Prognostic factors and predictive model in patients with advanced biliary tract adenocarcinoma receiving first-line palliative chemotherapy




Advanced biliary tract adenocarcinoma (BTA) has been a rare but fatal cancer. If unresectable, palliative chemotherapy improved the quality and length of life, but to the authors' knowledge, prognostic factors in such patients have not been well established to date. In the current study, prognostic factors were investigated in patients with advanced BTA receiving first-line palliative chemotherapy.


Data from 213 patients with advanced BTA who were in prospective phase 2 or retrospective studies from September 2000 through October 2007 were used.


With a median follow-up duration of 29.7 months, the median overall survival (OS) was 7.3 months (95% confidence interval [95% CI], 6.3 months-8.3 months). A Cox proportional hazards model indicated that metastatic disease (hazards ratio [HR], 1.521; P = .011), intrahepatic cholangiocellular carcinoma (HR, 1.368; P = .045), liver metastasis (HR, 1.845; P < .001), Eastern Cooperative Oncology Group performance status (HR, 1.707; P < .001), and alkaline phosphatase level (IU/L) (HR, 1.001; P < .001) were statistically significant independent predictors of poor prognosis. Patients were classified into 3 risk groups based on the prognostic index (PI), which was constructed using the regression coefficients of each variable. The median OS was 11.5 months (95% CI, 9.6 months-13.5 months) for the low-risk group (PI ≤ 1.5; n = 67), 7.3 months (95% CI, 5.7 months-8.9 months) for the intermediate-risk group (PI > 1.5 but ≤ 2.2; n = 75), and 3.6 months (95% CI, 2.9 months-4.1 months) for the high-risk group (PI > 2.2; n = 70 [P < .001]).


Five prognostic factors in patients with advanced BTA were identified. The predictive model based on PI appears to be promising and may be used for the management of individual patients and to guide the design of future clinical trials, although external validation is needed. Cancer 2009. © 2009 American Cancer Society.

Advanced biliary tract adenocarcinoma (BTA) is a rare but fatal cancer. BTA includes a heterogeneous group of cancers: intrahepatic cholangiocellular carcinoma (IHCCC), gallbladder cancer (GBC), extrahepatic bile duct carcinoma (EHBDC), and ampulla of Vater cancer (AoVC). In the United States, BTAs affect approximately 12,000 people annually and the incidence of IHCCC has increased in recent years.1 In Korea, BTA accounts for approximately 4.8% of all malignant neoplasms and is estimated to be the seventh most common cancer.2 Complete resection is the only cure for BTA, but only 10% of patients present with early stage disease and are candidates for surgical resection.3 In addition, among patients who undergo curative surgery, recurrence rates are extremely high. For patients with unresectable advanced or recurrent BTA, 5-fluorouracil-based palliative chemotherapy, which can improve the quality and length of life,4, 5 is the mainstay of treatment.3

To our knowledge, because of the rarity of BTA and the heterogeneity of the patient population, there are no well-designed, large-scale randomized controlled trials. Most trials published to date have been small phase 2 studies. Furthermore, only a minority of BTA patients treated with chemotherapy experience favorable responses, and few achieve long-term survival. Therefore, the confirmation of a standard therapeutic regimen and the identification of new chemotherapeutic agents is a high priority.

In designing and interpreting phase 2 or 3 clinical trials for the treatment of BTA, it is important to identify factors associated with patient survival. In particular, phase 3 trials that include survival as a primary endpoint must account for prognostic factors so that treatment groups are comparable and trial outcomes are properly interpreted. In addition, adequate assessment of patient survival will also aid patients and physicians in their clinical judgment. Several studies have investigated prognostic factors associated with survival from BTA, but the majority addressed only the risk of disease recurrence after curative resection.6-12 To our knowledge, only 1 small study to date has investigated prognostic factors for patients with advanced GBC and EHBDC who received palliative chemotherapy.13

The objective of the current study was to evaluate the prognostic factors that affect the overall survival (OS) of patients with BTA and to establish a prognostic prediction model for patients with advanced BTA who receive palliative chemotherapy.



