Asia–Pacific consensus recommendations for endoscopic and interventional management of hilar cholangiocarcinoma


  • Declarations of conflict of interest: All authors have nothing to disclose.
  • Author contributions: RR was involved with the consensus concept and design, acquisition of data, drafting of the manuscript, critical revision of the manuscript, obtain funding, and participation in electronic and face-to-face voting. PA and TR were involved with the consensus concept and design, acquisition of data, drafting of the manuscript, critical revision of the manuscript, and participation in electronic and face-to-face voting. CK, RP, JM, DS, LP, AS, PP, TA, DR, AM, SP, PK, MR, and MK were involved with drafting of the manuscript, critical revision of the manuscript, and participation in electronic and face-to-face voting. TI was involved with drafting of the manuscript, critical revision of the manuscript, and participation in electronic voting. NP, ST, SA, BD, HW, EO, DL, and PM were involved with critical revision of the manuscript, and participation in electronic and face-to-face voting. KG and BO were involved with critical revision of the manuscript and participation in electronic voting. SC and AJ were involved with study supervision and participation in electronic voting. VB was involved with study supervision.
  • Funding resources: An unrestricted education grant was donated from Boston Scientific, Asia–Pacific. The Thai Association of Gastrointestinal Endoscopy supported for all local logistics and pre-consensus preparation. The industrial sponsor did not participate in the literature collection, consensus discussion, voting, and manuscript preparation and it chose to remain anonymous.
  • The consensus statements have been presented during the Asian Pacific Digestive Week (APDW) 2012, in Bangkok, Thailand, and part of the manuscript has been published in the Journal of Gastroenterology and Hepatology supplement issue for APDW.


Dr Rungsun Rerknimitr, Division of Gastroenterology, Department of Internal Medicine, Faculty of Medicine, Chulalongkorn University, Bangkok 10310, Thailand. Email:


Hilar cholangiocarcinoma (HCCA) is one of the most common types of hepatobiliary cancers reported in the world including Asia–Pacific region. Early HCCA may be completely asymptomatic. When significant hilar obstruction develops, the patient presents with jaundice, pale stools, dark urine, pruritus, abdominal pain, and sometimes fever. Because no single test can establish the definite diagnosis then, a combination of many investigations such as tumor markers, tissue acquisition, computed tomography scan, magnetic resonance imaging/magnetic resonance cholangiopancreatography, endoscopic ultrasonography/intraductal ultrasonography, and advanced cholangioscopy is required. Surgery is the only curative treatment. Unfortunately, the majority of HCCA has a poor prognosis due to their advanced stage on presentation. Although there is no survival advantage, inoperable HCCA managed by palliative drainage may benefit from symptomatic improvement. Currently, there are three techniques of biliary drainage which include endoscopic, percutaneous, and surgical approaches. For nonsurgical approaches, stent is the most preferred device and there are two types of stents i.e. plastic and metal. Type of stent and number of stent for HCCA biliary drainage are subjected to debate because the decision is made under many grounds i.e. volume of liver drainage, life expectancy, expertise of the facility, etc. Recently, radio-frequency ablation and photodynamic therapy are promising techniques that may extend drainage patency. Through a review in the literature and regional data, the Asia–Pacific Working Group for hepatobiliary cancers has developed statements to assist clinicians in diagnosing and managing of HCCA. After voting anonymously using modified Delphi method, all final statements were determined for the level of evidence quality and strength of recommendation.


Hilar cholangiocarcinoma (HCCA) is one of the most common type of bile duct cancers reported in the world, and the Asia–Pacific region reported the highest incidence.[1] To date, there have been a few guidelines for investigations and management of HCCA.[2-4] After the latest guideline,[4] the techniques in the subject of endoscopy and interventional management have been evolved, but there has been no update in the consensus or guideline and only a handful number of reviews are available.[5, 6] The Asia–Pacific Working Group on hepatobiliary cancers was established in 2011 under the auspices of the scientific organizing committee for the Asian Pacific Digestive Disease Week 2012. The Working Group felt that HCCA is the unique type of Asian hepatobiliary cancer that needs to be addressed specifically. Therefore, the goal of this Consensus was to establish recommendations and managements of HCCA with specific relevance to Asian data on the course, standard approach, and recent advances in the management of HCCA. Because the role of curative surgery requires detailed explanation as described elsewhere[7, 8] and the techniques are so variable depending on expertise of each operator. After a comprehensive discussion, the group has considered to omit the statement on this part.


A modified Delphi process was performed to establish the consensus.[9] The process relied on a combination of the principles of evidence-based medicine through an anonymous voting system. The Consensus Panel opinions were convinced by a systemic literature review. The main stream of the issues was determined according to perceived clinical importance particular to the Asia–Pacific region. A planning group panel (RR, PA, ST, TR) generated a list of statements and distributed it electronically in advance to all Consensus members. The statements were divided into the topics of: epidemiology and nature, histology and tumor markers, cholangioscopy and image enhancement, image diagnosis and determining resectability, biliary drainage, and adjunctive therapies of HCCA. These statements were proposed to the Consensus Group panel for discussion, revision, and voting. A password-secured Web site was populated with relevant literature assembled by the literature review team (RR and PA). Systematic literature reviews, with defined inclusion and exclusion criteria, were conducted to identify and grade the available evidence to support each statement. Literature search was conducted in English language publications using MEDLINE, EMBASE, and the Cochrane Trials Register in human subjects. Relevant literature from the Asia–Pacific region was of particular interest. Categorization of evidence, classification of recommendation, and voting schema was modified from the Canadian Task Force on the Periodic Health Examination (Table 1).[10]

Table 1. Classification of recommendation and voting schema is modified from the Canadian Task Force on the Periodic Health Examination
Category and gradeDescription
  1. RCT, randomized controlled trial.
Quality of evidence
IEvidence obtained from at least 1 RCT
II-1Evidence obtained from well-designed control trials without randomization
II-2Evidence obtained from well-designed cohort or case–control study
II-3Evidence obtained from comparison between time or places with or without intervention
IIIOpinion of respected authorities, based on clinical experience and expert committees
Classification of recommendation
AThis is good evidence to support the statement
BThere is fair evidence to support the statement
CThere is poor evidence to support the statement but recommendation made on other ground
DThere is fair evidence to refute the statement
EThere is good evidence to refute the statement
Voting on recommendation
AAccept completely
BAccept with some reservation
CAccept with major reservation
DReject with reservation
EReject completely

The first vote was conducted electronically by email, without explanation or access to the relevant literature. The second vote was conducted electronically after Web-based access to the provided literature. All feedbacks were collated prior to the face-to-face meeting.

