiac, and superior mesenteric artery nodes. Nodal disease in the porta hepatis often causes common bile duct (CBD) obstruction and resultant jaundice, which is the first clinical symptom in 30% of patients. Jaundice may also be caused by tumors arising in the infundibulum, which may spread directly to the cystic duct and common hepatic duct. Although peritoneal metastases are frequent, distant extraperitoneal metastases are not.
Limited information exists regarding the genetic changes in gallbladder cancer. Recently reported gene abnormalities associated with gallbladder cancer include p53, K-ras, CDKN2 (9p21), FGFR, and PI3 kinase pathway aberrations.16–20 The finding that patients with an anomalous pancreaticobiliary junction have a greater frequency of K-ras mutations has led investigators to believe that reflux of pancreatic enzymes into the biliary tree may contribute to the development of cancer.16 Because of our limited knowledge of the sequence of molecular changes, there are no known detectors of early disease or of risk assessment. Clearly, this is an area that needs improvement, particularly in endemic areas.
The American Joint Committee on Cancer’s (AJCC) seventh edition of its tumor, node, metastasis (TNM) staging system (Table 63-1) reflects prognostic characteristics of tumor depth, regional nodal disease, and distant spread. The gallbladder differs histologically from the rest of the gastrointestinal tract in that it lacks a muscularis mucosa and submucosa. The gallbladder wall is composed of (a) a single layer of columnar cells, the mucosa, and lamina propria; (b) a fibromuscular layer; (c) a perimuscular, subserosal layer containing lymphatics and neurovascular structures; and (d) a serosal surface, except where the gallbladder is embedded in the liver. Because lymphatics are present in the subserosal layer only, tumors invading less than the full thickness of the muscular layer have minimal risk of nodal spread. Thus, disease invading into, but not through, the muscular layer of the gallbladder is T1b disease (stage I), whereas invasion into the perimuscular connective tissue is T2 (stage II). Stage III disease includes tumors that have perforated the serosa or have directly invaded the liver or other surrounding structures (T3), which are clearly more advanced but still potentially resectable. Tumors that invade the main portal vein, hepatic artery, or two or more extrahepatic organs/structures are classified as T4 (stage IVA) and are typically unresectable. Lymph node metastasis to regional, periportal nodes are N1 whereas those to more distant nodes outside of what would be included in a standard resection (celiac, periaortic, and superior mesenteric nodes) are N2. N1 nodal metastasis is classified as stage IIIB and N2 as stage IVB. In line with other pancreaticobiliary malignancies, the stage III grouping refers to locally advanced disease and stage IV indicates metastatic disease.
Clinical Findings and Diagnosis
Most patients are found to have gallbladder cancer during workup or treatment of cholelithiasis or choledocholithiasis. In patients with symptoms, abdominal pain consistent with biliary colic or acute cholecystitis is most common. Patients also present with jaundice, weight loss, anorexia, or an increase in abdominal girth secondary to ascites. Physical findings may include right upper quadrant tenderness or a palpable mass, hepatomegaly, and ascites. Laboratory investigation, if abnormal, is most often consistent with biliary obstruction. Because of its nonspecific presentation and the lack of reliable screening tests, gallbladder cancer is not diagnosed preoperatively in more than half the cases.
