18

F-fluorodeoxyglucose Positron
Emission Tomography Influences Management
Decisions in Patients with Biliary Cancer
Carlos U Corvera, MD, FACS, Leslie H Blumgart, MD, FACS, Timothy Akhurst, MD,
Ronald P DeMatteo, MD, FACS, Michael D’Angelica, MD, FACS, Yuman Fong, MD, FACS,
William Robert Jarnagin, MD, FACS

BACKGROUND: Although 18F-fluorodeoxyglucose positron emission tomography (PET) has widespread clinical
use, its role in cancers of the biliary tract is ill-defined. The aim of this study was to determine
if preoperative PET provided additional staging information in patients with biliary tract
cancer, beyond that obtained through conventional anatomic imaging. The role of PET in
detecting disease recurrence after resection was also examined.
STUDY DESIGN: Between March 2001 and October 2003, 126 patients with biopsy-proved or presumed biliary
tract cancer (intrahepatic or extrahepatic cholangiocarcinoma and gallbladder carcinoma) un-
derwent PET in addition to standard imaging evaluation. Histologic confirmation of the
diagnosis was used as the reference standard with which PET results were compared. Patient
followup information and serial imaging were reviewed for progression of lesions detected by
PET.
RESULTS: Of the 126 study patients, 93 (74%) underwent preoperative staging PET scans, the results of
which changed the stage and treatment in 22 patients (24%): 15 of 62 (24%) with cholangio-
carcinoma and 7 of 31 (23%) with gallbladder carcinoma. When used to assess for cancer
recurrence (n ⫽ 33), PET identified disease in 86% of patients but altered treatment in only
9%. So, of the entire study group, the findings of PET influenced management in 20% of
patients (24% preoperative staging and 9% cancer recurrence). The sensitivity of PET for
identifying the primary tumor was 80% overall: 78% for cholangiocarcinoma, 86% for gall-
bladder carcinoma.
CONCLUSIONS: Most biliary tract cancers are 18F-fluorodeoxyglucose avid tumors. In patients with potentially
resectable tumors based on conventional imaging, PET identified occult metastatic disease and
changed management in nearly one-fourth of all patients. PET also helped confirm recurrent
cancer after resection. (J Am Coll Surg 2008;206:57–65. © 2008 by the American College of
Surgeons)

Cancers of the biliary tract can arise anywhere along the
biliary epithelium. They most commonly involve the gall-
bladder (GBCA) or biliary confluence (hilar cholangiocar-
cinoma [CCA]) and, less commonly, the distal or intrahe-
patic bile ducts. Biliary tract cancers are different clinically
and biologically and have different patterns of spread and
prognoses.1 Nevertheless, most patients with biliary tract
Competing Interests Declared: None.
Received March 9, 2007; Revised July 2, 2007; Accepted July 2, 2007.
From the Departments of Surgery (Corvera, Blumgart, DeMatteo,
D’Angelica, Fong, Jarnagin) and Nuclear Medicine (Akhurst), Memorial
Sloan-Kettering Cancer Center, New York, NY.
Correspondence address: William R Jarnagin, MD, Department of Surgery,
Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY
10021.
cancer present with advanced disease, for which palliative
measures are most appropriate. But for patients with po-
tentially resectable tumors, an aggressive surgical approach
is warranted, because resection remains the most effective
therapy. In addition, the progressive reduction in mortality
associated with surgical resection has increasingly made it a
viable treatment option, when technically feasible.2-4
Preoperative imaging is clearly a critical component of
the initial assessment. Imaging studies are used to assess
local tumor-related factors that are critical for determining
resectability and, optimally, should also identify distant
metastatic disease, which generally contraindicates resec-
tion. The presence of regional nodal disease appears to have
an adverse impact on outcomes after resection, although
whether this finding should preclude a resection remains

© 2008 by the American College of Surgeons ISSN 1072-7515/08/$34.00

Published by Elsevier Inc. 57 doi:10.1016/j.jamcollsurg.2007.07.002
58 Corvera et al Positron Emission Tomography in Biliary Cancer
Abbreviations and Acronyms
CCA ⫽ cholangiocarcinoma
18
F-FDG ⫽ 18
F-fluorodeoxyglucose
GBCA ⫽ gallbladder carcinoma
MRCP ⫽ magnetic resonance cholangiopancreatography
MSKCC ⫽ Memorial Sloan-Kettering Cancer Center
PET ⫽ positron emission tomography
SUV ⫽ standardized uptake value
US ⫽ ultrasonography
controversial.5-8 Nevertheless, preoperative awareness of
such disease would be helpful for preoperative decision
making, because it may alter the risk/benefit ratio of resec-
tion for some, if not many, patients. Although general
progress in noninvasive abdominal imaging continues to
evolve rapidly, many patients with CCA or GBCA are
found to have unresectable disease only at the time of ex-
ploratory surgery. In fact, despite extensive preoperative
imaging evaluation, a potentially curative resection is pos-
sible in only 50% to 75% of these patients.9 Currently, the
best way to assess resectability is with a combination of
complementary imaging studies, including ultrasonogra-
phy, cross-sectional imaging with either CT or MRI, and
cholangiography. Staging laparoscopy is also helpful but
is invasive, requires general anesthesia, and is clearly
imperfect.
