F-18-Fluorodeoxyglucose Positron Emission
Tomography in Differentiating Malignant From
Benign Pancreatic Cysts: A Prospective Study
Cosimo Sperti, M.D., Claudio Pasquali, M.D., Giandomenico Decet, M.D.,
Franca Chierichetti, M.D., Guido Liessi, M.D., Sergio Pedrazzoli, M.D., F.A.C.S.
The differential diagnosis between benign and malignant pancreatic cystic lesions may be very difficult.
We recently found that F-18-fluorodeoxyglucose positron emission tomography (18-FDG PET) was
results. From February 2000 to July 2003, 50 patients with a pancreatic cystic lesion were prospectively
investigated with 18-FDG PET in addition to helical computed tomography (CT) and, in some instances,
magnetic resonance imaging (MRI). The validation of diagnosis was based on pathologic findings after
surgery (n ? 31), percutaneous biopsy (n ? 4), and according to follow-up in 15 patients. The 18-FDG
PET was analyzed visually and semiquantitatively using the standard uptake value (SUV). The accuracy
of FDG PET and CT was determined for preoperative diagnosis of malignant cystic lesions. Seventeen
patients had malignant cystic lesions. Sixteen (94%) showed increased 18-FDG uptake (SUV ?2.5),
including two patients with carcinoma in situ. Eleven patients (65%) were correctly identified as having
malignancy by CT. Thirty-three patients had benign tumors: two patients showed increased 18-FDG
uptake, and four patients showed CT findings of malignancy. Sensitivity, specificity, positive and negative
predictive value, and accuracy of 18-FDG PET and CT in detecting malignant tumors were 94%, 94%,
89%, 97%, and 94% and 65%, 88%, 73%, 83%, and 80%, respectively. 18-FDG PET is accurate in
identifying malignant pancreatic cystic lesions and should be used in combination with CT in the
preoperative evaluation of patients with pancreatic cystic lesions. A negative result with 18-FDG PET
mayavoid unnecessaryoperationinasymptomatic orhigh-risk patients.(J GASTROINTESTSURG2005;9:22–
29) ?2005 The Society for Surgery of the Alimentary Tract
KEY WORDS: Pancreas, cystic lesions, intraductal papillary mucinous tumor, mucinous cystadenoma,
serous cystadenoma, positron emission tomography, benign pancreatic cysts
Pancreatic cystic tumors include a variety of le-
sions with different biological behavior from benign
to premalignant or borderline and to frankly malig-
nant neoplasms.1Aggressive resection has been ad-
vised for most of the pancreatic cystic lesions, but
this attitude is now changing for several reasons:
(1) not all of these tumors require resection, and
some patients have comorbidconditions that increase
surgical risk, (2) increasing incidental detection of
pancreatic cystic lesions in asymptomatic patients has
been reported,2and (3) limited pancreatic resec-
tions are performed with increasing frequency for
Presented at the Forty-Fifth Annual Meeting of The Society for Surgery of the Alimentary Tract, New Orleans, Louisiana, May 15–19, 2004
From the Department of Medical and Surgical Sciences (C.S., C.P., G.D., S.P.), University of Padua, Padova, Italy; and Departments of
Nuclear Medicine (F.C.) and Radiology (G.L.), Castelfranco Hospital, Castelfranco Veneto, Treviso, Italy.
This study was supported by the Ministero Universita ` e Ricerca Scientifica (Cofin 2001068593-001), Rome, Italy.
Reprint request: Sergio Pedrazzoli, M.D., F.A.C.S., Universita ` degli Studi di Padova, Dipartimento di Scienze Mediche e Chirurgiche, Direttore
Sezione di Clinica Chirurgica IV, Ospedale Giustinianeo, Via Giustiniani, 2, 35128 Padova, Italy. e-mail: email@example.com
?2005 The Society for Surgery of the Alimentary Tract
Published by Elsevier Inc.
1091-255X/05/$—see front matter
benign or borderline cystic lesions. Therefore, a
correct preoperative identification is crucial for the
appropriate management of these lesions.
