SEMINAR

Seminar

Pancreatic cancer

Donghui Li, Keping Xie, Robert Wolff, James L Abbruzzese

Pancreatic cancer remains a major unsolved health problem, with conventional cancer treatments having little impact
on disease course. Almost all patients who have pancreatic cancer develop metastases and die. The main risk factors
are smoking, age, and some genetic disorders, although the primary causes are poorly understood. Advances in
molecular biology have, however, greatly improved understanding of the pathogenesis of pancreatic cancer. Many
patients have mutations of the K-ras oncogene, and various tumour-suppressor genes are also inactivated. Growth
factors also play an important part. However, disease prognosis is extremely poor. Around 15–20% of patients have
resectable disease, but only around 20% of these survive to 5 years. For locally advanced, unresectable, and
metastatic disease, treatment is palliative, although fluorouracil chemoradiation for locally advanced and gemcitabine
chemotherapy for metastatic disease can provide palliative benefits. Despite pancreatic cancer’s resistance to
currently available treatments, new methods are being investigated. Preoperative chemoradiation is being advocated,
with seemingly sound reasoning, and a wider role for gemcitabine is being explored. However, new therapeutic
strategies based on the molecular biology of pancreatic cancer seem to hold the greatest promise.

Pancreatic cancer is one of the most lethal human cancers

Risk factors for pancreatic cancer
and continues to be a major unsolved health problem at
the start of the 21st century. It has been estimated that
Demographic factors
this disease causes 30 000 deaths per year in the USA; this
Old age (most reliable and important predictor)
number has been quite steady over the past 3–5 years.1
Sex (more common in males than in females)
Despite efforts in the past 50 years, conventional treat-
Ethnic origin (mortality highest in black populations)
ment approaches, such as surgery, radiation, chemo-
therapy, or combinations of these, have had little impact Genetic factors and medical conditions
on the course of this aggressive neoplasm. Therefore, only Family history
by developing a detailed understanding of the molecular Hereditary pancreatitis
biology of pancreatic cancer will we be in a position to Hereditary non-polyposis colorectal cancer
effectively diagnose, prevent, and treat this disease. Close Ataxia-telangiectasia
to 100% of patients with pancreatic cancer develop Peutz-Jeghers syndrome
metastases and die because of the debilitating metabolic Familial breast cancer
effects of their unrestrained growth. Thus, a critical Familial atypical multiple mole melanoma
requirement for progress will be the development of Chronic pancreatitis
effective systemic treatments capable of reversing the Diabetic mellitus
aggressive biology of this disease. Gastrectomy
Deficiency in carcinogen metabolism and DNA repair
Epidemiology and molecular epidemiology Environmental and lifestyle factors
In many studies risk factors associated with pancreatic Cigarette smoking
cancer have been explored (panel). The only risk factors Occupational exposures
consistently reported are age and cigarette smoking.2 Low dietary intake of fruits and vegetables
Cigarette smoking is estimated to account for 25–29% of Food preparation and cooking methods (grilling or charring
pancreatic cancer incidence, with reported odds ratios confers the highest risk)
ranging from 1·6 to 5·4.3 Another risk factor for pancreatic
cancer might be family history.4,5 Several genetic
syndromes are associated with an increased risk of
pancreatic cancer, including hereditary pancreatitis, Search strategy and selection criteria
hereditary non-polyposis colorectal cancer, ataxia- We identified reports by MEDLINE search through the PubMed
telangiectasia, Peutz-Jeghers syndrome, familial breast database (1986–2002) by combining the keywords
cancer, and familial atypical multiple-mole melanoma.4 In “pancreatic cancer” with the following topics: carcinogenesis,
epidemiological studies of pancreatic cancer, a protective angiogenesis, progression, metastasis, pathology,
pathobiology, pathophysiology, molecular genetics, and
Lancet 2004; 363: 1049–57 genetics. We searched citation lists in retrieved papers to
identify additional references. Papers were selected on the
Department of Gastrointestinal Medical Oncology, University of basis of the best available evidence for each specific
Texas, M D Anderson Cancer Center, 1515 Holcombe Boulevard, question discussed. To limit the number of references, review
Box 426, Houston, TX 77030, USA (D Li PhD, K Xie MD, R Wolff MD, articles or the latest publications in a series of articles from
Prof J L Abbruzzese MD) the same laboratory were given preference. Only English
Correspondence to: Prof James L Abbruzzese language papers were included.
(e-mail: jabbruzz@mdanderson.org)

THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com 1049

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

role has been noted for diets high in fruits and cancer. Even though the mechanisms underlying the
vegetables.6,7 This effect might be related to dietary intake association between K-ras mutation and these
of folate and other methyl-donor groups.8 Occupational environmental factors are not fully understood, the
exposure to carcinogens has long been suspected as a opportunity exists to reduce pancreatic cancer incidence
causal factor for pancreatic cancer, but evidence is through dietary change and limiting exposure to
insufficient to identify any specific exposure as likely to carcinogens.
substantially increase the risk of pancreatic cancer.9–11 In
sum, the primary causal factors for pancreatic cancer are Pathophysiology and molecular biology
poorly understood. Research efforts aimed at quantifying Molecular pathology
risk factors and identifying individuals at high risk are In the past few years, our knowledge of the pathogenesis of
critical to the eventual prevention of this disease. pancreatic cancer has been significantly advanced due to
In the past few years, some major advances have been the rapid accumulation of our understanding of the
made in the understanding of environment-susceptibility molecular biology of the disease. Like many other
interactions in human cancers. Studies in pancreatic malignant diseases, pancreatic ductal carcinoma results
cancer have involved detection of DNA damage derived from the accumulation of acquired mutations (table 1).
from carcinogen exposure and endogenous metabolic The multigenic nature of most pancreatic ductal cancers is
processes. To test the hypothesis that exposure to reflected in the abnormalities of three broad classifications
carcinogens contributes to human pancreatic cancer, and of genes—ie, oncogenes, tumour-suppressor genes, and
to shed a light on the carcinogens involved, the variety of genomic maintenance genes.25,26 The accumulated
DNA damage in human pancreatic tissues has been mutations in such genes are believed to occur in a
assessed.12–15 For example, smoking-related aromatic DNA predictable time course. On the basis of the understanding
adducts,12 and other types of DNA damage13–15 have been of the histological and molecular genetic profiles, a
detected. These observations suggest that the human progression model has been developed that describes
pancreas is susceptible to carcinogen exposure and DNA pancreatic ductal carcinogenesis: the pancreatic ductal
damage, which might contribute to genetic mutation, and epithelium progresses from normal to increasing grades of
in turn cancer development. pancreatic intraepithelial neoplasia, to invasive cancer
In many studies of smoking-related human cancers, (figure 1).27
genetic variability in carcinogen metabolism and DNA A few pancreatic cancers aggregate in families and have
repair affects individual susceptibility to carcinogen aided our understanding of pancreatic tumorigenesis.
exposure and cancer risk. However, few such studies have Most pancreatic cancers, however, occur sporadically.
been done in pancreatic cancer. In small-scale case- Since the identification of the first notable genetic
control studies no significant association has been noted alteration—mutation of the K-ras oncogene mutation—
between the risk of pancreatic cancer and polymorphisms there has been an explosion in our understanding of
of drug-metabolising enzymes.16,17 In a large-scale pancreatic cancer genetics. More than 85% of pancreatic
population-based case-control study, however, a ductal cancers have an activating point mutation in the K-
significant interaction was reported between the GSTT1 ras gene at a very early stage of pancreatic-cancer
null genotype and cigarette smoking in pancreatic development.28 The detection of K-ras mutations in the
cancer.18 Evidence from the same study also showed a duodenal juice, pancreatic juice, and stool of patients with
significant interaction between heavy smoking and an pancreatic cancer has been proposed as an early detection
Arg399Gln polymorphism of the DNA repair gene strategy.29,30
XRCC1.19 Jiao and colleagues20 identified a significant The p16 tumour-suppressor gene is inactivated in
association between a functionally important codon 143 around 95% of pancreatic cancers.31,32 and typically occurs
(Ile/Val) polymorphism of the O6-alkylguanine DNA later in pancreatic carcinogenesis.33 The second most
transferase gene and risk of pancreatic cancer in a frequently inactivated tumour-suppressor gene is TP53, a
hospital-based case-control study. These data support the
hypothesis that individuals who have deficient carcinogen Chromosomal location Alteration frequency (%)
detoxification and DNA repair capacities are at increased Oncogenes
risk of pancreatic cancer. More studies are needed to K-ras 12p 75–100
elucidate the role of other genes and pathways that are HER2/neu 17q 65–70
AKT2 19q 10–20
important in pancreatic carcinogenesis. MYB 6q 10
In studies of the spectra of tumour-suppressor-gene
mutations, specific endogenous or exogenous mutagens Tumour-suppressor and genome-maintenance genes
TP53 17p 40–75
have induced characteristic patterns of DNA alteration. CDKN2A* 9p 27–98
These changes can be used as a fingerprint of exposure. CDKN2A† 9p 27–82
Pancreatic cancer has the highest frequency (>85%) of CDKN2B 9p 27–48
K-ras mutation among all human cancers, which has been MADH4 18q 50–55
associated with cigarette smoking or alcohol FHIT 3p 66–70
RBI 13q 0–10
consumption.21,22 The risk of mutation is three times BRCA2 13q 7–10
higher in alcohol drinkers than in non-drinkers. Serum STK11 19q 5
concentrations of organochlorine compounds are also MAP2K4 17p 4
significantly associated with K-ras mutations in pancreatic ALK5 9q 1
cancer.23 Furthermore, a specific association has been TGFBR2‡ 3p 1
TGFBR2§ 3p 3
noted between serum concentrations of DDT and DDE MLH1 3p 3
and G-to-T transversion at codon 12 of the K-ras gene.
*p16INK4a. †p19ARF. ‡MSI–. §MSI+. Modified from Mangray S, King TC. Molecular
The same investigators have also noted an association pathobiology of pancreatic adenocarcinoma. Front Biosci 1998; 3: D1148–60,
between occupational exposure to organic solvent and and Sohn TA, Yeo CJ. The molecular genetics of pancreatic ductal carcinoma: a
K-ras mutations in pancreatic cancer.24 These results review. Surg Oncol 2000; 9: 95–101.
support the hypothesis that K-ras mutations can be Table 1: Commonly altered oncogenes and tumour-suppressor
related to lifestyle and environmental factors in pancreatic genes in human pancreatic adenocarcinoma

1050 THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

Rights were not granted to include this image in electronic media.

Please refer to the printed journal.

Figure 1: Progression model of pancreatic cancer

PanIN=pancreatic intraepithelial neoplasia. Reproduced with permission from Hruban RH, Goggins M, Parsons J, Kern SE. Progression model for pancreatic
cancer. Clin Cancer Res 2000; 6: 2969–72. Artwork by Jennifer Parsons.

well-characterised tumour-suppressor located on showed that expression of vascular endothelial growth

chromosome 17p. Its inactivation is a late event in factor can be significantly up-regulated by low
tumorigenesis. The MADH4 gene (DPC4 or SMAD4) is extracellular pH, or acidosis, which occur frequently
inactivated in 55% of pancreatic adenocarcinomas.34 Like within the expanding tumour mass, particularly in
TP53, MADH4 inactivation is a late event in pancreatic regions surrounding necrotic areas within tumours. They
tumorigenesis. Other less-common genetic alterations showed also that acidosis activates the interleukin-8 gene.
continue to be described in pancreatic cancer.35–39 Vascular endothelial growth factor and interleukin 8 are
Rozenblum and colleagues,40 in a comprehensive key angiogenic molecules for pancreatic cancer. In
mutational analysis of 42 pancreatic ductal cancers, noted detailed molecular biology studies, up-regulation of these
that all tumours harboured mutations in the K-ras genes by acidosis can be mediated through transac-
oncogene. The individual mutational frequencies of tivation and cooperation of transcription factors, NF-␬B
tumour-suppressor genes p16, TP53, MADH4, and and AP-1.51
BRCA2 were 82%, 76%, 53%, and 10%, respectively. In many types of tumours, raised vascular endothelial
growth factor production can frequently be detected in
Molecular biology tumour cells located in the extreme periphery of the
At the genetic level, pancreatic cancer is a well- tumour, in which there is no apparent hypoxia and
characterised neoplasm. By contrast, the molecular acidosis. These observations are consistent with the
mechanisms linking the genetic changes to the aggressive finding that exogenous factors such as hormones,
nature of this disease remain poorly understood. The cytokines, and growth factors modulate expression of
biology of pancreatic cancer is thought to be related to vascular endothelial growth factor, thereby affecting
mutation and inactivation of oncogenes and tumour angiogenesis.52 Also, many tumour cells can constitutively
suppressor genes, as well as abnormalities in growth express vascular endothelial growth factor in vitro with no
factors and their receptors, which affect the downstream apparent external stimuli, which is consistent with the
signal transduction pathways involved in the control of finding that loss or inactivation of tumour-suppressor
growth and differentiation.39 These perturbations confer a genes and activation of oncogenes are associated with
tremendous survival and growth advantage to pancreatic- overexpression of vascular endothelial growth factor.53 In
cancer cells, as manifested by development of invasive and fact, analyses of vascular endothelial growth factor
metastatic phenotypes that are resistant to all promoters have revealed several potential transcription-
conventional treatments. factor binding sites, such as HIF-1, AP-1, AP-2, Egr-1,
Studies have established that human pancreatic cancer Sp1, and many others,54 which suggests that multiple
overexpresses many growth factors and their receptors, signal-transduction pathways are involved in transcription
including the epidermal growth factor family,41 vascular regulation of this growth factor. For example, the
endothelial growth factor,42,43 fibroblast growth factor,44 differential constitutive Sp1 activation is essential for
and many cytokines, such as transforming growth different degrees of vascular endothelial growth factor
factor ␤,45 interleukin 1,46 interleukin 6,47 tumour expression.43 Constitutively activated Stat3 also directly
necrosis factor ␣,48 and interleukin 8.49 The abundance of contributes to the constitutive vascular endothelial
growth-promoting factors and the disturbance of growth- growth factor expression in human pancreatic-cancer
inhibitory factors lead to evasion of programmed cell cells.55 Without apparent external stimuli, human
death, self-sufficiency in growth signals, angiogenesis, pancreatic cancer also constitutively expresses interleukin
and metastasis. 8 through constitutively activated NF-␬B and AP-1.51,56
Some mechanisms have already been identified for the All of these factors might contribute to the aggressive
aberrant expression of these cytokines. For example, growth and drug resistance characteristic of pancreatic
increasing evidence suggests that expression of vascular adenocarcinoma. An important focus of current
endothelial growth factor is regulated mainly by hypoxia, pancreatic-cancer research seeks to understand the
which is a common feature of most solid tumours, upstream molecular mechanisms leading to constitutive
including pancreatic cancers. Shi and colleagues50 activation of these transcription factors.

THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com 1051

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

Cytokines, growth less common manifestation is pancreatitis,

Activated oncogenes factors, hypoxia, which can occur if there is substantial
and inactivated tumour- acidosis, and obstruction of the pancreatic duct.58
suppressor genes free radicals Therefore, for patients with no known risk
factor for pancreatitis, a pancreatic
Constitutive Inducible neoplasm should be considered as a
(genetic) (epigenetic) potential underlying cause. For pancreatic-
AP-1, NF-␬B, Sp1, Stat3 tail lesions initial symptoms might be
related to the primary tumour, with pain in
the left side of the abdomen or left upper
Downstream genes expressed quadrant, but these patients more
commonly present with symptoms
attributable to metastatic disease.
Cell cycle Cell survival Adhesion Invasiveness Angiogenesis
Diagnosis and staging
For patients who present with painless
jaundice, the diagnostic work-up is
generally straightforward. CT of the
Tumour growth and metastasis abdomen is recommended as the first
diagnostic procedure rather than endo-
Figure 2: Molecular biology of pancreatic tumour growth and metastasis scopic retrograde cholangiopancreatog-
raphy because the appearance of the biliary
Aberrant expression of multiple-metastases-related tree and the pancreas are better defined before endoscopic
proteins, such as interleukin 8 and vascular endothelial retrograde cholangiopancreatography and stent place-
growth factor, might result from the alterations of several ment. Once the biliary tree has been manipulated,
transcription-factor activities. To synthesise these visualisation of small tumours might be obscured on CT
observations two potential pathways are proposed because of the presence of the stent or inflammatory
(figure 2). One mechanism relates to the genetic changes caused by instrumentation of the bile duct. After
mutations of oncogenes, suppressor genes, or both, such the pancreatic and peripancreatic anatomy has been
as K-ras and TP53, resulting in constitutive activation of defined on CT, endoscopic retrograde cholangiopan-
the transcription factors. This effect could be especially creatography with stent placement is appropriate to
true in the early stage of pancreatic cancer growth. In the manage obstructive jaundice.59 This approach is not,
late stage of pancreatic cancer development, however, however, without controversy. In a meta-analysis,
important stress factors, such as hypoxia and acidosis, Sewnath and colleagues60 suggested that preoperative
which are frequently encountered in the tumour biliary stenting carries no benefit and should not be done
microenvironment, further upregulate those metastases- routinely. This controversy is inextricably linked to the
related proteins through activation of many transcription uncertainty surrounding the relative usefulness of
factors. Therefore, at advanced stages, uncontrolled neoadjuvant compared with adjuvant treatment for
tumour growth and the consequent development of a resectable pancreatic cancer. A prospective trial of the role
stress environment might increase tumour angiogenesis, of preoperative biliary decompression in patients with
growth, and development of metastases. Understanding resectable pancreatic cancer will be needed before
the expression and regulation of these molecules might consensus can be achieved.61,62 Currently, for patients who
shed more light on the pathophysiology of pancreatic seem to have a localised, potentially resectable neoplasm,
cancer, and suggest new targets for preventive and placement of a plastic stent by an experienced
treatment approaches to pancreatic cancer. gastroenterology team and subsequent referral to a high-
volume pancreatic surgical centre or immediate surgical
Diagnosis and management intervention is recommended. If tumours are locally
For most patients diagnosed as having cancer of the advanced and unresectable or if metastatic disease is
exocrine pancreas life expectancy is measured in months. present, insertion of an expandable metal stent might be
Three factors underlie this poor outlook. First, pancreatic preferable for durability, compared with plastic stents.63
cancer disseminates to distant sites early in its natural When pain is the presenting symptom without jaundice,
history. Second, as the disease progresses it is associated the differential diagnosis possibilities are broader.
with substantial morbidity, characterised by cachexia and Nevertheless, CT scanning will generally be required
asthaenia. Third, pancreatic cancer is resistant to most ultimately to visualise pathology within the pancreas if
forms of treatment studied to date. other diagnostic assessments do not explain the patient’s
symptoms.
Symptoms Once a pancreatic mass has been identified, we prefer
For tumours located in the head and body of the pancreas to make a tissue diagnosis. Tissue can be obtained by CT-
symptoms are generally precipitated by compression of guided fine-needle aspiration, transabdominal ultrasound-
surrounding structures: the bile duct, mesenteric and guided fine-needle aspiration, or fine-needle aspiration
coeliac nerves, the pancreatic duct, and the duodenum.57 under endoscopic ultrasound guidance.64 For patients who
These effects eventually bring patients to medical have potentially resectable pancreatic cancers, endoscopic
attention and the diagnosis of pancreatic cancer can ultrasound-guided biopsy offers the opportunity to
generally be made quickly. When a pancreatic-head visualise the pancreatic mass, to judge its relation with the
tumour is quite small, painless jaundice might be the only surrounding vasculature, and to obtain a tissue diagnosis
sign. Many patients however, experience an antecedent without the risk of tumour seeding along the needle
period of abdominal or back pain, followed by the tract.65 With use of these techniques, pancreatic biopsy
development of obstructive jaundice. Other signs can be samples obtained during laparotomy are rarely required
the development of diabetes mellitus or malabsorption. A and should be discouraged. For patients presenting with