We evaluated data from 213 patients with advanced BTA who were enrolled in 2 prospective phase 2 studies or who were included in a retrospective cohort study from September 2000 through October 2007. The first prospective study was a phase 2 trial (capecitabine plus cisplatin protocol) in which 42 patients were treated with oral capecitabine (at a dose of 1250 mg/m2 twice per day from Day 1 to Day 14) and cisplatin (at a dose of 60 mg/m2 intravenously at Day 1 every 3 weeks).14 The second phase 2 study (gemcitabine plus capecitabine protocol) was a prospective study in which 38 patients received gemcitabine (at a dose of 1000 mg/m2 intravenously for 30 minutes on Days 1, 8, and 15) and oral capecitabine (at a dose of 830 mg/m2 twice per day from Day 1 to Day 21 every 4 weeks).15 The cohort study was a retrospective review of 133 patients who were treated with oral S-1 monotherapy (at a dose of 40 mg/m2 twice per day from Day 1 to Day 14 every 3 weeks) and who fulfilled the preset eligibility criteria.16 The inclusion criteria for both prospective studies were as follows: 1) histologically or cytologically confirmed adenocarcinoma of the biliary tract; 2) distant metastases or locally advanced disease that was not eligible for curative surgery; 3) ages 18 to 75 years; 4) Eastern Cooperative Oncology Group (ECOG) performance status of 2 or better; 5) presence of measurable lesions; 6) no serious or uncontrolled concomitant medical illness; and 7) adequate bone marrow and organ function (absolute neutrophil count ≥ 2000/μL, platelet count ≥ 100,000/μL, and serum creatinine ≤ 1.5 mg/dL). The eligibility criteria for the retrospective study were not significantly different. However, in the retrospective study, patients with elevated transaminase and hyperbilirubinemia (total bilirubin ≥ 2 mg/dL) associated with biliary obstruction caused by the disease were included so that the efficacy and safety of S-1 monotherapy for patients with impaired liver function could be assessed. In addition, patients with evaluable disease, but not measurable disease based on Response Evaluation Criteria in Solid Tumor criteria, were included because the primary endpoint of the retrospective study was OS.

Data Analysis and Statistical Consideration

The data analyzed included sex; age; ECOG performance status; macroscopic and microscopic characteristics of the tumors, including primary site (IHCCC, GBC, EHBDC, and AoVC); disease status (initially locally advanced, locoregionally recurrent, initially metastatic, and recurrence with distant metastasis); histologic differentiation; sites of metastases; number of metastases; baseline biochemical parameters, including white blood cell (WBC) count (/μL), hemoglobin concentration (Hb) (g/dL), platelet count (/μL), bilirubin (mg/dL), alkaline phosphatase (ALP) (reference value, 40-120 IU/L), calcium (mg/dL), alanine aminotransferase (ALT) (IU/L), aspartate aminotransferase (AST) (IU/L), albumin (g/dL), and carbohydrate antigen 19-9 (CA 19-9) (reference value, 0-37 U/mL); experience of biliary decompression procedures; chemotherapy regimens; date of treatment; and date and cause of death or date of the last follow-up.

The primary endpoint was OS, defined as the date from the first chemotherapeutic cycle to the date of death. The Kaplan-Meier algorithm was used to estimate OS. We compared survival using the Cox proportional hazards model according to clinicopathologic characteristics including age, sex, ECOG performance status, disease status at palliative chemotherapy, primary tumor site, metastasis site, laboratory values, chemotherapeutic regimen, and whether biliary obstruction was present. All potential prognostic factors with a probability value ≤ .20 on univariate analyses were entered into the multivariable Cox models. The final models were determined by backward elimination. Schoenfeld residuals and the log (-log [survival rate]) were used to verify that the proportional hazards assumptions were not violated. The Statistical Package for the Social Sciences (SPSS) for Windows (SPSS Inc, Chicago, Ill) and S-plus 2000 (Mathsoft Inc, Seattle, Wash) were used for statistical analyses.


