Article
Reference
Intrahepatic cholangiocarcinoma: an international multi-institutional
analysis of prognostic factors and lymph node assessment
DE JONG, Mechteld C, et al.
Abstract
To identify factors associated with outcome after surgical management of intrahepatic
cholangiocarcinoma (ICC) and examine the impact of lymph node (LN) assessment on
survival.
DE JONG, Mechteld C, et al. Intrahepatic cholangiocarcinoma: an international
multi-institutional analysis of prognostic factors and lymph node assessment. Journal of clinical
oncology, 2011, vol. 29, no. 23, p. 3140-5
DOI : 10.1200/JCO.2011.35.6519
PMID : 21730269
Available at:
http://archive-ouverte.unige.ch/unige:24979
Disclaimer: layout of this document may differ from the published version.
1 / 1
Intrahepatic Cholangiocarcinoma: An International
Multi-Institutional Analysis of Prognostic Factors and
Lymph Node Assessment
Mechteld C. de Jong, Hari Nathan, Georgios C. Sotiropoulos, Andreas Paul, Sorin Alexandrescu,
Hugo Marques, Carlo Pulitano, Eduardo Barroso, Bryan M. Clary, Luca Aldrighetti, Cristina R. Ferrone,
Andrew X. Zhu, Todd W. Bauer, Dustin M. Walters, T. Clark Gamblin, Kevin T. Nguyen, Ryan Turley,
Irinel Popescu, Catherine Hubert, Stephanie Meyer, Richard D. Schulick, Michael A. Choti,
Jean-Francois Gigot, Gilles Mentha, and Timothy M. Pawlik
Mechteld C. de Jong, Hari Nathan, Richard
D. Schulick, Michael A. Choti, and Timothy
M. Pawlik, Johns Hopkins University
School of Medicine, Baltimore, MD; Geor-
gios C. Sotiropoulos and Andreas Paul,
University Hospital Essen, Essen, Germa-
ny; Sorin Alexandrescu and Irinel Popescu,
Institute for Digestive Diseases and Liver
Transplantation Fundeni, Bucharest, Roma-
nia; Hugo Marques and Eduardo Barroso,
Curry Cabral Hospital, Lisbon, Portugal;
Carlo Pulitano and Luca Aldrighetti, Osped-
ale San Raffaele, Milan, Italy; Bryan M.
Clary and Ryan Turley, Duke Medical
Center, Durham, NC; Cristina R. Ferrone
and Andrew X. Zhu, Massachusetts
General Hospital, Boston, MA; Todd W.
Bauer and Dustin M. Walters, University of
Virginia, Charlottesville, VA; T. Clark
Gamblin and Kevin T. Nguyen, University of
Pittsburgh, Pittsburgh, PA; Catherine
Hubert and Jean-Francois Gigot, Cliniques
Universitaires Saint-Luc, Brussels, Belgium;
and Stephanie Meyer and Gilles Mentha,
Hoˆ pitaux Universitaires de Gene` ve,
Geneva, Switzerland.
Submitted March 2, 2011; accepted
May 19, 2011; published online ahead
of print at www.jco.org on July 5,
2011.
Presented orally at the 2011 Annual
Gastrointestinal Cancers Symposium,
January 20-22, 2011, San Francisco, CA
and 2011 Annual Meeting of the Soci-
ety of Surgical Oncology, March 2-5,
2011, San Antonio, TX.
Authors’ disclosures of potential conflicts
of interest and author contributions are
found at the end of this article.
Corresponding author: Timothy M. Pawlik,
MD, MPH, Associate Professor of Surgery
and Oncology, Department of Surgery,
Harvey 611, 600 N. Wolfe St, Baltimore,
MD 21287; e-mail: tpawlik1@jhmi.edu.
© 2011 by American Society of Clinical
Oncology
0732-183X/11/2923-3140/$20.00
DOI: 10.1200/JCO.2011.35.6519
ABSTRACT
Purpose
To identify factors associated with outcome after surgical management of intrahepatic cholangio-
carcinoma (ICC) and examine the impact of lymph node (LN) assessment on survival.