Imaging evaluation often reveals a thickened gallbladder wall or a mass within or replacing the gallbladder on ultrasound examination. Because polyps, asymmetric wall thickening from cholecystitis (especially xanthogranulomatous type), and carcinoma can have an echogenicity similar to the gallbladder wall, these lesions are often difficult to distinguish. This is even more difficult when inflammation is present from gallstones. At times, ultrasound can visualize invasion of the liver, adjacent adenopathy, and a dilated biliary tree. The ability of ultrasound to differentiate benign from neoplastic disease is enhanced using endoscopic ultrasound, and may be more specific than computed tomography (CT) or magnetic resonance imaging (MRI).21–23
A dynamic contrast-enhanced CT scan may identify a gallbladder mass or invasion into the liver parenchyma or adjacent organs. The classic finding in a patient with gallbladder cancer is asymmetric thickening of the gallbladder wall. Staging of gallbladder carcinoma using CT, however, is limited by poor sensitivity in identifying nodal spread.24
In patients who are jaundiced, direct cholangiography may be useful to delineate the extent of biliary involvement as well as to palliate symptoms of biliary obstruction. A mid–bile duct obstruction not caused by gallstones is gallbladder cancer until proved otherwise (Fig. 63-2). More recently, with improvements in MRI technology, diffusion-weighted imaging and magnetic resonance cholangiopancreatography (MRCP) have evolved into a single, noninvasive imaging modality that allows complete assessment of biliary, vascular, hepatic parenchymal, and nodal involvement, as well as involvement of adjacent organs (Fig. 63-3).25–27
Table 63-1 American Joint Committee on Cancer, 7th Edition, Staging System for Gallbladder Carcinoma
In patients who present with incidentally discovered gallbladder cancer following cholecystectomy, staging should be performed with high-quality CT or MRI of the chest, abdomen, and pelvis. There is no clear role for PET/CT in this setting.28
5 It is clear that the only curative option in patients with gallbladder cancer is complete surgical resection. It is essential for optimal patient care that patients with gallbladder cancer be recognized before laparoscopic cholecystectomy is performed, because of the risk of bile spillage, with its potential for subsequent carcinomatosis.29
Role of Staging Laparoscopy
Because a large percentage of patients with gallbladder cancer are found to have occult, unresectable disease at the time of surgical exploration, several authors have investigated the use of initial staging laparoscopy for this disease.30–33 Because gallbladder cancer has such a propensity to spread intra-abdominally, this tumor is ideal for detection of intra-abdominal metastases with laparoscopy. This is demonstrated by the fact that up to 50% of patients are found to have unresectable disease at the time of laparoscopy.34,35 Patients who are found to have unresectable disease at laparoscopy can begin other forms of systemic therapy earlier and may undergo the procedure as an outpatient. Particularly because patients with unresectable disease have a median survival of only 6 months, the impact on quality of life, including decreased length of stay in the hospital, cannot be overemphasized.
Figure 63-2. Endoscopic retrograde cholangiopancreatogram obtained from a patient with gallbladder cancer. Mid–bile duct obstruction (arrow) is caused by direct extension of tumor to the cystic and common hepatic duct.
Cholecystectomy With or Without Partial Hepatectomy
Gallbladder cancer, if not completely surgically removed, results in rapid local progression and death. In a collected review of 5,836 patients with gallbladder cancer, the overall mean survival was between 2 and 5 months, whereas the 5-year survival was 4%.5 The 5-year survival of patients undergoing resection with curative intent was 17%. Of the 2,115 patients with unresectable disease, only a single survivor was found at 5 years. Although surgical resection represents the treatment of choice and the only potentially curative therapy available, resection is possible in only 25% of patients at presentation because of the advanced nature of the disease.5
Figure 63-3. T1-weighted magnetic resonance imaging scan of a patient with gallbladder cancer (small arrows) with extension into the duodenum and the hepatic flexure of colon (large arrows).
There is little doubt that the results of treatment, as well as the scope of operation, are related to depth of tumor penetration (Table 63-2). For tumors limited to the muscular layer of the gallbladder (T1), there is near-universal agreement that simple cholecystectomy is adequate.15,36–38 T1 tumors have not yet invaded the subserosal layer, which contains lymphatics, and therefore lymphadenectomy is not required. Attesting to the fact that early gallbladder carcinoma is completely curable, simple cholecystectomy has resulted in 90% to 100% survival when early cancer is an incidental finding after elective cholecystectomy.15,47
Table 63-2 Five-Year Survival After Resection for Gallbladder Cancer
Difficulty can arise at the time of surgery in evaluating polypoid lesions of the gallbladder as either benign or early gallbladder cancer. Although it appears that frozen section diagnosis is fairly reliable in distinguishing whether lesions are malignant or benign (95% accurate), the accuracy in correctly assessing depth of invasion is only 70%.47 Thus, it may be difficult at the time of surgery to determine the extent of resection. Because of this, pursuing a more aggressive resection if the depth of invasion is in doubt is important for adequate tumor clearance.
