Cancer Therapy Vol 2, 329-344, 2004
Imaging of pancreatic cancer: a promise for early
diagnosis through targeted strategies
Research Article
Zdravka Medarova, Anna Moore*
Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, USA
__________________________________________________________________________________
*Correspondence: Anna Moore, Ph.D., MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Bldg. 149,
13th St, Rm. 2301, Charlestown, MA 02129; tel: (617)724-0540; fax: (617)726-7422; E-mail:amoore@helix.mgh.harvard.edu
Key words: Clinical imaging of pancreatic cancer, Molecular imaging of pancreatic cancer,
Abbreviations: Abdominal ultrasonography, (US); charged coupled device, (CCD); Computerized tomography, (CT); dextran coated iron oxide, (CLIO); Endoscopic retrograde
cholangiopancreatography, (ERCP); Endoscopic ultrasonography guided fine needle aspiration, (EUS-FNA); fluorescence mediated tomography, (FMT); Magnetic resonance cholangio
pancreatography, (MRCP); Magnetic resonance imaging, (MRI); matrix metalloproteinases, (MMPs); multidetector CT, (MDCT); Pancreatic intraepithelial neoplasia, (PanIN); Positron
emission tomography, (PET); single photon emission computed tomography, (SPECT)
Received: 24 August 2004; Accepted: 08 September 2004; electronically published: September 2004
Summary
Pancreatic cancer is a disease characterized by high invasiveness, acute resistance to chemo- and radiotherapy, and consequently
represents one of the most difficult malignancies to detect and treat. As a result, the diagnosis of pancreatic cancer is essentially a
death sentence. Current strategies for pancreatic cancer detection using non-invasive imaging techniques rely on non-targeted
morphological approaches. Combining the present knowledge about the molecular nature of the disease on the one hand, and the
capacity of imaging to provide a real-time, global view of processes within a living organism, on the other, molecular imaging has the
potential to greatly advance our ability for early detection of pancreatic cancer. This review will discuss a targeted imaging approach
to explore the specific molecular events during carcinogenesis.
I. Introduction
At the beginning of the 21st century, pancreatic cancer remains one of the most devastating human cancers, responsible for 30,000 deaths per
year in the USA . The lack of specific symptoms until late in the pathology, as well as the high proliferative and metastatic potential of the
disease hamper early diagnosis and consequently, effective treatment. As a result, disease prognosis is poor, with a 5-year survival rate of
only around 4% .
The only potentially curative treatment remains surgical resection. At the time of presentation, approximately 15-20% of patients have
resectable disease. Nevertheless, despite the possibility for radical intervention, due to the aggressive nature of the disease, only 20% of these
patients survive to 5 years. Adjuvant chemotherapy and radiation therapy only provide marginal palliative benefits due to the low
chemosensitivity of the disease with response rates to conventional agents of less than 10% .
Despite significant efforts to improve the prognosis associated with this disease, progress towards the management of pancreatic cancer over
the past 50 years has been limited and still close to 100% of patients diagnosed with pancreatic cancer will develop metastases and die. This
outlook, however, can be improved dramatically by the availability of early diagnostics. An essential element of diagnosis is represented by
non-invasive imaging. Imaging techniques are in active development both in the laboratory and the clinic, since the conception of novel
approaches ultimately leading to therapy for pancreatic cancer is an urgent priority.
II. Pancreatic cancer: molecular course
The progression to pancreatic cancer appears to follow the Vogelstein model described for colorectal cancers . Namely, tumorigenesis is the
result of the violation of multiple checkpoints controlling cell proliferation, differentiation, and death. This process, characterized by serial
perturbations in a variety of cell signaling pathways, ultimately leads to the acquisition of a survival and growth advantage by select cells
which escape the normal safeguards to uncontrollable proliferation and invasiveness, and consequently results in carcinogenesis .
Pancreatic cancer, in particular, is believed to originate from mutations in pancreatic duct cells. Pancreatic intraepithelial neoplasia (PanIN) is
defined as the increased incidence of abnormal ductal structures seen in patients with pancreatic carcinoma. The similarity in spatial
distribution between PanIN lesions and malignant tumors led to the hypothesis that these lesions in fact represent incipient pancreatic
adenocarcinomas. Morphologically, PanIN lesions are quite heterogeneous and can be classified based on graded stages of increasingly
dysplastic growth .