Patients and Methods
From an international multi-institutional database, 449 patients who underwent surgery for ICC
between 1973 and 2010 were identified. Clinical and pathologic data were evaluated using uni- and
multivariate analyses.
Results
Median tumor size was 6.5 cm. Most patients had a solitary tumor (73%) and no vascular
invasion (69%). Median survival was 27 months, and 5-year survival was 31%. Factors
associated with adverse prognosis included positive margin status (hazard ratio [HR], 2.20; P ⬍ .001),
multiple lesions (HR, 1.80; P ⫽ .001), and vascular invasion (HR, 1.59; P ⫽ .015). Tumor size
was not a prognostic factor (HR, 1.03; P ⫽ .23). Patients were stratified using the American
Joint Committee on Cancer/International Union Against Cancer T1, T2a, and T2b categories
(seventh edition) in a discrete step-wise fashion (P ⬍ .001). Lymphadenectomy was performed
in 248 patients (55%); 74 of these (30%) had LN metastasis. LN metastasis was associated
with worse outcome (median survival: N0, 30 months v N1, 24 months; P ⫽ .03). Although
patients with no LN metastasis were able to be stratified by tumor number and vascular
invasion (N0; P ⬍ .001), among patients with N1 disease, multiple tumors and vascular
invasion, either alone or together, failed to discriminate patients into discrete prognostic
groups (P ⫽ .34).
Conclusion
Although tumor size provides no prognostic information, tumor number, vascular invasion, and
LN metastasis were associated with survival. N1 status adversely affected overall survival and
also influenced the relative effect of tumor number and vascular invasion on prognosis.
Lymphadenectomy should be strongly considered for ICC, because up to 30% of patients will
have LN metastasis.
J Clin Oncol 29:3140-3145. © 2011 by American Society of Clinical Oncology
INTRODUCTION
Cholangiocarcinoma can be anatomically classified
into intrahepatic (ICC), hilar (Klatskin tumors),
and distal bile duct types according to their location
in the biliary tree.
1
Unlike extrahepatic bile duct
cancers, ICC occurs within the hepatic parenchyma,
where it frequently presents as a mass lesion in the
absence of jaundice or other constitutional symp-
toms.
2
ICC is the second most common primary
liver malignancy after hepatocellular carcinoma
(HCC).
3
Although ICC was historically considered
the least common of the bile duct cancers, incidence
of ICC has been increasing.
4,5
No distinction was
made in the sixth edition of the American Joint
Committee on Cancer (AJCC)/International Union
Against Cancer (UICC) staging manual between
ICC and HCC, in part because of the relative
rarity of the disease.
6
The sixth edition AJCC/
UICC staging system used tumor size, tumor
number, and presence of vascular invasion as ma-
jor prognostic criteria to establish the T-category
JOURNAL OF CLINICAL ONCOLOGY
ORIGINAL REPORT
VOLUME 29 䡠 NUMBER 23 䡠 AUGUST 10 2011
3140 © 2011 by American Society of Clinical Oncology
Universität Zürich) on October 20, 2011 from 129.195.0.205
Information downloaded from jco.ascopubs.org and provided by at SWISS CONSORTIUM (Hauptbibliothek
Copyright © 2011 American Society of Clinical Oncology. All rights reserved.
subgroups. However, combining ICC and HCC into a single stag-
ing system may be problematic, because ICC and HCC have dis-
tinct mechanisms of carcinogenesis, underlying risk factors, and
biologic and clinical behaviors.
7,8
Recently, our group used the Surveillance, Epidemiology, and
End Results (SEER) data set to assess the predictive accuracy of several
ICC staging systems, including the sixth edition of the AJCC/UICC
staging manual.
9
We reported that the sixth edition AJCC/UICC
T-category subgroups failed to accurately stratify patients with ICC.
Instead, we proposed a simplified T-category system more able to
risk-stratify patients into discrete prognostic groups. In part on the
basis of our work, the seventh edition of the AJCC/UICC staging
manual has incorporated a new distinct staging system for ICC based
on prognostic factors including tumor number and vascular invasion
but not tumor size.
