Volume 45, Issue 1, Pages 33-47 (January 2009)

7 of 26

ABSTRACT

FULL TEXT

FULL-TEXT PDF (217 KB)

Prognostic molecular markers in cholangiocarcinoma: A systematic review

Received 17 April 2008; received in revised form 6 August 2008; accepted 26 August 2008. published online 20 October 2008.

Abstract

The worldwide incidence of cholangiocarcinoma (CC) is steadily rising, with the incidence in United Kingdom (UK) now exceeding 1000 cases per year. It is an aggressive malignancy typified by unresponsiveness to the existing chemotherapy and radiotherapy regimes in the vast majority of cases. Surgery offers the only hope of a cure, though post-operative disease recurrence is common, with 5-year survival rates of less than 25% following resection. Developments in molecular techniques and improved understanding of the basis of carcinogenesis in CC has led to examination of the role of biomarkers in predicting poor outcome. This systematic review examines published evidence relating to the prognostic significance of these molecular markers in CC. Of the molecular markers which have been investigated to date, p53 mutation, cyclins, proliferation indices, mucins, CA19-9, CRP and aneuploidy appear to hold significant potential as predictors of outcome in CC. These and other biomarkers may themselves represent novel therapeutic targets for CC.

Keywords: Cholangiocarcinoma, Molecular markers, Prognosis, Survival

Article Outline

• Abstract

• 1. Introduction

• 2. Tumour suppressor genes

• 2.1. p53 []

• 2.2. Other tumour suppressor genes

• 3. Oncogenes

• 3.1. k-ras

• 3.2. KIT (CD117)

• 4. Apoptotic regulation

• 4.1. Bcl-2/Bax

• 4.2. DAP-kinase

• 4.3. Fas/FasL

• 4.4. Survivin

• 5. Cell cycle regulators []

• 5.1. Cyclin and cyclin-dependent kinases

• 5.2. Cyclin-dependent kinase inhibitors (CKI)

• 5.2.1. KIP/CIP family

• 5.2.2. INK4 family

• 6. Proliferation indices []

• 7. Growth factors and receptors

• 7.1. Epidermal growth factor receptor (EGFR) family

• 7.2. Tenascin

• 7.3. CCN proteins

• 8. Invasion and metastasis

• 8.1. Cadherins

• 8.2. Syndecan-1

• 8.3. CEACAM6

• 8.4. CD24

• 8.5. CD44

• 8.6. Matrix metalloproteinase 7 (MMP7)

• 9. Angiogenesis

• 10. Mucins []

• 10.1. MUC1

• 10.2. MUC2

• 10.3. MUC4

• 10.4. MUC5AC

• 11. Serum markers

• 11.1. Carbohydrate antigen 19-9 (CA19-9, sialyl Lewis serum antigen)

• 11.2. Carcinogenic embryonic antigen (CEA)

• 11.3. Bilirubin

• 11.4. C-reactive protein (CRP)

• 11.5. Sialic acid

• 12. Sialyl Lewis antigens in tissue

• 13. Genetic factors

• 13.1. Chromosomal instability/anueploidy

• 13.2. Loss of heterozygosity (LOH) and microsatellite instability (MSI)

• 13.3. Genetic repair

• 13.4. Genotype

• 14. Other markers

• 14.1. Cytokeratins

• 14.2. Neuroendocrine differentiation

• 14.3. Host immune response

• 14.4. Pancreatic secretory trypsin inhibitor (PSTI)

• 14.5. Cyclooxygenase-2 (COX-2)

• 15. Conclusions

• Conflict of interest statement

• References

• Copyright

1. Introduction

Cholangiocarcinoma (CC) is a primary malignancy of the epithelial lining of the biliary tree, which carries a poor prognosis despite combined therapeutic strategies. The current reported incidence is only 1–2 per 100,000 population per year in the United Kingdom (UK), however, it is an increasingly prevalent disease both here1, 2 and worldwide.3 Induction of biliary dysplasia and subsequent malignant transformation is thought to be partly due to local inflammation in the biliary tree, which may be caused by a number of disease states. Endemic factors, such as parasitic infestation, may account for the reported increasing incidence in southeast Asia.4 In western populations, the escalation in cases may be due to an increase in the incidence of hepatitis C virus and subsequent long-term biliary inflammation.3 Other known associated conditions include primary sclerosing cholangitis (PSC), chronic intra-ductal gallstone disease, choledochal cysts, congenital liver malformations, human immunodeficiency virus infection, smoking and exposure to certain chemicals (possibly related to employment in the aircraft, rubber and wood finishing industries).

CC may arise anywhere in the biliary tree, from the small, peripheral hepatic ducts to the distal common bile duct. Commonly used classification systems utilise anatomical location to group tumours into three main categories: intra-hepatic (20–25%), hilar (also known as Klatskin tumour – 50%) and extra-hepatic (20–25%).5 In the published studies, hilar tumours may be classified together with either intra-hepatic or extra-hepatic type. Gallbladder carcinoma is also considered in some series to represent a sub-type of CC, but is not included here.

Despite advances in chemo-, immuno-, photodynamic- and radio-therapy, CC carries a poor overall survival rate. Surgery offers the only hope of cure, and with careful selection and increasingly radical surgical strategies, 5-year survival as high as 41% has been reported.6, 7, 8, 9 However, existing pre-operative radiological, pathological and operative (laparoscopic) staging strategies do not allow the accurate determination of long-term prognosis in operable patients. The ability to predict tumour behaviour on the basis of molecular markers from either biopsy or serum samples would help inform the patient and clinician during the decision-making process. With rapid advances in the understanding of tumour biology, there is sufficient new evidence available to gain further insight into this disease. Furthermore, biomarkers associated with adverse outcome may themselves represent novel therapeutic targets.

The aim of this systematic review was to summarise the results of published studies looking at molecular biomarkers and their prognostic significance in CC. The PubMed and Web of Science databases were searched using the following keywords in varying combinations in order to identify published works in English language up to October 2007: cholangiocarcinoma, bile duct cancer, Klatskin tumour, prognosis, biomarkers and molecular markers. Individual biomarkers were also included in searches. All cited references were hand-searched in order to identify other important published works. Peri-ampullary and gallbladder carcinoma series were excluded from the survey. Case reports, studies published in the abstract form only, unpublished works and publications in non-peer-reviewed journals were excluded. In vitro investigations and animal studies were not included.

2. Tumour suppressor genes

2.1. p53 [Table 1]

p53, which is located on chromosome 17p13.1, is responsible for cell cycle regulation at the G1/S and G2/M checkpoints. It is known to induce apoptosis in response to severe damage to cellular DNA, through Bax activity regulation. Mutation of the p53 gene results in unchecked replication of defective DNA and possible progression to cancer.10 p53 gene mutation has been shown to be present in 28–61% of CCs.11, 12, 13, 14 Wild-type p53 protein has a short half-life, rendering it difficult to detect by immunohistochemical (IHC) methods. However, the mutated form is significantly more stable, possibly due to protein-stabilising conformational alteration, and this enables detection.15 Several studies of protein over-expression show no association with outcome.14, 16, 17, 18, 19, 20, 21, 22, 23, 24 The remainder demonstrate conflicting results. A number of groups found an association with decreased overall survival,11, 12, 25, 26, 27, 28 but only one of these has demonstrated independent prognostic significance.29 However, in a large series, Argani and colleagues showed an increased overall survival in their p53 positive group.30 A number of factors may account for these disparities including use of different cut-off values for positivity, antibodies and inclusion of varying anatomical and histological sub-types of CC. Indeed, two studies have noted increasing p53 expression in more distal tumours in comparison to proximal ones when a sub-analysis is undertaken.28, 30 This may indicate a more important role for p53 over-expression in extra-hepatic, distal bile duct carcinomas. The prognostic role of p53 protein over-expression therefore remains to be fully established.

Table 1.

Studies examining the prognostic role of p53 mutation and protein expression in cholangiocarcinoma


Study	Year	Patients (n)	Method	Cut-off for positivitya	% Positive cases	Prognostic role
Kuroda et al.16	2007	93	IHC	>10%	37	No significant prognostic role
Cheng et al.29	2007	112	IHC	>30%	33	p53-positivity associated with decreased OS on UVA and independently predicted decreased OS (P<0.01)
Limpaiboon et al.14	2002	55	DNA-PCR and electrophoresis	p53 MSI	0	No significant prognostic role
Limpaiboon et al.14	2002	55	DNA-PCR and electrophoresis	p53 LOH	32	No significant prognostic role
Hoang et al.17	2002	9	IHC	>30%	50	No significant prognostic role
Argani et al.30	2001	116	IHC	>30%	51	p53-positivity associated with increased OS on UVA, but NS on MVA
Li et al.18	2000	34	IHC	>20%	50	No significant prognostic role
Tannapfel et al.12	2000	41	IHC	Not reported	–	p53 mutation associated with decreased OS on UVA, NS on MVA
Tannapfel et al.12	2000	41	PCR-SSCP	p53 mutation	37	p53 mutation associated with decreased OS on UVA, NS on MVA
Havlik et al.11	2000	18	PCR-SSCP	p53 mutation	28	p53 mutation associated with decreased OS on UVA
Ahrendt et al.25	2000	12	IHC	>10%	54	p53-positivity associated with decreased OS on UVA
Albores- Saavedra et al.19	2000	9	IHC	>30%	42	No significant prognostic role (Papillary carcinomas)
Rijken et al.26	1999	47	IHC	>30%	40	p53-positivity associated with decreased OS on UVA
Arora et al.20	1999	28	IHC	>1%	86	No significant prognostic role
Kim et al.21	1999	28	IHC	>10%	36	No significant prognostic role
Shrestha et al.22	1998	32	IHC	>10%	56	No significant prognostic role
Suto et al.23	1998	54	IHC	>5%	44	No significant prognostic role
Takamori et al.27	1997	17	IHC	>5%		p53-positivity associated with decreased OS on UVA
Washington et al.24	1996	27	IHC	>30%	38	No significant prognostic role
Diamantis et al.28	1995	45	IHC	>10%	64	p53-positivity associated with a trend towards decreased OS

IHC, immunohistochemistry; OS, overall survival; UVA, univariate analysis; MVA, multivariate analysis; NS, not significant; PCR-SSCP, polymerase chain reaction-single-stranded conformational polymorphism; MSI, microsatellite instability; LOH, loss of heterozygosity.

Immunohistochemical cut-offs indicate the percentage of cells with positively staining nuclei unless stated otherwise.