Preoperative evaluation ofpancreatic cystic lesions
currently includes abdominal sonography, helical
computed tomography (CT), and magnetic reso-
nance imaging (MRI).3–5Recently, percutaneous6,7
or endoscopic ultrasonography (EUS)8,9aspiration
cytology and cyst fluid tumor marker determina-
tion have been proposed as useful tools helping in the
differential diagnosis. However, conflicting results
and pitfalls have been reported.10,11
Vol. 9, No. 1
2005 PET and Pancreatic Cystic Tumors
Positron emission tomography with F-18-fluoro-
deoxyglucose (18-FDG PET) shows an expanding
role in the evaluation of many solid tumors, includ-
ing pancreatic adenocarcinoma.12,1318-FDG PET is
based on the increased incorporation and metabolism
ofglucose bytumorcells compared with normalcells:
so, a focal uptake likely suggests malignancy. Re-
cently, we have reported in a preliminary study14that
18-FDG PET was very accurate in discriminating
between malignant and benign cystic lesions.
The purpose of this prospective study was to con-
firm our preliminary data with PET in a cohort of
patients with pancreatic cystic lesions and its rele-
vance on clinical management of these patients.
MATERIAL AND METHODS
From February 2000 through July 2003, 50 pa-
tients with suspected cystic tumor of the pancreas
(n ? 33) or intraductal papillary mucinous tumors
(IPMTs) (n ? 17) were prospectively investigated
with 18-FDG PET. All patients underwent helical
CT scanning and serum CA 19-9 tumor marker de-
reference ?37 U/mL). The preoperative evaluation
also included MRI (n ? 25) when the CT findings
were not clear or when an IPMT was suspected. 18-
FDG PET images were obtained using a dedicated
tomograph (Siemens ECAT EXACT 47) with a
field of view of 16.2 cm. After an overnight fast, 444
MBq (12 mCi) of 18-FDG was injected intravenously
to each patient. To avoid interferences due to hyper-
glycemia, blood glucose level was checked just before
the procedure and lowered to less than 120 mg/dL
with insulin administration whenever necessary. Two
transmission scans of the abdomen, for 15 minutes
each, were obtained with 68 Ge rod sources before
the FDG administration to obtain cross sections for
attenuation correction of the emission images. Then,
two emission scans, 15 minutes each, were acquired
starting 60 minutes after FDG administration. The
reconstruction was performed in a 128 × 128 matrix
with Hanning filter 0.3 cutoff. Transaxial, coronal,
and sagittal sections were obtained for visual analysis.
To perform a quantitative analysis, the standardized
uptake value (SUV) was calculated in the suspected
neoplastic foci (SUV ? tissue tracer concentration
per injected dose per body weight). For the SUV
analysis, a circular region of interest was placed over
the area of maximal focal FDG uptake suspected to
be a neoplastic focus, and the mean radioactivity
values were obtained. Based on a previous study,14a
focal uptake with an SUV of at least 2.5 was consid-
ered positive. The PET scan was interpreted by a
single observer (F.C.) without knowledge of the CT
scan results. Each CT scan was also interpreted by a
single reader (G.L.), and the diagnosis of malignancy
was based on general rules and on criteria suggested
for IPMTs.15Validation of diagnosis was based on
the pathologic findings of resected specimen, biopsy,
or follow-up. Pathologic classification of the pancre-
atic tumors was made according to World Health
Organization (WHO) histologic typing.1Sensitivity,
specificity, positive and negative predictive value, and
accuracy of 18-FDG PET and CT scan in differenti-
ating malignant from benign lesions were evaluated.
Our policy in the clinical management of patients
with pancreatic cystic lesions was to resect, whenever
possible, all of the symptomatic, or PET-positive,
cystic lesions. For PET-negative cystic lesions, sur-
gery was performed only when clinical and radiologic
features suggested mucinous tumors (cystadenomas
or IPMTs). Standard resection was the operation of
choice for malignancies, whereas more conservative
surgery-sparing pancreatic parenchyma or spleen,
was reserved for benign lesions. Follow-up was
tients, with high surgical risk or lesions located in
the head of the pancreas.