The extent of surgical resection for T2 or greater tumors is controversial, with recommendations ranging from simple cholecystectomy to radical excision, including hepatectomy. For advanced local disease, some groups have advocated radical resections, including hepatectomy and pancreatectomy. Whereas it is clear that major hepatic resection can be performed safely with a mortality less than 5% (Table 63-3) it has not been universally accepted that more aggressive resections, such as combined hepatectomy and pancreaticoduodenectomy, improve survival.14,15,36,37,40,48
To understand the rationale for extensive resections, it is also important to understand the pattern of spread of gallbladder cancer. Direct extension to the adjacent liver parenchyma often occurs first, followed by adjacent organ involvement, including duodenum, colon, and stomach (Fig. 63-2). Lymphatic spread of gallbladder cancer is routine, often involving nodes in the porta hepatis, peripancreatic region, celiac axis, and the aortocaval nodal basins. There is currently no consensus on the extent of lymphadenectomy that is appropriate for gallbladder cancer. Extensive lymphadenectomy including para-aortic nodes is advocated in the East, but is not universally accepted. Recently, a technique of intraoperative nodal staging was proposed that entails sampling the highest peripancreatic lymph node for frozen section. This is the lymph node situated between the N1 and N2 nodal stations at the junction of the bile duct and the superior border of the head of the pancreas (Fig. 63-4). Metastatic disease to this node has been shown to be a poor prognostic factor.49
Figure 63-4. Illustration of the highest peripancreatic lymph node.
Because the gallbladder is not surrounded by serosa where it is attached to the liver in the gallbladder fossa, even T2 tumors (full-thickness invasion of the muscular layer into the perimuscular connective tissue, no extension beyond serosa or into liver) can invade the normal plane of dissection in the gallbladder fossa during simple cholecystectomy. Therefore, T2 tumors cannot be completely removed with cholecystectomy alone, and a radical cholecystectomy, with resection of a 1- to 2-cm rim of normal liver around the gallbladder fossa, is the minimal resection that is required. Many authors, however, have found that segmental resection of segments IVb and V of the liver, which abut the gallbladder fossa, results in a more anatomically controlled dissection with less blood loss.50 An additional part of the definitive surgical treatment is regional lymphadenectomy, because about half the patients with T2 tumors are found to have nodal spread after resection.14 Dissection of lymph nodes should include all tissues from the bifurcation of the hepatic ducts to the distal CBD. Proponents of this approach advocate liver resection on the basis that it is the only way to obtain an adequate margin on the hepatic side of the gallbladder and resection of the regional nodes allows the best chance for complete tumor clearance. For all of these reasons, simple cholecystectomy is inadequate for T2 or greater tumors. When larger anatomic hepatic resections have been performed in patients with T2 tumors, it has increased the 5-year survival from 25% to 40% after simple cholecystectomy to 70% to 100% after radical resection.14,36–38,39,43,44,51,52
For T3 and T4 lesions, there is a high likelihood of intraperitoneal and hematogenous spread and significant morbidity from the radical procedures that are often necessary for excision of local disease. Recent series, however, support an aggressive approach to resection of these large tumors, particularly if no indication of nodal involvement is found (Table 63-3). For T3 and resectable T4 tumors, a minimal resection includes segments IVb and V, and in many cases an extended right hepatectomy (segments IV, V, VI, VII, and VIII) may be necessary to obtain complete resection. With aggressive resection, long-term survival can be achieved even for patients with these more advanced tumors.14,15,36,37,40,53
Surgical exploration should be performed for all patients without medical contraindications. If a T1 tumor is suspected, a cholecystectomy and biopsy of regional nodes should be performed after thorough examination of the abdominal cavity for any signs of tumor dissemination. The pathology and depth of penetration should be confirmed by frozen section, and the procedure terminated if a T1 tumor with negative margins is confirmed. For T2 lesions, either a radical cholecystectomy (wedge resection of the hepatic bed) or a segment IVb and V resection with lymphadenectomy should be performed.14 For T3 lesions, a segment IVb and V resection or extended right hepatectomy is performed. Finally, for T4 lesions, a more radical excision of the liver, such as extended right hepatectomy, usually must be performed for adequate tumor clearance.