Face-to-face meeting of the Consensus group was held on June 30 and July 1, 2012, in Pattaya, Chonburi, Thailand, to review and discuss the evidence for all statements. All statements were edited and finally agreed at the concluding plenary session. In addition, some overlapping statements were combined and rewritten before the final vote. Consensus was considered to be achieved when 80% or above of voting members indicated “accept completely” or “accept with some reservation.” A statement was refuted when 80% or above of voting members “reject completely” or “reject with some reservation.” Every accepted statement was then graded to indicate the level of evidence available and the strength of recommendation. Those statements that did not reach consensus were modified to compensate for the rejected reasons and underwent a revote. If the statement still failed to reach the consensus, that statement was dropped from the list. Discussed points on dropped statements are also reported in the most relevant accepted statements. Commentaries on statements were written by the chairmen (RR) and the persons assigned to present the statements during the face-to-face meeting. Co-authors were involved in the final editing of the commentaries.

Consensus statements

Epidemiology and nature of HCCA

  • 1.The incidence of cholangiocarcinoma (CCA) varies considerably depending on the geographic region due to the variation in risk factors. The highest incidence is reported in Eastern and Southeastern Asia, and the main risk factor in Asian countries is mostly linked to certain liver fluke infestation.
    • Level of agreement: a—100%, b—0%, c—0%, d—0%, e—0%
    • Quality of evidence: II-1
    • Classification of recommendation: A

There are markedly geographic variations in the incidence of CCA worldwide. The incidence of CCA in the West was reported as much lower (1–2 per 100 000) than in certain parts of Asia (5–71 per 100 000).[11] The highest incidence was reported from Northeastern Thailand (71 per 100 000 in men and 31 per 100 000 in women), followed by Eastern China (10 per 100 000 in men and 5 per 100 000 in women)[1] Table 2.

Table 2. Age-standardized incidence rates of ICCA and ECCA by gender
Registries (data for 1998–2002)MenWomen
ICCA (C22)ECCA (C24.0)ICCA (C22)ECCA (C24.0)
  1. ECCA, extrahepatic cholangiocarcinoma; ICCA, intrahepatic cholangiocarcinoma.
China, Qidong10.30.04.6
China, Guangzhou0.
China, Hong Kong2.
China, Shanghai7.
Japan, Hiroshima (1996–2000)
Japan, Osaka1.
Korea KCCR (1999–2002)
Korea, Busan5.
Korea, Daegu5.
Korea, Daejeon5.
Philippines, Manila1.
Singapore, Chinese1.
Thailand, Khon Kaen71.30.431.60.1
Thailand, Chiang Mai8.
Thailand, Bangkok2.
Thailand, Songkhla1.
Viet Nam, Hanoi0.
Italy, Brescia2.
UK, Scotland1.

There are several established risk factors for CCA such as liver fluke infestation (Opisthorchis viverrini and Clonorchis sinensis), primary sclerosing cholangitis (PSC), biliary duct cysts, hepatolithiasis, and toxins (Thorotrast) exposure.[12] The infestations with O. viverrini and C. sinensis have been classified by the International Agency for Research on Cancer as group I carcinogen for the development CCA.[13] The significantly high prevalence of liver fluke infestation in Asian countries correlates well with the high incidence of CCA.[13, 14] However, the cumulative CCA incidence in the regions with high rate of infestation still varies. Perhaps, other cofactors including the different patterns of lifestyle (e.g. tobacco and/or alcohol consumption) and the variations in genetic susceptibility may play additional role in the pathogenesis of CCA.[15]

  • 2.HCCA (Klatskin tumor) is the most common type of CCA reported in Asia and elsewhere in the world.
    • Level of agreement: a—89%, b—11%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

By using the second order of bile ducts as the reference anatomy, CCA is classified as intrahepatic cholangiocarcinoma (ICCA) and extrahepatic cholangiocarcinoma (ECCA). ECCA can be further divided into HCCA (Klatskin tumor) and distal CCA at the level of the cystic duct.[16, 17] In the update of International Classification of Diseases for Oncology (ICD-O-3), HCCA has been reclassified as ECCA.[18, 19] This in turn influenced in the observed changes in ICCA and ECCA incidence rates.[18] From data around the world, HCCA has been reported as the most common type of CCA, with the prevalence ranges from 46% to 97% and Thailand reported the highest prevalence of HCCA (97%) (Table 3).[20-24]

  • 3.The prognosis of HCCA is poor as the majority of patients present with advanced disease.
    • Level of agreement: a—100%, b—0%, c—0%, d—0%, e—0%
    • Quality of evidence: II-3
    • Classification of recommendation: A
Table 3. Proportion of CCA
First authorNumber of CCAPercentage of HCCAPercentage of distal CCA
  1. CCA, cholangiocarcinoma; HCCA, hilar cholangiocarcinoma.
Nakeeb A[20]2946727
Juttijudata P[21]61973
Nagorney DM[22]1714648
Tompkins RK[23]964944
Yusoff AR[24]695528