The distribution of patients according to pathol-
ogy is summarized in Table 1. The final pathologic
after percutaneous biopsy in 4. Fifteen PET negative
Table 1. Distribution of patients based on pathology
Type of lesionNo. of patients
IPMT ? intraductal papillary mucinous tumor; ACCD ? adenocar-
cinoma with retention cyst. Values in parentheses ? diagnosis of pa-
tients in follow-up based on morphologic appearance.
*Left adrenal lymphangioma, cystic dysplasia of the duodenum, pan-
creatic localization of Tangiers’s disease.
Sperti et al.
cystic lesions were put on follow-up (median, 12
months; range, 8–44 months). None of these lesions
developed changesinradiologic appearanceormalig-
nancy. There were 17 men and 33 women, with a
mean age of 58.1 years (range, 14–87 years). Thirty-
one patients (62%) were symptomatic: the most
common symptoms and signs were pain (n ? 23),
dyspepsia (n ? 5), jaundice (n ? 1), and palpable ab-
dominal mass (n ? 2). Twelve patients had one or
more attacks of acute pancreatitis. Nineteen patients
(38%) were asymptomatic and the pancreatic lesion
was incidentally discovered during investigations for
unrelated disease. Mean tumor diameter was 3.5 cm
(range, 1.0–10.0 cm). Fourteen patients had multiple
pancreatic cystic lesions.
Seventeen patients had malignant lesions: five cys-
tadenocarcinomas, two solid-papillary carcinomas,
one endocrine carcinoma, one adenocarcinoma with
mucinous carcinoma, two in situ and six invasive
types, according to WHO classification). Thirty-
three patients had benign lesions.
There were 11 women and 6 men with a mean age
of 58.3 years (range, 14–87 years). Fourteen patients
(82%) were symptomatic: nine had abdominal pain;
two, dyspepsia; two, palpable mass; and one, jaun-
dice. One patient underwent cystojejunostomy for
presumed pancreatic pseudocyst before referral to
our department; another patient was admitted with
a diagnosis of chronic pancreatitis and cholangitis.
Three patients experienced one or more attacks of
acute pancreatitis. Three patients were asymptomatic
gation for other disease (chronic hepatitis in two and
breast cancer in one). Four patients had diabetes.
Fig. 1. Positron emission tomography scan (transaxial projec-
tion). Peripheral F-18-fluorodeoxyglucose uptake in the wall
of a large pancreatic head cystic lesion (arrows). Malignant
intraductal papillary mucinous tumor.
TheCTscanshowedasolitarycysticmass(n ? 11)
with internal septa (n ? 6) or multiple cysts (n ? 5) or
a dilated main pancreatic duct (n ? 5). In one patient,
encasement of the superior mesenteric vein. Mean
tumor size was 4.3 cm (range, 1.5–10.0 cm). Clear
CT features of malignancy were found in 11 patients
(65%). PET scan was positive in 10 of them. Sixteen
of the 17 patients (94%) showed 18-FDG PET
uptake with an SUV range of 2.5–7.0. An isolated
uptake with central absence of metabolism was found
in 6 patients (Fig. 1). Both patients with IPMT
SUV of 2.5 and 5.0, respectively (Fig. 2). In three
patients, 18-FDG PET showed liver metastases:
other patients, PET detected lymph node metastases
that were confirmed at laparotomy.
duodenectomy in six, total pancreatectomy in one,
distal pancreatectomy and splenectomy in five), and
two underwent bypass operation (both for vascular
becauseofmultiplelivermetastases(n ? 2)ormesen-
teric vein involvement (n ? 1); in these patients, the
diagnosis was confirmed by percutaneous fine-
Among them, the single patient who showed
normal 18-FDG uptake was a woman with well-dif-
ferentiated cystadenocarcinoma of the head of the
pancreas that was treated with pancreaticoduodenec-
tomy. This tumor recurred 13 months later in the
peritoneum, and 18-FDG PET showed multiple foci
of pathologic uptake in the abdomen. Chemotherapy
Fig. 2. Positron emission tomography scan (transaxial projec-
tion). Isolated focal F-18-fluorodeoxyglucose uptake in the
pancreatic head. Malignant intraductal papillary mucinous
tumor (carcinoma in situ).