Patient Characteristics

Baseline clinical and laboratory characteristics are summarized in Table 1. The median age of the patients was 59 years (range, 21 years-83 years) and 120 patients (56.3%) were men. Primary sites of disease were IHCCC (85 patients; 39.9%), GBC (72 patients; 33.8%), EHBDC (39 patients; 18.3%), and AoVC (17 patients; 8.0%). Forty-nine patients (23.0%) had biliary obstruction and underwent drainage procedures, such as percutaneous transhepatic biliary drainage or endoscopic retrograde biliary stent. Fifty-nine patients (27.7%) received second-line chemotherapy.

Table 1. Characteristics of Patients With Advanced Biliary Tract Cancer (n = 213)
  1. IHCCC indicates intrahepatic cholangiocellular carcinoma; GBC, gallbladder cancer; EHBDC, extrahepatic bile duct cancer; AoVC, ampulla of Vater cancer; XP, capecitabine plus cisplatin; GX, gemcitabine plus capecitabine; ECOG, Eastern Cooperative Oncology Group; ALP, alkaline phosphatase, ALT, alanine aminotransferase; AST, aspartate aminotransferase; CA 19-9, carbohydrate antigen 19-9.

Age, median (range), y59 (21-83)
Initial status
 Recurrence, locoregional104.7
 Recurrence, distant metastasis5123.9
 Initially locally advanced2511.7
 Initially metastatic12759.6
Primary site
Metastasis site
 Abdominal lymph node (M1)10850.7
 Cervical lymph node167.5
No. of metastatic organ
 ≥ 4115.2
Chemotherapeutic regimen
Histologic grade
 Well differentiated125.6
 Moderately differentiated8941.8
 Poorly differentiated5626.3
Measurable disease19190.5
Bile duct obstruction4923.0
ECOG performance status
Laboratory tests, median (range)
 White blood cells, /μL7200 (2200-24400) 
 Hemoglobin, g/dL12.3 (8.5-16.1) 
 Total bilirubin, mg/dL0.9 (0.3-12.0) 
 ALP, IU/L146 (41-1543) 
 ALT, IU/L27 (5-322) 
 AST, IU/L32 (11-317) 
 Albumin, g/dL3.6 (1.9-4.8) 
 Calcium, mg/dL9.0 (7.5-12.4) 
 CA 19-9,U/mL72.1 (1.5-310000) 

Survival Distribution

At the time of analysis (June 2008), 192 patients had died (90.1%) and the median OS time was 7.3 months (95% CI, 6.3 months-8.3 months) (Fig. 1), with a median follow-up duration of 29.7 months (95% CI, 16.9 months-42.5 months). The 6-month survival rate was 57.5%, and the 1-year survival rate was 28.5%.

Figure 1.

Overall survival of patients with advanced biliary tract adenocarcinoma (n = 213) undergoing first-line palliative chemotherapy is shown.

Univariate Survival Analysis

A univariate analysis indicated that disease status (metastatic [initially metastatic or disease recurrence with distant metastasis] vs locally advanced [initially locally advanced or locoregionally recurrent]), primary site (IHCCC vs GBC or EHBDC or AoVC), peritoneal metastasis, liver metastasis, bone metastasis, lung metastasis, ECOG performance status, WBC counts, Hb levels, total bilirubin levels, ALP values, AST values, albumin levels, and corrected calcium levels17 were statistically associated with OS (Table 2). Age, gender, chemotherapy regimen, and CA 19-9 values were found to have no influence on patient survival.

