If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), San Sebastian, SpainNational Institute for the Study of Liver and Gastrointestinal Diseases, CIBERehd, “Instituto de Salud Carlos III” (ISCIII), Madrid, Spain
Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), San Sebastian, Spain
Department of Molecular Medicine (DMM), University of Padua School of Medicine, Padua, ItalyDigestive Disease Section, Yale University School of Medicine, New Haven, CT, USA
Department of Medical, Surgical and Experimental Sciences, Clinica Medica Unit, University of Sassari, Azienda Ospedaliero-Universitaria di Sassari, Sassari, Italy
National Institute for the Study of Liver and Gastrointestinal Diseases, CIBERehd, “Instituto de Salud Carlos III” (ISCIII), Madrid, SpainExperimental Hepatology and Drug Targeting (HEVEPHARM) Group, University of Salamanca, Biomedical Research Institute (IBSAL), Salamanca, Spain
Norwegian PSC Research Center, Department of Transplantation Medicine, Oslo University Hospital Rikshospitalet, Oslo, NorwayInstitute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, NorwayResearch Institute of Internal Medicine, Division of Surgery, Inflammatory Diseases and Transplantation, Oslo University Hospital Rikshospitalet, Oslo, NorwaySection of Gastroenterology, Department of Transplantation Medicine, Oslo University Hospital, Oslo, Norway
Department of Medicine II Saarland University Medical Center, Saarland University, Homburg, GermanyLaboratory of Metabolic Liver Diseases, Center for Preclinical Research, Department of General, Transplant and Liver Surgery, Medical University of Warsaw, Warsaw, Poland
Department of Gastroenterology and Hepatology with Internal Disease Unit, Teaching Hospital No 1 in Rzeszów, PolandMedical Department, University of Rzeszów, Poland
Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), San Sebastian, Spain
Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), San Sebastian, Spain
National Institute for the Study of Liver and Gastrointestinal Diseases, CIBERehd, “Instituto de Salud Carlos III” (ISCIII), Madrid, SpainExperimental Hepatology and Drug Targeting (HEVEPHARM) Group, University of Salamanca, Biomedical Research Institute (IBSAL), Salamanca, Spain
Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), San Sebastian, SpainNational Institute for the Study of Liver and Gastrointestinal Diseases, CIBERehd, “Instituto de Salud Carlos III” (ISCIII), Madrid, Spain
National Institute for the Study of Liver and Gastrointestinal Diseases, CIBERehd, “Instituto de Salud Carlos III” (ISCIII), Madrid, SpainLiver Unit. Barcelona Clinic Liver Cancer (BCLC) group. Hospital Clinic Barcelona. IDIBAPS. University of Barcelona, Spain
Corresponding author. Address: Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute – Donostia University Hospital, Paseo Dr. Begiristain s/n, E-20014, San Sebastian, Spain. Tel.: +34 943006067.
Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), San Sebastian, SpainNational Institute for the Study of Liver and Gastrointestinal Diseases, CIBERehd, “Instituto de Salud Carlos III” (ISCIII), Madrid, SpainDepartment of Biochemistry and Genetics, School of Sciences, University of Navarra, Pamplona, SpainIkerbasque, Basque Foundation for Science, Bilbao, Spain
CCA subtypes present different risk factors and tumor features.
•
CA19-9 shows low sensitivity in early stages but increased sensitivity in advanced disease.
•
Under surgery, positive margins and lymph node invasion compromise survival.
•
ECOG-PS, disease status and CA19-9 are independent prognostic factors.
Background & Aims
Cholangiocarcinoma (CCA) is a rare and heterogeneous biliary cancer, whose incidence and related mortality is increasing. This study investigates the clinical course of CCA and subtypes (intrahepatic [iCCA], perihilar [pCCA], and distal [dCCA]) in a pan-European cohort.
Methods
The ENSCCA Registry is a multicenter observational study. Patients were included if they had a histologically proven diagnosis of CCA between 2010-2019. Demographic, histomorphological, biochemical, and clinical studies were performed.
Results
Overall, 2,234 patients were enrolled (male/female=1.29). iCCA (n = 1,243) was associated with overweight/obesity and chronic liver diseases involving cirrhosis and/or viral hepatitis; pCCA (n = 592) with primary sclerosing cholangitis; and dCCA (n = 399) with choledocholithiasis. At diagnosis, 42.2% of patients had local disease, 29.4% locally advanced disease (LAD), and 28.4% metastatic disease (MD). Serum CEA and CA19-9 showed low diagnostic sensitivity, but their concomitant elevation was associated with increased risk of presenting with LAD (odds ratio 2.16; 95% CI 1.43-3.27) or MD (odds ratio 5.88; 95% CI 3.69-9.25). Patients undergoing resection (50.3%) had the best outcomes, particularly with negative-resection margin (R0) (median overall survival [mOS] = 45.1 months); however, margin involvement (R1) (hazard ratio 1.92; 95% CI 1.53-2.41; mOS = 24.7 months) and lymph node invasion (hazard ratio 2.13; 95% CI 1.55-2.94; mOS = 23.3 months) compromised prognosis. Among patients with unresectable disease (49.6%), the mOS was 10.6 months for those receiving active palliative therapies, mostly chemotherapy (26.2%), and 4.0 months for those receiving best supportive care (20.6%). iCCAs were associated with worse outcomes than p/dCCAs. ECOG performance status, MD and CA19-9 were independent prognostic factors.
Conclusion
CCA is frequently diagnosed at an advanced stage, a proportion of patients fail to receive cancer-specific therapies, and prognosis remains dismal. Identification of preventable risk factors and implementation of surveillance in high-risk populations are required to decrease cancer-related mortality.