Both tumor suppressor genes, responsible for arresting cell growth at critical points in the cell's life, and oncogenes, promoting cell division,
have been identified as possible culprits for the development of pancreatic cancer. Some of these have been reviewed in more detail .
One of the most frequently mutated genes in pancreatic cancer is the k-ras oncogene, altered in nearly 90% of pancreatic malignancies . k-ras
is a key element in cell-growth signaling cascades, linking growth factor and hormone receptors to downstream mediators of cell proliferation
and differentiation. In pancreatic carcinomas, k-ras mutations are generally the first detected alterations in the progression series. In fact, kras mutations are seen in about 30% of early pancreatic lesions and increase in frequency with disease progression . However, k-ras, is also
altered in some benign conditions, which makes it a poor candidate for a diagnostic marker.
The tumor suppressor gene, p16, which is part of the well-characterized Rb tumor-suppressive pathway, responsible for global cell-cycle
regulation, is also commonly mutated in pancreatic malignancy. p16 inactivation has been identified in 27% to 82% of primary tumors .
Furthermore, p16 inactivation has been linked to prognosis, altered more frequently in short term survivors (85% of the cases) compared to
long-term survivors (50% of the cases) .
Another tumor suppressor, p53, is inactivated in between 40% and 75% of pancreatic cancers . p53 is a mediator of DNA-damage induced
cell-cycle arrest and programmed cell death. Interestingly, in pancreatic carcinomas, p53 is often found mutated in association with k-ras
mutations, suggesting cooperativity between the two genes in pancreatic cancer tumorigenesis . Still, p53 is a later marker of tumorigenesis,
mutated in PanINs which demonstrate a high degree of dysplasia.
DPC4, a component of the TGF-beta signaling pathway, is deleted in approximately 50% of pancreatic adenocarcinomas . It has been
suggested that mutations in DPC4 correlate with invasiveness . Still, the precise role of this tumor suppressor in the development of
pancreatic cancer is poorly understood.
In addition to classical oncogenes and tumor-suppressor genes implicated in the development of pancreatic cancer, many growth factors and
growth factor receptors have been found to be overexpressed in carcinomas of the pancreas. Among them are: the epidermal growth factor ,
vascular endothelial growth factor , fibroblast growth factor , as well as multiple cytokines, including transforming growth factor beta ,
interleukin 1 , interleukin 6 , tumor necrosis factor alpha , and interleukin 8 reviewed in . For some of these molecules, a direct link has been
established between their upregulated expression and pancreatic cancer. Vascular endothelial growth factor for instance is a key angiogenic
factor regulated by hypoxia, common to most solid tumors, among which is pancreatic cancer . Angiogenesis is one mechanism conducive to
local and systemic expansion of the tumor mass and as a result is implicated in primary tumor growth and indirectly, metastasis.
Pancreatic cancer is characterized by its high invasivenes and metastatic potential. The molecules involved in the metastatic process fall into
two groups: cell adhesion molecules and extracellular proteases. One cell adhesion molecule mutated in pancreatic cancers is E-cadherin . Ecadherin couples adjacent cells via E-cadherin bridges and in that way is directly implicated in the transmission of antigrowth signals in
response to cell-cell contacts. Loss of E-cadherin function is thus one mechanism leading to cancer cell invasion and metastasis. Adhesion
molecules overexpressed in pancreatic cancer include ICAM-1, VCAM-1 , and integrins . Overexpression of these adhesion molecules
imparts to cancer cells the ability to migrate and home in to distant sites where they form metastatic lesions.
Another class of molecules also implicated in cancer invasiveness and metastasis are extracellular proteases. Extracellular proteases facilitate
the invasion of cancer cells into the adjacent stroma, across blood vessels and to a metastatic site by mediating the degradation of the
extracellular matrix. Studies on pancreatic cancer have clearly established the tendency for upregulation of proteases, downregulation of
protease inhibitors and conversion of inactive zymogens into active enzymes . Evidence exists for induction of urokinase expression in
stromal cells and overexpression of a variety of matrix metalloproteinases (MMPs) in pancreatic cancer reviewed in .