Table 2. Univariate Analysis of Prognostic Factors Affecting Overall Survival in Patients With Advanced Biliary Tract Adenocarcinoma
VariableP*HR (95% CI)
  • HR indicates hazards ratio; 95% CI, 95% confidence interval; GX, gemcitabine plus capecitabine; XP, capecitabine plus cisplatin; IHCCC, intrahepatic cholangiocellular carcinoma; LN, lymph node; ECOG, Eastern Cooperative Oncology Group; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; CA 19-9, carbohydrate antigen 19-9.

  • *

    P value was determined using Cox regression analysis.

  • Values are divided by 1000 to accentuate the HR.

Age, y.3290.992 (0.977-1.008)
Sex, male vs female.5271.097 (0.824-1.459)
 GX vs XP.0401.619 (1.023-2.563)
 S-1 vs XP.2571.240 (0.855-1.798)
Initial status, metastatic vs locally advanced<.0012.007 (1.491-2.701)
Primary site, IHCCC vs others<.0011.752 (1.310-2.343)
Peritoneal metastasis, yes vs no.0421.415 (1.013-1.977)
Liver metastasis, yes vs no<.0011.918 (1.421-2.587)
Abdominal LN metastasis, yes vs no.0771.294 (0.972-1.721)
Bone metastasis, yes vs no.0013.337 (1.625-6.851)
Lung metastasis, yes vs no.0511.470 (0.998-2.165)
ECOG performance status<.0011.626 (1.282-2.062)
Leukocyte counts, /μL<.0011.098 (1.043-1.156)
Hemoglobin, g/dL.0610.921 (0.845-1.004)
Bilirubin, mg/dL.041.130 (1.005-1.271)
ALP, IU/L<.0011.001 (1.001-1.002)
AST, IU/L<.0011.011 (1.006-1.016)
ALT, IU/L.4291.001 (0.998-1.005)
Albumin, g/dL<.0010.620 (0.483-0.797)
Corrected calcium, mg/dL<.0011.576 (1.265-1.964)
CA 19-9, U/mL.4951.000 (1.000-1.000)

Multivariable Analysis, Prognostic Index, and Risk Groups

Among the parameters considered, initial disease status, primary tumor site (IHCCC vs GBC or EHBDC or AoVC), ECOG performance status, liver metastasis, and ALP values were found to be significant prognostic factors (Table 3). Of the 213 patients, we had complete data for 212 patients (99.5%) with regard to 5 parameters and therefore these patients were included in the prognosis prediction model. For the clinical application of these findings, a prognostic index (PI) was calculated based on the regression coefficients derived from the 5 variables identified by multivariable analysis. The index equation was as follows: 0.42 X (0, locally advanced disease; 1, metastatic disease) + 0.31 × (0, extrahepatic biliary tract adenocarcinoma; 1, IHCCC) + 0.61 X (0, no liver metastasis; 1, liver metastasis) + 0.54 X (ECOG performance status) + 0.001 X (ALP [IU/L]). The individual PI values for the patients ranged from 0.47 to 3.63, with a median of 1.82. The patients were then reclassified into 3 risk groups according to the PI values (33%ile and 66%ile): the low-risk group had a PI ≤ 1.5 (n = 67), the intermediate-risk group had a PI of > 1.5 but ≤ 2.2 (n = 75), and the high-risk group had a PI > 2.2 (n = 70). The resulting Kaplan-Meier curves for the 3 groups (P < .001) (Fig. 2) indicated marked differences in survival. The median OS was 11.5 months (95% CI, 9.6 months-13.5 months) for the low-risk group (PI ≤ 1.5; n = 67), 7.3 months (95% CI, 5.7 months-8.9 months) for the intermediate-risk group (PI > 1.5 but ≤ 2.2; n = 75), and 3.6 months (95% CI, 2.9 months-4.1 months) for the high-risk group (PI > 2.2; n = 70). The 1-year survival rates for each group were 48.2%, 26.3%, and 4.2%, respectively. Compared with the low-risk group, the intermediate-risk group had a 2.1-fold increased risk of death (hazards ratio [HR], 2.08; 95% CI, 1.43-3.01), and the high-risk group had a 4.3-fold increased risk of death (HR, 4.27; 95% CI, 2.89-6.31). When patients with GBC were removed in this model, the difference in survival of each group remained significant (P < .001).