Lay summary
This is, to date, the largest international (pan-European: 26 hospitals and 11 countries) observational study, in which the course of cholangiocarcinoma has been investigated, comparing the 3 subtypes based on the latest International Classification of Diseases 11th Edition (ICD-11) (i.e., intrahepatic [2C12], perihilar [2C18], or distal [2C15] affected bile ducts), which come into effect in 2022. General and tumor-type specific features at diagnosis, risk factors, biomarker accuracy, as well as patient management and outcomes, are presented and compared, outlining the current clinical state of cholangiocarcinoma in Europe.
Cholangiocarcinoma (CCA) comprises a heterogeneous group of malignancies that can arise anywhere in the bile ducts. Despite being a rare disease, over the past 20 years, incidence and mortality rates have increased globally (0.3-6 cases per 100,000 inhabitants yearly in Western countries, and >6 cases in some East Asian regions).
Although different risk factors have been identified, many patients have no apparent cause at diagnosis, limiting the ability of early detection by surveillance programs.
Accordingly, a coherent international study to enhance granularity of the global clinical situation of CCA is pivotal to better understand the disease course, define the similarities and/or differences between CCA subtypes, describe outcomes after selected treatments, and identify challenges for future prospective analyses. This study aims to address this unmet need by describing the disease presentation, risk factors, treatment, and outcomes of patients with CCA in a large pan-European cohort.
Patients and methods
Study design and data collection
The ENSCCA Registry, endorsed by the European Network for the Study of Cholangiocarcinoma (ENSCCA), is a multicenter observational study of patient with histologically and/or cytologically confirmed CCA. Patients’ data was accrued based on contributions of 26 referral Healthcare Centers from 11 European countries (Austria, France, Germany, Italy, Lithuania, Netherlands, Norway, Poland, Romania, Spain, and United Kingdom) (Fig. S1A). Patients diagnosed with CCA over a 10-year period (January 1st 2010 to December 31st 2019) were included in the registry study. Fig. S1B summarizes the flow and number of patients based on the following inclusion criteria: i) diagnosis following the latest International Classification of Diseases 11th Edition (ICD-11),
in which CCA was categorized as intrahepatic (2C12), perihilar (2C18), or distal (2C15); and ii) histological and/or cytological confirmation of the diagnosis. Individual patient data was obtained from medical records by the participating hospitals. Information on patients’ demographics, documented risk factors and medical history, biochemical and clinical parameters, and treatments were included. In the ENSCCA Registry, clinical and pathological parameters were registered at the time of diagnosis using the 7th edition of the American Joint Committee of Cancer (AJCC)/Union for International Cancer Control TNM cancer staging manual.
Data were recorded using a de-identified format in an electronic case report form, collected and managed using the web-based application designed to support data capture for research studies “Research Electronic Data Capture” (REDCapTM) hosted at “Asociación Española de Gastroenterología” (AEG; www.aegastro.es), a non-profit Scientific and Medical Society focused on Gastroenterology research. Data export was performed in February 2020, and subjected to data harmonization and completeness check. Patients were excluded from the study when mandatory epidemiological and/or clinical data (i.e., type of CCA, date of diagnosis, and date of last follow-up or death) were missing. Moreover, patients without tissue-proven CCA (investigator-reported) or with undefined biliary location were also excluded after an internal investigator review process. Patients diagnosed from 2010 to 2017 (n = 1,962) were considered for survival analysis to ensure a minimum 2-year follow-up. The ENSCCA Registry Study protocol was approved by the Ethic Committee of Euskadi, Spain (Code: PI2016137), as coordinating Center. Additionally, each participating Center obtained a local ethical approval (or equivalent).
Data analysis
Patients were classified according to the anatomical location of the primary tumor within the bile ducts (i.e., iCCA, pCCA or dCCA) following the ICD-11
criteria and the experience of investigators within multidisciplinary teams. Positive lymph node invasion and/or tumor metastasis were identified by either histology or imaging techniques. As a result, patients were categorized by the disease status at diagnosis, as: i) local disease (LD), ii) locally advanced disease (LAD), or iii) metastatic disease (MD). LAD was stated as positive regional lymph node tumor invasion measuring above 1.5 cm in diameter (short axis) and classified as N+ (i.e., N1 for iCCA and dCCA; N1 and N2 for pCCA). According to the aforementioned staging guidelines, MD indicated distant involvement (M1), with the exception for liver dissemination of iCCA that is classified as multiple tumors (T2b), and thus, as LD. Based on local multidisciplinary team discussions, patients were divided into 2 groups, those with resectable vs. unresectable CCA following widely accepted international guidelines (e.g., from the European Society for Medical Oncology and/or the International Liver Cancer Association),
and taking into account multiparametric criteria based on performance status, tumor stage, underlying diseases, and comorbidities, among others. Accordingly, treatments were categorized as: 1) surgery (i.e., tumor resection or liver transplantation subdivided into i) resection margin R0 [negative margin tumor resection], ii) resection margin R1 [microscopic residual disease], and iii) resection margin R2 [gross residual disease], and (2) active palliative treatment (i.e., chemotherapy, hepatic artery-based therapies, radiation therapy, and/or ablation). Patients receiving staging laparoscopy or exploratory laparotomy were classified according to the subsequent therapeutic strategy.