Evidence from studies utilising PCR-based P53 gene mutation analysis is more consistent. Mutation appears to be associated with a decreased overall survival. However, with only two studies available in the literature examining 18 and 41 tumours, respectively,11, 12 further work to confirm this association is desirable.

2.2. Other tumour suppressor genes

DPC4 (Smad4) is a tumour suppressor gene whose protein regulates transduction of signals from the transforming growth factor-β (TGFβ) superfamily.17 TGFβ is a major cell proliferation inhibitor. Loss of DPC4 is thought to lead to progression in the cell cycle from G1 to S phase and subsequent increased proliferation.31 In two studies of DPC4 protein expression, loss was noted in 55% and 60% of extra-hepatic bile duct carcinomas. However, there was no correlation with survival.17, 30

The retinoblastoma gene encodes the Rb protein which, in its unphosphorylated state, blocks transition from G1 to S phase in the cell cycle and is necessary for prevention of cell replication when DNA has been damaged.32 In 51 intra-hepatic and 34 extra-hepatic CCs, no correlation between Rb expression and survival could be shown.33, 34

The p73 gene has significant homology with p53, though mutation in human tumours has not been found. However, over-expression of the p73 protein is thought to induce inhibition of cell growth and apoptosis through activation of p53-responsive genes.35 One study of 41 resected intra-hepatic CCs found that 32% of tumours over-expressed p73 protein (>10% nuclei positively stained) and this correlated with improved survival on univariate analysis, but lost significance upon multivariate analysis.36

3. Oncogenes

3.1. k-ras

The ras oncogenes encode a family of signal transduction proteins downstream of growth factor receptors. Mutation in the k-ras gene can permanently activate the protein leading to increased cellular proliferation and growth.37 In five studies examining k-ras mutation in CC, four showed no significant correlation with survival,38, 39, 40, 41 with one group demonstrating a significant decrease in survival.25 However, this study examined tissues from 12 patients with cancer arising on a background of primary sclerosing cholangitis; conceivably a different neoplastic entity.

3.2. KIT (CD117)

KIT is a type three receptor tyrosine kinase which binds mast cell growth factor and can effect cellular proliferation, growth and differentiation.42 Hong and colleagues examined a large series of 289 extra-hepatic bile duct carcinomas using IHC. Of these, 19% and 20% had cytoplasmic and nuclear reactivity respectively. Positive cytoplasmic expression significantly correlated with improved survival on both univariate and multivariate analyses (P=0.01 and P=0.04).43

4. Apoptotic regulation

Evasion of programmed cell death (apoptosis) is one of the key hallmarks of malignant growth. Furthermore, loss of the normal control of cell longevity is also thought to confer increased resistance to chemotherapeutic agents, many of which utilise these pathways to induce cell death.

4.1. Bcl-2/Bax

Members of the Bcl-2 protein family, such as the pro-apoptotic Bax, Bak and BAD, and the anti-apoptotic Bcl-2, Bcl-w and Bcl-xl, influence mitochondrial membrane permeability.44 Of this family, Bcl-2 and Bax have been investigated in CC. In a study of 28 resected tumours, no expression of Bcl-2 was found by IHC, though Bax expression was not evaluated.20 In 22 resected intra-hepatic CCs from Italy, Bax was expressed in 45%, but had no correlation with outcome and Bcl-2 was not assessed. Of note, however, Maspin, a multifunctional protein which induces apoptosis by modulation of Bcl proteins, was reported to be expressed in 77% of cases. Low expression correlated with increased depth of invasion (P=0.043), vascular invasion (P=0.027) and was an independent prognostic factor for poor outcome on multivariate analysis.45 The role of the Bcl-2 family remains to be fully established as proteins have only been evaluated on an individual basis in CC, and need to be assessed relative to each other in order to gauge apoptotic influence. Further studies examining Bcl-2, Bax and Maspin expression together would be of value.

4.2. DAP-kinase

Death-associated protein kinase (DAP-kinase) is a calcium/calmodulin-dependent protein which participates in several pro-apoptotic systems, with loss of expression leading to decreased cell death.46 DAP-kinase promoter hyper-methylation, and subsequent decreased protein expression, was found to be significantly associated with higher tumour grade (P=0.043), decreased overall survival (P=0.009) and an independent prognostic marker (P=0.024).47

4.3. Fas/FasL

Expression of Fas/FasLigand (FasL) is thought to mediate apoptosis by interaction with T lymphocytes in the immune system. Cross-linking of the Fas receptor with either FasL or activated antibody induces apoptosis of Fas-expressing cells. Investigators have hypothesised that tumours may induce Fas-mediated apoptosis in lymphocytes by expressing FasL, thus creating an immune-privileged site for increased growth.48 Two groups have assessed Fas and FasL by IHC in Japanese patients. In both study groups, tumours found to strongly express Fas had a significantly better prognosis on univariate analysis, possibly due to increased tumoural apoptosis. FasL expression did not correlate with outcome.49, 50

4.4. Survivin

Survivin is a strong negative regulator of apoptosis. By inhibiting or preventing the activation of caspases, it promotes cellular survival in otherwise apoptotic conditions.51 Over-expression has been associated with aggressive behaviour and poor outcome in a number of malignancies.52, 53, 54 Javle and colleagues found strong nuclear expression in 17% of tumours and correlated it with poor survival (P=0.033), and also found expression to be an independent poor prognostic factor.55

5. Cell cycle regulators [Table 2]

5.1. Cyclin and cyclin-dependent kinases

Cyclin/cyclin-dependent kinase complexes are key to the progression of cells through the G1/S phase of the cell cycle by inactivating Rb through phosphorylation. Over-expression is postulated to increase cell turnover and proliferative activity.56 Two studies examining immunohistochemical cyclin D1 in a total of 110 patients have found over-expression to be associated with poor survival on univariate analysis,57, 58 with one group finding it to be an independent prognostic factor.57 Over-expression of Cyclin B1, which acts at the G2/M phase checkpoint of the cell cycle, has also been shown to be a poor prognostic indicator on univariate analysis. In the same study, patients with decreased 14-3-3σ expression (an inhibitor of cyclin B1 inducible by p53 in response to DNA damage) had poorer overall survival, in both univariate and multivariate analyses.16

Table 2.

Studies examining cell cycle regulators in cholangiocarcinoma and their prognostic significance


Study	Year	Patients (n)	Marker (method)	+ve(+)/–ve(–) regulator	Cut-off for positivitya	% Positive cases	Prognostic role
Kuroda et al.16	2007	93	Cyclin B1 (IHC)	+	>4%	82	High Cyclin B1 associated with decreased OS on UVA
Kuroda et al.16	2007	93	14-3-3σ (IHC)	–	>30%	68	Low 14-3–3σ associated with decreased OS on UVA and independently prognostic on MVA
Hui et al.57	1999	34	Cyclin D1 (IHC)	+	>5%	68	High cyclin D1 associated with decreased OS on UVA and independently prognostic on MVA
Hui et al.57	1999	34	p27KIP1 (IHC)	–	>50%	44	Low p27KIP1 associated with decreased OS on UVA and independently prognostic on MVA
Sugimachi et al.58	2001	66	Cyclin D1 (IHC)	+	>10%	42	High Cyclin D1 associated with decreased OS on UVA
Li et al.18	2000	34	p21WAF1/CIP1 (IHC)	–	>30%	15	p21WAF1/CIP1 –ve and high expressors associated with increased recurrence and decreased DFS on UVA
Fiorentino et al.59	2001	47	p27 (IHC)	–	>50%	38	Low p27 associated with decreased OS on UVA and independently prognostic on MVA
Taguchi et al.60	2001	62	p27KIP1 (IHC)	–	>50%	73	Low p27KIP1 associated with decreased OS on UVA and independently prognostic on MVA
Shibahara et al.61	2004	27	p27 (IHC)	–	>50%	63	No significant prognostic role
Tamada et al.62	2006	70	p27 (IHC)	–	>20%	57	No significant prognostic role
Tannapfel et al.33	2002	51	INK4a(MS-PCR)	–	Methylation status	76	No significant prognostic role
Tannapfel et al.33	2002	51	ARF (MS-PCR)	–	Methylation status	25	No significant prognostic role
Tannapfel et al.41	2000	41	INK4a(MS-PCR)	–	Methylation status	83	No significant prognostic role
Tannapfel et al.41	2000	41	p16 (IHC)	–	Not reported	17	No significant prognostic role
Hong et al.64	2006	90	INK4a(MS-PCR)	–	Methylation status	77	No significant prognostic role
Hong et al.64	2006	90	p16 (IHC)	–	Any nuclear staining	48	No significant prognostic role
Ichikawa et al.65	2002	99	p16 (IHC)	–	Diffuse staining	19	No significant prognostic role
Cui et al.34	2000	34	p16 (IHC)	–	>1%	35	No significant prognostic role

IHC, immunohistochemistry; MS-PCR, methylation specific polymerase chain reaction; OS, overall survival; DFS, disease-free survival; UVA, univariate analysis; MVA, multivariate analysis.

Immunohistochemical cut-offs indicate the percentage of cells with positively staining nuclei unless stated otherwise.

5.2. Cyclin-dependent kinase inhibitors (CKI)

5.2.1. KIP/CIP family

Cyclin-dependent kinase inhibitors block G1/S phase transition, and decreased expression is thought to promote tumourigenesis.56 However, in CC evidence relating to their prognostic role is conflicting. Both over- and under-expression of p21WAF1/CIP1, a p53 inducible universal CKI, have been found to be associated with decreased disease-free survival.18 Low expression of another CKI, p27, has been shown to be a negative prognostic factor in several malignancies. In five studies of CC, all utilising IHC, three demonstrated a significant association between decreased p27 expression and poor prognosis,57, 59, 60 with the other two finding no such correlation.61, 62 The reasons for this disparity are unclear, as all these studies used similar evaluation methods, statistical analyses and involved similar numbers of patients. Further investigation is required to clarify the role of CKIs as prognostic indicators.

5.2.2. INK4 family

p16 suppresses cyclin-dependent kinase 4-Cyclin D complexes which control phosphorylation of Rb. Lack of p16 therefore up-regulates Rb phosphorylation allowing possible inappropriate progression to S phase.63 Hyper-methylation and subsequent silencing of the p16 promoter are common, being present in 76–83% of CCs.33, 41, 64 p16 protein expression has been assessed by IHC in CC by 4 groups, but no significant prognostic role has been found.33, 34, 41, 65

6. Proliferation indices [Table 3]

An indication of rate of tumour growth and aggressiveness can be gained from measurement of tumour proliferative activity. Several experimental methods exist for this purpose, including quantification of mitotic index, S phase cell fraction count, or immunohistochemical antibody detection of nuclear proteins (ki-67, its epitope MIB-1 or proliferating cell nuclear antigen (PCNA)).