Figure 2.

Kaplan-Meier survival curves according to prognostic groups are shown. The low-risk group (solid line) (prognostic index of ≤ 1.5; n = 67), intermediate-risk group (dash-dot line) (prognostic index of > 1.5 but ≤ 2.2; n = 75), and high-risk group (dotted line) prognostic index of > 2.2; n = 70) showed marked survival differences (P < .001).

Table 3. Multivariable Analysis Using the Cox Proportional Hazards Model in Patients With Advanced Biliary Tract Adenocarcinoma
VariableHR (95% CI)PCox Model Coefficient (β)
  1. HR indicates hazards ratio; 95% CI, 95% confidential interval; LA, locally advanced; IHCCC, intrahepatic cholangiocellular carcinoma; ECOG, Eastern Cooperative Oncology Group; ALP, alkaline phosphatase.

Initial status, metastatic vs LA1.521 (1.103-2.097).0110.419
Primary site, IHCCC vs others1.368 (1.007-1.858).0450.313
Liver metastasis, yes vs no1.845 (1.352-2.516)<.0010.612
ECOG performance status1.707 (1.308-2.228)<.0010.535
ALP, IU/L1.001 (1.001-1.002)<.0010.001


The identification of factors that predict survival from advanced BTA can help investigators predict life expectancy, guide treatment plans, analyze clinical studies, and design future clinical trials. To the best of our knowledge, previous studies of predictors of survival from BTA for the most part examined postoperative patients and were not confined to patients who were undergoing palliative chemotherapy.6-12 In the current study, we identified 5 independent prognostic factors for survival from advanced BTA: metastatic disease status, IHCCC, ECOG performance status, liver metastasis, and ALP elevation. We also calculated a simple PI based on the regression coefficient and devised prognostic groups, which indicated that patients in different risk groups had markedly different survival rates. The means of determining all 5 prognostic factors are readily available to treating physicians before chemotherapy, and we expect that their use will aid in clinical decision making and risk stratification. To our knowledge, this is the first report that systemically identifies prognostic factors in a large subset of patients with advanced BTA. However, future studies that use independent data sets are needed to verify the results of the current study.

BTA can arise anywhere in the biliary tree and includes IHCCC, GBC, EHBDC, and AoVC. There is an indication of location-related pathogenetic differences in BTAs, and these may be associated with different responses to chemotherapy and overall prognosis.18 However, there has been debate regarding differences in the prognoses of patients with different types of BTA. The results of the current study confirm that IHCCC predicts poor OS, as suggested in our previous report.14, 16 Several other studies reported results similar to ours,14, 19-22 but 1 study reported poorer outcomes for patients with EHBDC (especially GBC)23 and another study reported no prognostic differences in patients with different types of BTC.24

We identified metastasis to the liver as an independent factor that was predictive of poor prognosis. Similar results have been reported for gastric cancer and urothelial cancers.25-27 In the current study, elevated ALP was found to be an independent prognostic factor of survival from advanced BTA. In other malignancies such as breast, esophagogastric, and colorectal cancer, elevated ALP was also reported to be an independent prognostic factor.25, 28, 29 In the study of Chau et al,25 elevated ALP and liver metastasis were found to be independent prognostic factors in patients with advanced gastric cancer, as in the current study. It has been suggested that elevated ALP reflects metastatic tumor burden rather than the presence or absence of liver metastasis. Poor ECOG performance status is a simple metric, but is a robust prognostic factor for a variety of malignancies.13, 26, 30-32 As with those previous studies, we found that poor ECOG performance status predicts a low rate of survival in patients with advanced BTA.