Statistical analysis
Baseline demographics and risk factors were summarized using descriptive statistics. Continuous data were described as median (IQR), while categorical variables were summarized as n (%). Probability was calculated excluding cases with unknown information. Shapiro-Wilk test was used to test continuous variables for normal distribution. For multiple comparisons, parametric or non-parametric data were compared using one-way ANOVA or Kruskal-Wallis tests, respectively, and followed by Bonferroni post hoc test. Pairwise comparisons were calculated using Dunn’s method. Pearson’s χ2 test was used to compare categorical variables between the 3 subgroups. For pairwise comparison between CCA subtypes of categorical data, Fisher’s exact test was performed. Logistic regression analysis was carried out in variables previously dichotomized as “normal” vs. “high” based on the normality threshold to assess the risk of disease dissemination. Overall survival (OS) was assessed as the time from diagnosis to death or last medical visit, while post-treatment survival considered the treatment start date. Relapse-free survival was calculated as the time from tumor resection to the event of relapse or death. Patients with no information on survival, lost to follow-up or alive at last medical visit were censored at the date of the latest record. Survival analysis was performed with the Kaplan-Meier method and Cox regression (univariate and multivariable analysis including variables statistically significant in the univariate analysis, defined as p <0.05). The Log-rank test was used for comparisons of survival in Kaplan-Meier curves. Prognostic factors were related to hazard ratio (HR), 95% CIs, and p values.
Statistical analyses were performed with IBM SPSS Statistics Version 22.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism version 6.0 for Microsoft Windows, (GraphPad Software, La Jolla California, USA). All p values were obtained in 2-tailed tests and p <0.05 was considered statistically significant.
Results
Patients’ characteristics and CCA features at diagnosis
From the 3,039 patients initially included in the ENSCCA Registry, 2,234 (73.5%) were selected and further analyzed after fulfilling the inclusion criteria (Fig. S1B), including 1,243 (55.6%) with iCCA, 592 (26.5%) with pCCA and 399 (17.9%) with dCCA. Baseline patient characteristics, including laboratory tests and comorbidities, are listed in Table 1. The majority of patients were Caucasian (96.6%) with a median age at diagnosis of 66 years (IQR 58-73) and slight overrepresentation of males (56.4%). Most patients showed, at diagnosis, increased serum levels of alanine aminotransferase (ALT), markers of cholestasis (gamma-glutamyltransferase [GGT] and alkaline phosphatase [ALP]) and carbohydrate antigen 19-9 (CA19-9), particularly patients with pCCA or dCCA (Table 1). No significant abnormalities were observed in specific hematological and metabolic blood test measures (Table S2).
Table 1Baseline patient characteristics and concomitant conditions.
Statistic analyses (one-way ANOVA or Kruskal-Wallis tests for continous variables, and Pearson’s Chi-square test for categorical variables) were performed by comparing the 3 cholangiocarcinoma subtypes (iCCA vs. pCCA vs. dCCA).
Clinical thresholds for laboratory variables were the highest reported in the literature: ALT 45 UI/L, AST 40 UI/L, GGT 71 UI/L, ALP 129 UI/L, Albumin 5.2 g/dl, Bilirubin 1.3 mg/dl.
a Statistic analyses (one-way ANOVA or Kruskal-Wallis tests for continous variables, and Pearson’s Chi-square test for categorical variables) were performed by comparing the 3 cholangiocarcinoma subtypes (iCCA vs. pCCA vs. dCCA).
b Statistical test considered Caucasian vs. other ethnicities (Asian, African, Hispanic, Caribbean).
c Clinical thresholds for laboratory variables were the highest reported in the literature: ALT 45 UI/L, AST 40 UI/L, GGT 71 UI/L, ALP 129 UI/L, Albumin 5.2 g/dl, Bilirubin 1.3 mg/dl.
Considering patients’ comorbidities (Table 1), 55.1% were overweight/obese (BMI 25-30 kg/m2 [35.7%] or BMI ≥30 kg/m2 [19.4%]) at the time of diagnosis, this feature being more frequent in patients with iCCA (Table S1); 22.5% had diabetes, observed more frequently in patients with iCCA or dCCA compared to pCCA, and 39.9% had arterial hypertension. Of note, 15% of the patients with CCA were obese and diabetic. In addition, patients suffered from underlying biliary or liver diseases predisposing to CCA development, including primary biliary cholangitis (PBC: 3.3%, mainly iCCA), primary sclerosing cholangitis (PSC: 4.5%; mainly pCCA), bile duct stones (6.1%; mainly pCCA and dCCA), viral hepatitis (2.8% hepatitis C virus, 4.6% hepatitis B virus, and 0.1% concomitant infection; mainly iCCA) and cirrhosis (7.8%; mainly iCCA). In this registry cohort there was also a history of smoking or alcohol consumption in 33% and 19.8% of patients, respectively.
Table 2 summarizes patients’ fitness, measured as Eastern Cooperative Oncology Group performance status (ECOG-PS), and tumor-related features at diagnosis. The majority of patients with CCA had ECOG-PS of 0 (44.0%) or 1 (39.1%). Regarding tumor size and growth pattern, iCCAs were frequently larger lesions (>3 cm or multifocal) with a mass-forming pattern compared to pCCA and dCCA that in general were smaller lesions (<3 cm) with periductal infiltration (Table S3). Moderate grade of tumor differentiation was the most frequently observed in the 3 CCA subtypes. From 1,998 patients with available information on imaging, 6.2% had initial tumor staging based on MRI and/or magnetic resonance cholangiopancreatography (MRCP), 47.4% with CT, and 54.9% with both approaches (Table S4). Of note, 32.3% and 4.0% of all patients with MRI/MRCP/CT-based staging had an additional ultrasound or PET evaluation, respectively. Imaging findings elucidated that regional lymph node invasion and disseminated disease were present in 48.7% and 24.2% of patients, respectively. CCAs preferentially metastasized to lung, liver, distant lymph nodes, bone and peritoneum – including omentum – with significant differences between subtypes. iCCA was mainly found to disseminate into lung, distant lymph nodes, and bone, whereas pCCA and dCCA mainly metastasized into the liver or to the peritoneum (Fig. 1A). In most patients with MD at presentation, a single site of metastasis was found (80.8%). Fig. 1B shows the frequency of each metastatic site based on the CCA subtype.