Several studies have utilised the various nuclear proteins detectable by IHC to quantify proliferation and examine the relationship with survival (Table 3). Conflicting results exist but it is evident in several studies, including the majority of the larger study populations, that an association exists between increased proliferative activity and decreased overall survival.12, 16, 22, 23, 27, 39, 66, 67, 68 The conflicting results of some groups may be due to the small sample size investigated, between 6 and 9 patients in some studies,17, 19, 58 or the different criteria for proliferation indices which have been used.21, 69

Table 3.

Studies examining proliferation indices in cholangiocarcinoma and their prognostic significance


Study	Year	Patients (n)	Marker (method)	Cut-off for positivitya	% Positive cases	Prognostic role
Kuroda et al.16	2007	93	MIB-1 (IHC)	>10%	69	High MIB-1 LI associated with decreased overall survival on UVA, NS on MVA
Aishima et al.69	2006	100	Ki-67 (IHC)	>18%	51	No significant prognostic role
Hoang et al.17	2002	9	MIB-1 (IHC)	>30%	89	No significant prognostic role
Sugimachi et al.58	2001	6	Ki-67 (IHC)	Nuclear staining	Not reported	High ki-67 LI independently predicted OS on MVA (P=0.0036)
Albores-Saavedra et al.19	2000	9	MIB-1	Not reported	Not reported	No significant prognostic role (Papillary carcinoma)
Tannapfel et al.12	2000	41	MIB-1 and PCNA (IHC)	Nuclear staining	Not reported	No significant prognostic role
Kim et al.21	1999	28	Ki-67 (IHC)	Nuclear staining	100	No significant prognostic role
Rijken et al.68	1998	35	MIB-1 (IHC)	>20%	34	High MIB-1 LI associated with decreased OS on UVA
Shrestha et al.22	1998	32	MIB-1 (IHC)	>57%	Not reported	High MIB-1 LI associated with decreased OS on UVA
Suto et al.23	1998	54	MIB-1 (IHC)	>29%	46	High MIB-1 LI associated with decreased OS on UVA, NS on MVA
Nishida et al.66	1997	35	PCNA (IHC)	>45%	29	High PCNA associated with decreased OS and independently predicted decreased OS on MVA (P=0.012)
Takamori et al.27	1997	17	Ki-67 (IHC)	>20%	62	High ki-67 LI at tumour margin associated with decreased OS on UVA
Yamada et al.67	1995	21	MIB-1 (IHC)	>10%	Not reported	High MIB-1 LI associated with decreased OS on UVA

IHC, immunohistochemistry; OS, overall survival; UVA, univariate analysis; MVA, multivariate analysis; NS, not significant; LI, labelling index.

Immunohistochemical cut-offs indicate the percentage of cells with positively staining nuclei unless stated otherwise.

7. Growth factors and receptors

7.1. Epidermal growth factor receptor (EGFR) family

The trans-membrane tyrosine kinase receptors of the EGFR family, which include EGFR (ErbB-1, HER-1), HER-2/neu (ErbB-2), HER-3 (ErbB-3), and HER-4 (ErbB-4) play a significant role in cellular growth and proliferation signalling.70

EGFR expression was assessed by IHC in 24 patients with CC and was unrelated to survival.71 Similarly, Gwak and colleagues performed DNA sequencing for EGFR mutation and they also failed to find any correlation with outcome.72

In advanced breast cancer, HER-2/neu over-expression has attracted considerable interest due to correlations with prognosis and has subsequently been successfully targeted with monoclonal antibodies.73 In CC, however, four groups have examined HER-2/neu expression using IHC and found no correlation with survival.61, 62, 74, 75

7.2. Tenascin

The extra-cellular matrix glycoprotein Tenascin may directly bind the EGFR receptor, activating the EGFR kinase cascade and subsequently increasing the proliferative and migratory activity of tumour cells.76 Aishima and colleagues found the expression of Tenascin at the invasive front of tumours to correlate with decreased overall survival, though this was not significant on multivariate analysis.77

7.3. CCN proteins

The CCN protein family (named by the first letters of the first three proteins discovered – Cyr61, CTGF and Nov) includes connective tissue growth factor (CTGF), a highly fibrogenic protein, which may promote desmoplastic reactions, possibly decreasing the ability of tumours to invade and metastasise. However, over-expression has been shown in several epithelial and non-epithelial malignancies, but prognostic implications remain to be confirmed.78 In the only study in CC, Gardini and colleagues found over-expression in 60% of tumours, demonstrating a significantly improved disease-free and overall survival in this group compared with low-expressors. On multivariate analysis, CTGF expression independently predicted disease-free and increased overall survival.79

The Wnt-inducible secreted proteins (WISP1, WISP2 and WISP3) are three further members of the CCN family.80 These proteins appear to have complex roles in cell signalling and their prognostic implications in cancer remain to be fully established.81 In CC, the WISP1v splice variant has been to shown to be expressed in 49% of intra-hepatic tumours at the mRNA level and this correlated with reduced overall survival on univariate analysis.82

8. Invasion and metastasis

8.1. Cadherins

Mikami and colleagues examined expression of E-cadherin and the alpha- and beta-catenins in 55 extra-hepatic CCs. These important molecules in calcium-dependent intercellular adhesion have been shown to be down-regulated in several epithelial malignancies83 and are a hallmark of the epithelial-mesenchymal transition which is thought to play a pivotal role in cancer invasion and metastasis.84 Decreased E-cadherin expression in this group of CCs was significantly associated with poor prognosis on univariate, but not multivariate, analysis. Catenin expression was significantly lower in carcinoma tissues when compared to normal epithelium, but was unrelated to prognosis.85

8.2. Syndecan-1

The trans-membrane heparin sulphate proteoglycan molecule Syndecan-1 (CD138) is a mediator of cell–cell and cell–matrix adhesion and communication.86 Loss of expression was found to be an independent predictor of poor outcome in the only study to express the prognostic role of this factor in CC to date.87

8.3. CEACAM6

Over-expression of carcino-embryonic antigen-related cell adhesion molecule 6 (CEACAM6), a member of the immunoglobulin superfamily, has been associated with a variety of malignancies. Over-expression is thought to enhance metastatic and invasive ability via its ability to influence intra-cellular signalling cascades.88 The only study in CC demonstrated over-expression of CEACAM6 to correlate with decreased disease-free, but not overall, survival on univariate analysis.89

8.4. CD24

CD24 is a heavily glycosylated, mucin-like, cell-surface protein which can act as an alternative ligand for P-selectin. Binding of P-selectin may enhance the metastatic potential of CD24-expressing tumour cells by allowing them to ‘roll’ on the endothelium.90, 91 Su and colleagues found CD24 expression to be an independent predictor of shortened survival in 70 patients undergoing resection of CC (P=0.0113).92 Agrawal and colleagues also found over-expression of CD24 to correlate with decreased overall survival on univariate analysis (P=0.02).93

8.5. CD44

The polymorphic transmembrane glycoprotein CD44 is a cell adhesion molecule which may be expressed in several isoforms due to alternate splicing. It appears to have tissue-specific roles in inflammation and tumourigenesis.94 In a study of 25 extra-hepatic bile duct tumours, the CD44v6 splice variant was found to be absent in tumours with poor prognosis on univariate analysis.95

8.6. Matrix metalloproteinase 7 (MMP7)

Abnormal secretion of matrix metalloproteinases enables tumour invasion and metastasis.96 In the only study to examine matrix metalloproteinase expression in CC, MMP7 over-expression was found to correlate with decreased survival (P=0.003) on univariate analysis.97

9. Angiogenesis

The ability to stimulate new blood vessel formation and so increase nutrient supply is an essential characteristic of any growing tissue, not least malignancy. However, only two studies have investigated this area in relation to its prognostic implications in CC. Park and colleagues found high vascular endothelial growth factor-C (VEGF-C) expression to correlate with shortened overall and disease-free survival in 36 patients undergoing resection of intra-hepatic CC (P<0.01 and <0.05, respectively), with strong VEGF-C expression representing an independent prognostic indicator (P=0.028).98 In contrast, Tang and colleagues investigated several angiogenesis-related factors, including VEGF, micro-vessel density, angiopoietin-1, angiopoietin-2 and thrombospondin-1 expression, with IHC in CC, but found no correlation with survival.99

10. Mucins [Table 4]

Mucins are large proteins synthesised by epithelial cells in many organs. They may be secreted to form a protective barrier at the mucosal surface or may act as transmembrane proteins.100 There is considerable interest in their possible role as targets for chemotherapy.100 Table 4 displays identified studies in CC.

Table 4.

Studies examining mucins in cholangiocarcinoma and their prognostic significance


Study	Year	Patients (n)	Marker (method)	Cut-off for positivitya	% Positive cases	Prognostic role
Shibahara et al.61	2004	27	MUC1 (IHC)	>5%	81	MUC1 positivity associated with decreased OS on UVA
Shibahara et al.61	2004	27	MUC4 (IHC)	>5%	37	MUC4 positivity associated with decreased OS on UVA and independently prognostic on MVA
Tamada et al.62	2006	70	MUC1 (IHC)	>5%	87	MUC1 positivity associated with decreased OS on UVA
Tamada et al.62	2006	70	MUC4 (IHC)	>5%	51	MUC4 positivity associated with decreased OS on UVA and independently prognostic on MVA
Matsumura et al.102	2002	50	MUC1 (IHC)	>5%	76	MUC1 positivity associated with decreased OS on UVA and independently prognostic on MVA
Takao et al.104	1999	73	MUC1 (IHC)	>25%	69	MUC1 positivity associated with decreased OS on UVA and independently prognostic on MVA
Yuan et al.106	2005	37	MUC1 (IHC)	>25%	71	MUC1 positivity associated with decreased OS on UVA (includes HCC, no sub-group analysis)
Higashi et al.101	1999	39	MUC1/DF3	>50%	44	MUC1/DF3 or MUC1/MY.1E12 positivity associated with decreased OS on UVA
			MUC1/MY.1E12	>50%	56
			MUC2 (all IHC)	>50%	18	MUC2 positivity associated with increased OS on UVA
Suh et al.103	2002	30	MUC1 (IHC)	>5%	80	No significant prognostic role
Suh et al.103	2002	30	MUC2 (IHC)	>5%	40	MUC2 positivity associated with increased OS on UVA
Tamada et al.105	2002	60	MUC1/DF3	>5%	77	MUC1/DF3 or MUC1/MY.1E12 positivity associated with decreased OS on UVA
			MUC1/MY.1E12	>5%	78
			MUC2 (all IHC)	>5%	19	MUC2 positivity associated with increased OS on UVA
Boonla et al.107	2005	87	MUC1 (IHC)	>25%	39	MUC1 positivity associated with decreased OS on UVA and independently prognostic on MVA
Boonla et al.107	2005	87	MUC5AC (IHC)	>25%	53	No significant prognostic role
Hong et al.108	2005	193	MUC2 (IHC)	>5%	42	No significant prognostic role
Yonezawa et al.109	1997	16	MUC2 (mRNA)	>1%	38	MUC2 positivity associated with increased OS on UVA
Aishima et al.69	2006	100	MUC2 (IHC)	>1%	9	Combined mucin phenotype associated with OS on UVA and independently prognostic on MVA
			MUC5AC (IHC)	>10%	40
			MUC6 (IHC)	>10%	21
Boonla et al.110	2003	179	MUC5AC (serum)	>1%	63	MUC5AC serum positivity associated with decreased OS on UVA and independently prognostic on MVA

IHC, immunohistochemistry; OS, overall survival; UVA, univariate analysis; MVA, multivariate analysis; NS, not significant; HCC, hepatocellular carcinoma.