Tumor location may be important in determining the extent of resection. If the tumor arises in the gallbladder infundibulum, the CBD is often involved with tumor, either by direct extension or external invasion of the hepatoduodenal ligament. In this case, an extended liver resection and removal of a portion of the CBD should be performed. Reconstruction is then performed by Roux-en-Y hepaticojejunostomy. Tumor arising in the fundus of the gallbladder, however, can be treated with limited hepatic resection without excision of the CBD. Complete regional lymphadenectomy should be performed, skeletonizing the CBD, hepatic artery, and portal vein. Short-term postoperative outcomes following resection of gallbladder cancer have remained relatively stable over time, with postoperative morbidity occurring in 30% to 40% of patients, and mortality occurring in <5% (Table 63-4).
Incidentally or Laparoscopically Discovered Gallbladder Cancer
Gallbladder cancer is often discovered during pathologic examination after cholecystectomy for presumed benign disease. Because of the popularization of laparoscopic cholecystectomy in the past decade, an increasing number of patients with gallbladder cancer are found incidentally. Patients with T1b or greater tumors and no signs of distant disease should be offered exploration and further resection to eradicate all diseases. In a series of 135 patients who underwent reexploration following laparoscopic cholecystectomy, 61% of all and 36% of those with T1b primary tumors were found to have residual disease.54 In this study, survival following reresection of residual locoregional disease was extremely poor and was not different from that of patients with metastatic disease. Improved survival has been demonstrated following reresection to achieve negative margins, however, especially in patients with T2 or T3 disease.51,55,56 While laparoscopic port-site recurrence has been reported, it is associated with peritoneal-based disease. Port-site excision at the time of reexploration is no longer recommended as recent data have shown that it is not associated with decreased recurrence or improved survival.57,58
Reresection after recent cholecystectomy is often technically challenging. Postoperative inflammation in the right upper quadrant often hinders distinction of tumor from normal tissue. Bile spillage at the time of the initial operation may result in carcinomatosis.59 Determination of ductal or nodal involvement by tumor is always difficult at the time of reoperation. In addition, postoperative fibrosis often encases the right hepatic artery, which crosses behind the bile duct in most patients. Because of this, during a second operation for incidentally discovered gallbladder cancer, an extended right hepatectomy along with excision of the extrahepatic biliary tree and portal lymphadenectomy is often necessary. This resection allows adequate exposure for lymphadenectomy and greater confidence of a negative margin on the bile duct, and also permits biliary reconstruction to only one side of the liver. The disadvantage is that a large portion of normal liver parenchyma is sacrificed, and consequently, transient postoperative liver dysfunction is common. Although it may be more difficult to curatively resect disease in patients with incidentally discovered gallbladder cancer after laparoscopic cholecystectomy, there is no difference in overall survival between patients with incidentally discovered gallbladder cancer who are submitted to curative resection and those patients who undergo initial curative resection.50
When a patient presents with T1a gallbladder cancer discovered after simple cholecystectomy, the pathology should be reviewed to determine if the entire gallbladder has been removed and if the cystic duct margin is clear of tumor. If the cystic duct margin is positive, the patient requires bile duct excision. If all margins are negative, no further therapy is warranted. If the tumor is proved to be T1b or greater, complete staging should be performed. In the absence of metastatic disease, patients should be counseled regarding reexcision to attempt complete resection, chemotherapy with or without radiation, or observation. Patients with a known or suspected early gallbladder carcinoma should be referred to an experienced center where curative-intent resection can be performed at the initial operation.
Adjuvant therapy for gallbladder cancer remains a controversial and unproved consideration and is rarely utilized.45 Very few randomized trials have been completed, and the conclusions that can be drawn from them are limited given the small sample sizes. Given the relative rarity of these malignancies in the United States, large-scale, randomized trials are feasible only in the context of a multi-institutional or cooperative group setting.