(A) Most commonly found metastatic locations stratified by CCA subtype, expressed as number and percentage [n (%)], and (B) classification of patients with disseminated CCA depending on their sites of metastasis, as single, two, three or more sites of metastasis. †Significant Pearson's chi-squared test for Bone (χ2p <0.05): iCCA vs. pCCA, p <0.05; DLN (χ2p <0.001): iCCA vs. pCCA, p <0.0001; Lung (χ2p <0.001): iCCA vs. pCCA, p <0.001, and iCCA vs. dCCA, p <0.01. CCA, cholangiocarcinoma; dCCA, distal CCA; DLN, distant lymph nodes; iCCA, intrahepatic CCA; pCCA, perihilar CCA. (This figure appears in color on the web.)
Sensitivity of serum carcinoembryonic antigen (CEA) and CA19-9 tumor biomarkers
The sensitivity of CEA (cut-off value: 5 IU/ml) and CA19-9 (≥37 IU/ml) was evaluated according to the disease stage. Serum CEA was above the upper reference limit in 30.9% of patients, correlating with disease severity (for LAD: odds ratio [OR] 1.71; 95% CI 1.16-2.51; for MD: OR 3.03; 95% CI 2.11-4.35) (Fig. 2B and Table S5-6). Increased CA19-9 was found in 59.1% of cases, particularly with LAD or MD (Fig. 2B and Table S5). Of note, CA19-9 above the cut-off value was associated with an increased risk of tumor spread (for LAD: OR 1.99; 95% CI 1.47-2.70; for MD: OR 3.04; 95% CI 2.21-4.17) (Fig. 2B and Table S6). The elevation of a single serum marker (i.e., CEA or CA19-9) was slightly associated with LAD (OR 1.72; 95% CI 1.16-2.53) or MD (OR 2.53; 95% CI 1.56-4.10) at diagnosis, whereas the concomitant elevation of both considerably increased the odds for LAD (OR 2.16; 95% CI 1.43-3.27) and for MD (OR 5.86; 95% CI 3.69-9.25) (Fig. 2C).
(A) Serum CEA levels depicted for disease stage (i.e., LD, LAD, or MD) for overall CCAs and subtypes. Odds value of CEA as potential tumor biomarker in the identification of tumor spread compared to LD. CEA cut-off value established at 5 IU/ml. (B) Serum CA19-9 levels depicted for tumor spread stage (i.e., LD, LAD, or MD) for overall CCAs and subtypes. Odds value of CA19-9 as potential tumor biomarker in the identification of tumor spread compared to LD. CA19-9 cut-off value established at 37 IU/ml. (C) Patients were classified into 3 comparison groups: (1) with both circulating tumor biomarkers (CA19-9 and CEA) below the established threshold, (2) one of both biomarkers over the cut-off value, or (3) both, CA19-9 and CEA, over the cut-off value. Odds value of the combination of CA19-9 and CEA in the prediction of CCA staging compared to LD. One-way ANOVA Kruskal-Wallis test was used for multiple comparisons. CA19-9, carbohydrate antigen 19-9; CCA, cholangiocarcinoma; CEA, carcinoembryonic antigen; dCCA, distal CCA; iCCA, intrahepatic CCA; LAD, locally advanced disease; LD, local disease; MD, metastatic disease; OR, odds ratio; pCCA, perihilar CCA. (This figure appears in color on the web.)
Patients with CCA often present with tumor-mediated biliary obstruction, requiring biliary drainage prior to starting any therapeutic regimen. In particular, 40.3% of the patients received biliary drainage, from whom 42.4% required subsequent stenting, with a median time interval of 1.8 months (Table S7). Fig. 3A represents a flow chart summarizing the first therapeutic strategy for patients with CCA following initial diagnosis. Of note, biliary drainage was performed prior to surgery or systemic therapy in 32.2% and 35.0% of patients, respectively (Fig. 3A). Moreover, 61.8% of all patients not receiving anti-cancer therapy had biliary drainage as part of best supportive care (BSC).
Fig. 3Cholangiocarcinoma clinical management and outcome.
(A) Diagram of classification of patients with CCA divided by type of treatment strategy (i.e. surgery, active palliative treatment or BSC) together with the corresponding median overall survival and Cox regression analysis between groups. Kaplan-Meier analysis and multivariable Cox regression models for the assessment of long-term outcome of patients with CCA after tumor resection, with (B), tumor relapse as primary endpoints of patients after tumor resection, and (C), lymph node invasion-associated mortality for patients under tumor resection. BSC, best supportive care; CCA, cholangiocarcinoma; Cis, cisplatin; dCCA, distal CCA; Gem, gemcitabine; HR, hazard ratio; iCCA, intrahepatic CCA; mOS, median overall survival; mRFS, median relapse-free survival; N, lymph node invasion; Ox, oxaliplatin; pCCA, perihilar CCA; R0, null margin tumor resection; R1, microscopic residual disease tumor resection; R2, gross residual disease tumor resection. (This figure appears in color on the web.)