The clinical presentations of CCA depend on the stage of tumor. Early HCCA is usually silent or associated with nonspecific symptoms.[25] When complete hilar obstruction develops, the patient classically presents with jaundice (80–90%), pale stools, dark urine, pruritus, abdominal pain, and sometimes fever.[20, 22, 24, 26, 27] Unfortunately, these presentations usually indicate an advanced HCCA.[25] Therefore, HCCA is difficult to diagnose early, and only 20–30% of HCCA patients are amenable to complete resection (R0).[20, 24, 28-30] The median survival of patients who achieved R0 resection ranged from 1–4 years, whereas the median survival of patients with unresectable tumor was only 5–9 months.[28, 31, 32]

Histology and tumor markers

  • 4.Individual technique of obtaining pathological specimens from hilar biliary strictures has limited sensitivity; combining sampling techniques increases yields.
    • Level of agreement: a—87%, b—13%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

Tissue diagnosis of HCCA can be difficult to achieve; in three large series operated for presumptive HCCA, about 10% had benign disease.[33-35] Despite the development of tissue acquisition techniques, their overall diagnostic sensitivity remains suboptimal.[36, 37] The yield from brush cytology is variable, and positive diagnosis ranges from 44–80%.[36, 38] To date, there has been no prospective control study on the yield of tissue acquisition in HCCA. Pooled data from over 800 CCA patients reported sensitivity of 42%, specificity of 98%, and positive predictive value (PPV) of 98% among patients with confirmed cancer.[36] It has been reported that at least five brush passes, removal of the brush and catheter together, and inclusion of washings from the brush catheter may increase yield.[39] Intraductal fine-needle aspiration (FNA) had a sensitivity of just 34%, with specificity of 100% and PPV of 100%.[36] Although intraductal biopsies have shown the highest yield for detection of malignancy, with a pooled sensitivity of 56%, specificity of 97%, and PPV of 97%,[36, 39] intraductal biopsy in HCCA stricture is a cumbersome technique and may result in a lower diagnostic yield than the result reported in all CCAs. A “smash prep” protocol showed the overall sensitivity of 76% for all cancers with 100% specificity. The highest diagnostic yields for tissue sampling at Endoscopic retrograde cholangiopancreatography (ERCP) were obtained by using a combination of two or three standard techniques at the same setting. Ponchon et al. found that combining brush cytology (35% sensitivity) and forceps biopsy (43% sensitivity) yielded a sensitivity of 86%.[40] The Indiana group reported a sensitivity of 73% in CCA subset using triple samplings with brush cytology, FNA, and forceps biopsy. The addition of a 2nd or 3rd sampling modality consistently increased diagnostic yield.[41] Therefore, we recommend at least a combination of two techniques such as brushing and forceps biopsy for all suspicious strictures.

  • 5.Carbohydrate antigen 19-9 (CA 19-9) and carcinoembryonic antigen (CEA) are moderately specific for CCA. The presence of cholestasis and cholangitis lower the specificity of serum CA 19-9
    • Level of agreement: a—63%, b—37%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: B

CA 19-9 and CEA are the two markers best studied with respect to CCA, but their utility is limited by poor sensitivity in early stage malignancy, and marginally elevated levels (> 100U/mL) may be associated with benign conditions.[42-47] Many studies have looked at their diagnostic utility in the setting of both PSC-related CCA and non-PSC-related CCA.[42-47] By using the serum cutoff value of more than 180 U/mL in some large series,[42-49] the sensitivity was moderate at 53%–79% and the specificity was fair to excellent at 83%–98%. However, the specificity of CA 19-9 in diagnosing biliary malignancy is reduced by the presence of either cholangitis or cholestasis. A Korean study found a cutoff value of ≥ 37 U/mL to be 78% sensitive and 83% specific for malignant disease in patients without cholangitis; in the presence of cholangitis, the specificity dropped to just 42% at the same cutoff level.[48] Less data are available for the utility of serum CEA in CCA diagnosis. The Pittsburgh group reported that serum CEA level of > 5.2 ng/mL had a sensitivity of 68% and specificity of 82%.[44] One study showed that in patients with CCA, the biliary CEA level was about five times that of patients with benign strictures.[50] A combined index of serum CA 19-9 and CEA (CA 19-9 + [CEA × 40]) has been reported to correctly identified 10 of 15 patients with CCA, including 6 of 11 with radiographically occult disease; and without false positive.[46] This corresponded to an accuracy of 86% for CCA detection. A subsequent study however suggested that this score was no better than CA 19-9 alone in predicting the presence of CCA.[51]

Role of cholangioscopy and image enhancement

  • 6.Cholangiography with tissue acquisition has been the traditional technique to diagnose HCCA. Cholangioscopy may be performed to increase the diagnostic yield.
    • Level of agreement: a—80%, b—20%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: B

Endoscopic tissue acquisition during ERCP can be performed either via papilla under fluoroscopic guidance or via mother–baby cholangioscopy. In general, the diagnostic sensitivity of transpapillary biopsy under fluoroscopic guidance for CCA ranges from 40% to 60%.[40, 52-55] Transpapillary biopsy enables the collection of a relatively large amount of tissue because of the use of standard biopsy forceps. In contrast, mother–baby or spyglass or percutaneous cholangioscopy-assisted targeted biopsy cannot always collect a sufficient amount of tissue because of the use of forceps with small jaws. In fact, the diagnostic sensitivity of cholangioscopy-targeted biopsy alone was previously reported to be suboptimal (49%).[56] However, the advantage of cholangioscopy is that it may additionally provide a cholangioscopic impression to better clarify some indeterminate cholagiographies. The pool data demonstrated that mother–baby cholangioscopy plus targeted biopsy can improve the sensitivity to detect biliary malignancy to 89–100% with the specificity of 87–96%.[57-59]