The use of positron emission tomography (PET), a
functional study for preoperative assessment of cancer pa-
tients, is rapidly evolving and may be useful for patients
with biliary tract cancer. PET, which relies on the metabolic
profiles of tumors, can potentially detect metastatic disease
that is unsuspected or equivocal on cross-sectional imaging
studies, changing the stage of disease and the planned ther-
apy. To date, PET has proved useful for the preoperative
staging of a variety of cancers10,11 and is approved for use as
a staging modality in some of these diseases. But the role of
PET in patients with CCA and GBCA remains unclear.
Early studies indicate that PET may be useful in preventing
unnecessary exploration by detecting otherwise unsus-
pected metastatic disease.12,13 The primary aim of this
study was to determine whether preoperative PET pro-
vided additional staging information, beyond that ob-
tained through conventional anatomic imaging, in pa-
tients with biliary tract cancer. We also examined the
role of PET in detecting disease recurrence after curative
resection.
METHODS
General
Patients with a clinical diagnosis of CCA (extrahepatic [hi-
lar or distal] or intrahepatic) or GBCA who underwent
J Am Coll Surg
PET imaging evaluation, either pre- or postoperatively,
were identified from a prospectively maintained database
andanalyzedretrospectively.Dataforanalysisincludeddem-
ographics, clinical history, histopathology, imaging studies
performed and results, operative findings, and followup
information.
Patients
Patients were typically referred after undergoing prelimi-
nary investigations at other institutions. After referral, ad-
ditional cross-sectional imaging evaluation was conducted
at Memorial Sloan-Kettering Cancer Center (MSKCC) in
the form of CT, MRI, or both. Magnetic resonance cholan-
giopancreatography (MRCP) and duplex ultrasonography
were used extensively, primarily to assess biliary and local
tumor extent in patients with hilar cholangiocarcinoma;
ultrasonography was often used to assess vascular involvement
as an adjunct to other studies. Our general approach to pre-
operative evaluation and criteria of resectability and resection
has been previously described.2,14 In this report, data from the
imaging studies were analyzed in light of these resectability
criteria, with particular attention on metastatic disease to dis-
tant sites or regional lymph node basins that might have
changed the treatment recommendation. Metastatic disease to
distant sites typically constituted irresectability; involvement
of regional lymph node basins did not necessarily contraindi-
cate an attempt at resection. But the decision to proceed in the
face of such findings was made on a case-by-case basis. Patients
with incidental GBCA identified after laparoscopic cholecys-
tectomy, who represent the majority patients with potentially
resectable disease, were included in this study. Patients and
imaging studies were reviewed at a biweekly multidisciplinary
disease management conference attended by surgeons, radiol-
ogists, medical oncologists, and gastroenterologists.
A subset of patients in this study underwent PET to
assess cancer recurrence after resection. Such patients are
typically followed clinically and with periodic cross-
sectional imaging studies (CT or MRI) at 6-month inter-
vals, although evaluations are done any time there is a
change in clinical status. Patients with documented recur-
rence are typically referred for systemic therapy, if appro-
priate, and selected patients with localized disease may be
candidates for chemoradiation therapy.