Vol. 9, No. 1
2005 PET and Pancreatic Cystic Tumors
was started and the patient is still alive 19 months
after primary operation.
An additional four patients had one or more PET
scans during the follow-up after resection; three had
tumor recurrence in the liver, liver and peritoneum,
and para-aortic lymph nodes, respectively. CT iden-
tified only liver metastases in two patients, while
18-FDG PET correctly showed all sites of tumor
of lymph node recurrence.
Among the patients with benign tumors, there
were 22 women and 11 men with a mean age of 58.0
years (range, 17–86 years). Five patients (15%) had
a pseudocyst (in two cases, multiple pseudocyst) with
CT features resembling a cystic tumor, without his-
tory of acute or chronic pancreatitis. Seventeen pa-
tients (52%) were symptomatic: the most common
complaints were abdominal pain (n ? 14) and dys-
pepsia (n ? 3). In nine patients, there was a history
of one or more bouts of acute pancreatitis, and five
had diabetes. Sixteen patients (48%) were asymptom-
atic and their lesion was incidentally found during
investigations for unrelated disease.
CT showed a solitary cystic mass in 25 patients
a dilated main pancreatic duct in 8. The mean tumor
size was 3.2 cm (range, 1.0–10.0 cm). Four patients
showed CT features suggesting a malignant tumor
(Fig. 3). In 31 of 33 patients (94%), no uptake of 18-
FDG was shown. Two patients showed a peripheral
uptake with a central area of absent tracer concentra-
tion: a malignant cystic tumor in the tail of the pan-
creas was diagnosed and resected. Final pathologic
examination showed a pseudocyst opened into the
spleen (SUV ? 2.6) and a pancreatic localization of
Tangier’s disease (SUV ? 3.0), respectively.
Seven patients underwent distal pancreatectomy
(five with spleen preservation) (Fig. 3); three patients,
pylorus-preserving pancreaticoduodenectomy; two
patients, duodenum-preserving pancreatic head re-
section; one patient, median pancreatectomy; two
patients, tumor enucleation; and two patients,
cystojejunostomy. Sixteen patients were not operated
on: only one of them underwent percutaneous aspira-
tion biopsy and cyst fluid tumor marker determi-
nation without evidence of malignancy (possible
serous cystic tumor). All 16 patients had follow-up
(median follow-up, 12 months; range, 8–44 months):
logic findings of their lesion.
Sensitivity, specificity, positive and negative pre-
dictive values, and accuracy of 18-FDG PET in
detecting malignant cystic lesions were 94%, 94%,
were 65%, 88%, 73%, 83%, and 80%, respectively.
The differential diagnosis of cystic lesions of the
pancreas remains a challenge. Resection has been ad-
vocated forall suspectedcystic tumors ofthepancreas
to minimize diagnostic errors.16On the other hand,
the increasing number of these lesions seen in the
clinical practice17and the large number of studies
concerning pancreatic cystic tumors require some
considerations. First, more asymptomatic patients
with pancreatic cystic mass are now detected, as in
reports.2,18Second, some pancreatic cystic lesions are
invariably benign16; so, it is unlikely that asymptom-
atic patients experience benefit from tumor removal.