Surgical resection was performed in 50.3% of patients showing a median OS (mOS) of 33.4 months. A total of 35.8% of patients had a negative-resection margin (R0) after surgery; these patients achieved a mOS of 45.1 months and 1-, 3-, 5-year survival rates of 84.5%, 56.3% and 43.3%, respectively. Microscopic residual disease (R1) after tumor resection was associated with an increased risk of relapse compared to R0 (HR 1.65; 95% CI 1.35-2.02), with a median relapse-free survival of 10.7 and 19.1 months, respectively (Fig. 3B). Moreover, R1 surgery achieved a mOS of 24.7 months and 1-, 3-, 5-year survival rates of 81.1%, 29.4% and 13.7%, respectively. Patients with R1 after surgery were at an increased risk of death compared to patients who had R0 resection (HR 1.92; 95% CI 1.53-2.41), despite not showing survival differences compared to those with gross residual disease (R2) (HR 1.37; 95% CI 0.76-2.48).
Lymph node invasion (N+) also compromised the OS of patients after resection (Fig. 3C). Worse outcomes were observed in patients with N+ compared to N0, both after R0 or R1 tumor resections (HR 2.13; 95% CI 1.55-2.94, and HR 1.61; 95% CI 1.08-2.38), respectively. These differences were also observed in mOS: 52.2 months for R0/N0, 23.3 months for R0/N+, and 29.3 months and 21.8 months for R1/N0 and R1/N+, respectively. Notably, 25.9% of R0- and 34.4% of R1-resected patients received adjuvant treatment, which did not improve the mOS when compared to patients not receiving any adjuvant therapy (Fig. 3A, Fig. S2).
Out of the 816 (49.6%) patients with unresectable disease at diagnosis, the majority (477; 29.0%) received active palliative therapy (i.e. chemotherapy (26.2% of whole cohort), locoregional therapy (1.5%) and combined chemo- and locoregional therapies (1.3%)), with mOS and 1- and 3-year survival rates from time of treatment initiation of 10.6 months, and 45.2% and 8.4%, respectively. In total, 92.2% of patients under palliative treatment showed tumor progression before death. In patients receiving palliative chemotherapy, gemcitabine plus cisplatin was the most common regimen used (70.4%) and led to a significantly reduced risk of death compared to BSC (HR 2.24; 95% CI 1.87-2.67) or gemcitabine alone (HR 1.66; 95% CI 1.22-2.28)(Fig. S3). Patients under active palliative treatment had reduced mOS compared to those undergoing surgery with curative intent (R0/R1) (for R0: HR 4.33; 95% CI 3.60-5.21; for R1: HR 2.25; 95% CI 1.82-2.77) (Fig. 3C). No significant survival differences were found between active palliative therapy and R2 tumor resection (HR 1.62; 95% CI 0.91-2.89). Of note, 20.6% of patients received only BSC resulting in a mOS of 4.0 months.
This comparative analysis of patient management and outcomes revealed certain differences between the 3 CCA subtypes (Fig. 4). In particular, significant differences in survival were observed between CCA subtypes receiving BSC (Fig. 4D). Patients with pCCA received BSC more often (37.3%); however, patients with iCCA showed the poorest prognosis with mOS of 2.8 months compared to the 7.0 and 7.7 months found in pCCA and dCCA, respectively (Table S7). On the other hand, comparable mOS were obtained between CCA subtypes in patients undergoing tumor resection or active palliative treatment (Fig. 4A–C, Table S7).
Fig. 4Therapeutic approaches and outcomes in cholangiocarcinoma subtypes.
Kaplan-Meier analysis for the assessment of long-term outcome of patients with CCA after (A) R0 tumor resection, (B) R1 tumor resection, (C) active palliative treatment, and (D) best supportive care. dCCA, distal cholangiocarcinoma; iCCA, intrahepatic cholangiocarcinoma; pCCA, perihilar cholangiocarcinoma; R0, null margin tumor resection; R1, microscopic residual disease tumor resection. (This figure appears in color on the web.)
The univariable analysis between clinical and demographic variables at diagnosis and OS showed significant associations for CCA subtype, age, sex, ECOG-PS, disease status, and the serum levels of CA19-9, CEA, ALT, aspartate aminotransferase, GGT, ALP, albumin, or bilirubin (Table 3). However, a stepwise multivariable Cox regression analysis indicated that ECOG-PS, MD, and elevated CA19-9 levels were independent prognostic factors (HRs 1.52, 4.03, 2.79, respectively; Table 3). Thus, patient outcomes based on these 3 independent variables were further depicted.
Table 3Univariableand multivariableCox regression analysis of variables at diagnosis.
Fig. 5A represents the OS for patients stratified according to ECOG-PS scores (0, 1, 2, 3-4) at diagnosis with pronounced differences between groups and with a mOS of 25.2 months (reference), 14.8 months (HR 1.57; 95% CI 1.38-1.78), 8.7 months (HR 2.76; 95% CI 2.32-3.28) and 3.0 months (HR 4.65 (95% CI 3.64-5.95), respectively. Besides, the disease stage (LAD, single-site MD, ≥2-site MD) progressively increased the risk of death compared to patients with LD (reference: mOS = 30.9 months), with mOS of 16.2 months (HR 1.94; 95% CI 1.67-2.26) for LAD, 8.1 months (HR 3.75 (95% CI 3.18-4.42) for single-site MD, and 6.1 months (HR 5.02; 95% CI 3.86-6.54) for ≥2-site MD (Fig. 5B). Moreover, CA19-9 had intrinsic prognostic value (HR 2.02; 95% CI 1.71-2.37), with mOS of 31.0 and 12.7 months for normal or elevated (>37 IU/ml) CA19-9 levels, respectively (Fig. 5C).