Understanding of the genetic alterations contributing to the progression to pancreatic cancer is an essential starting point for future research,
which would hopefully lead to the construction of a detailed map of the intricate molecular interactions characterizing the disease. It is clear
now that the aggressive nature of the disease necessitates early intervention as the most promising treatment strategy. Commonly used
clinical tests for pancreatic cancer tumor markers are serum-based immunoassays for tumor-specific antigens, such as CA19-9, which is a
mucin-associated marker . This test, however, is not specific to pancreatic cancer, since elevated marker levels are also present in pancreatitis
and other cancers. Therefore, timely diagnosis and characterization of the malignancy through non-invasive imaging of the pancreas is an
essential first step towards disease management. Targeted imaging which utilizes the molecular alterations specific to pancreatic cancer, has
the potential to advance our capacity for identification and characterization of the malignancy in its initial stages.
III. Imaging of pancreatic cancer
A. Clinical imaging of pancreatic cancer
Currently pancreatic cancer imaging in the clinic relies on nontargeted morphologically-based modalities. Ultrasonographic methods, such as
abdominal ultrasonography, endoscopic ultrasonography, as well as helical CT and MRI have evolved as the major tools for pancreatic
cancer detection and staging . More recently, these techniques have been complemented by the biochemically-based detection procedure of
FDG-PET.
1. Abdominal ultrasonography (US)
Ultrasonography is the most widely used clinical imaging modality because of its low cost, availability, and safety. Ultrasound images are
obtained when high-frequency (>20-kHz) sound waves are emitted from a transducer placed against the skin and the ultrasound is reflected
back from the internal organs under examination.
Contrast in the images obtained depends on the imaging algorithm used, backscatter, attenuation of the sound, and sound speed. Some of the
drawbacks of ultrasound imaging, however, extend from the presence of bone and air artefacts, due to the tendency of air and bone to not
transmit sound waves. Consequently ultrasound is characterized by limited depth penetration.
Abdominal ultrasonography represents a good initial screening tool. However, US has a fairly low sensitivity and specificity of 67% and
40%, respectively for pancreatic cancer . Furthermore, US is operator-dependent and affected by local artefacts. Nevertheless, US is suitable
for detection of tumors over 2 cm in diameter, dilated biliary and pancreatic ducts, and extrapancreatic spread .
2. Endoscopic ultrasonography
Endoscopic ultrasonography is a sensitive and reliable modality for the detection, staging, and surgical evaluation of pancreatic cancer. It uses
a high-frequency sonographic transducer that is introduced into the gastrointestinal tract by a side-viewing endoscope. This setup permits the
operator to obtain real-time cross-sectional images of the gastrointestinal wall and soft-tissue . Endoscopic ultrasound overcomes some of the
drawbacks of abdominal ultrasonography based on the endoscopic positioning of the ultrasound probe adjacent to the pancreas. This allows
the detailed assessment of local anatomy and permits the identification of small lesions normally not detected by abdominal ultrasound.
The reported sensitivity, specificity, and accuracy of endoscopic ultrasonography are significantly higher than conventional CT, with an
accuracy of 85-100%, compared to 64-66% for CT and 61-64% for abdominal ultrasound . The superiority of endoscopic ultrasonography
over CT is particularly evident for lesions smaller than 3 cm in size.
Furthermore, endoscopic ultrasound is a good staging tool for pancreatic cancer. Generally, staging is based on a TNM classification. T stage
refers to tumor characteristics including vascular involvement. N stage reflects regional lymph node involvement, and M stage assesses
metastatic spread. Endoscopic ultrasound is reliable for the identification of certain types of vascular involvement of the portal and splenic
veins into the tumor. However, despite its sensitivity in terms of T and N staging, endoscopic ultrasound is not suitable for assessing M stage
due to its limited tissue penetration .
3. Endoscopic ultrasonography guided fine needle aspiration (EUS-FNA)
EUS-FNA combines the ability to visualize primary tumors, lymph nodes, and the liver with the capacity to obtain a tissue sample during the
diagnostic procedure. Its sensitivity ranges from 45% to 100%. Its specificity approaches 100% . Still the diagnostic benefits of the procedure
are limited by the ability to visualize the lesion in the first place, since endoscopic ultrasound cannot differentiate between inflamed and
metastasis-bearing lymph nodes as well as focal pancreatitis from a tumor .