1052 THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

liver metastases and an obvious pancreatic mass, liver gemcitabine, and the Radiation Therapy Oncology Group
biopsy is an appropriate alternative, and if positive for trial (R9704) in the USA, in which patients were
adenocarcinoma, is acceptable as evidence of metastatic randomised to postoperative gemcitabine or fluorouracil
pancreatic cancer. in conjunction with fluorouracil-based chemoradiation,
Thin-cut dynamic multiphase helical CT scan of the will shed more light on this controversial area of oncology
abdomen and pelvis is the most important staging study.66 practice.
Such imaging can generally show the tumour and its
relation to the surrounding structures, including the Locally advanced or unresectable pancreatic cancer
superior mesenteric artery and vein, the portal vein, and Locally advanced pancreatic cancer is defined as a tumour
the coeliac axis. Tumours invading the superior that encases a vascular structure, such as the superior
mesenteric artery or coeliac axis are unresectable since mesenteric artery, coeliac axis, or superior mesenteric
surgery with curative intent (negative margins) is vein-portal vein confluence. Tumours associated with
impossible. Moreover, a positive surgical margin, whether bulky peripancreatic lymphadenopathy are also deemed
gross or microscopic, predicts survival similar to that of unresectable. However, there should be no evidence of
patients who have locally advanced disease. Thus, tumour distant metastatic disease to the chest, liver, or
debulking in this disease has minimum impact on survival, peritoneum. The standard care for unresectable
and is frequently associated with substantial morbidity.67 pancreatic cancer is fluorouracil-based chemoradiation.
Abdominal CT scan is also useful in revealing hepatic This intervention provides a survival advantage and offers
metastases, peritoneal implants, regional adenopathy, and a palliative benefit. In general, chemoradiation should be
ascites. Laparoscopy is frequently recommended to rule considered for patients with adequate performance status
out the presence of small liver or peritoneal metastases for to undergo a course of treatment. In the early to mid-
patients who seem to have resectable disease on the basis 1980s the Gastrointestinal Tumor Study Group
of preoperative imaging studies.68 Currently, with the (GITSG)78 did a three-treatment-group randomised trial
effective use of the diagnostic and staging studies, surgical in patients with locally advanced pancreatic cancer.
procedures, such as exploratory laparotomy or diagnostic Patients received radiation alone to a dose of 60 Gy,
laparotomy, should be avoided. fluorouracil plus intermediate-dose radiation to 40 Gy,
or fluorouracil plus radiation to 60 Gy. Those receiving
Management fluorouracil with radiation had a median survival of
In developing treatment algorithms, pancreatic cancer 42–44 weeks; those undergoing radiation alone had a
should be separated into different patient populations median survival of 23 weeks. Thus, the combination of
based on the extent of disease at presentation. We fluorouracil with radiation doubles survival duration
recommend classifying pancreatic cancer into localised compared with radiation alone. Patients who have locally
and potentially surgically resectable, locally advanced and advanced pancreatic cancer generally experience notable
surgically unresectable, or metastatic. pain. Although the palliative benefit of radiation has not
been extensively studied, radiation seems to provide pain
Resectable disease relief in 50–85% of patients.79 Unfortunately, despite the
Resectable pancreatic cancer is defined, based on potential for palliation with chemoradiation for locally
preoperative work-up, as a pancreatic tumour without advanced disease, this intervention rarely controls
evidence of involvement of the superior mesenteric artery pancreatic cancer. Within months of completing this
or coeliac axis, a patent superior mesenteric-portal venous treatment, patients frequently have evidence of local
confluence, and no evidence of distant metastatic disease. tumour progression (biliary obstruction, relapsing pain,
Around 15–20% of patients have resectable pancreatic gastric outlet obstruction) or new metastatic disease to the
cancer. In optimally staged patients, the 5-year survival liver or peritoneum. Further systemic treatment at that
rate is 20%.69 Recognised prognostic factors include juncture might provide some transient benefit, but the
surgical margin status, nodal status, and tumour size.69 clinical course of such patients is generally poor.
Analysis of large medical databases has established that For these reasons, patients who have locally advanced
the operative mortality of this procedure is significantly disease are an attractive population to consider for
less in high-volume referral centres (⭐4%) compared with treatment with novel agents. These patients harbour
hospitals in which pancreaticoduodenectomy is done subclinical metastatic disease that will eventually become
infrequently.70 In the USA, adjuvant treatment with manifest. With use of molecular treatments the onset of
fluorouracil-based chemoradiation is frequently overt metastatic disease might be delayed and survival
recommended.71,72 However, in the ESPAC-1 trial73 the prolonged without resorting to more toxic conventional
role of chemoradiation as a component of adjuvant agents after chemoradiation.
therapy was questioned, and the researchers supported
chemotherapy instead. Despite its size, the ESPAC-1 trial Metastatic pancreatic cancer
is itself controversial because of the use of different Metastatic pancreatic cancer is a progressive, debilitating
randomisation options based on physicians’ preferences. disease that is characterised by pain, asthaenia, anorexia,
Patients were also allowed to receive additional treatment, and cachexia. Sudden changes in a patient’s clinical status
termed background treatment, which adds additional are common, and patients can have continuing problems
uncertainty to the study’s conclusions. Although the with pain, thromboembolic events, and biliary or gastric
ESPAC-1 approach is argued to represent real-life clinical outlet obstruction. In addition, peritoneal carcinomatosis
practice and its results to be applicable to a wide with intestinal dysmotility or intractable ascites is common
population of patients,74 the combination of rigorous and difficult to manage. Survival is dependent on tumour
clinical research and standard clinical practice make this burden and performance status at presentation. Chemo-
trial difficult to interpret. Other issues including therapy is never curative for metastatic disease, and its
pathology, surgical, and radiation quality control have potential palliative benefit must be carefully weighed
been raised.75,76 Hopefully, the ESPAC-3 trial,77 in which against toxic effects. In the era of accurate cross-sectional
990 patients have been randomly assigned postoperative imaging, single-agent chemotherapy produces low
chemotherapy with fluorouracil and leucovorin or objective response rates.

THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com 1053

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

Gemcitabine is a deoxycytidine analogue with improves locoregional control compared with immediate
structural and metabolic similarities to cytarabine. As a surgery.
prodrug, gemcitabine must be phosphorylated to its
active metabolites gemcitabine diphosphate and gem- Gemcitabine as a radiation sensitiser
citabine triphosphate. Gemcitabine diphosphate inhibits Gemcitabine is a potent radiation sensitiser and has better
ribonucleotide reductase and depletes intracellular pools systemic activity than fluorouracil. Clinical trials of
of all the deoxynucleotide triphosphates necessary for gemcitabine combined with radiation as a component of
DNA synthesis. Gemcitabine triphosphate may be preoperative treatment for patients with locally advanced
incorporated into an elongating DNA chain and leads to and potentially resectable disease are now underway.
premature chain termination. In addition, the Although early results of gemcitabine-based chemo-
triphosphate form of gemcitabine can impede normal radiation regimens are encouraging, this approach
DNA repair, which may explain the drug’s radio- remains investigational, with no clear dose or schedule of
sensitising properties, and its apparent synergy with chemotherapy or radiation currently accepted as
other chemotherapeutic agents. In a randomised trial, standard.86–89
weekly gemcitabine was compared to weekly bolus
fluorouracil in previously untreated patients.80 Survival Gemcitabine as an adjuvant treatment
among patients treated with gemcitabine improved The role of radiation therapy as a component of adjuvant
slightly compared with that among those treated with therapy remains debatable, but it is generally accepted
fluorouracil (median survival 5·6 vs 4·4 months). In that gemcitabine is a better systemic agent than
addition, more clinically meaningful effects on disease- fluorouracil in advanced disease. Therefore, the Radiation
related symptoms (pain, performance status, weight) Therapy Oncology Group is doing a randomised trial of
were seen with gemcitabine than with fluorouracil (24 vs fluorouracil compared with gemcitabine as systemic
4%). Treatment with gemcitabine was associated with a therapy after pancreaticoduodenectomy for pancreatic
survival advantage at 1 year (18%) compared with cancer. All patients have been randomised to either
fluorouracil (2%). Clinical benefit has also been systemic weekly gemcitabine or infusional fluorouracil
documented for patients who were treated with before and after fluorouracil-based chemoradiation. The
gemcitabine after experiencing disease progression while preliminary results are anticipated within the next
receiving fluorouracil.81 However, overall survival for 1–2 years.
these patients receiving gemcitabine as a second-line
treatment is poor. Gemcitabine is currently the standard Gemcitabine-based combination chemotherapy
care for metastatic pancreatic cancer. Since gemcitabine alone has little activity in pancreatic
cancer, gemcitabine in combination with other cytotoxic
Future directions in clinical management agents is being investigated. These studies include
Preoperative chemoradiation for resectable pancreatic combining gemcitabine with docetaxel, cisplatin, oxali-
cancer platin, fluorouracil, and irinotecan.90–98 In phase II trials to
In single-institution and multicentre trials 20–30% of date, the response rate for gemcitabine combinations have
patients who undergo pancreaticoduodenectomy are not generally been higher than with gemcitabine alone.
eligible for postoperative chemoradiation. Therefore, However, no randomised phase III trial has yet
preoperative chemoradiation for patients with resectable established a survival benefit for combination therapy
disease has been advocated by some groups. The rationale compared with gemcitabine alone.
for this approach is sound. First, all patients who have
potentially resectable disease receive the treatment. Targeted therapy
Second, a subset of patients will develop overt metastatic Although new conventional cytotoxic agents or other
disease in the 8–12 weeks of preoperative treatment and, gemcitabine combinations might improve survival for
therefore, become non-surgical patients and will be spared patients with pancreatic cancer, they are likely to produce
the morbidity of a pancreaticoduodenectomy. Finally, small incremental advances. Thus, intense interest has
findings from trials of preoperative treatment suggest focused on the emerging molecular biology of pancreatic
lower rates of locoregional failure with this approach. cancer. Molecular defects correlating with tumour
Most preoperative regimens consist of treatment with growth, resistance, invasion, and angiogenesis have been
combined methods, with use of various radiosensitisers, elucidated. As molecular targets are identified,
including fluorouracil, cisplatin, paclitaxel, and interventions with specific agents might be targeted to
gemcitabine.82–85 The toxic effects of this treatment improve tumour control (table 2).99
approach is notable, with hospital admission being Trials have been done of several biological agents for
required for up to 50% of patients. However, although a pancreatic cancer. Although early efforts with farnesyl
survival benefit has not been established for patients transferase inhibitors have proved disappointing,100 more
undergoing resection, such an approach does keep promising examples of this strategy include inhibition of
unnecessary surgery to a minimum, and probably members of the ErbB family of receptors, ErbB-1

Activity Drug
Target
Epidermal growth-factor receptor Signal transduction or proliferation, ? radioprotection Monoclonal antibody: C225; receptor tyrosine
and angiogenesis kinase inhibitors: ZD1839 and OSI-774
HER2/neu Signal transduction or proliferation Monoclonal antibody: trastuzumab
Vascular endothelial growth factor Angiogenesis Monoclonal antibody: bevacizumab; receptor
tyrosine kinase inhibitors: SU5416, SU6668, and others
K-ras oncogene Signal transduction or proliferation Farnesyl transferase inhibitors and antisense oligonucleotides
Cox-2 Prostaglandin synthesis Celecoxib and rofecoxib
Matrix metalloproteinases Invasion Matrix metalloproteinases inhibitors: marimastat and others
Table 2: Molecular targets for potential exploitation with novel agents