  • 7.Cholangioscopy with image enhancement systems and possible targeted biopsy or probe-based confocal laser endomicroscopy (pCLE) may improve the accuracy of HCCA diagnosis.
    • Level of agreement: a—73%, b—21%, c—6%, d—0%, e—0%
    • Quality of evidence: III
    • Classification of recommendation: C

Although peroral cholangioscopy such as conventional mother–baby scope and spyglass system can be useful for detecting CCA, the images' quality is still suboptimal because of the limitation in resolution of the fiber-optic choledochoscope. With the advent in video endoscope and the image enhancement technology such as narrow band imaging, the vascular pattern of neoplasm in the bile duct can be better characterized by a new video choledochoscope.[60] However, the standardization of the insertion technique and the indications for a use in HCCA remain controversial. To date, there are only a handful of reports[60, 61] to support the feasibility of this technology. In addition, peroral cholangioscopy appears to be associated with a significantly higher rate of cholangitis, possibly because of the intermittent intraductal irrigation required during the procedure.[62] pCLE is a new imaging technique that provides real-time microscopic information on the tissue during ERCP.[63] Several investigators have reported the usefulness of pCLE in the diagnosis of CCA.[63-65] Recently, Miami criteria for the diagnosis of malignant biliary stricture have been proposed.[66] Thick dark and white bands, dark clumps, visible epithelium, and fluorescein leakage were criteria indicating malignancy. Although the diagnostic sensitivity was excellent, the specificity was still suboptimal (67%).[66] The criteria may need some refinement and pilot them in a larger set of indeterminate biliary strictures before recommendation as a standard approach.

Image diagnosis and determining resectability (Fig. 1)

Figure 1.

Algorithm 1 guideline for the management of hilar cholangiocarcinoma. CBD, common bile duct; EUS, endoscopic ultrasonography; FNA, fine-needle aspiration; HCCA, hilar cholangiocarcinoma; MDCT, multidetector computed tomography; MRCP, magnetic resonance cholangiopancreatography; MRI, magnetic resonance imaging; PTBD, percutaneous biliary drainage; PVE, portal vein embolization; US, ultrasonography.

  • 8.Abdominal ultrasonography (US) is frequently the initial imaging modality performed to evaluate patients with suspected biliary obstruction. Other imaging modalities are required for further characterization and staging of HCCA.
    • Level of agreement: a—90%, b—10%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

Abdominal US is practically performed to confirm the presence of biliary obstruction, to identify the extent of obstruction, and possibly to determine the cause of the obstruction.[67] In HCCA, US can demonstrate dilation of bilateral intrahepatic ducts. Occasionally, intraluminal masses may be discovered in papillary type HCCA, and US in a patient with infiltrative-type HCCA may show periductal thickening of bile ducts.[67] However, the sensitivity of US to identify the etiology of the obstruction is lower than other modalities such as CT scan, magnetic resonance imaging (MRI), and direct cholangiography.[68, 69] Therefore, further delineation of HCCA for the detail of tumor characterization, vascular involvement, staging, and variation in biliary anatomy by other modalities is required.

  • 9.Multidetector computed tomography (MDCT) and MRI/magnetic resonance cholangiopancreatography (MRCP) are the two best imaging modalities for diagnosis and staging of HCCA, as well as for determining its resectability. The role of positron emission tomography (PET)/computed tomography (CT) is not clearly defined.
    • Level of agreement: a—74%, b—21%, c—0%, d—5%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

The recent staging and registry for HCCA relies on the extent of the disease in the biliary system, the involvement of the hepatic vasculatures, the involvement of lymph nodes, distant metastases, and the volume of the future hepatic remnant (FLR) after resection.[17] CT scan and MRI are the two most practical imagings that serve this purpose. MRI combined with MRCP is comparable with MDCT with direct cholangiography for assessment of biliary involvement, vascular encasement, lymph node metastasis, and tumor resectability.[70-72] The accuracy of CT and MRI with MRCP for prediction of the extent of ductal involvement, hepatic arterial invasion, portal vein invasion, and lymph node metastasis is in the range of 84–91%, 83–93%, 86–98%, and 74–84%, respectively.[73, 74] Although PET combined with CT (PET/CT) has been recommended to evaluate the metastasis of many intra-abdominal malignancies, it is premature to state the routine use of PET/CT in HCCA. The sensitivity rate of detecting non-nodal distant metastases by PET and PET/CT in patients with CCA was in the range of 70–100%, while the sensitivity of regional lymph node metastases was only about 12%.[75, 76]

  • 10.Endoscopic ultrasonography (EUS) with FNA in combination with other modalities may improve the diagnostic accuracy for HCCA.
    • Level of agreement: a—47%, b—35%, 12%, d—6%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: B

Although CT and MRI are the standard imaging tools to evaluate the presence and resectability of CCA, they may miss some small lesions.[70, 77, 78] EUS has been proven to detect those small lesions and may help to predict the unresectability of CCA.[79] However, its sensitivity is significantly higher in distal CCA than in HCCA.[79] Although it is technically difficult due to the tumor's anatomical position, EUS-FNA in brush-negative HCCA patients has been practiced in many advanced endoscopy centers.[80-82] Original reports in suspected HCCA patients with an inconclusive tissue diagnosis demonstrated that the overall diagnostic accuracy, sensitivity, specificity, PPV, and negative predictive value for EUS-FNA in diagnosing HCCA were 91%, 89%, 100%, 100%, and 67%, respectively.[80-82] Therefore, EUS may be considered, where available, to confirm HCCA diagnosis and to evaluate the resectability in those with inconclusive results after the standard evaluation.