The major advantage of this procedure lies in its relative noninvasiveness. Traditionally, the high mortality and morbidity associated with
pancreatidoduodenectomy has discouraged the progress to surgery without tissue diagnosis. The morbidity associated with percutaneous FNA
has been minimal, despite fears that percutaneous biopsy would lead to peritoneal seeding of tumor cells along the needle tract . Obtaining a
tissue sample would allow one to distinguish pancreatitis from malignancy, and in cases where adjuvant chemo- or radiotherapy is
implemented, biopsy of a suspected lesion is necessary and beneficial.
4. Computerized tomography (CT)
CT is the most commonly used modality for the initial diagnosis, staging, and evaluation of response to therapy of pancreatic cancer. The
reported accuracy of CT in determining that a tumor is unresectable approaches 100%. However, about a third of the cases considered
resectable based on CT, in fact are unresectable .
Signal in computed tomography (CT) results from differential absorption of X-rays by component tissues and media, namely, bone, air, fat,
and water. Volumetric data are collected as an X-ray source and a detector rotate around the subject. Limiting factors affecting the level of
resolution of this imaging modality include the pixel sampling size, the size of the X-ray source, and blurring in the phosphor screen which
constitutes an element in the signal detector system . A major drawback of CT is the poor soft tissue contrast, which necessitates the
administration of iodinated contrast agents which pass through different tissues at different rates. CT has a relatively high spatial resolution
(50 m) and is characterized by fast acquisition times . CT is a commonly applied clinical imaging modality and is traditionally used as a
cancer diagnostic tool.
More recently modifications of the CT procedure have shown improved sensitivity and diagnostic accuracy. Helical CT provides thin-section,
motion-free images. It permits imaging of the entire pancreas and tissues adjacent to it in different circulatory phases. The different phases
are defined by variability in scan delay. The pancreatic phase (scan delay of 40 s) has an enhanced capacity to differentiate between
pancreatic parenchyma and blood vessels, compared to portal-vein phase imaging (scan delay 60-70 s) . On the other hand, imaging during
the portal-vein phase is best for imaging of liver metastases. Therefore, the assessment of tumor stage and metastasis is derived from a "dualphase" technique. However, circulation times vary between patients, which is a source of error in the use of this application. The reported
accuracy of helical CT for T staging is 77%, for N staging 58%, and for M staging 79% .
Thin section helical CT reduces the obscuring impact of volume averaging on the detection of small lesions. Typically slices ranging between
3 and 5 mm in thickness are obtained using this procedure .
In general, helical CT can reveal the presence of the tumor and its location in relation to surrounding structures, such as the superior
mesenteric artery and vein, the portal vein, and the coeliac axis. The evaluation of resectability is therefore dependent on the capacity of CT
to determine whether the tumor is invading the superior mesenteric artery and the coeliac axis, as well as to detect liver and distant
metastases, which is an indicator of unresectability .
The latest advance in CT imaging of the pancreas combines volume rendering of CT data with a three-dimensional display and is referred to
as multidetector CT (MDCT). The advantages of this technique lie in the ability of the operator to optimize the visualization of structures,
which allows key elements of the anatomy to be enhanced. Still, the accuracy and reliability of this procedure remain to be determined.
Furthermore, its demand for extensive computer memory and relative expense make it less popular .
5. Endoscopic retrograde cholangio pancreatography (ERCP)
ERCP with stent placement is a relatively invasive procedure which identifies visual symptoms of biliary and pancreatic duct stenosis. It is
recommended for patients who present with symptoms of obstructive jaundice, i.e., renal failure or cholangitis, and is a way of alleviating
biliary obstruction.
ERCP has several drawbacks as a diagnostic tool. Due to the indirect determination of parenchymal abnormalities, a normal pancreatogram
does not exclude the possibility for the presence of a tumor. Chronic pancreatitis and pancreatic cancer cannot be differentiated predictably.
Lesions in certain areas of the pancreas are less likely to be detected by this method .