Third, pancreatic resection in old, high–surgical risk
patients seems to be justified only for malignant or
symptomatic tumors. Finally, the resection of tumors
located in the head of the pancreas suggests caution
even in younger patients. Therefore, we need a
simple, reproducible,noninvasive method able to dif-
ferentiate malignant from benign cystic lesions. De-
spite previous descriptions of typical clinical and
radiologic features, conventional imaging modalities
such as CT and MRI do not reliably distinguish be-
tween benign and malignant cystic lesions.5,19EUS
has an expanding role in the preoperative evaluation
of patients with pancreatic tumors.20,21EUS also has
been advocated for the evaluation of pancreatic cystic
masses,8,22–24but EUS alone seems not to improve
CT results in distinguishing malignant from benign
pancreatic cystic lesions.9,25Fine-needle aspiration
cytology and cyst fluid analysis for enzyme and tumor
marker determination appear to be logical adjunctive
tests to better define a cystic mass of the pan-
creas.9,25,26However, cytology often shows false-neg-
ative or inconclusive results, and the wide overlap of
tumor marker values makes the differentiation diffi-
cult.9,10,25Mucin determination in the aspirated cyst
fluid seems to improve accuracy in detecting mu-
method with a risk, although low, of complications
also during EUS examination.9,25
In recent years, 18-FDG PET imaging has been
increasingly used in the diagnosis, staging, and post-
During the process of malignant transformation, the
increased glucose transport and utilization. The en-
hanced glucose uptake explains why 18-FDG PET
Sperti et al.
Fig. 3. Top right and left, Magnetic resonance image of a suspected cystadenocarcinoma of the body-tail
of the pancreas in a 32-year-old woman. The large collateral veins draining the spleen suggest the
obstruction ofthe splenic vein. Bottomleft, Positron emission tomography scan is negativefor malignancy.
Bottom right, Borderline mucinous cystadenoma resected with a spleen-preserving procedure.
canfunctionally identify malignanttissues. This prin-
PET in the differential diagnosis of cystic lesions of
the pancreas, particularly in distinguishing malignant
from benign pancreatic cystic lesions.14This prelimi-
nary study in 56 patients with suspected pancreatic
cystic tumors showed that PET correctly identified
16 of 17 malignant and 38 of 39 benign cystic lesions
with a specificity, positive and negative predictive
value, and accuracy of 97%, 94%, 97%, and 96%,
respectively, in detecting malignant tumors. Since
then, we have continued to use 18-FDG PET, when-
ever possible, in the preoperative work-up of all pa-
tients with suspected cystic tumor of the pancreas.
To our knowledge there are few reports dealing with
PET imaging and cystic diseases of the pancreas.
Yoshioka et al.28reported a high 18-FDG uptake in
two patients with IPMT and invasive carcinoma.
On the other hand, McHenry et al.29found that EUS
fine-needle aspiration was more accurate (71%) than
However, this study included only 13 valuable pa-
tients,suggesting furtherexperiences are necessary to
assess the role of 18-FDG PET in cystic lesions of
the pancreas. Our current study results confirm the
excellent results that we previously reported14: 18-
FDG PET was able to detect 16 of 17 malignant
cystic lesions (sensitivity of 94%) with a specificity of
93%. Interestingly, PET imaging showed increased
uptake of 18-FDG, also in both patients with IPMT
Vol. 9, No. 1
2005 PET and Pancreatic Cystic Tumors
and carcinoma in situ, in whom CT and MRI did
not show any sign of malignancy. The identification
of initial malignant transformation in the course of
carcinogenesis is obviously crucial for the treatment
of neoplasms, especially for those lesions that repre-
sent potentially curable tumors. Some30emphasized
the importance of EUS biopsy in detecting malignant
IPMTs, with an accuracy of 91% for invasive cancers.