  • 11.EUS has a limited role in local staging of HCCA but may be useful in detecting nodal disease.
    • Level of agreement: a—42%, b—42%, c—16%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: B

Complete staging of HCCA with EUS is challenging because of the limited depth of visualization and T staging may be inadequate. EUS is able to detect locoregional lymph nodes in the hepatic hilum and in the coeliac axis, as well as para-aortic lymph nodes.[83] In a study of 47 patients, EUS correctly identified lymph nodes in all the patients and confirmation of malignancy by FNA precluded liver transplantation in 17%, implying that EUS-FNA for regional lymph node staging should be further considered in all resectable HCCA patients predicted by CT or MRI to avoid unnecessary surgery.[84] Intraductal ultrasonography (IDUS) is useful in the evaluation of CCA from inside out. IDUS was found to be superior to EUS for T staging (78% vs 54%).[85] The accuracy of IDUS in assessing vascular invasion of the right hepatic artery and portal vein was 92–100%.[85-89] In contrast, the accuracy of EUS in assessing portal vein invasion was only 57%.[90] However, to recommend IDUS for an evaluation of HCCA before surgery is not recommended because tumor resection can still be performed in a HCCA patient with limited vascular involvement at the periphery.

  • 12.Staging laparoscopy with or without laparoscopic ultrasonographic examination should be considered before attempting a curative resection to avoid unnecessary laparotomy.
    • Level of agreement: a—79%, b—14%, c—7%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

Staging laparoscopy has been a traditional approach prior to attempting a curative surgery in HCCA. The role of laparoscopy is for detecting liver and peritoneal metastasis.[91, 92] However, locally advanced tumor and nodal disease could be missed.[91, 92] More extensive dissection during laparoscopy could have discovered locally advanced conditions. However, the risk and cost of longer and more aggressive approach have to be considered. Subsequently, laparoscopic ultrasonographic examination has been added in the protocol in some centers to compensate for this limitation. Unfortunately, the diagnostic yield did not differ from laparoscopy alone in majority of many reports.[92-95] The overall diagnostic yield of laparoscopy with or without laparoscopic ultrasonographic examination was reported in the range of 25–42%.[92-95] Recently, the role of laparoscopic staging has been challenged with many new non-invasive imaging modalities such as PET/CT, EUS, and IDUS. A recent report from the tertiary center in Netherlands demonstrated that the diagnostic yield of staging laparoscopy decreased to 14%.[96] The result may be associated with the increased use of PET/CT and other better imaging during the last 3 years of their study.[96]

Role biliary drainage (Fig. 2)

Figure 2.

Algorithm 2: Palliative biliary drainage in HCCA Bismuth II-IV. HCCA, hilar cholangiocarcinoma.

  • 13.Preoperative biliary drainage (PBD) in HCCA should be performed in selected patients but may increase risk of postoperative complications.
    • Level of agreement: a—69%, b—19%, c—12%, d—0%, e—0%
    • Quality of evidence: II-3
    • Classification of recommendation: B

PBD is definitely indicated in an HCCA patient with acute cholangitis, but a routine use of PBD is controversial. Obstructive jaundice might be associated with hepatic and renal dysfunction and coagulopathy.[97, 98] In an effort to improve the outcome, PBD has been advocated as a mean of improving the functional status of the FLR and reducing the rate of postoperative hepatic insufficiency.[99] In addition, PBD may be indicated in HCCA patients with severe pruritus and/or impeding renal failure However, PBD can increase risk of postoperative infectious complications[100, 101] and procedure-related complications such as hemobilia, cholangitis, and neoplastic seeding.[102, 103] At present, there are only a handful of randomized controlled trials (RCTs) or meta-analyses performed to evaluate the value of PBD before the major resection of HCCA. The results showed that the postoperative complications in the PBD group were significant higher than that in the no-PBD group.[104] Therefore, the routine use of PBD cannot be justified at the moment. However, certain countries (Japan and South Korea) preferred to perform PBD in HCCA patient via percutaneous approach.[105, 106] The advantages of this approach are: (i) to assess the function of residual hepatic parenchyma before an extensive hepatic resection, and (ii) to reserve time for hepatic-hypertrophy induction after selective portal vein embolization in an HCCA patient with inadequate FLR.

  • 14.In Bismuth II-IV HCCA patients with a predicted survival of longer than 3 months, metallic stent performance is superior to plastic stenting for palliation with respect to outcomes and cost-effectiveness.
    • Level of agreement: a—80%, b—20%, c—0%, d—0%, e—0%
    • Quality of evidence: I
    • Classification of recommendation: A

The current available biliary stents are plastic stent (PS) and self-expandable metallic stent (SEMS). Although PS is much less expensive than SEMS, its disadvantage is a high occlusion rate. The PS median patency time is 1.4–3 months,[107-109] whereas a larger diameter SEMS provides a longer patency at 6–10 months.[109, 110] The additional benefit of SEMS in HCCA is that the mesh allows drainage of the side branch ducts. Sangchan et al. randomly inserted either PS or SEMS in 108 Bismuth II-IV HCCA patients.[111] They reported that the successful drainage rate in the SEMS group was significantly higher than in the PS group (70% vs 46%) and the median survival times of both groups were 126 and 49 days, respectively.[111] The same group also used model-based cost utility analysis and demonstrated that SEMS was more cost-effective than PS (99%).[112] Reasons for this conclusion are the higher drainage efficacy, less occlusion rate, longer survival, and higher quality of life in the SEMS group.[112]

  • 15.For the palliation of advanced HCCA (Bismuth III and IV), the outcomes of percutaneous stenting are superior to endoscopic stenting.
    • Level of agreement: a—25%, b—55%, c—15%, d—5%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