18
Technique of F-fluorodeoxyglucose PET
PET scans at MSKCC were performed on dedicated con-
ventional, full-ring, high-resolution scanners (GE Ad-
vanced, GEMS). Patients were injected with 10 to 15 mCi
of 18F-FDG after fasting for at least 6 hours. All images
were interactively reconstructed using postemission trans-
mission attenuation corrected datasets. Region of interest
analysis tools were used to calculate the maximal FDG
Vol. 206, No. 1, January 2008 Corvera et al
concentration within the primary tumor mass. Standard-
ized uptake values (SUV max) were obtained by correcting
for the injected dose and the patient’s weight; results were
dichotomized as either positive or negative for detection of
the primary tumor, metastatic disease, or both. The SUV
cut-off for an abnormal PET reading was a value greater
than two. For patients submitted to PET scans at other
centers, data were obtained from the study report; the same
SUV cut-off value for abnormal FDG uptake was used. For
the purposes of this study, only uptake in the primary tu-
mor site and possible metastatic sites were analyzed. In
addition, preoperative PET staging refers to a diagnostic
study performed before exploration for a potentially cura-
tive resection at MSKCC, so this definition includes pa-
tients who had previously undergone operations before
referral.
Data analysis and statistics
Decisions about potential resectability were made based on
data from cross-sectional imaging studies and included
analysis of local tumor-related factors in addition to assess-
ment for the presence of metastatic disease.2,14 The decision
to proceed with PET in this patient subgroup reflected a
general change in practice, such that the large majority of
patients with potentially resectable tumors were included
in this study. Although PET was obtained in some patients
for further evaluation of equivocal findings on cross-
sectional imaging studies, the majority of patients were not
necessarily selected for the procedure based on any partic-
ular imaging or clinical findings. Patients with evidence of
locally advanced, unresectable tumors were included in this
analysis to assess the ability of PET to correctly identify the
primary tumor and occult metastatic disease, the presence
of which would necessarily have important treatment im-
plications. For example, chemoradiation therapy may be
used for locally advanced hilar cholangiocarcinoma and,
less commonly, for gallbladder cancer but would be con-
traindicated in patients with distant metastatic disease.
Likewise, patients with unresectable intrahepatic cholan-
giocarcinoma confined to the liver might be eligible for
regional chemotherapy on protocol or hepatic artery em-
bolization but, again, not with evidence of distant spread
outside the liver. Additionally, the few patients ultimately
proved to have benign disease were included in the final
analyses for PET sensitivity and specificity. PET scans per-
formed to evaluate for possible disease recurrence were ob-
tained based on imaging findings that were considered in-
conclusive in this regard or, in some patients, because of a
change in clinical status without suggestive radiographic
changes.
In the majority of patients, histologic confirmation of
either CCA or GBCA was used as the reference standard
Positron Emission Tomography in Biliary Cancer 59
with which PET results were compared. In a small number
of patients, primarily those undergoing evaluation for met-
astatic disease, documented cross-sectional imaging pro-
gression of PET-positive lesions was accepted as confirma-
tory. The clinical utility of PET scanning for patients with
biliary tract cancer was determined by how often it pro-
vided diagnostic information beyond that obtained with
conventional anatomic imaging and changed the treatment
plan. This assessment was made after review of all clinical
information and included discussion with the attending
surgeon in many cases. The sensitivity and specificity of
PET for detecting the primary tumor and metastatic le-
sions were calculated. In patients with GBCA, residual dis-
ease in the gallbladder fossa after a noncurative cholecys-
tectomy was considered primary tumor, and these patients
were included in the primary tumor detection analysis. Dif-
ferences between groups were analyzed by chi-square; p values
less than 0.05 were considered statistically significant. Analy-
ses were performed using version 4.0 of the JMP statistical
package from SAS Institute. Numerical data are expressed as
mean ⫾ standard deviation (SD) unless otherwise noted. This
study was approved by the Institutional Review Board at
MSKCC and is compliant with the Health Insurance Porta-
bility and Accountability Act (HIPAA).
RESULTS
General
Between March 2001 and October 2003, we identified 126
patients with presumed biliary tract malignancies who ful-
filled the study criteria as outlined previously. More than
twice as many patients had a presumptive diagnosis of cho-
langiocarcinoma as had gallbladder carcinoma (n ⫽ 85,
67% versus n ⫽ 41, 33%). Cholecystectomy or attempted
resection was performed in 23 GBCA patients before refer-
ral. Four patients were ultimately proved to have benign
disease (three CCA and one GBCA). The gender ratio was
nearly equal (48% women), and the median age was 62
years (range 23 to 84 years). PET scans were done before
referral in 21 (16%) patients, with the remaining scans
having been performed at MSKCC.
PET scans were obtained as part of preoperative staging
in 93 patients (74%) and to evaluate suspected recurrence
after curative resection in 33 patients (26%, Fig. 1). Pa-
tients in the preoperative staging group underwent a median
of 3 additional imaging studies to assess extent of disease (CT,
100%; MRI/MRCP, 65%; ultrasonography, 89%; direct
cholangiography, 64%). The percentage of patients who un-
derwent PET for preoperative staging was approximately the
same for CCA as for GBCA (73% versus 75%).