Immunohistochemical cut-offs indicate the percentage of cells with positively staining cytoplasm unless stated otherwise.

10.1. MUC1

MUC1 is a trans-membrane mucin thought to be particularly important in the invasive and metastatic potential of CC. Of the nine studies that have examined the prognostic value of MUC1 in CC, all except one have demonstrated a significant correlation between MUC1 expression and decreased survival, both on univariate and multivariate survival analyses.61, 62, 101, 102, 103, 104, 105, 106, 107

10.2. MUC2

In contrast to MUC1, MUC2 acts as a protective protein and is associated with tumours of mucinous type, both in the biliary and pancreatic systems, which carry a more favourable prognosis. The majority of studies in CC show a significantly better prognosis in MUC2 expressing tumours, and a tendency for MUC2 expression to be associated with extra-hepatic rather tumours than with intra-hepatic tumours.101, 103, 105, 108, 109

10.3. MUC4

MUC4 is a trans-membrane mucin which may act as a ligand for receptor tyrosine kinase ErbB2, and is thought to play a role in the regulation of p27, which is involved in control of the G1/S phases of the cell cylcle. High expression has been shown by two groups to significantly correlate with poor survival and act as an independent prognostic factor on multivariate analysis.61, 62

10.4. MUC5AC

This gel-forming secreted mucin has been studied by three groups. Following IHC analysis of 100 resected tumours, a combined mucin phenotype, including MUC2, MUC5AC and MUC6 expression, was found to be an independent prognostic factor on multivariate analyses (P=0.0042).69 Another analysis of 179 serum samples taken pre-operatively from CC patients demonstrated MUC5AC positive patients to have a 2.5-fold higher risk of death on multivariate analysis (P<0.001).110

11. Serum markers

11.1. Carbohydrate antigen 19-9 (CA19-9, sialyl Lewisa serum antigen)

Carbohydrate antigens may be expressed on the tumour cell membrane and are thought to play an important role in extravasation by mediating adhesion to endothelial cells expressing E-selectin.111 Elevation of serum levels has been demonstrated in several malignancies.112 Of the studies relating to prognostic implications in CC, three groups have demonstrated a significant difference in survival when a high cut-off level (>1000ng/ml113, >100ng/ml114, 115) is used to compare groups. Wahab and colleagues also found elevated CA19-9 to be associated with poor survival after resection, both on univariate and multivariate analysis, but did not define their cut-off level.116 At lower cut-off values (>37ng/ml) two studies found no significant association with survival.21, 97 Indeed, Miwa and colleagues, in a further analysis after their 2002 paper, found CA19-9 to be a significant prognostic factor on univariate analysis, but found jaundice confounded elevated CA19-9 on multivariate analysis.117 The conflicting evidence within these studies is likely to be related to the non-standardised cut-off values used in the analyses. In addition, biliary obstruction, whether due to benign or malignant disease, is known to cause elevation of CA19-9, and few groups have allowed for this in multivariate analysis. Overall, high levels of CA19-9 appear to be a poor prognostic indicator.

11.2. Carcinogenic embryonic antigen (CEA)

Serum CEA levels have been examined in five studies. Four of these found that a raised pre-operative CEA had no significant relationship with survival.21, 97, 115, 117 In the remaining study, which examined 35 resected intra-hepatic CCs, there was a significant correlation between high CEA level and decreased survival on univariate and multivariate analyses (P=0.021 and P=0.029, respectively).118 Overall, the evidence suggests that serum CEA measurement is less useful than CA19-9 as a significant prognostic marker.

11.3. Bilirubin

Elevated serum bilirubin has been evaluated as a prognostic indicator and seems to indicate a worse overall survival in the groups analysed.116, 117, 119 This should perhaps be interpreted with caution as, in at least one study, jaundiced patients had a tendency to receive more palliative procedures rather than resection procedures, which may obviously influence survival.119 Further evaluation of serum bilirubin as a prognostic indicator is needed.

11.4. C-reactive protein (CRP)

C-reactive protein (CRP) has been shown to be of prognostic value in many malignancies.120 In 98 peri-hilar CCs, Gerhardt and colleagues found an elevated CRP to be an independent predictor of worse survival (P=0.029).121

11.5. Sialic acid

In a single study of serum sialic acid concentration, Wongkham and colleagues found elevated levels in 75% of CC patients from Thailand, but this had no prognostic association.122

12. Sialyl Lewis antigens in tissue

In addition to the measurement of serum sialyl Lewisa (CA19-9), IHC examination of tissue specimens for various sialyl Lewis antigens has been performed. Further supporting the role of serum sialyl Lewisa as predictive of prognosis, Juntavee and colleagues found sialyl Lewisa expression in 60% of the 110 specimens they examined, are found it to be an independent prognostic indicator on multivariate analysis (P=0.001).123 In contrast, Takao and colleagues examined the sialyl Lewisx and sialosyl-Tn antigens in 73 extra-hepatic bile duct carcinomas and found no relationship with survival.104

13. Genetic factors

13.1. Chromosomal instability/anueploidy

Chromosomal instability and aneuploidy are characteristic features of most aggressive human malignancies.124 Defining features include the loss or gain of the whole or a part of a chromosome and/or gross chromosomal rearrangements.125 Two studies in CC have examined tumours using DNA flow-cytometry, both confirming a significant association between aneuploid tumours and poor survival.126, 127 Iachino and colleagues found aneuploidy to be an independent prognostic marker.127

13.2. Loss of heterozygosity (LOH) and microsatellite instability (MSI)

LOH at genetic loci has been investigated in CC, but the prognostic implications remain unclear. In three studies, no correlation with survival has been shown,128 though Muenphon and colleagues demonstrated a significant association between allelic imbalance and survival at three individual loci.129

Microsatellites are short tandem repeat sequences found throughout the human genome. Differences in the lengths of repeat sequences between normal and tumour DNA from the same patient are used to assess the instability of these microsatellites.130 Increased MSI is thought to be due to abnormalities in the function of DNA mismatch repair genes. Studies examining a single chromosomal locus in CC have found varying MSI rates, though only MSI at chromosomal locus 1p36 D1S228 has been shown to correlate with prognosis.128, 130 However, when more loci are taken together or combined with LOH, the representation of overall tumour genetic instability may give a more accurate indication of aggressive behaviour.14 As in other malignancies,131 newer techniques for genome wide evaluation of damage, such as array comparative genomic hybridisation, may assist in assessing the prognostic implications of genomic instability as a whole.

13.3. Genetic repair

As part of the mismatch repair pathway, O6-methylguanine-DNA methyltransferase (MGMT) participates in the repair of DNA which has been incorrectly transcribed during replication.132 In CC, MGMT expression has been found to be a good prognostic indicator on univariate analysis.133 In a further study, methylation of the MGMT promoter (which closely correlated with decreased protein expression) was significantly associated with poor survival.134 Loss of MGMT expression is thought to be one of the factors responsible for increased MSI and overall genetic instability.

13.4. Genotype

The GNAS1 locus encodes the Gαs subunit of G protein, which is involved in the cAMP pathway, participating in control of apoptosis and proliferation. Schmitz and colleagues determined the GNAS1 T393C genotype in 87 resected intra-hepatic CCs and found the TT genotype to significantly associate with decreased survival on univariate and multivariate analyses.135 The proliferation indices in TT genotype tumours were significantly elevated, suggesting a reduced function of the G protein in these patients. Further studies are needed to fully clarify the implications of these results.

14. Other markers

14.1. Cytokeratins

Cytokeratins are intermediate keratin filaments which form part of the intra-cytoplasmic cytoskeleton of epithelial cells, and there is increasing interest in their application as serum biomarkers.136 Keratin 903 monoclonal antibody can be used to stain tissues to assess expression of the 1, 5, 10 and 14 cytokeratins. In a study of 73 resected mass-forming CCs, patients with Keratin 903 reactivity were found to have significantly improved survival on univariate and multivariate analyses.137

14.2. Neuroendocrine differentiation

Neuroendocrine differentiation has been shown to be of prognostic importance in several malignancies.138 In CC, synaptophysin has been found to be a significant independent negative prognostic marker.139

14.3. Host immune response

One study of dendritic and T-cell localisation in CC found a significant improvement in survival of patients with positive dendritic cell markers (CD83 and CD1a) and T-cell markers (CD8 and CD4), as opposed to their negative counterparts, presumably denoting an increased anti-tumour response140.

14.4. Pancreatic secretory trypsin inhibitor (PSTI)

Having previously been shown to be an independent prognostic marker in bladder141, ovarian142 and renal143 carcinomas, PSTI expression was investigated in CC specimens. High expression was associated with decreased recurrence-free survival on univariate analysis.144

14.5. Cyclooxygenase-2 (COX-2)

The involvement of COX in cancer is now well established and its potential as a target for chemo-preventive and/or chemotherapeutic inhibitors is under investigation.145 In CC the role of COX-2 remains unclear, with one study suggesting a significant association between over-expression and decreased overall survival,146 but two others finding no relationship.71, 147 Different methods of quantifying COX-2 expression may account for the discrepancies between these studies. Without further investigation the prognostic value of COX-2 remains a matter for debate.