However, only 40% of noninvasive cancers (e.g., car-
cinoma in situ) were detected by examination of
Furthermore, in our series, 18-FDG PET added
new information about tumor extension in 3 of 17
patients, showing liver and lymph node metastases
not detected by traditional imaging. The single false-
negative result occurred in a patient with cystadeno-
carcinomaof theheadof thepancreas, correctly iden-
tified by CT. Thirteen months after resection, PET
scan showed peritoneal recurrences that were not
detected by CT scan. The false-positive results in our
series occurred in a patient with pseudocyst partially
invading the spleen and in a patient with a very rare
pancreatic localization of Tangier’s disease mimick-
ing a malignant cystic tumor. A false-positive result
of FDG PET in acute inflammations is a well-known
event31; the false-positive result in such a rare meta-
bolic disease (Tangier’s) is explained by abundant
infiltration of hystiocytes and macrophages and their
18-FDG incorporation, like inflammatory granulo-
no FDG uptake. On the basis of previous experi-
ence,14PET-negative cystic lesions were regarded as
benign. Therefore, they were resected in low-risk
patients when a premalignant tumor was diagnosed
or when disabling symptoms were present. Further-
more, a negative PET scan prompted a more con-
servative pancreatic resection (n ? 7) or avoided
unnecessary splenectomy (n ? 5). According to this
policy, 16 asymptomatic patients underwent follow-
operated patients were checked at 6 months and
thereafter once a year. None showed changes in cyst
diameter or appearance (all patients underwent ab-
dominal sonography and/or CT or MRI; three pa-
tients had PET repeated). A limitation of 18-FDG
PET remains the inability of this functional imaging
modality to replace anatomic imaging in the assess-
ment of tumor resectability; thus, 18-FDG PET is a
sensitive and specific adjunct to CT in the differential
diagnosis of cystic tumors of the pancreas. In the near
future, when the PET-CT scan33becomes widely
available, a single procedure probably can provide
more precise information, both functional and mor-
18-FDG PET is a very useful technique for the
preoperative work-up of patients with suspected
cystic tumors of the pancreas. The high FDG uptake
suggest a malignant tumor that require aggressive
resection, whereas a negative PET scan likely identi-
fies a benign tumor that may be treated with more
tance for the clinical management, especially in
asymptomatic high-risk patients, is clearly relevant.
The authors gratefully acknowledge Tania Lazzarin for helping
with the manuscript.
1. Kloppel G, Solcia E, Longnecker DS, et al. Histological
Typing of Tumours of the Exocrine Pancreas (World Health
Organization), 2nd ed. Berlin, Springer, 1996.
2. Fernandez del Castillo C, Targarona J, Thayer SP, Rattner
DW, Brugge WR, Warshaw AL. Incidental pancreatic cysts.
Clinicopathologic characteristics and comparison with symp-
tomatic patients. Arch Surg 2003;138:427–434.
3. Demos TC, Posniak HV, Harmath C, Olson MC, Aranha G.
Cystic lesions of the pancreas. AJR Am J Roentgenol 2002;
4. Kalra MK, Maher MM, Sahani DV, Digmurthy S, Saini S.
Current status of imaging in pancreatic diseases. J Comput
Assist Tomogr 2002;26:661–675.
5. Kalra MK, Maher MM , Boland GW, Saini S, Fischman AJ.
Correlation of positronemission tomographyand CT in eval-
uating pancreatic tumors: Technical and clinical implications.
AJR Am J Roentgenol 2003;181:387–393.
6. Lewandrowski KB, Lee J, Southern JF, et al. Cyst fluid analy-
sis in the differential diagnosis of pancreatic cysts: A new
approach to the preoperative assessment of pancreatic cystic
lesions. AJR Am J Roentgenol 1995;164:815–819.
7. Sperti C, Pasquali C, Guolo P, et al. Serum tumor markers
andcyst fluid analysisareusefulfor thediagnosisofpancreatic
cystic tumors. Cancer 1996;78:237–243.
8. Van Dam J. EUS in cystic lesions of the pancreas. Gastroint-
est Endosc 2002;56(Suppl 4):S91–S93.
9. Hernandez LV, Mishra G, Forsmark C, et al. Role of endo-
scopic ultrasound (EUS) and EUS-guided fine needle aspira-
tion in the diagnosis and treatment of cystic lesions of the
pancreas. Pancreas 2002;25:222–228.
10. Sand JA, Hyoty MK, Mattila J, Dagorn JC, Nordback JH.
Clinical assessment compared with cyst fluid analysis in the
differential diagnosis of cystic lesions of the pancreas. Surgery
11. Ahmad NA, Kochman ML, Lewis JD, Ginsberg GG. Can
EUS alone differentiate between malignant and benign cystic
lesions of the pancreas? Am J Gastroenterol 2001;96:3295–
12. Van Heertum RL, Fawwaz RA. The role of nuclear medicine
in the evaluation of pancreatic disease. Surg Clin North Am
13. Pham KH, Ramaswamy MR, Hawkins RA. Advances in posi-
tron emission tomography imaging for the GI tract. Gas-
trointest Endosc 2002;55(Suppl 7):S53–S63.