In the entire cohort, PET identified additional disease
not seen on cross-sectional imaging and changed manage-
60 Corvera et al Positron Emission Tomography in Biliary Cancer
Figure 1. Flow diagram showing the breakdown of patients staged with 18FDG-PET. The number and proportion of
patients whose management was altered by the PET scan are indicted by the asterisks. 18FDG-PET,
18
Fluorodeoxygluose-positron emission tomography.
ment in 20% of patients. The yield of PET was greatest
when used for preoperative staging compared with evalua-
tion for recurrence (24% versus 9%, Fig. 1).
Effectiveness of preoperative PET staging for CCA
and GBCA
Overall, preoperative PET found unsuspected additional
disease and changed management in 22 (24%) patients
(Fig. 1). Among patients with CCA (extrahepatic [hilar
and distal], n ⫽ 41; intrahepatic, n ⫽ 21), the PET find-
ings changed management in 24% (15 of 62, Table 1) by
demonstrating disease that either avoided operation in
patients initially thought to have resectable tumors
(n ⫽ 12) or avoided use of locoregional therapies in pa-
tients initially thought to have locally advanced tumors
(n ⫽ 3). In eight additional patients, PET results sug-
gested the possibility of metastatic disease to nodal or
distant sites that was ultimately confirmed at explora-
tion. PET showed no evidence of metastatic disease in
39 patients, 13 of whom were found to have unresect-
able, locally advanced tumors at exploration; 1 of these
13 patients also had unsuspected regional nodal disease.
So, in patients with CCA, PET provided conclusive
(n ⫽ 15) or suggestive (n ⫽ 8) evidence of additional
disease, not seen on standard imaging, in 23 patients
(37%). Sites of metastatic disease identified by PET in-
cluded regional lymph nodes in 10 patients (16%, Fig.
2) and distant sites in 18 (29%); 6 patients had both
findings. PET also identified unsuspected second malig-
nancies in two patients (non-small cell lung cancer and
renal cell carcinoma).
J Am Coll Surg
The sensitivity of PET was 78% for detecting the pri-
mary tumor and 96% for detecting metastatic disease;
specificities were 75% and 89%, respectively. But the over-
all sensitivity of PET for detecting intrahepatic CCA was
95% but only 69% for extrahepatic CCA (Table 2).
Among patients with GBCA (n ⫽ 31), PET changed
management in 7 (23%) by showing metastatic disease not
seen on other imaging studies, preventing an unnecessary
exploration. In another 7 patients, suggestive evidence of
metastatic disease was noted and ultimately proved at op-
eration (Table 1). In 17 patients, PET showed disease con-
fined to the primary tumor, supporting the decision to
proceed with surgical exploration, but 5 patients had un-
resectable, locally advanced tumors. Two of the five unre-
sectable patients were found to have small-volume perito-
neal disease not detected by PET. Overall, in patients with
GBCA, PET provided conclusive (n ⫽ 7) or suggestive ev-
idence (n ⫽ 7) of additional disease in 14 of 31 patients
(45%). The sites of metastatic disease identified by PET
included regional disease (local lymph nodes) in 12 pa-
tients (39%) and distant sites in 11 (35%) patients (peri-
toneal or port site implants [3]; intrahepatic [2]; other,
bone [3], lung [2], mediastinum [1]); and 7 patients (23%)
had both. The sensitivity of PET for GBCA was 86% for
detecting the primary tumor (gallbladders in situ) or de-
tecting residual disease at the operative bed (in postchole-
cystectomy patients) and was 87% for detecting metastatic
GBCA (Table 2). In 3 of the 11 patients with distant dis-
ease, the abnormal FDG signal was located within the ab-
domen and thought to be related to peritoneal metastases
or laparoscopic port site recurrence (Fig. 3).