Fig. 5Independent prognostic value of ECOG performance status, tumor dissemination and CA19-9 in CCA.
Kaplan-Meier analysis and multivariable Cox regression models for the assessment of long-term outcomes of patients with CCA, with all-cause mortality as primary endpoints for (A) ECOG, (B) disease status, and (C) CA19-9 serum tumor biomarker. CA19-9, carbohydrate antigen 19-9; CCA, cholangiocarcinoma; ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; LAD, locally advanced disease; LD, local disease; MD, metastatic disease; mOS, median overall survival. (This figure appears in color on the web.)
To our knowledge, this study constitutes the largest and most comprehensive international analysis evaluating the presentation, clinical management, and outcome of patients with CCA, with special focus on differences between CCA subtypes.
Although a significant proportion of CCAs arise within an apparently healthy liver, population-based studies have identified different risk factors.
Expert consensus document: cholangiocarcinoma: current knowledge and future perspectives consensus statement from the European Network for the Study of Cholangiocarcinoma (ENS-CCA).
Herein, we show that more than 50% of patients with CCA were overweight/obese and 20% were diabetic at diagnosis. Evidence suggests obesity, and the metabolic syndrome in particular, as a major risk factor for cancer in general, but also for CCA.
In fact, the obesity pandemic in the adult population has grown rapidly since the 1970s, which preceded the increased incidence of iCCA observed since the 1980s.
Moreover, recent studies suggest that non-alcoholic fatty liver disease (NAFLD), part of the metabolic syndrome, might be a major risk factor, alone or in association with obesity, for cancer, and in particular for hepatocellular carcinoma and CCA.
we could expect a considerable prevalence of NAFLD in our patients with CCA. In addition, other pathologic conditions that have traditionally been associated with CCA development seem to have a subtype-dependent impact. For instance, viral hepatitis B and C infections inferred greater risk for iCCA than p/dCCA. Of note, during the last decades, the prevalence of viral hepatitis has decreased due to vaccination programs or new effective treatments,
while metabolic-associated conditions are altering the epidemiological setting of CCA. This raises the need for the involvement of primary care to carry out awareness policies based on lifestyle prevention.
Furthermore, 12.6% of the patients with iCCAs arose on a cirrhotic background, most probably in association with the previously reported viral infection, alcohol consumption, or NAFLD. Interestingly, and according to our data, although end-stage cirrhosis in patients with PBC is a well-known risk factor for the development of hepatocellular carcinoma, it may also predispose to iCCA development. This confirms previous data from the Surveillance, Epidemiology and End Results (SEER) program of the National Cancer Institute in the United States, in which PBC was associated with iCCA, but not extrahepatic CCA.
was confirmed in our study. These data support the need to raise CCA awareness and for specific screening programs for clearly identified high-risk populations (i.e., choledochal cysts, biliary stones, cirrhosis, biliary diseases [Caroli´s, PSC], viruses [HBV, HCV]).
Beyond the features shared by all CCAs, increasing evidence indicates that CCA subtypes might differ in clinical presentation, etiology, natural history, management, and prognosis and thus, should be regarded as distinct entities.
Expert consensus document: cholangiocarcinoma: current knowledge and future perspectives consensus statement from the European Network for the Study of Cholangiocarcinoma (ENS-CCA).
This study suggests that perihilar and distal tumors are detected earlier than intrahepatic lesions, mainly because they usually cause obstructive jaundice at an early stage. Consequently, iCCAs appeared as larger or multifocal lesions, and predominantly as moderate-to-poorly differentiated tumors. Most iCCAs showed a mass-forming growth pattern, whereas pCCA and dCCA were mostly flat or periductal infiltrating, and less frequently intraductal, supporting previous observations.
Nonetheless, no differences in disease stage were found between CCA subtypes. This may be, at least in part, because hepatic dissemination of iCCA is not formally considered metastasis according to the current AJCC guidelines. In this regard, we have recently shown that patients with iCCA and cancer spread within the liver, with or without lymph node invasion, have worse prognosis than patients with local iCCA, strongly encouraging the establishment of a new specific code (M1) for these patients.
CCAs are usually diagnosed by a sequential protocol comprising imaging approaches and assessment of non-specific tumor biomarkers in serum, followed by biopsy or cytology, when feasible.
Our data showed that CEA and CA19-9 are increasingly elevated as the disease progresses, supporting previous reports in which preoperative elevation of serum CA19-9 appears as a predictor of nodal invasion.
Impact of nodal involvement on surgical outcomes of intrahepatic cholangiocarcinoma: a multicenter analysis by the Study Group for Hepatic Surgery of the Japanese Society of Hepato-Biliary-Pancreatic Surgery.
Consequently, future prospective studies should determine the potential utility of CEA and CA19-9 to identify patients who would benefit from a more detailed staging, using, for instance, PET with 18F-fluorodeoxyglucose (FDG-PET). Our results highlight the underuse of FDG-PET for the staging of CCA, even though it may help in the identification of occult nodal and distant metastatic status.
Of note, the multivariable Cox regression analysis revealed that elevation of CA19-9, but not of CEA, is an independent prognostic factor for CCA, a finding of translational relevance for patient stratification and design of clinical trials. Indeed, our findings showed independent prognostic value not only for CA19-9, but also for both the ECOG-PS, and disease status. These results share some similarities with a previous work that proposed a new clinical-based staging system for pCCA that includes ECOG-PS, tumor size and number, vascular encasement, tumor dissemination and CA19-9 as stratification factors.