Sperti et al.
14. Sperti C, Pasquali C, Chierichetti F, Liessi G, Ferlin G,
Pedrazzoli S. Value of 18-fluorodeoxyglucose positron emis-
sion tomography in the management of patients with cystic
tumors of the pancreas. Ann Surg 2001;234:675–680.
15. YamaoK, OhashiK, NakamuraT, et al.Evaluationofvarious
imaging methods in the differential diagnosis of intraductal
papillary mucinous tumor (IPMT) of the pancreas. Hepato-
16. Sheehan MK, Beck K, Pckleman J, Aranha GV. Spectrum of
cystic neoplasms of the pancreas and their surgical manage-
ment. Arch Surg 2003;138:657–662.
Current state of the art and unanswered questions. J GAS-
TROINTEST SURG 2003;7:417–428.
18. Bassi C, Salvia R, Molinari E, Biasiutti C, Falconi M, Peder-
zoli P. Management of 100 consecutive cases of pancreatic
serous cystadenoma: Wait for symptoms and see at imaging
or vice versa? World J Surg 2003;27:319–323.
19. Sahani D, Prasad S, Saini S, Mueller P. Cystic pancreatic
neoplasms evaluation by CT and magnetic resonance cholan-
giopancreatography. Gastrointest Endosc Clin North Am
20. Mertz HR, Sechopoulos P, Delbeke D, Leach SD. EUS,
PET, and CT scanningfor evaluationof pancreaticadenocar-
cinoma. Gastrointest Endosc 2000;52:367–371.
21. DelFrateC, Zanardi R,Mortele K,etal.Advancesin imaging
for pancreatic disease. Curr Gastroenterol Rep 2002;4:140–
22. Bounds BC, Brugge WR. EUS diagnosis of cystic lesions of
the pancreas. Int J Gastrointest Cancer 2001;30:27–31.
23. Song MH, Lee SK, Kim MH, et al. EUS in the evaluation
of pancreatic cystic lesions. Gastrointest Endosc 2003;57:
24. Ahmad N, Kochman ML, Brensinger C, et al. Interob-
server agreement among endosonographers for the diagnosis
Dr. Nathaniel Soper (Chicago, IL): Did you look
at any other modalities that have been reported for
such as aspiration techniques and assessing cyst fluid
for CA 19-9, etc.? And second, did you find any
differences between this prospective study compared
to the retrospective study that you previously pub-
lished in the Annals of Surgery?
Dr.Pedrazzoli: Asto thefirst question,weworked
for several years trying to find a specific marker for
malignancy for a differential diagnosis but could not
find such a reliable marker. The problem is also
that when you do aspiration, if you have a malignant
lesion, you can also seed malignant cells. So it is
not safe when you have a suspected malignancy and
you need to decide the feasibility or not. So that if
you do a PET, it is a noninvasive modality.
About the difference between the first study and
the second study, the accuracy was the same, exactly
the same. In the first study, we decided to have histol-
ogy in all patients. In the second study, after the first
of neoplastic versus non-neoplastic pancreatic cystic lesions.
Gastrointest Endosc 2003;58:59–64.
25. Walsh RM, Henderson JM, Vogt D,et al. Prospective preop-
erative determination of mucinous pancreatic cystic neo-
plasms. Surgery 2002;132:628–634.
26. Frossard JL, Amouyal P, Amouyal G, et al. Performance of
endosonography-guided fine needle aspiration and biopsy in
the diagnosis of pancreatic cystic lesions. Am J Gastroenterol
27. Delbeke D. Oncological applications of FDG PET im-
aging. J Nucl Med 1999;40:1706–1715.
28. Yoshioka M, Sato T, Furuya T, et al. Positron emission to-
mography with 2-doxy-2-fluor d-glucose for diagnosis of in-
traductal papillary mucinous tumors of the pancreas with
parenchymal invasion. J Gastroenterol 2002;38:1189–1193.