10
Like other staging systemsfor solid malignancies,
the seventh edition AJCC/UICC staging system also requires ascer-
tainment of nodal status. However, the role of lymphadenectomy for
ICC remains controversial, with many surgeons not performing
lymph node (LN) evaluation.
11
Given that the seventh edition ICC
staging system was only recently published, data to support its wide-
spread adoption and use are lacking. Furthermore, data on the role of
lymphadenectomy for ICC and its impact on prognosis are also inad-
equate. As such, the objective of the current study was to identify
factors associated with outcome after surgical management of ICC.
More specifically, we sought to validate the new seventh edition AJCC/
UICC ICC T-category scheme as well as examine the impact of LN
assessment and nodal status on survival.
PATIENTS AND METHODS
Using an international multi-institutional database, 449 patients with ICC
who underwent surgical resection with curative intent between October 1973
and February 2010 at one of 11 institutions (Johns Hopkins School of Medi-
cine, Baltimore, MD; Duke Medical Center, Durham, NC; University of Pitts-
burgh School of Medicine, Pittsburgh, PA; Massachusetts General Hospital,
Boston, MA; University of Virginia, Charlottesville, VA; University Hospital
Essen, Essen, Germany; Fundeni Clinical Institute of Digestive Disease, Bucha-
rest, Romania; Hoˆpitaux Universitaires de Gene`ve, Geneva, Switzerland; Os-
pedale San Raffaele, Milan, Italy; Cliniques Universitaires Saint Luc, Brussels,
Belgium; and Curry Cabral Hospital, Lisbon, Portugal) were identified. The
institutional review board of each respective institution approved this study.
Only patients with histologically confirmed ICC who received their initial
treatment for ICC at a study center were included.
Data Collection
Standard demographic and clinicopathologic data were collected, in-
cluding sex, age, and primary tumor characteristics. Specifically, data were
collected on primary tumor location, size, and number as well as morphologic
subtype and presence of vascular invasion, defined as minor and/or major.
Data on treatment-related variables, such as type of surgery, receipt of
lymphadenectomy, and adjuvant therapy, were also obtained. Resection
was classified as less than hemihepatectomy, hemihepatectomy, or ex-
tended hepatectomy.
12
Margin and nodal status were ascertained based on
final pathologic assessment. Date of last follow-up and vital status were
collected on all patients.
Statistical Analyses
Summary statistics were obtained using established methods and pre-
sented as percentages, mean, or median values. Overall survival time was
calculated from date of surgery to date of last follow-up. Cumulative event
rates were calculated using the Kaplan-Meier method.
13
Univariate analyses
were performed using the
2
or log-rank test to compare differences between
categorical groups and the Mann-Whitney U test for continuous variables.
Cox proportional hazards models
14
were developed using relevant clinico-
pathologic variables to determine the association of each with overall survival.
Relative risks were expressed as hazard ratios (HRs) with 95% CIs. Significance
levels were set at P ⬍ .05; all tests were two sided. All statistical analyses were
performed using SPSS version 17.0 (Chicago, IL).
RESULTS
Patient and Primary Tumor Characteristics
Table 1 lists the clinicopathologic features of the 449 patients with
pathologically confirmed ICC who were included in the study. A
majority of patients presented with a solitary tumor (n ⫽ 329; 73.3%),
and the median size of the largest lesion was 6.5 cm (range, 7 to 25.0
cm). Only a minority of patients were treated with adjuvant systemic
chemotherapy (n ⫽ 125; 27.8%) or radiation therapy (n ⫽ 31; 6.9%)
after surgical resection.