15. Conclusions

Research in CC has yielded a multitude of molecular markers which have a significant role in predicting outcome. These biomarkers hold promising application for the future, both in terms of patient counselling with stratification of risk and individualised treatment strategies, and the potential to target some of these molecules with novel therapeutic agents. Individualised treatment holds obvious advantages, allowing patients at high risk of recurrence to receive aggressive adjuvant therapy, and those with very low risk to avoid the possible morbidity associated with chemotherapeutic drugs. Follow-up, normally based around a generalised protocol for all patients, may also be individualised according to the molecular profile of the tumour, potentially allowing serum monitoring alone to guide imaging and further therapy.

In this review, we have identified several biomarkers of particular interest which appear to carry prognostic significance. Of these, p53 mutation, cyclins, proliferation indices, mucins, CA19-9, CRP and aneuploidy appear to hold potential as predictors of outcome. In other types of hepatobiliary malignancy angiogenesis factors, growth factors and cyclooxygenase have been shown to be of value and these should be evaluated as a priority in CC.148, 149, 150 Few studies have examined circulating serum markers in CC. The use of simple serum analysis holds obvious benefits to the patient over the need to obtain tissue samples by biopsy or at operation and would facilitate earlier prediction of outcome. However, given the continual increase in the plethora of potential molecular markers in CC, application of genomic and proteomic micro-array technology should allow measurement and analysis of a number of these markers simultaneously in the future. Application of these techniques will produce gene or protein profiles, allowing much more accurate stratification of individual risk and the recognition of utility of each biomarker in individual cases. The first steps have already been taken in this area and although direct prognostic implications have not been assessed, initial studies indicate correlations with increased stage and presence of lymph node metastases, well-recognised markers of poor outcome.151, 152

The majority of studies to date in CC have utilised retrospective analyses, with differing methodologies. These factors may explain inconsistencies in results from individual markers. In addition, CC is a complex, heterogeneous disease with a wide variety of aetiological factors in different parts of the world. A large proportion of the available literature comes from studies in southeast Asian populations. Biological behaviour of these tumours at the molecular level may be significantly different to other populations due to the increased incidence of pro-inflammatory conditions, such as hepatitis and liver fluke infestation. However, given the higher incidence of CC in these populations, these studies investigate significant numbers of patients and provide some of the best evidence available to date.

Allied to this idea of varying epidemiological conditions in different parts of the world is the predisposition of some patients with chronic conditions, such as primary sclerosing cholangitis, to CC. Patients with presumed sporadic CC may represent a different biological and pathological type of disease to those in whom CC arises on a background of chronic disease. It remains to be elucidated whether the patients in these sub-groups express different biomarkers which may indicate more or less aggressive disease.

It is widely accepted in the literature that CC from different anatomical locations may represent different sub-groups of biological disease. Indeed, due to the relatively low incidence of CC, many studies have grouped these tumours together; though in studies where sub-group analyses have been performed clear differences are apparent. In future investigations the authors would suggest that it is desirable to standardise the approach to studying molecular prognostic markers in CC. However, given the relatively low incidence of CC, multiple sub-groups are undesirable. One strategy, as utilised in several studies, would be to clearly divide CC into intra-hepatic and extra-hepatic tumours for analyses, placing hilar tumours in the intra-hepatic group. In addition, the intra-hepatic group appears to contain three distinct histological types (mass-forming, intra-ductal growth and peri-ductal-infiltrating types) and sub-analyses of these groups would be of interest.

In summary, we have reviewed the available evidence with regard to prognostic molecular biomarkers in CC. Several of these biomarkers appear to hold significant potential both as prognostic indicators and as targets for future therapeutic agents. The application of these markers in clinical practice possibly holds the key to significant advances in this disease.

Conflict of interest statement

None declared.

References

1. 1West J, Wood H, Logan RF, Quinn M, Aithal GP. Trends in the incidence of primary liver and biliary tract cancers in England and Wales 1971–2001. Br J Cancer. 2006;94(11):1751–1758. MEDLINE | CrossRef

2. 2Taylor-Robinson SD, Toledano MB, Arora S, et al. Increase in mortality rates from intrahepatic cholangiocarcinoma in England and Wales 1968–1998. Gut. 2001;48(6):816–820. MEDLINE | CrossRef

3. 3Patel T. Increasing incidence and mortality of primary intrahepatic cholangiocarcinoma in the United States. Hepatology. 2001;33(6):1353–1357. MEDLINE | CrossRef

4. 4Malhi H, Gores GJ. Cholangiocarcinoma: modern advances in understanding a deadly old disease. J Hepatol. 2006;45(6):856–867. Abstract | Full Text | Full-Text PDF (1225 KB) | CrossRef

5. 5Khan SA, Davidson BR, Goldin R, et al. Guidelines for the diagnosis and treatment of cholangiocarcinoma: consensus document. Gut. 2002;51(Suppl. 6):VII-9.

6. 6Su CH, Tsay SH, Wu CC, et al. Factors influencing postoperative morbidity, mortality, and survival after resection for hilar cholangiocarcinoma. Ann Surg. 1996;223(4):384–394. MEDLINE | CrossRef

7. 7Cameron JL, Pitt HA, Zinner MJ, Kaufman SL, Coleman J. Management of proximal cholangiocarcinomas by surgical resection and radiotherapy. Am J Surg. 1990;159(1):91–97[discussion 97–8]. MEDLINE | CrossRef

8. 8Nimura Y, Hayakawa N, Kamiya J, Kondo S, Shionoya S. Hepatic segmentectomy with caudate lobe resection for bile duct carcinoma of the hepatic hilus. World J Surg. 1990;14(4):535–543[discussion 544]. MEDLINE | CrossRef

9. 9Nakeeb A, Tran KQ, Black MJ, et al. Improved survival in resected biliary malignancies. Surgery. 2002;132(4):555–563[discussion 563–64]. Abstract | Full Text | Full-Text PDF (88 KB) | CrossRef

10. 10Feng Z, Hu W, Rajagopal G, Levine AJ. The tumor suppressor p53: cancer and aging. Cell Cycle. 2008;7(7):842–847.

11. 11Havlik R, Sbisa E, Tullo A, et al. Results of resection for hilar cholangiocarcinoma with analysis of prognostic factors. Hepatogastroenterology. 2000;47(34):927–931. MEDLINE

12. 12Tannapfel A, Weinans L, Geissler F, et al. Mutations of p53 tumor suppressor gene, apoptosis, and proliferation in intrahepatic cholangiocellular carcinoma of the liver. Dig Dis Sci. 2000;45(2):317–324. MEDLINE | CrossRef

13. 13Liu XF, Zhang H, Zhu SG, et al. Correlation of p53 gene mutation and expression of P53 protein in cholangiocarcinoma. World J Gastroenterol. 2006;12(29):4706–4709. MEDLINE

14. 14Limpaiboon T, Krissadarak K, Sripa B, et al. Microsatellite alterations in liver fluke related cholangiocarcinoma are associated with poor prognosis. Cancer Lett. 2002;181(2):215–222. Abstract | Full Text | Full-Text PDF (115 KB) | CrossRef

15. 15Jenkins JR, Rudge K, Chumakov P, Currie GA. The cellular oncogene p53 can be activated by mutagenesis. Nature. 1985;317(6040):816–818. MEDLINE | CrossRef

16. 16Kuroda Y, Aishima S, Taketomi A, et al. 14-3-3sigma negatively regulates the cell cycle, and its down-regulation is associated with poor outcome in intrahepatic cholangiocarcinoma. Hum Pathol. 2007;38(7):1014–1022. Abstract | Full Text | Full-Text PDF (1526 KB)

17. 17Hoang MP, Murakata LA, Katabi N, Henson DE, Albores-Saavedra J. Invasive papillary carcinomas of the extrahepatic bile ducts: a clinicopathologic and immunohistochemical study of 13 cases. Mod Pathol. 2002;15(12):1251–1258. MEDLINE | CrossRef

18. 18Li X, Hui A, Takayama T, Cui X, Shi Y, Makuuchi M. Altered p21(WAF1/CIP1) expression is associated with poor prognosis in extrahepatic bile duct carcinoma. Cancer Lett. 2000;154(1):85–91. Abstract | Full Text | Full-Text PDF (446 KB) | CrossRef

19. 19Albores-Saavedra J, Murakata L, Krueger JE, Henson DE. Noninvasive and minimally invasive papillary carcinomas of the extrahepatic bile ducts. Cancer. 2000;89(3):508–515.

20. 20Arora DS, Ramsdale J, Lodge JP, Wyatt JI. p53 but not bcl-2 is expressed by most cholangiocarcinomas: a study of 28 cases. Histopathology. 1999;34(6):497–501. MEDLINE | CrossRef

21. 21Kim HJ, Yun SS, Jung KH, Kwun WH, Choi JH. Intrahepatic cholangiocarcinoma in Korea. J Hepatobiliary Pancreat Surg. 1999;6(2):142–148. MEDLINE | CrossRef

22. 22Shrestha ML, Miyake H, Kikutsuji T, Tashiro S. Prognostic significance of Ki-67 and p53 antigen expression in carcinomas of bile duct and gallbladder. J Med Invest. 1998;45(1–4):95–102. MEDLINE

23. 23Suto T, Sugai T, Nakamura S, et al. Assessment of the expression of p53, MIB-1 (Ki-67 antigen), and argyrophilic nucleolar organizer regions in carcinoma of the extrahepatic bile duct. Cancer. 1998;82(1):86–95.

24. 24Washington K, Gottfried MR. Expression of p53 in adenocarcinoma of the gallbladder and bile ducts. Liver. 1996;16(2):99–104. MEDLINE

25. 25Ahrendt SA, Rashid A, Chow JT, Eisenberger CF, Pitt HA, Sidransky D. p53 overexpression and k-ras gene mutations in primary sclerosing cholangitis-associated biliary tract cancer. J Hepatobiliary Pancreat Surg. 2000;7(4):426–431. MEDLINE | CrossRef

26. 26Rijken AM, Offerhaus GJ, Polak MM, Gouma DJ, van Gulik TM. P53 expression as a prognostic determinant in resected distal bile duct carcinoma. Eur J Surg Oncol. 1999;25(3):297–301. Abstract | Full-Text PDF (352 KB) | CrossRef

27. 27Takamori S, Yoshimoto J, Kojima K, Matsumoto T, Beppu T, Futagawa S. Prognostic implication and intratumoral heterogeneity of p53, Ki-67, CEA and CA19-9 expressions in cholangiocellular carcinoma. Hepatol Res. 1997;8(2):81–94.