Vol. 206, No. 1, January 2008 Corvera et al Positron Emission Tomography in Biliary Cancer 61

Table 1. Utility of Preoperative Positron Emission Tomography for Cholangiocarcinoma and Gallbladder Carcinoma
PET changed
Disease n Findings on PET Outcomes management, n (%)
CCA 62 15 (24)
12 Metastatic disease identified Avoided operation
9 Isolated disease at primary tumor site supporting exploration Resected
2 Normal PET, supporting exploration; locally advanced tumor at Explored, not resected
exploration*
10 Isolated disease at primary tumor site, supporting exploration; Explored, not resected
locally advanced tumor at exploration
3 Isolated disease at primary tumor site but locally unresectable on Explored, HAIP
standard imaging, supporting HAIP placement placed
2 Normal PET, supporting exploration; benign histology Resected
1 Isolated disease at primary tumor site, supporting exploration; Resected
benign histology
1 Isolated disease at primary tumor site and second primary tumor Resected
identified
1 Isolated disease at primary tumor site and second primary tumor Explored, not resected
identified; locally advanced tumor at exploration
13 Normal PET, supporting exploration; malignant histology Resected
8 Metastatic disease suggested Explored, not resected
GBCA 31 7 (23)
7 Metastatic disease identified Avoided operation
6 Isolated disease at primary tumor site, supporting exploration Resected
5 Isolated disease at primary tumor site, supporting exploration; Explored, not resected
locally advanced tumor at exploration†
4 Normal PET, supporting exploration; malignant histology Resected
1 Normal PET, supporting exploration; no residual cancer Resected
1 Normal PET, patient refused surgery
7 Metastatic disease suggested Explored, not resected
Total 93 22 (24)
*One of these two patients also had unsuspected regional nodal disease.
†
Two of these five patients also had small-volume peritoneal disease.
CCA, cholangiocarcinoma; GBCA, gallbladder carcinoma; HAIP, hepatic arterial infusion pump; PET, positron emission tomography.

Of the 23 GBCA patients submitted to operation before
referral cholecystectomy (n ⫽ 17) or exploration (n ⫽ 6),
PET changed planned management in 3 patients (13%)
compared with 4 of 8 patients (50%) undergoing evalua-
tion for primary resection (p ⬍ 0.05). Among the 17 pa-
tients who had an earlier cholecystectomy, PET correctly
predicted persistent disease in the gallbladder fossa in 14
(82%) and metastatic disease in 5 (29%).
Effectiveness of PET for detecting
cancer recurrence
PET identified recurrent cancer in 25 (76%) of the 33
patients with biliary tract cancer who were evaluated for
suspected recurrence after resection (Fig. 1). In two pa-
tients with CCA and one patient with GBCA, PET
changed management by demonstrating recurrent disease
not seen on other studies. But in the remaining 22 patients,
recurrence was identified on cross-sectional imaging, and
the PET studies were merely confirmatory. For detecting
recurrent CCA and GBCA, the sensitivity and specificity of
PET were 89% and 100%, respectively.
DISCUSSION
Although significant advances in the surgical treatment of
biliary tract cancer have been made over the past several
years, it remains a difficult group of diseases to treat. The
critical anatomic location and biologic aggressiveness of
this cancer dictate that longterm survival is generally pos-
sible only in patients who undergo complete resection of
early-stage tumors. Unfortunately, most patients present
late in the course of disease when resection is no longer
possible, with perhaps the only exception being early
GBCA in patients diagnosed incidentally after elective cho-
lecystectomy for symptomatic gallstones. The clinical di-
lemma remains that a great proportion of patients with
biliary cancer undergo exploratory surgery, only to be
62 Corvera et al Positron Emission Tomography in Biliary Cancer J Am Coll Surg

Figure 2. Positron emission tomography scans from a patient with hilar cholangiocarcinoma. (A) The
primary tumor (arrowhead) and (B) regional nodal metastases (arrow) are indicated.

found to have unresectable locally advanced disease or un-
suspected metastatic disease. PET holds the promise of
significantly reducing the number of patients submitted to
operation unnecessarily, allowing care to be expeditiously
and appropriately refocused (ie, systemic chemotherapy,
experimental chemotherapy protocols, palliation).