1054 THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

(epidermal growth factor receptor) and ErbB-2 (HER2), factors and pancreatic cancer risk in a cohort of male smokers.
with monoclonal antibodies. Cetuximab, a humanised Am J Epidemiol 2001; 153: 680–87.
monoclonal antibody to epidermal growth factor receptor, 9 Kauppinen T, Partanen T, Degerth R, et al. Pancreatic cancer and
occupational exposures. Epidemiology 1995; 6: 498–502.
and trastuzumab, an antibody to HER2, have been 10 Garabrani DH, Held J, Langholz B, et al. DDT and related
combined with gemcitabine in patients with advanced compounds and risk of pancreatic cancer. J Natl Cancer Inst 1992;
disease.101,102 Early results suggest increased activity with 84: 764–71.
these two drugs, and with cetuximab a longer period of 11 Fryzek J, Garabrant DH, Harlow SD, et al. A case-control study of
disease stabilisation. Other small molecules that inhibit self-reported exposures to pesticides and pancreas cancer in Southern
Michigan. Int J Cancer 1997; 72: 62–67.
the epidermal growth factor receptor, such as ZD1839
12 Wang MY, Abbruzzese JL, Friess H, et al. DNA adducts in human
and OSI-774 are also being investigated for activity in pancreatic tissues and their potential role in carcinogenesis. Cancer Res
advanced pancreatic cancer. Another group of drugs with 1998; 58: 38–41.
potential therapeutic application include the matrix 13 Li D, Firozi PF, Zhang WQ, et al. DNA adducts, genetic
metalloproteinase inhibitors, which are designed to inhibit polymorphisms, and K-ras mutation in human pancreatic cancer.
Mutat Res 2002; 513: 37–48.
the degradative enzymes central to pancreatic cancer cell
14 Thompson PA, Seyedi F, Lang NP, et al. Comparison of DNA
invasion, and possibly tumour angiogenesis. The initial adduct levels associated with exogenous and endogenous exposures in
trials of first-generation metalloproteinase inhibitors have human pancreas in relation to metabolic genotype. Mutat Res 1999;
been disappointing,103 but development of these and other 424: 263–74.
novel agents alone and in combination with gemcitabine 15 Kadlubar FF, Anderson KE, Häussermann S, et al. Comparison of
continues to be a high priority. DNA adduct levels associated with oxidative stress in human
pancreas. Mutat Res 1998; 405: 125–33.
16 Bartsch H, Malaveille C, Lowenfels AB, et al. Genetic polymorphism
The way forward of N-acetyltransferases, glutathione S-transferase M1 and
Pancreatic cancer will remain a challenging problem into NAD(P)H:quinone oxidoreductase in relation to malignant and
the 21st century. However, improvements in early benign pancreatic disease risk: the International Pancreatic Disease
Study Group. Eur J Cancer Prev 1998; 7: 215–23.
detection, screening, and staging of patients will be
17 Liu G, Ghadirian P, Vesprini D, et al. Polymorphisms in GSTM1,
expected to facilitate progress in the management of GSTT1 and CYP1A1 and risk of pancreatic adenocarcinoma.
patients with this disease. Most promising, is the potential Br J Cancer 2000; 82: 1646–49.
to base treatment on our rapidly evolving understanding 18 Duell EJ, Holly EA, Bracci PM, Liu M, Wiencke JK, Kelsey KT.
of the molecular biology of pancreatic cancer. Already A population-based, case-control study of polymorphisms in
various agents are being developed that target signal- carcinogen-metabolizing genes, smoking, and pancreatic
adenocarcinoma risk. J Natl Cancer Inst 2002; 94: 297–306.
transduction pathways or nuclear transcription factors. In 19 Duell EJ, Holly EA, Bracci PM, Wiencke JK, Kelsey KT. A
addition to the agents we describe, farnesyl transferase population-based study of the Arg399Gln polymorphism in X-ray
inhibitors, that target the RAS oncoprotein, Raf-1 repair cross-complementing group 1 (XRCC1) and risk of pancreatic
inhibitors, NF-␬B inhibitors, and Sp-1 inhibitors are adenocarcinoma. Cancer Res 2002; 62: 4630–36.
being studied and developed. Additional progress in 20 Jiao L, Firozi PF, Connor T, Li, D. Codon 143 polymorphism of
AGT gene in pancreatic cancer. Proc Am Assoc Cancer Res 2001; 42:
understanding the nature and sequence of the molecular A1844.
events in the development of carcinoma of the pancreas 21 Malats N, Porta M, Corominas JM, et al. K-ras mutations in exocrine
ultimately will permit the development of an array of pancreatic cancer: association with clinico-pathological characteristics
treatments that will inhibit specific pathways that mediate and with tobacco and alcohol consumption. Int J Cancer 1997; 70:
the aggressive biology of pancreatic cancer. 661–67.
22 Berger DH, Chang H, Wood M, et al. Mutational activation of K-ras
in non-neoplastic exocrine pancreatic lesions in relation to cigarette
Conflict of interest statement smoking status. Cancer 1999; 85: 326–29.
None declared.
23 Porta M, Malats N, Jariod M, et al. Serum concentrations of
organochlorine compounds and K-ras mutations in exocrine
Acknowledgments pancreatic cancer. Lancet 1999; 354: 2125–29.
All authors derived some support from the Lustgarten Foundation for 24 Alguacil J, Porta M, Malats N, et al. Occupational exposure to organic
Pancreatic Cancer Research, the Eli Lilly Research Foundation, and the solvents and K-ras mutations in exocrine pancreatic cancer.
Topfer Family Fund for Pancreatic Cancer Research. The funding source Carcinogenesis 2002; 23: 101–06.
had no influence on the decision to submit this manuscript for publication 25 Sohn TA, Yeo CJ. The molecular genetics of pancreatic ductal
or the writing of the report. carcinoma: a review. Surg Oncol 2000; 9: 95–101.
26 Sakorafas GH, Tsiotos GG. Molecular biology of pancreatic cancer:
potential clinical implications. BioDrugs 2001; 15: 439–52.
References 27 Hruban RH, Goggins M, Parsons J, Kern SE. Progression model for
1 Jemal A, Murray T, Samuels A, et al. Cancer statistics, 2003. pancreatic cancer. Clin Cancer Res 2000; 6: 2969–72.
CA Cancer J Clin 2003; 53: 5–26. 28 Almoguerra C, Shibata D, Forrester K, et al. Most human carcinomas
2 Anderson KE, Potter JD, Mack TM. Pancreatic cancer. In: of the exocrine pancreas contain mutant c-K-ras genes. Cell 1988; 53:
Schottenfeld D, Fraumeni JF Jr, eds. Cancer epidemiology and 549–54.
prevention. Oxford: Oxford University Press, 1996: 725–71. 29 Wenger FA, Zieren J, Peter FJ, Jacobi CA, Muller JM. K-ras
3 Fryzek JP, Garabrant DH, Greenson JK, Schottenfeld D. A review mutations in tissue and stool samples from patients with pancreatic
of the epidemiology and pathology of pancreas cancer. cancer and chronic pancreatitis. Langenbecks Arch Surg 1999; 384:
Gastrointest Cancer 1997; 2: 99–110. 181–86.
4 Hruban RH, Petersen GM, Ha PK, Kern SE. Genetics of pancreatic 30 Uehara H, Nakaizumi A, Tatsuta M, et al. Diagnosis of pancreatic
cancer: from genes to families. Surg Oncol Clin N Am 1998; 7: cancer by detecting telomerase activity in pancreatic juice:
1–23. comparison with K-ras mutations. Am J Gastroenterol 1999; 94:
5 Schenk M, Schwartz AG, O’Neal E, et al. Garabrant familial risk of 2513–18.
pancreatic cancer. J Natl Cancer Inst 2001; 93: 640–44. 31 Caldas C, Hahn SA, da Costa LT, et al. Frequent somatic mutations
6 Ji BT, Chow WH, Gridley G, et al. Dietary factors and the risk of and homozygous deletions of the p16 (MTS1) gene in pancreatic
pancreatic cancer: a case-control study in Shanghai, China. adenocarcinoma. Nat Genetics 1994; 8: 27–32.
Cancer Epidemiol Biomarkers Prev 1995; 4: 885–93. 32 Schutte M, Hruban RH, Geradts J, et al. Abrogation of the Rb/p16
7 Ghadirian P, Baillargeon J, Simard A, et al. Food habits and tumor-suppressive pathway in virtually all pancreatic carcinomas.
pancreatic cancer: a case-control study of the francophone Cancer Res 1997; 57: 3126–30.
community in Montreal, Canada. Cancer Epidemiol Biomarkers Prev 33 Wilentz RE, Iacobuzio-Donahue CA, Aragani P, et al. Loss of
1995; 4: 895–99. expression of DPC4 in pancreatic intrapithelial neoplasia (PanIN):
8 Stolzenberg-Solomon RZ, Pietinen P, Barrett MJ, Taylor PR, evidence that DPC4 inactivation occurs late in neoplastic progression.
Virtamo J, Albanes D. Dietary and other methyl-group availability Cancer Res 2000; 60: 2002–06.

THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com 1055

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

34 Hahn SA, Schutte M, Hoque ATMS, et al. DPC4, a candidate efficacy of preoperative biliary drainage for tumors causing obstructive
tumor-suppressor gene at 18q21.1. Science 1996; 271: 350–53. jaundice. Ann Surg 2002; 236: 17–27.
35 Su GH, Hilgers W, Shekher MC, et al. Alterations in pancreatic, 61 Pisters PW, Lee JE, Vauthey JN, Evans DB. Comment and
biliary, and breast carcinomas support MKK4 as a genetically perspective on Sewnath and colleagues’ recent meta-analysis of the
targeted tumor suppressor gene. Cancer Res 1998; 58: 2339–42. efficacy of preoperative biliary drainage for tumors causing obstructive
36 Su GH, Hruban RH, Bansal RK, et al. Germline and somatic jaundice. Ann Surg 2003; 237: 594–95.
mutations of the STK11/LKB1 Peutz-Jeghers gene in pancreatic and 62 Gouma DJ, Obertop H. Comment and perspective on Sewnath and
biliary cancers. Am J Pathol 1999; 154: 1835–84. colleagues’ recent meta-analysis of the efficacy of preoperative biliary
37 Ozkelik H, Schmocker B, Di Nicola N, et al. Germline BRCA2 drainage for tumors causing obstructive jaundice. Ann Surg 2003;
6174delT mutations in Ashkenazi Jewish pancreatic cancer patients. 237: 595–96.
Nat Genet 1997; 16: 17–18. 63 Davids PH, Groen AK, Rauws EA, et al. Randomized trial of self-
38 Goggins M, Schutte M, Lu J, et al. Germline BRCA2 gene mutations expanding metal stents versus polyethylene stents for distal malignant
in patients with apparently sporadic pancreatic carcinomas. biliary obstruction. Lancet 1992; 340: 1488–92.
Cancer Res 1996; 56: 5360–64. 64 Di Stasi M, Lencioni R, Solmi L, et al. Ultrasound-guided fine needle
39 Goggins M, Hruban RH, Kern SE. BRCA2 is inactivated late in the biopsy of pancreatic masses: results of a multicenter study.
development of pancreatic intraepithelial neoplasia: evidence and Am J Gastroenterol 1998; 93: 1329–33.
implications. Am J Pathol 2000; 156: 1767–71. 65 Gress F, Gottlieb K, Sherman S, Lehman G. Endoscopic
40 Rozenblum E, Schutte M, Goggins M, et al. Tumor-suppressive ultrasonography-guided fine-needle aspiration biopsy of suspected
pathways in pancreatic carcinoma. Cancer Res 1997; 57: 1731–34. pancreatic cancer. Ann Intern Med 2001; 134: 459–64.
41 Korc M. Role of growth factors in pancreatic cancer. 66 Fuhrman GM, Charnsangavej C, Abbruzzese JL, et al. Thin-section
Surg Oncol Clin N Am 1998; 7: 25–41. contrast-enhanced computed tomography accurately predicts the
resectability of malignant pancreatic neoplasms. Am J Surg 1994;
42 Luo J, Guo P, Matsuda K, et al. Pancreatic cancer cell-derived
167: 104–11.
vascular endothelial growth factor is biologically active in vitro
and enhances tumorigenicity in vivo. Int J Cancer 2001; 92: 67 Millikan KW, Deziel DJ, Silverstein JC, et al. Prognostic factors
361–69. associated with resectable adenocarcinoma of the head of the
pancreas. Am Surg 1999; 65: 618–24.
43 Shi Q, Le X, Peng Z, et al. Constitutive Sp1 activity is essential for
differential constitutive expression of vascular endothelial growth 68 Jimenez RE, Warshaw AL, Fernandez-Del Castillo C. Laparoscopy
factor in human pancreatic adenocarcinoma. Cancer Res 2001; 61: and peritoneal cytology in the staging of pancreatic cancer.
4143–54. J Hepatobiliary Pancreat Surg 2000; 7: 15–20.
44 Yamanaka Y, Friess H, Buchler M, et al. Overexpression of acidic 69 Yeo CJ, Cameron JL, Lillemore KD, et al. Pancreaticoduodenectomy
and basic fibroblast growth factors in human pancreatic cancer for cancer of the head of the pancreas: 201 patients. Ann Surg 1995;
correlates with advanced tumor stage. Cancer Res 1993; 53: 5289–96. 221: 721.
45 Kleeff J, Ishiwata T, Friess H, et al. The TGF-␤ signaling inhibitor 70 Lieberman MD, Kilburn H, Lindsey M, et al. Relation of
Smad7 enhances tumorigenicity in pancreatic cancer. Oncogene 1999; perioperative deaths to hospital volume among patients undergoing
pancreatic resection for malignancy. Ann Surg 1995; 222: 638–45.
18: 5363–72.
71 Gastrointestinal Tumor Study Group. Further evidence of effective
46 Blanchard JA 2nd, Barve S, Joshi-Barve S, Talwalker R, Gates LK Jr.
adjuvant combined radiation and chemotherapy following curative
Cytokine production by CAPAN-1 and CAPAN-2 cell lines.
resection of pancreatic cancer. Cancer 1987; 59: 2006–10.
Dig Dis Sci 2000; 45: 927–32.
72 Klinkenbijl JH, Jeekel J, Sahmoud T, et al. Adjuvant radiotherapy
47 Saito K, Ishikura H, Kishimoto T, et al. Interleukin-6 produced by
and 5-fluorouracil after curative resection for the cancer of the
pancreatic carcinoma cells enhances humoral immune responses
pancreas and peri-ampullary region: phase III trial of the EORTC
against tumor cells: a possible event in tumor regression. Int J Cancer
Gastrointestinal Tract Cancer Cooperative Group. Ann Surg 1999;
1998; 75: 284–89.
230: 776–84.
48 Watanabe N, Tsuji N, Kobayashi D, et al. Endogenous tumor
73 Neoptolemos JP, Dunn JA, Stocken DD, et al. Adjuvant
necrosis factor functions as a resistant factor against hyperthermic
chemoradiotherapy and chemotherapy in resectable pancreatic
cytotoxicity in pancreatic carcinoma cells via enhancement of the
cancer: a randomised controlled trial. Lancet 2001; 358: 1576–85.
heart shock element-binding activity of heat shock factor 1.
Chemotherapy 43: 406–14, 1997. 74 Neoptolemos JP, Stocken D, Dunn JA. ESPAC-1 trial of adjuvant
therapy for resectable adenocarcinoma of the pancreas. Ann Surg
49 Shi Q, Abbruzzese J, Huang S, Fidler IJ, Xie K. Constitutive and
2002; 236: 694–96.
inducible interleukin-8 expression by hypoxia and acidosis renders
75 Evans DB, Hess KR, Pisters PW. ESPAC-1 trial of adjuvant therapy
human pancreatic cancer cells more tumorigenic and metastatic.
for resectable adenocarcinoma of the pancreas. Ann Surg 2002; 236:
Clin Cancer Res 1999; 5: 3711–21.
694–96.
50 Shi Q, Le X, Wang B, et al. Regulation of vascular endothelial growth
76 Abrams RA, Lillemoe KD, Piantadosi S. Continuing controversy
factor expression by acidosis in human cancer cells. Oncogene 2001;
over adjuvant therapy of pancreatic cancer. Lancet 2001; 358:
20: 3751–56.
1565–66.
51 Xie K. Interleukin-8 and human cancer biology. Cytokine Growth
77 Neoptolemos JP, Cunningham D, Friess H, et al. Adjuvant therapy in
Factor Rev 2001; 12: 375–91.
pancreatic cancer: historical and current perspectives. Ann Oncol
52 Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z. Vascular 2003; 14: 675–92.
endothelial growth factor (VEGF) and its receptors. FASEB J 1999;
78 Moertel CG, Frytak S, Hahn RG, et al. Therapy of locally
13: 9–22.
unresectable pancreatic carcinoma: a randomized comparison of high
53 Rak J, Filmus J, Finkenzeller G, Grugel S, Marme D, Kerbel RS. dose (6000 rads) radiation alone, moderate dose radiation (4000 rads
Oncogenes as inducers of tumor angiogenesis. Cancer Metastasis Rev + 5-fluorouracil), and high dose radiation + 5-fluorouracil. Cancer
1995; 14: 263–77. 1981; 48: 1705–10.
54 Tischer E, Mitchell R, Hartman T, et al. The human gene for 79 Minsky BD, Hilaris B, Fuks Z. The role of radiation therapy in the
vascular endothelial growth factor: multiple protein forms are control of pain from pancreatic carcinoma. J Pain Symptom Manage
encoded through alternative exon splicing. J Biol Chem 1991; 266: 1988; 3: 199–205.
11947–54. 80 Burris HA III, Moore MJ, Andersen J, et al. Improvements in survival
55 Wei D, Le X, Zheng L, et al. Stat3 activation regulates the expression and clinical benefit with gemcitabine as first-line therapy for patients
of vascular endothelial growth factor and human pancreatic cancer with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997;
angiogenesis and metastasis. Oncogene 2003, 22: 319–29. 15: 2403–13.
56 Le X, Shi Q, Wang B, et al. Molecular regulation of constitutive 81 Rothenberg ML, Moore MJ, Cripps MC, et al. A phase II trial of
expression of interleukin-8 in human pancreatic adenocarcinoma. gemcitabine in patients with 5-FU-refractory pancreas cancer.
J Interferon Cytokine Res 2000; 20: 1532–40. Ann Oncol 1996; 7: 347–53.
57 Evans DB, Abbruzzese JL, Rich TR. Cancer of the pancreas. In: 82 Hoffman JP, Lipsitz S, Pisansky T, Weese JL, Solin L, Benson AB.
DeVita VT, Hellman S, Rosenberg SA, eds. Cancer principles and Phase II trial of preoperative radiation therapy and chemotherapy for
practice of oncology, 5th edn. Philadelphia: Lippincott Raven, 1997: patients with localized, resectable adenocarcinoma of the pancreas: an
1054–87. Eastern Cooperative Oncology Group Study. J Clin Oncol 1998; 16:
58 Lin A. Pancreatic carcinoma as a cause of unexplained pancreatitis: 317–23.
report of ten cases. Ann Intern Med 1990; 113: 166–67. 83 White R, Lee C, Anscher M, et al. Preoperative chemoradiation for
59 Pisters PW, Hudec WA, Hess KR, et al. Effect of preoperative biliary patients with locally advanced adenocarcinoma of the pancreas.
decompression on pancreaticoduodenectomy-associated morbidity in Ann Surg Oncol 1999; 6: 38–45.
300 consecutive patients. Ann Surg 2001; 234: 47–55. 84 Spitz FR, Abbruzzese JL, Lee JE, et al. Preoperative and
60 Sewnath ME, Karsten TM, Prins MH, et al. A meta-analysis on the postoperative chemoradiation strategies in patients treated with