Table 1. Patient Demographics and Clinicopathologic Characteristics
Variable
Patients (N ⫽ 449)
No. %
Age, years
Median 61
Range 23-85
Sex
Male 209 46.5
Race
White 412 91.9
Bilateral involvement 135 30.1
Tumor size, cm
Median 6.5
Range 0.7-25.0
Presence of multiple tumors 120 26.7
Concomitant extrahepatic disease 22 4.9
Type of liver resection
⬍ Hemihepatectomy 110 24.4
Hemihepatectomy 189 42.1
Extended hemihepatectomy 139 31.0
Central hepatectomy 8 1.8
Uknown 3 0.7
Lymphadenectomy performed 248 55.2
No. of lymph nodes harvested
Median 3
Range 1-76
Resection margin
R0 364 81.1
R1 70 15.6
R2 12 2.7
Unknown 3 0.6
Lymph node disease 74 16.5
No. of lymph node metastases
Median 1
Range 1-25
Invasion
Vascular 140 31.2
Perineural 52 11.6
Biliary 57 12.7
Treatment for Intrahepatic Cholangiocarcinoma
www.jco.org
© 2011 by American Society of Clinical Oncology 3141
Universität Zürich) on October 20, 2011 from 129.195.0.205
Information downloaded from jco.ascopubs.org and provided by at SWISS CONSORTIUM (Hauptbibliothek
Copyright © 2011 American Society of Clinical Oncology. All rights reserved.
At the time of surgical resection, the extent of hepatic resection
was less than a hemihepatectomy (n ⫽ 110; 24.5%), central hepatec-
tomy (n ⫽ 8; 1.8%), hemihepatectomy (n ⫽ 189; 42.1%), and ex-
tended hepatectomy (n ⫽ 139; 31.0%). On final pathologic analysis,
12 patients (2.7%) had a macroscopically positive margin (R2); mar-
gin status was microscopically positive (R1) in 70 patients (15.6%) and
microscopically negative (R0) in 364 patients (81.1%). On final patho-
logic analysis, 140 patients (31.2%) had vascular invasion, whereas 57
patients (12.7%) had biliary invasion.
Overall Survival: Prognostic Factors and Assessment
of Seventh Edition AJCC/UICC T Category
Median overall survival after surgical resection of ICC was 27.3
months. One-, 3-, and 5-year overall survival was 77.5%, 44.3%, and
30.7%, respectively. On univariate analysis, factors influencing sur-
vival included tumor number (HR, 1.82; 95% CI, 1.40 to 2.39;
P ⬍ .001) and presence of vascular invasion (HR, 1.69; 95% CI, 1.28 to
3.53; P ⬍ .001). Median survival for patients with solitary ICC was 36.0
months compared with 19.0 months for patients with multiple ICC
lesions (P ⬍ .001). Similarly, presence of vascular invasion was asso-
ciated with worse survival; patients who had no vascular invasion had
a median survival of 41.0 months versus 20.0 months for those pa-
tients with vascular invasion (P ⬍ .001). Tumor size was also associ-
ated with survival on univariate analysis (HR, 1.05; 95% CI, 1.02 to
1.08; P ⫽ .019). Other factors, such as presence of biliary invasion or
direct invasion of adjacent organs, were not associated with survival
(P ⬎ .05 for both). On multivariate analysis, tumor number and
presence of vascular invasion remained associated with poor outcome
(multiple tumors: HR, 1.80; 95% CI, 1.28 to 2.52; P ⫽ .001; presence of
vascular invasion: HR, 1.59; 95% CI, 1.10 to 2.32; P ⫽ .015). In
contrast, tumor size had no impact on survival after surgical resection
of ICC (HR, 1.03; 95% CI, 0.98 to 1.07; P ⫽ 0.23; Table 2).
We then examined the new T categories of the seventh edition
AJCC/UICC staging system relative to overall survival and prognosis.
Patients were unevenly distributed across T categories, with most
patients classified as T1 (n ⫽ 140; 31.2%), T2a (n ⫽ 64; 14.3%), or T2b
(n ⫽ 102; 22.7%). A small number of patients in the study cohort were
classified as T3 or T4, and therefore, these subgroups were not further
analyzed. Analysis of the AJCC/UICC T1, T2a, and T2b subgroups
stratified patients with regard to prognosis. Patients with T1 tumors
(ie, those with a solitary tumor plus no vascular invasion) had a 5-year
survival of 46.7% versus 25.0% and 12.0% for patients with T2a (ie,
those patients with solitary tumor plus vascular invasion) and T2b
tumors (ie, those patients with multiple tumors with or without vas-
cular invasion), respectively (Fig 1).