PET has gained widespread use in oncology for tumor
staging and detection of recurrence, but only a few case
reports and small series have evaluated the clinical utility of
PET in patients with biliary cancers.12,13,15-20 Among the
largest of these reports is the study by Anderson and col-
leagues,12 which included patients with CCA (n ⫽ 36) and

Table 2. Preoperative* Positron Emission Tomography Findings for Patients with Cholangiocarcinoma and Gallbladder Cancer
Intrahepatic Extrahepatic All CCA Gallbladder cancer†
Preoperative PET (n ⴝ 21) (n ⴝ 41) (n ⴝ 62) (n ⴝ 31)
Primary tumor
True positive, n 19 27 46 24
True negative, n 1 2 2 2
False positive, n 0 1 1 1
False negative, n 1 12 13 4
Median SUV 8.6 6.4 7.1 6.7
Sensitivity, % 95 69 78 86
Specificity, % 100 67 75 50
Metastatic disease (nodal and distant)
True positive, n 9 12 21 14
True negative, n 17 24 41 16
False positive, n 1 4 5 2
False negative, n 0 1 1 2
Median SUV 7.6 5.1 7.6 6.0
Sensitivity, % 100 93 95 87
Specificity, % 94 86 89 89
PET findings changed treatment, n (%) 7 (33)‡ 8 (20)‡ 15 (24) 7 (23)
The values shown in this table represent all abnormal fluorodeoxyglucose uptake sites (several patients had multiple sites, ie, primary tumor, local nodal, and
distant metastasis disease). Primary tumor site includes the operative bed in patients with gallbladder cancer who had been treated by cholecystectomy prior to
referral.
*Preoperative PET is defined as a diagnostic study performed before exploration for a potentially curative resection at Memorial Sloan-Kettering Cancer Center.
†
Includes patients who underwent operations prior to referral. False-positive values represent patients who had a positive PET but had a benign pathology or failed
to demonstrate disease progression on followup imaging. False-negative values were patients who had a negative PET but were found to have cancer in the
specimen or on biopsy.
‡
p ⫽ 0.03 for intrahepatic versus extrahepatic cholangiocarcinoma.
CCA, cholangiocarcinoma; PET, positron emission tomography; SUV, standardized uptake value.
Vol. 206, No. 1, January 2008
Figure 3. Positron emission tomography scan from a patient with gallbladder cancer discovered inciden-
tally after laparoscopic cholecystectomy. (A) Arrowhead and (B) Arrowpoint points to an abdominal wall
(port site) metastasis.
GBCA (n ⫽ 14). In this report, PET detected primary
CCA and GBCA with overall sensitivities of 61% and
78%, respectively. But when the authors examined the
CCA group by tumor morphology, it was determined that
the sensitivity of PET for detecting only nodular-type CCA
rose to 85%. In this series, preoperative PET resulted in an
overall sensitivity of 78% for CCA and 86% for GBCA.
Our analysis of CCA found a significant difference in PET
sensitivity between the intrahepatic (95%) and extrahe-
patic types (69%). Differences in tumor size may account
for much of this discrepancy, because intrahepatic CCAs
are generally much larger at the time of presentation. This
study also demonstrated that preoperative PET changed
treatment plans in nearly one-quarter of patients—a find-
ing that is comparable to the Anderson study (30%).
The results of this study must be interpreted within the
context of the limitations pertinent to all retrospective
analyses. Additionally, although all patients in this series
underwent extensive preoperative imaging evaluation to
assess resectability, many of which were performed at
MSKCC, a blinded re-review of all imaging was not per-
formed. In addition, 16% of PET studies were performed
at other centers, and there might have been slight differ-
ences in technique. The relatively small number of patients
analyzed also represents somewhat of a limitation, particu-
larly with respect to subgroup analysis. Last, the study
end point of “change in treatment plan” was determined
based on a comprehensive review of all clinical data (in-
cluding discussion with the attending surgeon in many
cases), but such an assessment is clearly imperfect in a ret-
rospective study. Despite these issues, the data showed that
preoperative staging PET adds important new information
in a subset of patients, information that is not always ap-
Corvera et al Positron Emission Tomography in Biliary Cancer 63
parent on cross-sectional imaging studies and that influ-
ences management. The results also highlight the limita-
tions of PET as a staging tool for biliary tract malignancy,
specifically, in providing anatomic detail related to local
tumor extent that is a major determinant of resectability.
Evidence for this is seen in the significant proportion of
patients with no evidence of metastatic disease on PET
who were subsequently found to have unresectable, locally
advanced disease at exploration (12 of 40).