1056 THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.
 SEMINAR

pancreaticoduodenectomy for adenocarcinoma of the pancreas. responses in patients with previously untreated advanced pancreatic
J Clin Oncol 1997; 15: 928–37. cancer. J Clin Oncol 2002; 20: 1182–91.
85 Pisters PW, Wolff RA, Janjan NA, et al. Preoperative paclitaxel and 95 Colucci G, Giuliani F, Gebbia V, et al. Gemcitabine alone or with
concurrent rapid-fractionation radiation for resectable pancreatic cisplatin for the treatment of patients with locally advanced and/or
adenocarcinoma: toxicities, histologic response rates, and event-free metastatic pancreatic carcinoma: a prospective, randomized phase III
outcome. J Clin Oncol 2002; 20: 2537–44. study of the Gruppo Oncologia dell’Italia Meridionale. Cancer 2002;
86 Wolff RA, Evans DB, Crane CH, et al. Initial results of pre- 94: 902–10.
operative gemcitabine-based chemoradiation for resectable 96 Berlin JD, Catalano P, Thomas JP, et al. Phase III study of
pancreatic adenocarcinoma. Proc Am Soc Clin Oncol 2002; 21: gemcitabine in combination with fluorouracil versus gemcitabine alone
130a (abstr 516). in patients with advanced pancreatic carcinoma: Eastern Cooperative
87 Wolff RA, Evans DB, Gravel DM, et al. Phase I trial of gemcitabine Oncology Group Trial E2297. J Clin Oncol 2002; 20: 3270–75.
combined with radiation for the treatment of locally advanced 97 Stehlin JS, Giovanella BC, Natelson EA, et al. A study of
pancreatic adenocarcinoma. Clin Cancer Res 2001; 7: 2246–53. 9-nitrocamptothecin (RFS 2000) in patients with advanced pancreatic
88 Blackstock AW, Bernard SA, Richards F, et al. Phase I trial of cancer. Int J Oncol 1999; 14: 821–31.
twice-weekly gemcitabine and concurrent radiation in patients with 98 Sharma S, Kemeny N, Schwartz GK, et al. Phase I study of
advanced pancreatic cancer. J Clin Oncol 1999; 17: 2208–12. topoisomerase I inhibitor exatecan mesylate (DX-8951f) given as
89 McGinn CJ, Zalupski MM, Shureiqi I, et al. Phase I trial of radiation weekly 24-hour infusions three of every four weeks. Clin Cancer Res
dose escalation with concurrent weekly full-dose gemcitabine in 2001; 7: 3963–70.
patients with advanced pancreatic cancer. J Clin Oncol 2001; 19: 99 Fan Z, Mendelsohn J. Therapeutic application of anti-growth factor
4202–08. receptor antibodies. Curr Opin Oncol 1998; 10: 67–73.
90 Ryan DP, Kulke MH, Fuchs CS, et al. A phase II study of 100 Van Cutsem E, Karasek P, Oettle H, et al. Phase III trial comparing
gemcitabine and docetaxel in patients with metastatic pancreatic gemcitabine plus R115777 (Zarnestra) versus gemcitabine plus
carcinoma. Cancer 2002; 94: 97–103. placebo in advanced pancreatic cancer (PC). Proc Ann Meeting Am Soc
91 Philip PA, Zalupski MM, Vaitkevicius VK, et al. Phase II study of Clin Oncol 2002, 21: 130 (abstr).
gemcitabine and cisplatin in the treatment of patients with advanced 101 Abbruzzese JL, Rosenberg A, Xiong Q, et al: Phase II study of
pancreatic carcinoma. Cancer 2001; 92: 569–77. anti-epidermal growth factor receptor cetuximab (IM-C225) in
92 Louvet C, Andre T, Lledo G, et al. Gemcitabine combined with patients with advanced pancreatic cancer. Proc Am Soc Clin Oncol
oxaliplatin in advanced pancreatic adenocarcinoma: final results of a 2001; 20: 130a (abstr 518).
GERCOR multicenter phase II study. J Clin Oncol 2002; 20: 102 Safron H, Ramanathan R, Schwartz J, et al. Herceptin and
1512–18. gemcitabine for metastatic pancreatic cancers that overexpress
93 Louvet C, Andre T, Hammel P, et al. Phase II trial of bimonthly HER-2/neu. Proc Am Soc Clin Oncol 2001; 20: 130a (abstr 517).
leucovorin, 5-fluorouracil and gemcitabine for advanced pancreatic 103 Bramhall SR, Schulz J, Nemunaitis J, et al. A double-blind placebo-
adenocarcinoma (FOLFUGEM). Ann Oncol 2001; 12: 675–79. controlled, randomized study comparing gemcitabine and marimastat
94 Rocha Lima CM, Savarese D, Bruckner H, et al. Irinotecan plus with gemcitabine and placebo as first line therapy in patients with
gemcitabine induces both radiographic and CA 19-9 tumor marker advanced pancreatic cancer. Br J Cancer 2002; 87: 161–67.

THE LANCET • Vol 363 • March 27, 2004 • www.thelancet.com 1057

For personal use. Only reproduce with permission from The Lancet.