Percutaneous and endoscopic stentings have been established as effective and less invasive approach for biliary drainage in unresectable HCCA when compared with surgical biliary bypass.[113] The advantage of percutaneous approach is the precise lobar selection for drainage. Hypothetically, this approach should yield a lower rate of cholangitis.[114] However, pain at the puncture site and two-step requirement in some cases are the main concern. In patients with low-grade hilar obstruction (Bismuth I and II), endoscopic stenting is considered as a less invasive approach with acceptable outcome.[108] In contrast, patients with advanced hilar obstruction (Bismuth III and IV), endoscopic stenting had a lower success rate of cholestasis palliation and a higher rate of post ERCP cholangitis.[108] A multicenter retrospective study from South Korea reported the outcomes of endoscopic and percutaneous stenting with SEMS in 85 and 41 HCCA patients (Bismuth III and IV), respectively.[115] They found that the successful biliary drainage was significantly higher in the percutaneous group than in the endoscopic group (93% vs 77%, P = 0.049). However, the overall rates of complication and median survival of the successfully drained patients were similar.[115]

  • 16.The goal of palliative stenting of HCCA is drainage of adequate liver volume (50% or more), irrespective of unilateral, bilateral, or multisegmental stenting.
    • Level of agreement: a—40%, b—60%, c—0%, d—0%, e—0%
    • Quality of evidence: II-A
    • Classification of recommendation: A

It is well accepted that in Bismuth I HCCA, only one stent in the common duct is appropriate. However, there is no consensus with regard to bilateral versus unilateral drainage in beyond Bismuth I HCCA. De Palma et al. reported on the more efficient drainage with unilateral stenting, however, one third of patients in their series were Bismuth I.[116] In contrast, a retrospective study by Chang et al. demonstrated that successful bilateral drainage provided longer survival advantage (225 days vs 145 days).[117] However, they reported on the drawback of failed bilateral drainage as a higher rate of cholangitis (32% vs 6%) and shorter survival of the patients (225 days vs 46 days).[117] A prolonged manipulation of the devices in the undrained lobe was blamed for the poor results in the failed group. Previously, it was assumed that draining 25% of liver volume is enough to relief jaundice.[118] Recently, a retrospective study by Vienne A et al. reported that HCCA patients who had more than 50% of their liver volume achieved more efficient drainage than those with lower volume drained (82% vs 45–55%).[119] Generally, right lobe of the liver covers 55–60% of the liver volume, while left lobe and caudate lobe cover 30–35% and 10% of the liver volume, respectively.[120] Draining more than 50% of liver volume frequently requires more than one stent, whether bilateral stenting or multisegmental stenting, which depends on the individual anatomy. In addition, atrophic segment and aberrant ductal anatomy need to be assessed by non-invasive imaging(s) before attempting biliary drainage.[121]

  • 17.MRCP or/and volumetry assessed by MDCT or MRI currently is (are) a good imaging modality for selecting the appropriate segment(s) for drainage and determining its effectiveness.
    • Level of agreement: a—74%, b—26%, c—0%, d—0%, e—0%
    • Quality of evidence: II-3
    • Classification of recommendation: B

Volume assessment of liver and its segment can be measured by the technique called “volumetry.” This technique calculates the volume from the drawing contour of the interpolated liver images obtained by MDCT or MRI.[122, 123] The summation of volume from multiple segments can be further calculated for drainage purpose based on the anatomy of main duct. Practically, drainage of at least two liver segments can support close to 50% of liver volume and mostly achieved by more than one stents. Multiple stenting usually requires bilateral stenting. However, multisegmental stenting can be performed unilaterally in the right lobe. MRCP can add the information on advanced subsegmental occlusion that precludes a complete drainage.[124] Moreover, when unilateral stenting with one stent is planned, MRCP can guide for dominant lobe drainage.[124] A group from Minneapolis reported on the usefulness of unilateral stenting suggested by MRCP to be efficient in 77 % of their HCCA patients and no further intervention was needed in 71%.[124] Moreover, Harewood and Baron reported that the MRCP-guided strategy seems to be more cost-effective than a routine bilateral stenting.[125]

  • 18.Endoscopic biliary drainage for advanced HCCA should be performed by an experienced biliary endoscopist with multidisciplinary backup.
    • Level of agreement: a—88%, b—12%, c—0%, d—0%, e—0%
    • Quality of evidence: III
    • Classification of recommendation: C

Endoscopic metallic stenting for a high-grade HCCA is a procedure requiring experienced professions.[126] According to Schutz and Abbott, this procedure is classified as grade 5 which is the most difficult level.[127] Schutz and Abbott reported that 35% of grade 5 ERCP procedures in their series were unsuccessful (16 of 46), compared with only 4% failure rate in the less difficult procedures (grade 1 to 4 [5 of 138, P < 0.001]). Although there was no statistical difference on the complication rate, there was a trend of higher complication rate in grade 5 ERCP than the lower grades (9% vs 4%). Therefore, endoscopic biliary drainage for HCCA should be performed by an experienced biliary endoscopist. In addition, multidisciplinary backup is needed when performing this level of ERCP complexity. For instance, when duct opacification without complete drainage happens, another approach, such as prompt percutaneous drainage, is mandatory,[110, 128, 129] otherwise post ERCP cholangitis may develop.[117]

  • 19.Bilateral biliary drainage using metallic stents for HCCA can be performed with side-by-side or stent-in-stent methods.
    • Level of agreement: a—88%, b—12%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

Endoscopic bilateral or multisegmental stenting with SEMS is technically challenging. After the initial stenting of the intrahepatic duct in one segment (or side), a second stent can be placed either using a “side-by-side” method, i.e. the second stent is deployed parallel to the initial stent, or using “stent-in-stent,” i.e. the second stent is deployed by crossing through the mesh within the initial stent.[129-131] To date, there is not enough data to support on which technique is preferable. Previously, contralateral stenting through the mesh of the first SEMS for “stent-in-stent” method was technically difficult because of the narrow mesh design of the first stent. Recently, an open-cell-design SEMS made “stent-in-stent” method become more practical.[132-134] However, there may be a problem when occlusion of the first SEMS develops. In contrast, “side-by-side” technique allows distal ends of both SMES to be left in duodenum, thus a selective cannulation to the occluded SEMS is technically possible. For this purpose, the length of SEMS has to be long enough (at least 8–12 cm). In addition, complete insertion of the two stents before deployment of any stent is mandatory for certain SEMS insertion technique (i.e. Zilver stent), otherwise, the insertion of second SEMS is impossible.[135]