In patients with GBCA, PET identified unsuspected
metastatic disease in half (4 of 8) of the patients without
earlier cholecystectomy, resulting in a change in disease
stage and treatment plan. In patients who underwent pre-
vious cholecystectomy or exploratory surgery before refer-
ral for definitive resection, information from PET resulted
in a change in treatment plan in only 13% (3 of 23) of
patients. This difference was not entirely unexpected, be-
cause these patients often have small-volume peritoneal
disease that is below the limits of detection. Additionally,
problems related to coexisting inflammation, resulting in
increased glucose metabolism, could confound the results
in some patients. This limitation of PET is important to
recognize, because most patients with GBCA are diagnosed
after undergoing cholecystectomy for benign disease and
often have varying degrees of postoperative inflammation
associated with normal healing. This possibility clearly in-
fluenced the decision to proceed with exploration in some
GBCA patients who had previously undergone cholecys-
tectomy. Likewise, in patients with extrahepatic CCA, the
possibility of inflammatory change related to biliary drain-
age catheters affected the decision to proceed with explora-
tion, despite abnormal PET studies. Nonetheless, given the
very small number of false-positive results observed, our study
64 Corvera et al Positron Emission Tomography in Biliary Cancer
showed that PET can reliably distinguish FDG uptake related
to inflammation from that related to malignancy.
The main goal of resection for patients diagnosed with
biliary cancer is to obtain histologically clear margins,
which often requires extended liver resection. Unfortu-
nately, the proximity of these bile duct tumors to major
vessels frequently results in irresectablity because of direct
invasion of these critical structures. In such cases, the ben-
efit of preoperative PET showing an area of focal uptake
restricted to the primary tumor site can support planned
exploratory surgery, but only if other imaging studies sug-
gest that the tumor is resectable. This is because PET does
not provide sufficient anatomic detail about local resect-
ability. So, the patients most likely to benefit from pre-
operative PET are those with evidence of increased
tracer activity in regional lymph nodes, peritoneum, or
other extraabdominal sites suggesting metastatic disease,
for which resection would be contraindicated. Patients
with metastatic disease to regional lymph nodes remain
a vexing problem. Although some studies have suggested
the possibility of longterm survival after resection in
such situations,5-8 others have not.2,21 This issue remains
an area of controversy in the management of biliary tract
malignancy. It is our opinion that patients with limited
nodal disease within the hepatoduodenal ligament may
benefit from resection; those with involvement of more
distant nodal basins (ie, retropancreatic, celiac, paraaor-
tic) likely do not. As imaging technology continues to
improve, PET will likely be able to more precisely delin-
eate the location and extent of metastatic disease in these
areas.
Although unresectable disease found at the time of op-
eration is common with biliary tract cancer (as high as 50%
for hilar tumors1,4), the true impact of PET can be deter-
mined only with a prospective study of patients subjected
to cross-sectional imaging studies of similar type and qual-
ity. A similar argument may also be made for the use of PET
in identifying cancer recurrence. In the latter circumstance,
it must also be acknowledged that even if PET did provide
earlier detection, the impact on clinical outcomes would
likely be marginal at best.
In conclusion, our study demonstrated that the majority
of biliary cancers are 18F-FDG avid tumors. The sensitivity
of PET for detecting primary tumors and metastatic disease
in both CCA and GBCA was high. In patients with poten-
tially resectable tumors based on conventional imaging,
PET identified occult metastatic disease and resulted in
changes in treatment in nearly one-quarter. Among pa-
tients with CCA, PET was a more sensitive test for those
with intrahepatic than for extrahepatic tumors. In patients
with GBCA, PET appeared to be more useful in those who
J Am Coll Surg
had not undergone previous cholecystectomy. PET was
also helpful in confirming recurrent cancer after resection,
but the data had apparently less impact on treatment
decisions.
Author Contributions
Study conception and design: Corvera, Blumgart, Akhurst,
Fong, Jarnagin
Acquisition of data: Corvera, Akhurst, Blumgart, Jarnagin
Analysis and interpretation of data: Corvera, Blumgart,
Jarnagin
Drafting of manuscript: Corvera, Jarnagin
Critical revision: Fong, DeMatteo, D’Angelica, Akhurst,
Blumgart, Jarnagin
Acknowledgment: We gratefully acknowledge Douglas R
DeCorato, MD, for his assistance in identifying patients for
this study. We thank Pamela Derish for her critical review of
the article.
REFERENCES
1. Jarnagin WR, Klimstra DS, Hezel M, et al. Differential cell
cycle-regulatory protein expression in biliary tract adenocarci-
noma: correlation with anatomic site, pathologic variables, and
clinical outcome. J Clin Oncol 2006;24:1152–1160.
2. Jarnagin WR, Fong Y, DeMatteo RP, et al. Staging, resectability,
and outcome in 225 patients with hilar cholangiocarcinoma.
Ann Surg 2001;234:507–517; discussion 517–519.
3. Ouchi K, Sugawara T, Ono H, et al. Diagnostic capability and
rational resectional surgery for early gallbladder cancer. Hepato-
gastroenterol 1999;46:1557–1560.