Improvements in ERCP utilize the capacity of the procedure to obtain tissue specimens from the location of interest. ERCP with secretin
stimulation and brush biopsy have been utilized in order to collect material for further analysis . However, the use of EUS-FNA as a
diagnostic/sampling tool will likely replace ERCP due to its limited invasiveness and enhanced specificity.
6. Magnetic resonance imaging (MRI)
The principle behind MRI is founded upon the tendency of unpaired nuclear spins (dipoles), e.g. hydrogen atoms in water and organic
molecules, to align themselves along an externally applied magnetic field. This external field is produced by a strong magnet surrounding the
subject. Following the magnetic pulse delivered by the magnet, the dipoles return to their baseline orientation. That event is detected as a
change in electromagnetic flux and is characterized by a differential rate of magnetic relaxation depending on the local environment. For
example, fat and hydrocarbon-rich environments have short relaxation times, whereas aqueous environments have relatively long relaxation
times. The measurement of dipole relaxation is translated into an MR signal with contrast provided by the differential nature of relaxation rate
.The most commonly used timing parameters are known as T1 and T2 and reflect the differential relaxation of the dipoles in the longitudinal
and transverse directions, respectively.
Whereas the resolution of MRI is high (10-100  m), its sensitivity is quite low (10-3-10-5 moles/L). Nevertheless, the capacity to derive both
anatomical and molecular/physiologic information simultaneously through MRI make it one of the most promising imaging modalities. Its
application in the clinic is expanding despite its relatively high cost. Furthermore, as a research tool, MRI has been used to image specific
molecular interactions by the use of chemical agents capable of altering MR signal intensity. Paramagnetic metal cations, such as gadolinium
or superaparamagnetic iron oxide nanoparticles have been used as targeted MRI probes . Still, the low sensitivity of MRI makes it necessary
to deliver very high concentrations of probe at the target site in order to achieve sufficient contrast for reliable imaging. Nevertheless,
magnetic resonance imaging has been used for multiple applications included but not limited to cell trafficking , and imaging of gene
expression .
Originally, conventional MRI had a limited diagnostic value for pancreatic cancer due to motion artifacts (respiratory, vascular, peristaltic).
Recently, however, the use of more advanced MRI techniques, such as dynamic contrast enhanced MRI, has led to considerable sensitivity
levels, surpassing even those reported for dual-phase helical CT. A comparative study found that MRI had an accuracy of 96% for predicting
resectability vs. 81% for helical CT . Compared to endoscopic ultrasound, MRI was found to have a positive predictive value of 77% vs. 69%
for EUS. In determining resectability, the negative predictive value (defining unresectability) was 76% . Furthermore, MR imaging is
reported to have higher sensitivity for small liver metastases compared to CT . Additional advantages of MRI over CT derive from the fact
that it offers better soft-tissue contrast, prior to the administration of iodinated contrast agents and that images can be acquired in multiple
planes. However, CT reportedly provides higher spatial resolution .
The imaging protocols for detection of pancreatic malignancy involve both T1 and T2-weighted sequences, or dynamically-enhanced T1weighted sequences. On T1-weighted images, the normal pancreas has higher signal intensity than any other abdominal organ. Its short T1
relaxation time has been attributed to the abundant protein and rough endoplasmic reticulum contained within it. Fat-saturated T1-weighted
sequences are useful for distinguishing normal from abnormal pancreatic parenchyma . Also, T1-weighted sequences have been shown to be
more efficient at reducing motion artifacts than T2-weighted sequences . T2-weighted sequences are typically used to differentiate between
benign and malignant liver lesions. Dynamic contrast-enhanced MRI estimates blood flow using a computational algorithm. Its capability to
identify pancreatic malignancy rests on the fact that pancreatic cancers are hypovascular relative to normal pancreas .
With its high sensitivity and tissue contrast, specifically in applications involving contrast agents, e.g. gadolinium as a T1 contrast agent, and
its capability to simultaneously provide anatomical and functional information, as well as by the diversity of its applications, i.e.
pancreatography by means of MRCP and angiography by means of dynamic contrast-enhanced MRI, the magnetic resonance imaging
modality is likely to replace other imaging tools, such as helical CT for instance, as the method of choice in the diagnosis of pancreatic
cancer.