28. 28Diamantis I, Karamitopoulou E, Perentes E, Zimmermann A. p53 protein immunoreactivity in extrahepatic bile duct and gallbladder cancer: correlation with tumor grade and survival. Hepatology. 1995;22(3):774–779. MEDLINE | CrossRef

29. 29Cheng Q, Luo X, Zhang B, Jiang X, Yi B, Wu M. Distal bile duct carcinoma: prognostic factors after curative surgery. A series of 112 cases. Ann Surg Oncol. 2007;14(3):1212–1219. MEDLINE

30. 30Argani P, Shaukat A, Kaushal M, et al. Differing rates of loss of DPC4 expression and of p53 overexpression among carcinomas of the proximal and distal bile ducts. Cancer. 2001;91(7):1332–1341.

31. 31Ross S, Hill CS. How the smads regulate transcription. Int J Biochem Cell Biol. 2008;40(3):383–408. CrossRef

32. 32Sun A, Bagella L, Tutton S, Romano G, Giordano A. From G0 to S phase: a view of the roles played by the retinoblastoma (Rb) family members in the Rb-E2F pathway. J Cell Biochem. 2007;102(6):1400–1404. CrossRef

33. 33Tannapfel A, Sommerer F, Benicke M, et al. Genetic and epigenetic alterations of the INK4a-ARF pathway in cholangiocarcinoma. J Pathol. 2002;197(5):624–631. MEDLINE | CrossRef

34. 34Cui X, Hui AM, Li X, et al. Alterations of retinoblastoma protein and p16INK4 protein expression in extrahepatic bile duct carcinomas. Hepatogastroenterology. 2000;47(35):1216–1220. MEDLINE

35. 35Buhlmann S, Putzer BM. DNp73 a matter of cancer: mechanisms and clinical implications. Biochim Biophys Acta. 2008;1785(2):207–216.

36. 36Tannapfel A, Engeland K, Weinans L, et al. Expression of p73, a novel protein related to the p53 tumour suppressor p53, and apoptosis in cholangiocellular carcinoma of the liver. Br J Cancer. 1999;80(7):1069–1074. MEDLINE | CrossRef

37. 37Rodriguez-Viciana P, Tetsu O, Oda K, Okada J, Rauen K, McCormick F. Cancer targets in the Ras pathway. Cold Spring Harb Symp Quant Biol. 2005;70:461–467. MEDLINE | CrossRef

38. 38Isa T, Tomita S, Nakachi A, et al. Analysis of microsatellite instability, K-ras gene mutation and p53 protein overexpression in intrahepatic cholangiocarcinoma. Hepatogastroenterology. 2002;49(45):604–608. MEDLINE

39. 39Rijken AM, van Gulik TM, Polak MM, Sturm PD, Gouma DJ, Offerhaus GJ. Diagnostic and prognostic value of incidence of K-ras codon 12 mutations in resected distal bile duct carcinoma. J Surg Oncol. 1998;68(3):187–192. MEDLINE | CrossRef

40. 40Lee CS. Ras p21 protein immunoreactivity and its relationship to p53 expression and prognosis in gallbladder and extrahepatic biliary carcinoma. Eur J Surg Oncol. 1997;23(3):233–237. Abstract | Full-Text PDF (1431 KB) | CrossRef

41. 41Tannapfel A, Benicke M, Katalinic A, et al. Frequency of p16(INK4A) alterations and K-ras mutations in intrahepatic cholangiocarcinoma of the liver. Gut. 2000;47(5):721–727. MEDLINE | CrossRef

42. 42Miettinen M, Lasota J. KIT (CD117): a review on expression in normal and neoplastic tissues, and mutations and their clinicopathologic correlation. Appl Immunohistochem Mol Morphol. 2005;13(3):205–220. MEDLINE

43. 43Hong SM, Hwang I, Song DE, Choi J, Yu E. Clinical and prognostic significances of nuclear and cytoplasmic KIT expressions in extrahepatic bile duct carcinomas. Mod Pathol. 2007;20(5):562–569. MEDLINE | CrossRef

44. 44Del Gaizo Moore V, Letai A. Rational design of therapeutics targeting the BCL-2 family: are some cancer cells primed for death but waiting for a final push?. Adv Exp Med Biol. 2008;615:159–175. CrossRef

45. 45Romani AA, Soliani P, Desenzani S, Borghetti AF, Crafa P. The associated expression of Maspin and Bax proteins as a potential prognostic factor in intrahepatic cholangiocarcinoma. BMC Cancer. 2006;6:255. MEDLINE | CrossRef

46. 46Gozuacik D, Kimchi A. DAPk protein family and cancer. Autophagy. 2006;2(2):74–79. MEDLINE

47. 47Tozawa T, Tamura G, Honda T, et al. Promoter hypermethylation of DAP-kinase is associated with poor survival in primary biliary tract carcinoma patients. Cancer Sci. 2004;95(9):736–740. MEDLINE | CrossRef

48. 48Whiteside TL. The role of death receptor ligands in shaping tumor microenvironment. Immunol Invest. 2007;36(1):25–46. MEDLINE | CrossRef

49. 49Murakami M, Sasaki T, Miyata H, Yamasaki S, Kuwahara K, Chayama K. Fas and Fas ligand: expression and soluble circulating levels in bile duct carcinoma. Oncol Rep. 2004;11(6):1183–1186.

50. 50Shimonishi T, Isse K, Shibata F, et al. Up-regulation of fas ligand at early stages and down-regulation of Fas at progressed stages of intrahepatic cholangiocarcinoma reflect evasion from immune surveillance. Hepatology. 2000;32(4 Pt 1):761–769. MEDLINE | CrossRef

51. 51Pennati M, Folini M, Zaffaroni N. Targeting survivin in cancer therapy. Expert Opin Ther Targets. 2008;12(4):463–476. CrossRef

52. 52Miyachi K, Sasaki K, Onodera S, et al. Correlation between survivin mRNA expression and lymph node metastasis in gastric cancer. Gastric cancer. 2003;6(4):217–224. MEDLINE | CrossRef

53. 53Schultz IJ, Kiemeney LA, Witjes JA, et al. Survivin mRNA expression is elevated in malignant urothelial cell carcinomas and predicts time to recurrence. Anticancer Res. 2003;23(4):3327–3331. MEDLINE

54. 54Cohen C, Lohmann CM, Cotsonis G, Lawson D, Santoianni R. Survivin expression in ovarian carcinoma: correlation with apoptotic markers and prognosis. Mod Pathol. 2003;16(6):574–583. MEDLINE | CrossRef

55. 55Javle MM, Tan D, Yu J, et al. Nuclear survivin expression predicts poor outcome in cholangiocarcinoma. Hepatogastroenterology. 2004;51(60):1653–1657. MEDLINE

56. 56Sharma PS, Sharma R, Tyagi R. Inhibitors of cyclin dependent kinases: useful targets for cancer treatment. Curr Cancer Drug Targets. 2008;8(1):53–75. CrossRef

57. 57Hui AM, Cui X, Makuuchi M, Li X, Shi YZ, Takayama T. Decreased p27(Kip1) expression and cyclin D1 overexpression, alone and in combination, influence recurrence and survival of patients with resectable extrahepatic bile duct carcinoma. Hepatology. 1999;30(5):1167–1173. MEDLINE | CrossRef

58. 58Sugimachi K, Aishima S, Taguchi K, et al. The role of overexpression and gene amplification of cyclin D1 in intrahepatic cholangiocarcinoma. J Hepatol. 2001;35(1):74–79. Abstract | Full Text | Full-Text PDF (687 KB) | CrossRef

59. 59Fiorentino M, Altimari A, D’Errico A, et al. Low p27 expression is an independent predictor of survival for patients with either hilar or peripheral intrahepatic cholangiocarcinoma. Clin Cancer Res. 2001;7(12):3994–3999. MEDLINE

60. 60Taguchi K, Aishima S, Asayama Y, et al. The role of p27kip1 protein expression on the biological behavior of intrahepatic cholangiocarcinoma. Hepatology. 2001;33(5):1118–1123. MEDLINE | CrossRef

61. 61Shibahara H, Tamada S, Higashi M, et al. MUC4 is a novel prognostic factor of intrahepatic cholangiocarcinoma-mass forming type. Hepatology. 2004;39(1):220–229. MEDLINE | CrossRef

62. 62Tamada S, Shibahara H, Higashi M, et al. MUC4 is a novel prognostic factor of extrahepatic bile duct carcinoma. Clin Cancer Res. 2006;12(14 Pt 1):4257–4264. MEDLINE | CrossRef

63. 63Canepa ET, Scassa ME, Ceruti JM, et al. INK4 proteins, a family of mammalian CDK inhibitors with novel biological functions. IUBMB Life. 2007;59(7):419–426. CrossRef

64. 64Hong SM, Choi J, Ryu K, Ro JY, Yu E. Promoter hypermethylation of the p16 gene and loss of its protein expression is correlated with tumor progression in extrahepatic bile duct carcinomas. Arch Pathol Lab Med. 2006;130(1):33–38.

65. 65Ichikawa K, Imura J, Kawamata H, Takeda J, Fujimori T. Down-regulated p16 expression predicts poor prognosis in patients with extrahepatic biliary tract carcinomas. Int J Oncol. 2002;20(3):453–461. MEDLINE

66. 66Nishida T, Nakao K, Hamaji M, Nakahara MA, Tsujimoto M. Prognostic significance of proliferative cell nuclear antigen in carcinoma of the extrahepatic bile duct. World J Surg. 1997;21(6):634–639. MEDLINE | CrossRef

67. 67Yamada Y, Yamauchi H, Kakizaki K, Suzuki H. Expression of Ki-67 by monoclonal antibody MIB-1 in carcinoma of the middle and lower bile duct. Tohoku J Exp Med. 1995;177(4):375–377. MEDLINE | CrossRef

68. 68Rijken AM, Umezawa A, van Gulik TM, et al. Prognostic value of cell proliferation (Ki-67 antigen) and nuclear DNA content in clinically resectable, distal bile duct carcinoma. Ann Surg Oncol. 1998;5(8):699–705. MEDLINE | CrossRef

69. 69Aishima S, Kuroda Y, Nishihara Y, et al. Gastric mucin phenotype defines tumour progression and prognosis of intrahepatic cholangiocarcinoma: gastric foveolar type is associated with aggressive tumour behaviour. Histopathology. 2006;49(1):35–44. MEDLINE | CrossRef

70. 70Uberall I, Kolar Z, Trojanec R, Berkovcova J, Hajduch M. The status and role of ErbB receptors in human cancer. Exp Mol Pathol. 2008;84(2):79–89. CrossRef

71. 71Javle MM, Yu J, Khoury T, et al. Akt expression may predict favorable prognosis in cholangiocarcinoma. J Gastroenterol Hepatol. 2006;21(11):1744–1751. MEDLINE | CrossRef