4. Fong Y, Jarnagin W, Blumgart LH. Gallbladder cancer: compar-
ison of patients presenting initially for definitive operation with
those presenting after prior noncurative intervention. Ann Surg
2000;232:557–566.
5. Matsumoto Y, Fujii H, Aoyama H, et al. Surgical treatment of
primary carcinoma of the gallbladder based on the histologic
analysis of 48 surgical specimens. Am J Surg 1992;163:239–
245.
6. Muratore A, Polastri R, Capussotti L. Radical surgery for gall-
bladder cancer: current options. Eur J Surg Oncol 2000;26:
438–443.
7. Nimura Y, Hayakawa N, Kamiya J, et al. Hepatopancreatoduo-
denectomy for advanced carcinoma of the biliary tract. Hepato-
gastroenterology 1991;38:170–175.
8. Shirai Y, Ohtani T, Tsukada K, Hatakeyama K. Pancreaticoduo-
denectomy for gallbladder cancer with peripancreatic nodal me-
tastases. Hepatogastroenterology 1997;44:376–377.
9. Bartlett DL, Fong YM, Fortner JG, et al. Long-term results after
resection for gallbladder cancer—implications for staging and
management. Ann Surg 1996;224:639–646.
10. Akhurst T, Fong Y. Positron emission tomography in surgical
oncology. Adv Surg 2002;36:309–331.
11. Reske SN, Bares R, Bull U, et al. [Clinical value of positron
emission tomography (PET) in oncologic questions: results
Vol. 206, No. 1, January 2008 Corvera et al Positron Emission Tomography in Biliary Cancer 65

of an interdisciplinary consensus conference]. Schirmerre-
schaft der deutschen gesellschaft for nuklearmedizin kli-
nische wertigkeit der positronen-emissions-tomographie (PET) bei
onkologischen fragestellungen: ergebnisse einer interdisziplinaren
konsensuskonferenz. Nuklearmedizin 1996;35:42–52.
12. Anderson CD, Rice MH, Pinson CW, et al. Fluorodeoxyglucose
PET imaging in the evaluation of gallbladder carcinoma and
cholangiocarcinoma. J Gastrointest Surg 2004;8:90–97.
13. Kim YJ, Yun M, Lee WJ, et al. Usefulness of 18F-FDG PET in
intrahepatic cholangiocarcinoma. Eur J Nucl Med Mol Imaging
2003;30:1467–1472.
14. Burke EC, Jarnagin WR, Hochwald SN, et al. Hilar cholangio-
carcinoma: patterns of spread, the importance of hepatic resec-
tion for curative operation, and a presurgical clinical staging
system. Ann Surg 1998;228:385–394.
15. Fritscher-Ravens A, Bohuslavizki KH, Broering DC, et al.
FDG-PET in the diagnosis of hilar cholangiocarcinoma. Nucl
Med Commun 2001;22:1277–1285.
16. Kato T, Tsukamoto E, Kuge Y, et al. Clinical role of (18)F-FDG
PET for initial staging of patients with extrahepatic bile duct
cancer. Eur J Nucl Med Mol Imaging 2002;29:1047–1054.
17. Keiding S, Hansen SB, Rasmussen HH, et al. Detection of chol-
angiocarcinoma in primary sclerosing cholangitis by positron emis-
sion tomography. Hepatology 1998;28:700–706.
18. Kluge R, Schmidt F, Caca K, et al. Positron emission tomog-
raphy with [(18)F]fluoro-2-deoxy-D-glucose for diagnosis
and staging of bile duct cancer. Hepatology 2001;33:1029–
1035.
19. Koh T, Taniguchi H, Yamaguchi A, et al. Differential diagnosis
of gallbladder cancer using positron emission tomography with
fluorine-18-labeled fluoro-deoxyglucose (FDG-PET). J Surg
Oncol 2003;84:74–81.
20. Widjaja A, Mix H, Wagner S, et al. Positron emission tomogra-
phy and cholangiocarcinoma in primary sclerosing cholangitis. J
Gastroenterol 1999;37:731–733.
21. Rea DJ, Munoz-Juarez M, Farnell MB, et al. Major hepatic
resection for hilar cholangiocarcinoma: analysis of 46 pa-
tients. Arch Surg 2004;139:514–523; discussion 523–525.

MOVING? MOVED?
Please go to www.efacs.org
Provide your new address,
telephone number, fax number,
or email changes.