In 2002, a randomized phase III trial of adjuvant chemotherapy with 5-fluorouracil and mitomycin C versus surgery alone for patients with pancreaticobiliary malignancies having resection found that in the subset of patients with gallbladder cancer (n = 112), the 5-year survival rate was significantly better in the adjuvant group (26%) versus the control group (14%).60 Similarly, the 5-year disease-free survival rate was 20.3% versus 11.6%, clearly favoring the adjuvant therapy group. A recent review of the SEER database also reported an association between adjuvant chemoradiation therapy and improved survival in patients with locoregionally advanced disease.61 This finding prompted a multi center phase II trial of adjuvant capecitabine and gemcitabine followed by concurrent capecitabine and radiation therapy for patients with resected gallbladder cancer or extra hepatic cholagiocarcinoma. The results of this study were published in 2015 and showed promising findings of 65% two-year survival.62 These studies suggest that patients with gallbladder cancer with high risk for recurrence (T2 or greater, node-positive, or margin-positive) should be considered for adjuvant treatment with systemic chemotherapy or chemoradiation.
Table 63-4 Complications Postoperative Morbidity Following Resection of Gallbladder Cancer Over Time in the United States56
The 5-year survival rate for all patients with gallbladder cancer is less than 5% in most series, with a median survival of 6 months. This is primarily because most patients present with unresectable disease. Of those patients undergoing resection, survival is dependent on depth of penetration and nodal status. Nearly 100% survival is reported after simple cholecystectomy for T1a disease, whereas patients with T2 and T3 tumors without nodal disease have a 5-year survival greater than 50%.14,15,40,52 Node positivity is an ominous finding, with few series reporting 5-year survivors.
Follow-up after Resection for Gallbladder Cancer
The most common sites of recurrence after resection of gallbladder cancer include carcinomatosis, intrahepatic metastases, or nodal recurrence in the retroperitoneum. For most tumors, local recurrence is found synchronously with diffuse intra-abdominal spread. Therefore, surgical treatment of recurrence has little potential for cure. The main goal of surgery after recurrence of resected gallbladder cancer is to provide palliation of symptoms such as pruritus or cholangitis associated with jaundice, or bowel obstruction associated with carcinomatosis. When jaundice or cholangitis is the presenting symptom of possible recurrence, a nonsurgical palliative approach using percutaneous transhepatic cholangiogram (PTC) and stenting is usually favored unless a benign postsurgical stricture is suspected. Because of the rapid growth of tumor in patients with recurrence, the hospitalization and recovery time from a surgical bypass is usually not justified for recurrences resulting in biliary obstruction.
The routine follow-up of a patient after resection of gallbladder cancer includes office visits every 3 months with physical examination and measurement of liver function tests, and cross-sectional imaging every 3 to 6 months for the first 2 years. In patients who remain free of disease at 2 years, follow-up should be continued on an individualized basis. Although CA19-9 may be elevated in patients with gallbladder cancer, the sensitivity and specificity are poor and, thus, should not be used for screening patients for recurrence.63 If recurrence is identified, systemic therapy with gemcitabine and cisplatin should be considered as it has been shown to prolong survival in the setting of metastatic disease.64 For patients who cannot tolerate this regimen, alternatives include single-agent gemcitabine or gemcitabine plus capecitabine.65
Issues for the Future
Clearly, improving our ability to recognize early gallbladder cancer in high-risk geographic areas would have an important impact on outcome in these patients. This will likely require implementation of screening programs in high-risk areas, which could result in prophylactic cholecystectomy.66 Additionally, standardization of minimum pathologic assessment of gallbladder specimens in high-risk areas is important to allow for accurate diagnosis, staging, and treatment of patients.
Further studies on characterization of molecular aberrations in gallbladder cancer by next generation sequencing and other technologies may lead to the discovery of targetable mutations and lead to the development of novel therapies. Additionally, there is a theoretical role for neoadjuvant therapy for patients with locoregionally advanced gallbladder cancer given their poor prognosis following surgery. Clinical trials in this area are needed.
BILE DUCT CARCINOMA
Bile duct carcinomas may arise within the liver (intrahepatic cholangiocarcinoma [ICC] or peripheral cholangiocarcinoma), at the liver hilum (hilar cholangiocarcinoma or Klatskin tumor), or in the extrahepatic bile duct (extrahepatic cholangiocarcioma [ECC]). While all arise from the biliary epithelium, the location of these tumors affects prognosis as well as the potential for curative resection.