In our dataset, 1/5 patients did not receive active palliative treatment (just BSC), probably due to late diagnosis and deterioration of patients’ ECOG-PS. However, our cohort study confirmed longer survival in patients who received some form of anti-cancer treatment for unresectable disease,
In this regard, future studies should compare locoregional – with current limited experience in CCA – vs. systemic therapies for the treatment of unresectable CCAs. According to our data, tumor resection is the best therapeutic option, even though lymph node involvement is an important determinant of clinical outcome. However, the decision to perform tumor resection is a difficult trade-off between short-term risk (i.e., post-surgical mortality) and potential long-term benefit. According to the CCA subtypes, patients with iCCA showed the worst OS under BSC, probably due to the progression of associated chronic liver diseases (e.g., cirrhosis, viral hepatitis, NAFLD). In contrast, similar outcomes were shown for all 3 CCA subtypes under active palliative treatment and tumor resection. Altogether, these data reinforce the need for adequate investment in early diagnosis of CCA, the shortening of time to surgery, and the systemic treatment of advanced disease when feasible, as the best strategies to improve the outcome of patients.
This study has several limitations. It shows novel data on the course of CCAs in European reference centers, but it cannot be interpreted as a demographical study; therefore, caution is required when extrapolating the results. Selection bias related to clinical specialties of participating centers (hepatologists, gastroenterologists, medical oncologists, surgeons) could explain the differences between CCA subtypes and disease stages at diagnosis. Besides, the diagnosis and classification by CCA subtypes were based on investigator-reported data following data harmonization. Even though no external audit was performed, an internal review was conducted by each center in order to double-check the included data. Nevertheless, the absence of a central reading should not have a major impact on the conclusions drawn from this work. In fact, the expected number of cases with undistinguishable location would be very low as they are retrieved by investigators affiliated to referral hospitals with large experience in the management of CCA. Clinical approaches related to the diagnostic work-up and disease monitoring programs may diverge between hospitals and specific departments of specialization. In addition, differences in terms of disease phenotype and incidence of risk conditions for CCA may exist between countries. In a separate matter, the percentage of patients receiving BSC is probably underestimated as invasive methods for histological/cytological disease confirmation required patients to be eligible for this study, and thus, are often not performed for those unfit for anti-cancer treatment.
In conclusion, our study including more than 2,200 patients with CCA from 11 European countries provides a comprehensive analysis of diagnostic, prognostic and therapeutic aspects of the complex CCA landscape, showing that CCA is still diagnosed at an advanced stage, a significant proportion of patients fail to receive any cancer-specific therapy, and therefore, the prognosis is dismal. Accordingly, the promotion of awareness campaigns and education programs aimed to prevent lifestyle-related risk factors and the implementation of surveillance for early detection of CCA in high-risk populations are urgently required in order to decrease cancer-related mortality. In this regard, this study represents valuable knowledge for future comparisons with new targeted therapies and for the design of next-generation personalized clinical trials.
Abbreviations
AEG, "Asociación Española de Gastroenterología"; AJCC, American Joint Committee of Cancer; ALP, alkaline phosphatase; ALT, alanine aminotransferase; BSC, best supportive care; CA19-9, carbohydrate antigen 19-9; CCA, cholangiocarcinoma; CEA, carcinoembryonic antigen; dCCA, distal CCA; ECOG-PS, Eastern Cooperative Oncology Group performance status; ENS-CCA, European Network for the Study of CCA; FDG-PET, positron emission tomography with 18F-fluorodeoxyglucose; GGT, gamma-glutamyltransferase; HR, hazard ratio; IBD, inflammatory bowel disease; iCCA, intrahepatic CCA; ICD, International Classification of Diseases; INR, international normalized ratio; LAD, locally advanced disease; LD, local disease; M, distant metastasis; MD, metastatic disease; mOS, median overall survival; MRCP, magnetic resonance cholangiopancreatography; N, lymph node invasion; NAFLD, non-alcoholic fatty liver disease; OR, odds ratio; OS, overall survival; Ox, oxaliplatin; PBC, primary biliary cholangitis; pCCA, perihilar CCA; PSC, primary sclerosing cholangitis; R0, null resection margin; R1, microscopic residual resection margin; R2, macroscopic residual resection margin.
Financial support
The ENSCCA Registry is competitively funded by the European Association for the Study of the Liver (EASL; Registry grant awards 2016 and 2019) and Incyte Pharma (grant award 2020 to JMB). JMB also received grants from Spanish Carlos III Health Institute (ISCIII) [FIS PI18/01075,PI21/00922 and Miguel Servet Program CPII19/00008), CIBERehd (ISCIII), Department of Health of the Basque Country (2017111010), BIOEF (Basque Foundation for Innovation and Health Research: EiTB Maratoia BIO15/CA/016/BD), “Fundación Científica de la Asociación Española Contra el Cáncer” (AECC Scientific Foundation) and the European Union’s Horizon 2020 Research and Innovation Program [grant number 825510, ESCALON: to JMB, AL, AV and JWV]. AL also received funding from The Christie Charity. AF received grant support from ISCIII (PI18/00542). CB is recipient of the Lord Kelvin Adam Smith Leadership fellowship at University of Glasgow. ALl was funded by “Fondazione AIRC per la Ricerca sul Cancro” (IG2019 project number 23408). RIRM received grant support from ISCIII (PI20/00189), co-funded ERDF/ESF, "A way to make Europe"/"Investing in your future" and “Centro Internacional sobre el Envejecimiento” (0348_CIE_6_E), Spain. JJGM received grant support from ISCIII (PI19/00819), CIBERehd (ISCIII), “Fundació la Marató de TV3” (201916-31) and Junta de Castilla y León (SA074P20), Spain. NBJ received grant support from Cancer Research UK (grant number A25813), and is a recipient of the Lord Kelvin Adam Smith Leadership fellowship at the University of Glasgow.