  • 20.With respect to the percutaneous approach, metal stenting is preferable to catheter drainage or internal plastic stenting for the palliation of jaundice.
    • Level of agreement: a—57%, b—43%, c—0%, d—0%, e—0%
    • Quality of evidence: II-2
    • Classification of recommendation: A

Percutaneous biliary drainage for HCCA has certain advantages over the endoscopic approach i.e. selection of intrahepatic duct for the drainage is more feasible and the technique requires less sedation in an unstable patient. However, the disadvantages of the external approach include pain at the puncture site, bile leak, and external bile loss.[136] Percutaneous approach can provide both external and internal drainage. Generally, a single-step approach is more preferred; however, a two-step approach may be required in a patient with severe biliary sepsis or when a stricture could not be traversed at the initial attempt. Similar to the endoscopic stenting, percutaneous stenting can be achieved with either PS or SEMS. Hii MW et al. reported a longer survival times (213 vs 142 days) and lower complication rates (44 vs 64%) in patients with SEMS placed than patients with PS placed.[137] Almost similar to the endoscopic bilateral stenting with SEMS in Y-configuration, the percutaeous stenting can be performed either with Y- or T-configuration. A group from Korea inserted a T-configuration SEMS in their 30 HCCA patients. They found that the median survival and stent patency times were 334 days (range, 195.6–472.4 days) and 279 days (range, 194.7–363.3 days), respectively.[138] Another series of Y-configuration SEMS for HCCA reported by the same group showed the similar median survival and stent patency at 218 and 375 days, respectively.[139]

  • 21.EUS-guided biliary drainage is emerging as an experimental alternative technique in patients with HCCA when transpapillary and percutaneous drainage have failed or are not possible.
    • Level of agreement: a—76%, b—18%, c—6%, d—0%, e—0%
    • Quality of evidence: III
    • Classification of recommendation: C

EUS-guided biliary drainage may be a good alternative for draining HCCA if initial ERCP attempt fails or percutaneous approach is contraindicated. EUS-guided biliary drainage can be achieved by various methods; EUS-guided hepaticogastrostomy,[140-142] EUS-guided rendezvous procedure with the conventional ERCP access[143, 144] and EUS-guided stent placement across stricture segment.[145, 146] Transgastrohepatic route is commonly used. This technique started from EUS-guided cholangiography.[147-149] After EUS-guided puncture of left intrahepatic duct, a guidewire can be inserted into bile duct and rendezvous procedure can be attempted if the guidewire passes into distal bile duct and duodenum. A guidewire-assisted rendezvous ERCP seems to be more physiologic because it does not create any fistula. In patients with duodenal obstruction or in cases in which a guidewire passage into duodenum is impossible, EUS-guided stent placement across hilar stricture can be used. If a guidewire cannot pass through hilar stricture, EUS-guided hepaticogastrostomy is the option left. There has been no comparison data regarding the superiority of each method. There are different types of stent which have been used for EUS-guided biliary drainage; PS, bare, partially covered, and fully covered SEMS.[141, 145, 150-153] When performed by experienced endosonographers, technical success rate of EUS-guided biliary drainage ranges from 70∼98%.[140, 143, 144, 151, 154-157] The overall complication rates of EUS-guided biliary drainage were reported as up to 20%[142, 151, 153, 158-160] and higher than that of standard ERCP. Most common complications were bile leakage and peritonitis.[142, 151, 153, 158-160] Therefore, we consider EUS-guided biliary drainage as experimental because the current technique is afflicted with a high complication rate.

  • 22.Palliative surgical bypass may be considered in selected patients, or when laparotomy discovers an unresectable locally advanced tumor.
    • Level of agreement: a—62%, b—38%, c—0%, d—0%, e—0%
    • Quality of evidence: II-3
    • Classification of recommendation: C

Palliative biliary bypass in HCCA are segment III cholangiojejunostomy, right sectoral duct bypass, and transtumoral tube placement. Segment III cholangiojejunostomy is the most preferred bypass technique. Earlier studies reported that jaundice resolution could be achieved in 70% of HCCA patients and the median survival was 6.3 months.[113, 161-163] Because surgical drainage procedures is not superior to nonsurgical one with respect to procedure-related mortality and survival,[113] then non-operative biliary stenting is regarded as the first choice. However, surgical bypass may be considered in HCCA patients with a good estimated life expectancy, where endoscopic and/or percutaneous stenting has failed[164] or when laparotomy that aimed for R0 discovers an unresectable locally advanced tumor.[165]

Adjunctive therapies chemo-radiation, photodynamic therapy (PDT), radio-frequency ablation (RFA)

  • 23.PDT in combination with stenting is an optional technique to improve duct patency. It may improve survival and quality of life of patients with inoperable HCCA.
    • Level of agreement: a—32%, b—58%, c—10%, d—0%, e—0%
    • Quality of evidence: I
    • Classification of recommendation: A

PDT is a technique for palliation of unresectable HCCA. PDT incorporates the use of a photosensitizing agent, which selectively accumulates in proliferating tissue such as malignant tumors. Photoactivation with a red laser light generates reactive oxygen species, leading to selective tumor-cell death. Ortner et al.[166] published the first RCT that confirmed dramatic increase of median survival time after PDT compared with HCCA patients receiving only stent therapy. Survival time of PDT patients was 16 months compared with 3 months for stenting alone. Treatment with PDT and stenting also led to improvement of cholestasis and quality of life compared with endoscopic stenting alone. Another study by Zoepf et al.[167] showed the median survival time after randomization in HCCA patients was 7 months for the control group and 21 months for the PDT group (P = 0.0109). Likewise, Cheon et al. demonstrated a longer median survival in the PDT group when compared with stenting alone (18 months vs 10 months).[168] Although these studies support the benefit of PDT for HCCA palliation, the use of PDT has been limited only in nontropical Asian countries where skin photosensitivity has been affected less by the sun.[168-170]