Lymphadenectomy: Incidence of LN Metastasis and
Impact of Nodal Status
Of the 449 patients who underwent surgical resection for ICC,
248 (55.2%) had a lymphadenectomy performed. In contrast, 201
patients (44.8%) did not have the locoregional LN basin evaluated
(Nx) at time of surgery. Of those patients who underwent LN evalua-
tion, the median number of LNs harvested was three (range, one to
76). Of the 248 patients who underwent a lymphadenectomy, 74 had
metastatic nodal disease. Therefore, among patients who had LN
evaluation, the incidence of N1 disease was 29.8%.
Both vascular (HR, 2.89; 95% CI, 1.56 to 5.35; P ⫽ .001) and
biliary invasion (HR, 4.03; 95% CI, 1.94 to 8.36; P ⬍ .001) were
strongly associated with increased risk of N1 disease. Of note, how-
ever, was the finding that incidence of LN metastasis was still 9.1% and
20.7%, respectively, among patients with no vascular or biliary
invasion. Other factors, such as tumor number and size as well as
Table 2. Factors Associated With Overall Survival Stratified by Nodal Status
Prognostic Factor
All Patients Irrespective of N Status
(N ⫽ 449) Patients With N0 Status (n ⫽ 165) Patients With N1 Status (n ⫽ 63)
HR 95% CI P HR 95% CI P HR 95% CI P
Tumor size 1.03 0.98 to 1.07 .23 1.00 0.94 to 1.07 .97 1.03 0.90 to 1.19 .66
Multiple tumors 1.80 1.28 to 2.52 .001 1.53 1.18 to 7.65 .021 3.01 0.94 to 2.47 .09
Positive resection margin 2.20 1.52 to 3.17 ⬍ .001 2.53 1.53 to 4.18 ⬍ .001 1.15 0.26 to 4.98 .85
Vascular invasion 1.59 1.10 to 2.32 .015 2.11 1.30 to 3.42 .003 1.22 0.36 to 4.17 .75
Direct invasion of adjacent organs 1.13 0.65 to 1.96 .15 0.69 0.31 to 1.52 .36 3.31 0.97 to 11.24 .055
Biliary invasion 0.70 0.42 to 1.15 .15 0.81 0.36 to 1.85 .62 0.98 0.33 to 2.95 .97
Abbreviation: HR, hazard ratio.
0
P < .001
Proportion Surviving
Time (months)
1.0
0.8
T1, solitary tumor without vascular invasion
T2a, solitary tumor with vascular invasion
T2b, multiple tumors ± vascular invasion
0.6
0.4
0.2
12 24 36 48 60
Fig 1. Overall survival stratified by T1 (patients with solitary tumor without
vascular invasion), T2a (patients with solitary tumor with vascular invasion),
and T2b (patients with multiple tumors with or without vascular invasion),
seventh edition American Joint Committee on Cancer/International Union
Against Cancer categories.
de Jong et al
3142
© 2011 by American Society of Clinical Oncology
J
OURNAL OF CLINICAL ONCOLOGY
Universität Zürich) on October 20, 2011 from 129.195.0.205
Information downloaded from jco.ascopubs.org and provided by at SWISS CONSORTIUM (Hauptbibliothek
Copyright © 2011 American Society of Clinical Oncology. All rights reserved.
mass-forming morphology and direct invasion of adjacent organs
were not associated with increased risk of LN metastasis (P ⬎ .05
for all; Table 3).
The finding of N1 disease affected overall survival, with N0 pa-
tients having a median survival of 30.1 months versus 22.9 months for
patients with N1 disease (P ⫽ .03). Using the three independent
variables associated with outcome—tumor number, presence of vas-
cular invasion, and N1 disease—patients were stratified with regard to
prognosis. Patients lacking all three risk factors had longer median
survival (46.9 months) compared with patients with either one factor
(29.5 months) or two or three risk factors (20.3 months; P ⫽ .002).