72. 72Gwak GY, Yoon JH, Shin CM, et al. Detection of response-predicting mutations in the kinase domain of the epidermal growth factor receptor gene in cholangiocarcinomas. J Cancer Res Clin Oncol. 2005;131(10):649–652. MEDLINE | CrossRef

73. 73Vogel CL, Cobleigh MA, Tripathy D, et al. Efficacy and safety of trastuzumab as a single agent in first-line treatment of HER2-overexpressing metastatic breast cancer. J Clin Oncol. 2002;20(3):719–726. CrossRef

74. 74Kim HJ, Yoo TW, Park DI, et al. Gene amplification and protein overexpression of HER-2/neu in human extrahepatic cholangiocarcinoma as detected by chromogenic in situ hybridization and immunohistochemistry: its prognostic implication in node-positive patients. Ann Oncol. 2007;18(5):892–897. MEDLINE | CrossRef

75. 75Ogo Y, Nio Y, Yano S, et al. Immunohistochemical expression of HER-1 and HER-2 in extrahepatic biliary carcinoma. Anticancer Res. 2006;26(1B):763–770. MEDLINE

76. 76Orend G. Potential oncogenic action of tenascin-C in tumorigenesis. Int J Biochem Cell Biol. 2005;37(5):1066–1083. MEDLINE | CrossRef

77. 77Aishima S, Taguchi K, Terashi T, Matsuura S, Shimada M, Tsuneyoshi M. Tenascin expression at the invasive front is associated with poor prognosis in intrahepatic cholangiocarcinoma. Mod Pathol. 2003;16(10):1019–1027. MEDLINE | CrossRef

78. 78Yeger H, Perbal B. The CCN family of genes: a perspective on CCN biology and therapeutic potential. J Cell Commun Signal. 2007;1(3–4):159–164.

79. 79Gardini A, Corti B, Fiorentino M, et al. Expression of connective tissue growth factor is a prognostic marker for patients with intrahepatic cholangiocarcinoma. Dig Liver Dis. 2005;37(4):269–274. Abstract | Full Text | Full-Text PDF (331 KB) | CrossRef

80. 80Bleau AM, Planque N, Perbal B. CCN proteins and cancer: two to tango. Front Biosci. 2005;10:998–1009. CrossRef

81. 81Davies SR, Watkins G, Mansel RE, Jiang WG. Differential expression and prognostic implications of the CCN family members WISP-1, WISP-2, and WISP-3 in human breast cancer. Ann Surg Oncol. 2007;14(6):1909–1918. MEDLINE | CrossRef

82. 82Tanaka S, Sugimachi K, Kameyama T, et al. Human WISP1v, a member of the CCN family, is associated with invasive cholangiocarcinoma. Hepatology. 2003;37(5):1122–1129. MEDLINE | CrossRef

83. 83Hirohashi S, Kanai Y. Cell adhesion system and human cancer morphogenesis. Cancer Sci. 2003;94(7):575–581. MEDLINE | CrossRef

84. 84Hugo H, Ackland ML, Blick T, et al. Epithelial–mesenchymal and mesenchymal–epithelial transitions in carcinoma progression. J Cell Physiol. 2007;213(2):374–383. CrossRef

85. 85Mikami T, Saegusa M, Mitomi H, Yanagisawa N, Ichinoe M, Okayasu I. Significant correlations of E-cadherin, catenin, and CD44 variant form expression with carcinoma cell differentiation and prognosis of extrahepatic bile duct carcinomas. Am J Clin Pathol. 2001;116(3):369–376. MEDLINE | CrossRef

86. 86Inki P, Jalkanen M. The role of syndecan-1 in malignancies. Ann Med. 1996;28(1):63–67. MEDLINE | CrossRef

87. 87Harada K, Masuda S, Hirano M, Nakanuma Y. Reduced expression of syndecan-1 correlates with histologic dedifferentiation, lymph node metastasis, and poor prognosis in intrahepatic cholangiocarcinoma. Hum Pathol. 2003;34(9):857–863. Abstract | Full Text | Full-Text PDF (301 KB) | CrossRef

88. 88Kuespert K, Pils S, Hauck CR. CEACAMs: their role in physiology and pathophysiology. Curr Opin Cell Biol. 2006;18(5):565–571. MEDLINE | CrossRef

89. 89Ieta K, Tanaka F, Utsunomiya T, Kuwano H, Mori M. CEACAM6 gene expression in intrahepatic cholangiocarcinoma. Br J Cancer. 2006;95(4):532–540. MEDLINE | CrossRef

90. 90Friederichs J, Zeller Y, Hafezi-Moghadam A, Grone HJ, Ley K, Altevogt P. The CD24/P-selectin binding pathway initiates lung arrest of human A125 adenocarcinoma cells. Cancer Res. 2000;60(23):6714–6722. MEDLINE

91. 91Aigner S, Ramos CL, Hafezi-Moghadam A, et al. CD24 mediates rolling of breast carcinoma cells on P-selectin. Faseb J. 1998;12(12):1241–1251. MEDLINE

92. 92Su MC, Hsu C, Kao HL, Jeng YM. CD24 expression is a prognostic factor in intrahepatic cholangiocarcinoma. Cancer Lett. 2006;235(1):34–39. Abstract | Full Text | Full-Text PDF (356 KB) | CrossRef

93. 93Agrawal S, Kuvshmoff BW, Khoury T, et al. CD24 expression is an independent prognostic marker in cholangiocarcinoma. J Gastrointest Surg. 2007;11(4):445–451. CrossRef

94. 94Yasuda M, Nakano K, Yasumoto K, Tanaka Y. CD44: functional relevance to inflammation and malignancy. Histol Histopathol. 2002;17(3):945–950. MEDLINE

95. 95Yokoyama Y, Hiyama E, Murakami Y, Matsuura Y, Yokoyama T. Lack of CD44 variant 6 expression in advanced extrahepatic bile duct/ampullary carcinoma. Cancer. 1999;86(9):1691–1699.

96. 96Chambers AF, Matrisian LM. Changing views of the role of matrix metalloproteinases in metastasis. J Natl Cancer Inst. 1997;89(17):1260–1270. MEDLINE

97. 97Miwa S, Miyagawa S, Soeda J, Kawasaki S. Matrix metalloproteinase-7 expression and biologic aggressiveness of cholangiocellular carcinoma. Cancer. 2002;94(2):428–434.

98. 98Park BK, Paik YH, Park JY, et al. The clinicopathologic significance of the expression of vascular endothelial growth factor-C in intrahepatic cholangiocarcinoma. Am J Clin Oncol – Cancer Clin Trials. 2006;29(2):138–142.

99. 99Tang D, Nagano H, Yamamoto H, et al. Angiogenesis in cholangiocellular carcinoma: expression of vascular endothelial growth factor, angiopoietin-1/2, thrombospondin-1 and clinicopathological significance. Oncol Rep. 2006;15(3):525–532.

100. 100Sasaki M, Ikeda H, Nakanuma Y. Expression profiles of MUC mucins and trefoil factor family (TFF) peptides in the intrahepatic biliary system: physiological distribution and pathological significance. Prog Histochem Cytochem. 2007;42(2):61–110. CrossRef

101. 101Higashi M, Yonezawa S, Ho JJ, et al. Expression of MUC1 and MUC2 mucin antigens in intrahepatic bile duct tumors: its relationship with a new morphological classification of cholangiocarcinoma. Hepatology. 1999;30(6):1347–1355. MEDLINE | CrossRef

102. 102Matsumura N, Yamamoto M, Aruga A, Takasaki K, Nakano M. Correlation between expression of MUC1 core protein and outcome after surgery in mass-forming intrahepatic cholangiocarcinoma. Cancer. 2002;94(6):1770–1776.

103. 103Suh KS, Chang SH, Lee HJ, Roh HR, Kim SH, Lee KU. Clinical outcomes and apomucin expression of intrahepatic cholangiocarcinoma according to gross morphology. J Am Coll Surg. 2002;195(6):782–789. Abstract | Full Text | Full-Text PDF (179 KB) | CrossRef

104. 104Takao S, Uchikura K, Yonezawa S, Shinchi H, Aikou T. Mucin core protein expression in extrahepatic bile duct carcinoma is associated with metastases to the liver and poor prognosis. Cancer. 1999;86(10):1966–1975.

105. 105Tamada S, Goto M, Nomoto M, et al. Expression of MUC1 and MUC2 mucins in extrahepatic bile duct carcinomas: its relationship with tumor progression and prognosis. Pathol Int. 2002;52(11):713–723. MEDLINE | CrossRef

106. 106Yuan SF, Li KZ, Wang L, et al. Expression of MUC1 and its significance in hepatocellular and cholangiocarcinoma tissue. World J Gastroenterol. 2005;11(30):4661–4666. MEDLINE

107. 107Boonla C, Sripa B, Thuwajit P, et al. MUC1 and MUC5AC mucin expression in liver fluke-associated intrahepatic cholangiocarcinoma. World J Gastroenterol. 2005;11(32):4939–4946. MEDLINE

108. 108Hong SM, Cho H, Moskaluk CA, Frierson HF, Yu E, Ro JY. CDX2 and MUC2 protein expression in extrahepatic bile duct carcinoma. Am J Clin Pathol. 2005;124(3):361–370. MEDLINE | CrossRef

109. 109Yonezawa S, Sueyoshi K, Nomoto M, et al. MUC2 gene expression is found in noninvasive tumors but not in invasive tumors of the pancreas and liver: its close relationship with prognosis of the patients. Hum Pathol. 1997;28(3):344–352. Abstract | Full-Text PDF (15321 KB) | CrossRef

110. 110Boonla C, Wongkham S, Sheehan JK, et al. Prognostic value of serum MUC5AC mucin in patients with cholangiocarcinoma. Cancer. 2003;98(7):1438–1443.

111. 111Takada A, Ohmori K, Yoneda T, et al. Contribution of carbohydrate antigens sialyl Lewis A and sialyl Lewis X to adhesion of human cancer cells to vascular endothelium. Cancer Res. 1993;53(2):354–361. MEDLINE

112. 112Kannagi R. Carbohydrate antigen sialyl Lewis a – its pathophysiological significance and induction mechanism in cancer progression. Chang Gung Med J. 2007;30(3):189–209.