Resection of biliary neoplasms, particularly hilar cholangiocarcinoma, often requires radical resections and complex biliary reconstructions that should be performed only at high-volume, experienced centers. There is also a role for surgery in palliation for these cancers by providing biliary bypass for jaundiced patients with unresectable tumors. Because disease is often diagnosed late in the course and because complex operative techniques are required for potentially curative resection, these tumors represent one of the greatest challenges for definitive treatment. Adding to this is that there are no proven effective options for adjuvant therapy.
The incidence of ICC in the United States is approximately 0.7/100,000 with a similar mortality. During the last 30 years, it appears that both the incidence and mortality in the United States are increasing.64 The overall incidence of hilar cholangiocarcinoma, the most common type, is 1.0/100,000 per year in the United States, although rates are higher in other geographic regions such as Israel and Japan.68 Recent population studies have noted a trend toward a relative increased incidence of ICC compared to ECC.4,69,70 Using the Surveillance, Epidemiology and End Results-Medicare databases, Welzel et al.70 noted HCV infection, chronic nonalcoholic liver disease and obesity, and smoking being associated only with ICC and not ECC, possibly explaining the divergent trends in incidence. Cholangiocarcinomas arise slightly more often in males, with a male-to-female ratio of 1.3:1 and an average age of 50 to 70 years.
Known risk factors for cholangiocarcinoma include primary sclerosing cholangitis, ulcerative colitis, choledochal cysts, and biliary tract infection, either with Clonorchis or in chronic typhoid carriers.71 Some industrial chemicals (e.g., nitrosamines, dioxin, asbestos, and polychlorinated biphenyls) have also been implicated in the pathogenesis of cholangiocarcinoma. Although there has been some suggestion of an increased risk of cholangiocarcinoma arising after transduodenal sphincteroplasty,72 it is difficult to determine if this is caused by the surgical intervention or the underlying disease leading to sphincteroplasty.
Pathology and Staging
More than 90% of these bile duct cancers are adenocarcinomas. They are morphologically described as nodular, which is the most common, scirrhous, diffusely infiltrating, or papillary. Histologic subtypes include acinar, ductular, trabecular, alveolar, and papillary. Papillary tumors have an improved outcome. Much less common types of bile duct tumors include cystadenocarcinomas, hemangioendotheliomas, and mucoepidermoid carcinomas.
Table 63-5 American Joint Committee on Cancer, 7th Edition, Staging System for Perihilar Bile Duct Carcinoma
In patients with ICC, negative prognostic signs include vascular invasion, multiple tumors, and lymph node metastases, but not tumor size.73 These tumors can be sclerotic, masslike, or cystic lesions.
Historically, ECCs have been classified according to their location in the upper (60%), middle (15% to 20%), or lower third (15% to 20%) of the bile duct. Middle-third lesions arise between the cystic duct and the superior border of the duodenum. Lower-third lesions are found below the superior border of the duodenum but above the ampulla. The problem with this classification is that the anatomic landmarks are somewhat arbitrary and not clinically useful. Most mid–bile duct malignant obstructions are caused by gallbladder cancer. A more useful classification is to divide these lesions into upper-half or lower-half tumors, based on the location of the cystic duct as it enters the common duct (in the case of normal anatomy). The usefulness of this classification scheme is that it allows the surgeon to delineate whether a hepatic or pancreatic resection will be required for clearance of tumor. The AJCC TNM staging system (seventh edition) for bile duct cancers is described in Tables 63-5 and 63-6.
Other staging systems have been created that attempt to incorporate clinically important indicators of resectability for hilar cholangiocarcinoma that are defined preoperatively, including hepatic lobe atrophy or portal vein involvement.74 With the increasing acceptance of major hepatic resection for these tumors, this preoperative staging system attempts to define whether there is ipsilateral involvement alone, because tumors with bilateral extension past the primary biliary radicles are not resectable.
Table 63-6 American Joint Committee on Cancer, 7th Edition, Staging System for Distal Bile Duct Carcinoma