Authors’ contributions
LI-S: study concept and design; analysis and interpretation of data; statistical analysis; drafting of the manuscript. ALl, AL-M, AS, ASc, AV, CB, FL-L, GLG, GV, HJK, JA, JI, JK, JPJ, KE, KU, LB, LF, LK, MK, MM, MP-S, NJ, RB, RIRM, SB, TF, VC, VS, WH, ZS, JIE: data acquisition; proof-reading. AS-L: study concept and design; data acquisition; proof-reading. GC, JBA, JJGM, DA: funding; proof-reading. AF, AL, ALC, BGK, JWV: data acquisition, analysis and interpretation; drafting of the manuscript; proof-reading. JMB: study concept and design; analysis and interpretation of data; statistical analysis; drafting of the manuscript; funding.
Data availability statement
Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so data should remain confidential.
Conflict of interest
AF declares lecture fees (from Bayer HealthCare, Gilead, Roche, MSD and Boston Scientific), and consultancy fees (from Bayer HealthCare, AstraZeneca, Roche, Guerbert, SIRTEX and Exact Science). AL reported travel and educational support (from Ipsen, Pfizer, Bayer, Advanced Acceletaro Applications, SirtEx, Novartis, Mylan and Delcath), speaker honoraria (from Merck, Pfizer, Ipsen, Incyte and AAA), advisory honoraria (from EISAI, Nutricia Ipsen, QED and Roche), and she is Member of the Knowledge Network and NETConnect Initiatives funded by Ipsen. ALl declares lecture fees (from Intercept Pharma, Abbvie, Gilead, AlfaSigma, and MSD) and consultancy fees (from Intercept and Alfa Sigma). AL-M declares travel and educational support (from Pfizer, Roche, Sanofi, Rovi, Pharma Mar and Merck). CB declares consultancy honoraria (from Incyte). GV declares consultancy fee (from Bayer HealthCare). JBA declares consulting role (for QED Therapeutics and SEALD). JMB declares research grants (from Incyte), personal fees for lecturer (from Bayer and Intercept), and consulting role (for QED Therapeutics, Albireo Pharma and OWL Metabolomics). JWV reports personal fees (from Agios, AstraZeneca, Baxter, Genoscience Pharma, Hutchison Medipharma, Imaging Equipment Ltd (AAA), Incyte, Ipsen, Mundipharma EDO, Mylan, QED, Servier, Sirtex, Zymeworks), and grants, personal fees and non-financial support (from NuCana). LF declares research grants (from Progetto di Ricerca Dipartimentale), honoraria (from Obesity 4C Complex Clinical Cases Contest), leadership role (for HCERES, NCN), and travel expenses (from AASLD, AISF and EASL). MP-S declares honoraria (from Roche and Servier), consulting or advisory role (for AstraZeneca) and travel expenses (from Roche, BMS and Incyte). KU declares fees for advisory role (BMS). ALC, AS, ASc, AS-L, AV, BGK, DA, FL-L, GC, GLG, HJK, JA, JI, JIE, JJGM, JK, JPJ, KE, LB, LI-S, LK, MK, MM, NJ, RB, RIRM, SB, TF, VC, VS, WH, and ZS declare no competing interests. Please refer to the accompanying ICMJE disclosure forms for further details.
Acknowledgement
Authors thank to patients and their families for the participation in this study, to the "Asociación Española de Gastroenterología" (AEG) for the REDCap database license and training, to Ms. Miren Maitane Gaztañaga Otermin for the kind donation to the Registry, and to Dr. Ioana Riaño (Biodonostia Health Research Institute) for the support on ethical aspects. This article is based upon work from European Horizon 2020 COST Action CA18122 European Cholangiocarcinoma Network (Euro-Cholangio-Net) supported by COST (European Cooperation in Science and Technology: www.cost.eu), in collaboration with the European Network for the Study of Cholangiocarcinoma (ENSCCA: http://www.enscca.org/), the International Primary Sclerosing Cholangitis Study Group (iPSCSG: https://www.ipscsg.org/), the European Reference Network on Rare Liver Diseases (ERN-Rare Liver: https://rare-liver.eu/) and the European Reference Network on Rare Adult Cancers (solid tumors; EURACAN:https://euracan.eu/).
Supplementary data
The following are the supplementary data to this article:
Expert consensus document: cholangiocarcinoma: current knowledge and future perspectives consensus statement from the European Network for the Study of Cholangiocarcinoma (ENS-CCA).
Impact of nodal involvement on surgical outcomes of intrahepatic cholangiocarcinoma: a multicenter analysis by the Study Group for Hepatic Surgery of the Japanese Society of Hepato-Biliary-Pancreatic Surgery.
02252-2&id=ga1.jpg){kind=link}
02252-2&id=gr1.jpg){kind=link}
02252-2&id=gr2.jpg){kind=link}
02252-2&id=gr3.jpg){kind=link}
02252-2&id=gr4.jpg){kind=link}
02252-2&id=gr5.jpg){kind=link}