113. 113Ohtsuka M, Ito H, Kimura F, et al. Results of surgical treatment for intrahepatic cholangiocarcinoma and clinicopathological factors influencing survival. Br J Surg. 2002;89(12):1525–1531. MEDLINE | CrossRef

114. 114Heimbach JK, Gores GJ, Haddock MG, et al. Predictors of disease recurrence following neoadjuvant chemoradiotherapy and liver transplantation for unresectable perihilar cholangiocarcinoma. Transplantation. 2006;82(12):1703–1707. MEDLINE | CrossRef

115. 115Huang JL, Biehl TR, Lee FT, Zimmer PW, Ryan JA. Outcomes after resection of cholangiocellular carcinoma. Am J Surg. 2004;187(5):612–617. Abstract | Full Text | Full-Text PDF (78 KB) | CrossRef

116. 116Abdel Wahab M, Fathy O, Elghwalby N, et al. Resectability and prognostic factors after resection of hilar cholangiocarcinoma. Hepatogastroenterology. 2006;53(67):5–10. MEDLINE

117. 117Miwa S, Miyagawa S, Kobayashi A, et al. Predictive factors for intrahepatic cholangiocarcinoma recurrence in the liver following surgery. J Gastroenterol. 2006;41(9):893–900. MEDLINE | CrossRef

118. 118Hirohashi K, Uenishi T, Kubo S, et al. Macroscopic types of intrahepatic cholangiocarcinoma: clinicopathologic features and surgical outcomes. Hepatogastroenterology. 2002;49(44):326–329. MEDLINE

119. 119Yi B, Zhang BH, Zhang YJ, et al. Surgical procedure and prognosis of hilar cholangiocarcinoma. Hepatobiliary Pancreat Dis Int. 2004;3(3):453–457. MEDLINE

120. 120Mahmoud FA, Rivera NI. The role of C-reactive protein as a prognostic indicator in advanced cancer. Curr Oncol Rep. 2002;4(3):250–255. MEDLINE | CrossRef

121. 121Gerhardt T, Milz S, Schepke M, et al. C-reactive protein is a prognostic indicator in patients with perihilar cholangiocarcinoma. World J Gastroenterol. 2006;12(34):5495–5500. MEDLINE

122. 122Wongkham S, Bhudhisawasdi V, Chau-in S, et al. Clinical significance of serum total sialic acid in cholangiocarcinoma. Clin Chim Acta. 2003;327(1–2):139–147. MEDLINE | CrossRef

123. 123Juntavee A, Sripa B, Pugkhem A, Khuntikeo N, Wongkham S. Expression of sialyl Lewis(a) relates to poor prognosis in cholangiocarcinoma. World J Gastroenterol. 2005;11(2):249–254. MEDLINE

124. 124Ellis CN, Burnette JJ, Sedlak R, Dyas C, Blakemore WS. Prognostic applications of DNA analysis in solid malignant lesions in humans. Surg Gynecol Obstet. 1991;173(4):329–342. MEDLINE

125. 125Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature. 1998;396(6712):643–649. MEDLINE | CrossRef

126. 126Lindberg B, Enochsson L, Tribukait B, Arnelo U, Bergquist A. Diagnostic and prognostic implications of DNA ploidy and S-phase evaluation in the assessment of malignancy in biliary strictures. Endoscopy. 2006;38(6):561–565. CrossRef

127. 127Iachino C, Catsicojannis N, Dallera F. Prognostic significance of tumor DNA content in carcinoma of the hepatic duct confluence. J Exp Clin Cancer Res. 1998;17(4):425–430. MEDLINE

128. 128Kawaki J, Miyazaki M, Ito H, et al. Allelic loss in human intrahepatic cholangiocarcinoma: correlation between chromosome 8p22 and tumor progression. Int J Cancer. 2000;88(2):228–231. MEDLINE | CrossRef

129. 129Muenphon K, Limpaiboon T, Jearanaikoon P, Pairojkul C, Sripa B, Bhudhisawasdi V. Amplification of chromosome 21q22.3 harboring trefoil factor family genes in liver fluke related cholangiocarcinoma is associated with poor prognosis. World J Gastroenterol. 2006;12(26):4143–4148. MEDLINE

130. 130Limpaiboon T, Tapdara S, Jearanaikoon P, Sripa B, Bhudhisawasdi V. Prognostic significance of microsatellite alterations at 1p36 in cholangiocarcinoma. World J Gastroenterol. 2006;12(27):4377–4382. MEDLINE

131. 131De Angelis PM, Stokke T, Beigi M, Mjaland O, Clausen OP. Prognostic significance of recurrent chromosomal aberrations detected by comparative genomic hybridization in sporadic colorectal cancer. Int J Colorectal Dis. 2001;16(1):38–45. MEDLINE | CrossRef

132. 132Casorelli I, Russo MT, Bignami M. Role of mismatch repair and MGMT in response to anticancer therapies. Anticancer Agents Med Chem. 2008;8(4):368–380.

133. 133Kohya N, Miyazaki K, Matsukura S, et al. Deficient expression of O(6)-methylguanine-DNA methyltransferase combined with mismatch-repair proteins hMLH1 and hMSH2 is related to poor prognosis in human biliary tract carcinoma. Ann Surg Oncol. 2002;9(4):371–379. MEDLINE | CrossRef

134. 134Koga Y, Kitajima Y, Miyoshi A, et al. Tumor progression through epigenetic gene silencing of O(6)-methylguanine-DNA methyltransferase in human biliary tract cancers. Ann Surg Oncol. 2005;12(5):354–363. MEDLINE | CrossRef

135. 135Schmitz KJ, Lang H, Frey UH, et al. GNAS1 T393C polymorphism is associated with clinical course in patients with intrahepatic cholangiocarcinoma. Neoplasia. 2007;9(2):159–165. CrossRef

136. 136Linder S. Cytokeratin markers come of age. Tumour Biol. 2007;28(4):189–195. CrossRef

137. 137Aishima S, Asayama Y, Taguchi K, et al. The utility of keratin 903 as a new prognostic marker in mass-forming-type intrahepatic cholangiocarcinoma. Mod Pathol. 2002;15(11):1181–1190. MEDLINE | CrossRef

138. 138Erickson LA, Lloyd RV. Practical markers used in the diagnosis of endocrine tumors. Adv Anat Pathol. 2004;11(4):175–189. MEDLINE | CrossRef

139. 139Hong SM, Kim MJ, Pi DY, Jo D, Yu E, Ro JY. Neuroendocrine differentiation in extrahepatic bile duct carcinomas and its prognostic significance. Hum Pathol. 2005;36(7):732–740. Abstract | Full Text | Full-Text PDF (376 KB) | CrossRef

140. 140Takagi S, Miyagawa S, Ichikawa E, et al. Dendritic cells, T-cell infiltration, and Grp94 expression in cholangiocellular carcinoma. Hum Pathol. 2004;35(7):881–886. Abstract | Full Text | Full-Text PDF (672 KB) | CrossRef

141. 141Kelloniemi E, Rintala E, Finne P, Stenman UH. Tumor-associated trypsin inhibitor as a prognostic factor during follow-up of bladder cancer. Urology. 2003;62(2):249–253. Abstract | Full Text | Full-Text PDF (108 KB) | CrossRef

142. 142Paju A, Vartiainen J, Haglund C, Itkonen O, von Boguslawski K, Leminen A, et al. Expression of trypsinogen-1, trypsinogen-2, and tumor-associated trypsin inhibitor in ovarian cancer: prognostic study on tissue and serum. Clin Cancer Res. 2004;10(14):4761–4768. MEDLINE | CrossRef

143. 143Paju A, Jacobsen J, Rasmuson T, Stenman UH, Ljungberg B. Tumor associated trypsin inhibitor as a prognostic factor in renal cell carcinoma. J Urol. 2001;165(3):959–962. Abstract | Full Text | Full-Text PDF (65 KB) | CrossRef

144. 144Tonouchi A, Ohtsuka M, Ito H, et al. Relationship between pancreatic secretory trypsin inhibitor and early recurrence of intrahepatic cholangiocarcinoma following surgical resection. Am J Gastroenterol. 2006;101(7):1601–1610. MEDLINE | CrossRef

145. 145Wang MT, Honn KV, Nie D. Cyclooxygenases, prostanoids, and tumor progression. Cancer Metastasis Rev. 2007;26(3–4):525–534. CrossRef

146. 146Schmitz KJ, Lang H, Wohlschlaeger J, Reis H, Sotiropoulos GC, Schmid KW, et al. Elevated expression of cyclooxygenase-2 is a negative prognostic factor for overall survival in intrahepatic cholangiocarcinoma. Virchows Archiv. 2007;450(2):135–141.

147. 147Kim HJ, Lee KT, Kim EK, et al. Expression of cyclooxygenase-2 in cholangiocarcinoma: correlation with clinicopathological features and prognosis. Journal of Gastroenterology and Hepatology. 2004;19(5):582–588. MEDLINE | CrossRef

148. 148Garcea G, Neal CP, Pattenden CJ, Steward WP, Berry DP. Molecular prognostic markers in pancreatic cancer: a systematic review. Eur J Cancer. 2005;41(15):2213–2236. Abstract | Full Text | Full-Text PDF (265 KB) | CrossRef

149. 149Mann CD, Neal CP, Garcea G, Manson MM, Dennison AR, Berry DP. Prognostic molecular markers in hepatocellular carcinoma: a systematic review. Eur J Cancer. 2007;43(6):979–992. Abstract | Full Text | Full-Text PDF (252 KB) | CrossRef

150. 150Neal CP, Garcea G, Doucas H, Manson MM, Sutton CD, Dennison AR, et al. Molecular prognostic markers in resectable colorectal liver metastases: a systematic review. Eur J Cancer. 2006;42(12):1728–1743. Abstract | Full Text | Full-Text PDF (258 KB) | CrossRef

151. 151Bonney GK, Craven RA, Prasad R, Melcher AF, Selby PJ, Banks RE. Circulating markers of biliary malignancy: opportunities in proteomics?. Lancet Oncol. 2008;9(2):149–158. Abstract | Full Text | Full-Text PDF (547 KB) | CrossRef

152. 152Obama K, Ura K, Li M, et al. Genome-wide analysis of gene expression in human intrahepatic cholangiocarcinoma. Hepatology. 2005;41(6):1339–1348. MEDLINE | CrossRef

a Cancer Biomarkers and Prevention Group, Department of Cancer Studies and Molecular Medicine, Biocentre, University of Leicester, University Road, Leicester LE1 7RH, United Kingdom

b Department of Hepatobiliary and Pancreatic Surgery, Leicester General Hospital, University Hospitals of Leicester, Gwendolen Road, Leicester LE5 4PW, United Kingdom

Corresponding author: Tel.: +44 07980748257; fax: +44 0116 2231840.

PII: S0959-8049(08)00682-5

doi:10.1016/j.ejca.2008.08.024

7 of 26