RFA has been accepted as an excellent tool to ablate solid tumors in the liver.[171] Recently, a lower power setting in a dedicate RFA probe has been introduced for the use in bile duct to induce local coagulative necrosis.[172] In a recent pilot study of 21 patients, Steel et al. demonstrated the safety and efficacy of RFA within the bile duct by using a bipolar RFA catheter in 22 patients with malignant obstructive jaundice (16 pancreatic cancers and 6 CCAs) without any major complication. Endobiliary RFA may add to the endoscopic armamentarium for the treatment of HCAA. However, further RCTs specifically to HCCA population are needed.


Other members who greatly contributed to our work during the consensus meeting and manuscript preparation are:

  1. Sombat Treeprasertsuk, MD, MSc, Department of Medicine, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
  2. Pradermchai Kongkam, MD, MSc, Department of Medicine, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
  3. Hsiu-Po Wang, MD, Department of Internal Medicine, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan
  4. Maylene Kok, MD, Kuala Lumpur General Hospital, Malaysia
  5. Evan Ong, MD, Department of Gastroenterology, Metropolitan Medical Center, Manila, the Philippines
  6. Dong Ki Lee, MD, Division of Gastroenterology, Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea
  7. Pisaln Mairiang, MD, Department of Medicine, Faculty of Medicine, KhonKaen University, KhonKaen, Thailand
  8. Bancha Ovartlarnporn, MD, NKC institute of Gastroenterology and Hepatology, Faculty of Medicine, Prince of Songkla University, Songkhla, Thailand
  9. Suraphol Churnrattanakul, MD, Pramongkutklao hospital, Bangkok, Thailand
  10. Akkawat Janchai, MD, Department of Radiology, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
  11. Vajarabhongsa Bhudhisawasdi, MD, Department of Surgery, Faculty of Medicine, KhonKaen University, KhonKaen, Thailand

Author contributions

RR was involved with the consensus concept and design, acquisition of data, drafting of the manuscript, critical revision of the manuscript, obtain funding, and participation in electronic and face-to-face voting. PA and TR were involved with the consensus concept and design, acquisition of data, drafting of the manuscript, critical revision of the manuscript, and participation in electronic and face-to-face voting. CK, RP, JM, DS, LP, AS, PP, TA, DR, AM, SP, PK, MR, and MK were involved with drafting of the manuscript, critical revision of the manuscript, and participation in electronic and face-to-face voting. TI was involved with drafting of the manuscript, critical revision of the manuscript, and participation in electronic voting. NP, ST, SA, BD, HW, EO, DL, and PM were involved with critical revision of the manuscript, and participation in electronic and face-to-face voting. KG and BO were involved with critical revision of the manuscript and participation in electronic voting. SC and AJ were involved with study supervision and participation in electronic voting. VB was involved with study supervision.

Appendix: Appendix

Electronic voting participants

Surgeon—Amit Maydeo, Thawatchai Akaraviputh, Thawee Rattanachu-ek, Vajarabhongsa Bhudhisawasdi

Gastroenterologist—Rungsun Rerknimitr, Pises Pisespongsa, Takao Itoi, Christopher J L Khor, Ryan Ponnudurai, Jong H Moon, Dong Wan Seo, Duvvuru Nageshwar Reddy, Apichat Sangchan, Pradermchai Kongkam, Nonthalee Pausawasdi, Siriboon Attasaranya, Dong Ki Lee, Bancha Ovartlarnporn, Suraphol Churnratanakul, Kean-Lee Goh, Benedict Devereaux, Hsiu-Po Wang, Evan Ong, Sombat Treeprasertsuk, Maylene Kok

Diagnostic radiologist—Linda Pantongrag-Brown

Interventional radiologist—Sundeep Punamiya, Akkawat Janchai

Face-to-face meeting participants

Surgeon—Amit Maydeo, Thawatchai Akaraviputh, Thawee Rattanachu-ek

Gastroenterologist—Rungsun Rerknimitr, Pises Pisespongsa, Christopher J L Khor, Ryan Ponnudurai, Jong H Moon, Dong Wan Seo, Apichat Sangchan, Pradermchai Kongkam, Nonthalee Pausawasdi, Siriboon Attasaranya, Dong Ki Lee, Benedict Devereaux, Hsiu-Po Wang, Evan Ong, Sombat Treeprasertsuk, Mohan Ramchandani, Maylene Kok

Diagnostic radiologist—Linda Pantongrag-Brown

Interventional radiologist—Sundeep Punamiya

(All face-to-face meeting participants participated in the first two electronic votes)

Membership of the consensus group

Members of the group were selected by the following criteria:

  1. Demonstration of knowledge/expertise in HCCA by publication/research or participate in national or regional guidelines.
  2. Geographical representation of the Asia–Pacific countries/region.
  3. Diversity of clinical views on management of HCCA (limited to clinicians).

Representative countries were Australia, Malaysia, India, Taiwan, the Philippines, South Korea, Japan, Singapore, and Thailand. Representatives from China, New Zealand, and Indonesia were invited but did not participate.

Dissemination strategies

This consensus statement which is tailored for the Asian population will be disseminated by the following means: first, presentations at APDW 2012 meetings held in Bangkok, Thailand, as forums to an audience of predominantly Asian clinicians and second, copies of this statement will be sent to national societies or associations of gastroenterology and endoscopy for their iteration and dissemination. It is hoped that the consensus strategies will be updated regularly and adopted by health authorities to improve the care of patients with HCCA in many other countries outside the Asia–Pacific region as well.