29. McHenry L Jr, Fletcher JW, Tann M, et al. Cystic tumors
ron emission tomography (PET) and endoscopic ultrasound-
guided fine needle aspiration (EUS-FNA). Presented at the
Annual Meeting of Digestive Disease Week, Orlando, Flor-
ida, May 17–22, 2003, W1204 (abstract).
tive predictors of malignancy in pancreatic intraductal papil-
lary mucinous neoplasms. Arch Surg 2003;138:610–618.
31. Delbeke D, Martin WH. Positron emission tomography im-
aging in oncology. Radiol Clin North Am 2001;39:893–917.
32. Kubota R, Yamada S, Kubota K, Ishiwata K, Tamahashi N,
Ido T. Intratumoral distribution of fluorine-18-fluorodeox-
yglucose in vivo: High accumulation in macrophages and
granulation tissues studied by microautoradiography. J Nucl
33. Hany TF, Steinert HC, Gerres GW, Buck A, von Schul-
thess GK. PET diagnostic accuracy: improvement with in-
line PET-CT system: Initial results. Radiology 2002;225:
few patients, 16 patients had only follow-up, and
most of them are asymptomatic patients who were
followed up from 12 months to 4 years, without
Dr. Henry Pitt (Milwaukee, WI): This analysis
was very nice and you are to be congratulated for
finding a way to differentiate the benign and malig-
nant tumors. Clearly, we have not had anything as
accurate as PET. It was interesting to me, however,
you had two groups that were benign and malignant;
but when you came to your conclusion, you had three
groups, benign, premalignant, and malignant. It
would be ideal if we had a way of differentiating
benign versus premalignant and malignant, because
most of us believe that the premalignant lesions
should come out before they become malignant. Un-
fortunately, PET does not give us that additional
Dr. Pedrazzoli: Unfortunately, I was wrong in my
conclusion, but premalignant are still benign lesions,
Vol. 9, No. 1 Download full-text
2005 PET and Pancreatic Cystic Tumors
and should be considered still a benign lesion not yet
transformed; for this reason they wereincluded in the
benign group. Usually, conventional radiology (CT
or MRI) is able to distinguish serous cystoadenomas,
always benign lesions, from the other types of
tumors. So the differential diagnosis between benign
and premalignant and malignant lesions is not so
difficult. However, it is important to diagnose a ma-
lignant lesion in old patients who need surgical treat-
ment or a benign premalignant lesion in old patients
who do not need surgical resection because they are
old and they would not die of the disease.
Dr. Stephen Vogel (Gainesville, FL): As you know,
manypatients with smallcystictumorsinthis country
undergo endoscopic ultrasound with aspiration, and
either the endoscopist samples the fluid in his or her
hands, finds it sticky, and then makes a diagnosis of
mucin or the pathologist will stain for mucin. Now,
we are finding small mucin-producing cystic lesions.
Do you have a series of small mucin-producing cystic
lesions with a negative PET scan that you followed,
in other words, you did not operate on, for some
period of time?
As you know, our pathologists may not diagnose
malignancy in a mucinous tumor. They will just call
it a mucinous tumor based on the fact, as Dr. Pitt
mentioned, that there is a “potential” forit to become
malignantsomewhere downtheroad. So my question
is, have you followed small mucin-producing cystic
lesions for a period of time with negative PET scans?
Dr. Pedrazzoli: I agree with you that a young
with a limited resection. It is easier. If you wait, it
pancreas. We have four PET-negative patients with
mucinous cystoadenoma currently on follow-up. But
the real problem is the intraductal papillary mucinous
tumor, which commonly occurs in the older popula-
tion. In the older population, if you have a benign
or borderline lesion, we would place the patient into
I forgot to tell that two patients with carcinoma
in situ were PET positive, while old patients with
borderline or only low-grade or medium-grade dys-
plasia were negative.
So I believe that PET scan is one of the best
ways to do the differential diagnosis in the uncertain
cases, not in all cases. When calculating the cost-
effectiveness of PET, we must consider not only the
high cost of the procedure, but also the cost of