Saisho et al13 analyzed prognostic factors for survival in 65 patients with advanced BTA (with the exception of IHCCC) who were receiving palliative chemotherapy. Their univariate analysis indicated that primary tumor location of EHBDC (but not GBC), poor performance status, anemia, hypoalbuminemia, lactate dehydrogenase (LDH) elevation, and C-reactive protein (CRP) elevation were significant. Their multivariate analysis indicated that poor performance status, CRP elevation, hypoalbuminemia, LDH elevation, and male gender were significant. There are several reasons why some of the results of the Saisho et al13 study differ from those of the current study. First, they did not include metastatic site and ALP as variables. Second, their study included no patients with IHCCC, whereas IHCCC was the most common BTA subtype in our patient population.

Contrary to our expectation, we found that CA 19-9 elevation was not an independent prognostic factor. We also dichotomized this variable by applying an upper normal limit of 37 U/mL or its median value (76 U/mL), but the results indicated that CA 19-9 elevation was not a significant prognostic factor (data not shown), as in the study of Saisho et al13

S-1 is a novel oral fluoropyrimidine that is comprised of tegafur, 5-choloro-2,4-dihydroxypyrimidine (CDHP), and potassium oxonate (Oxo). S-1 has shown potent antitumor activity against various solid tumors in clinical studies.33 A recent multicenter phase 2 trial of S-1 monotherapy as first-line treatment in patients with advanced BTA achieved an overall response rate of 35%, a stable disease rate of 42.5%, and a median OS of 9.4 months. Toxicities were generally mild.34 A retrospective analysis of our center's experience with S-1 monotherapy in patients with advanced biliary cancer16 revealed an overall response rate of 13.3% and an overall disease control rate of 55.8%. The median OS was 6.9 months.

Although the limited number of patients in our study (n = 213) hinders proper interpretation because of the lack of adequate statistical power, we found that chemotherapeutic regimens did not affect OS (P = .12). A systemic review of 2810 patients with advanced BTAs who were enrolled in clinical trials (predominantly phase 2) indicated that combination chemotherapy was better than monotherapy, and that a 2-drug combination of fluoropyrimidines or gemcitabine plus a platinum compound was better than other combinations.35 Gemcitabine plus a platinum agent for BTA patients was reported to be the most effective treatment. However, this pooled analysis of heterogeneous phase 2 clinical trials (which had different eligibility criteria) did not consider primary tumor site and differences in treatment schedules, and therefore the conclusions should be viewed with caution.35 To the best of our knowledge, no standard regimen for the treatment of patients with advanced BTA has been established to date, and therefore chemotherapeutic regimens must be compared in future prospective trials. Our center is currently performing randomized phase 2 screening trials that compare S-1 plus cisplatin versus gemcitabine plus cisplatin for treatment of patients with advanced BTA.

There are several limitations to the current study. First, we used retrospective analysis and therefore we could not include information that had not been precisely documented (eg, weight loss before treatment, presence and intensity of cancer pain, cancer cachexia, and quality of life). Some of these factors may significantly affect survival from BTA. For example, cancer cachexia is a well-known negative prognostic factor for lung cancer.32 Second, we could only include laboratory tests that were routinely performed. For example, LDH and CRP levels have been reported to have prognostic value in other malignancies,8, 13, 32 but they were not assessed in the current study.


For advanced BTA patients receiving palliative chemotherapy, we identified 5 independent prognostic factors for OS: metastatic disease status, IHCCC, liver metastasis, ECOG performance status, and ALP values. We suggest that these factors could be used in prognostic models to determine treatment strategy and to aid in the design of future clinical trials, but these findings need to be validated in an independent data set before application.

Conflict of Interest Disclosures

This study was supported by a grant from the Korea Health 21 R&D Project, Ministry for Health, Welfare and Family Affairs, R.O.K. (A030001).