Five-year survival for patients with none, one, or two to three risk
factors was 38.3%, 27.3%, and 18.1%, respectively (P ⬍ .001). When
patients were then stratified according to nodal status, tumor number
and presence of vascular invasion were able to stratify patients with no
LN metastasis (N0) with regard to prognosis (P ⬍ .001; Fig 2A). In
contrast, among patients with N1 disease, presence of multiple tumors
or vascular invasion either alone or together failed to discriminate
patients into discrete prognostic groups (P ⫽ .34; Fig 2B). Specifically,
when we examined the impact of LN status on tumor number, tumor
number was only a predictor of survival among patients with N0
disease (P ⫽ .004; Fig 3A). In contrast, patients with N1 disease had the
same overall survival whether they had multiple tumors or a solitary
lesion (P ⫽ .45). The same effect was seen with vascular invasion.
Although vascular invasion was a predictor of outcome among pa-
tients with N0 disease (P ⫽ .009), it failed to act as a prognostic marker
among patients with LN metastasis (P ⫽ .30; Fig 3B).
DISCUSSION
Although its incidence has been increasing over the last three decades,
ICC has historically been a relatively uncommon disease with a poor
prognosis.
4
ICC now accounts for 5% to 30% of all primary liver
malignancies, and some reports have noted an improved trend in the
prognosis of patients with ICC who undergo surgical resection.
4,15
In
part, as a consequence, there has been increasing clinical interest in
ICC as well as a greater focus on research surrounding this dis-
ease.
16
In the sixth edition of the AJCC/UICC staging manual, ICC
was staged identically to HCC. In the newly released seventh edition of
the AJCC/UICC staging manual, ICC now has a separate, unique
staging system,
10
guided in part by data derived from the SEER data set
published by our group.
9
Unfortunately, there remains a paucity of
prognostic data for ICC, with most data derived from small single-
institution studies or administrative data sets.
9,17,18
The current study
is important, because it reports one of the largest multi-institutional
experiences on the surgical management of ICC. We report that over-
all survival after surgical resection of ICC was approximately 30% to
35%. Although tumor number and vascular invasion both signifi-
cantly affected prognosis, tumor size was not a relevant prognostic
factor, consistent with the newly proposed seventh edition AJCC/
UICC T-category schema. In addition, we report that incidence of
LN metastasis associated with surgically resected ICC was 20% to
30%. N1 status not only adversely affected overall survival but also
influenced the relative effect of tumor number and vascular inva-
sion on prognosis.
Our previous ICC study that used the SEER data set was note-
worthy, because it was the first, to our knowledge, that was aimed at
Table 3. Factors Associated With Increased Risk of Lymph Node
Metastasis (n ⫽ 248)
ⴱ
Prognostic Factor OR 95% CI P
Size of largest lesion (continuous) 0.99 0.92 to 1.07 .80
Multiple tumors 1.56 0.81 to 3.02 .19
Vascular invasion 2.89 1.56 to 5.35 .001
Direct invasion of adjacent organ 1.74 0.68 to 4.47 .25
Perineural invasion 1.87 0.78 to 4.49 .16
Billiary invasion 4.03 1.94 to 8.36 ⬍ .001
Morphologic subtype
Mass forming Reference
Papillary 0.65 0.15 to 2.74 .55
Periductal infiltrating 0.19 0.02 to 1.56 .12
Mass forming plus periductal infiltrating 0.15 0.65 to 2.74 .55
Abbreviation: OR, odds ratio.
ⴱ
Univariate analysis.
A
B
0
P < .001
Proportion Surviving
Time (months)
N0
N1
1.0
0.8
0
1
2
0.6
0.4
0.2
12 24 36 48 60
0
P = .34
Proportion Surviving
Time (months)
1.0
0.8
0
1
2
0.6
0.4
0.2
12 24 36 48 60
Fig 2. Impact of tumor number and presence of vascular invasion stratified by
nodal status. Although tumor number and presence of vascular invasion were
able to stratify patients with (A) no lymph node metastasis (N0), these factors
either alone or together failed to discriminate (B) N1 patients into discrete
prognostic groups.
Treatment for Intrahepatic Cholangiocarcinoma
www.jco.org
© 2011 by American Society of Clinical Oncology 3143
Universität Zürich) on October 20, 2011 from 129.195.0.205
Information downloaded from jco.ascopubs.org and provided by at SWISS CONSORTIUM (Hauptbibliothek
Copyright © 2011 American Society of Clinical Oncology. All rights reserved.