7. Magnetic resonance cholangio pancreatography (MRCP)
MRCP is a noninvasive procedure, which is replacing ERCP for diagnosis of the biliary and pancreatic ducts. It is based on magnetic
resonance imaging and utilizes T2 weighted imaging with long echo times to deliver optimal contrast between the hyperintense signal of
pancreatic juice and bile and the hypointense signal produced by blood and solid organs .
The sensitivity of MRCP for diagnosis of pancreatic and biliary duct abnormalities is 93-100% . As a result, MRCP is suitable for the
assessment of obstructive jaundice. Still, its limited diagnostic potential necessitates the complementary use of alternative imaging
modalities.
8. Positron emission tomography (PET)
One of the most sensitive imaging modalities is positron emission tomography (PET). The sensitivity of PET ranges between 10 -11 and 10-12
mole/L and is independent of the location depth of the contrast-producing probe. This makes PET a very attractive modality for
metabolic/physiological characterization of the tumor microenvironment. The principle of PET relies on the labeling of biological molecules
with a positron-emitting isotope, such as 15O, 13N, 11C, 18F, 14O, 64Cu, 124I, 76Br, 82Rb, and 68Ga. This positron-emitting isotope is capable of
generating two  -rays by releasing a positron from its nucleus. The released positron subsequently annihilates with an electron in its vicinity
which results in the production of two  rays located at 180 apart. These emitted rays are detected using scintigraphic equipment, which
converts the energy of the  rays into visible light .
Alternatively  emitting isotopes, such as 99mTc, 111In, 123I, and 131I, can be used for imaging but require different equipment, namely gamma
cameras, which can generate tomographic information by rotating around the subject. This modality is known as single photon emission
computed tomography (SPECT). SPECT is at least a log less sensitive than PET, less quantitative compared to PET but allows the
simultaneous detection of multiple molecular events since it is capable of detecting several isotopes with different energy  rays .
Unlike the imaging modalities listed above, which largely rely on morphological parameters to detect and assess, PET delivers
biochemical/metabolic information about tumor biology. For detection of pancreatic cancer, PET traditionally uses FDG, a glucose analogue,
labeled with the radioisotope 18F. The principle behind the preferential uptake of this contrast agent by cancers is the enhanced metabolic
activity associated with malignancy. FDG enters cells in the same manner as glucose, and is trapped there after being phosphorylated by
endogenous kinases to a form, which cannot be further metabolized.
On PET, pancreatic cancer appears as an intense region of radiotracer uptake. The reported values for PET sensitivity and specificity vary
greatly and range between 64% and 100% for specificity and 71% to 100% for sensitivity . The main advantage of PET over other imaging
modalities is its enhanced capacity to identify metastatic disease and clarify uncertain CT findings in the liver . PET can detect lesions less
than 2 cm in diameter .
One major drawback of PET imaging is its poor spatial resolution and anatomic accuracy compared to MRI and CT. Novel combined PETCT scanners, however, can overcome that weakness. A case study in which FDG-PET results were superimposed on CT-generated scans
permitted the identification of a region of atrophy visible on CT as a tumor by virtue of its increased FDG uptake .
The application of PET to pancreatic cancer imaging is an exciting and very promising new strategy, particularly in view of the recent
progress in developing multimodal PET-CT technology for image collection and analysis. The sensitivity of PET combined with the good
spatial resolution of CT could become one of the leading strategies for diagnosis and staging of this malignancy.
B. Molecular imaging of pancreatic cancer
Molecular imaging is a novel field which attempts to combine the global anatomical/physiologic scale of in vivo imaging with the detailed
molecular/cellular scale of biochemistry and cell and molecular biology in order to obtain a visual representation and characterization of
biological processes at the cellular/sub-cellular level in living subjects . This approach would allow the unraveling of complex disease
pathways, the diagnosis of disease at the earliest causative stages characterized by the first signs of metabolic or molecular disturbance, and
the noninvasive real-time monitoring of disease progression as well as response to therapy in authentic physiologic environments.
There are very few described attempts at imaging pancreatic cancer using targeted molecular approaches. All of these studies have been
conducted in murine models and will be discussed below.
1. Nuclear imaging of pancreatic cancer
The majority of studies investigating the potential of nuclear imaging for pancreatic cancer targeting have focused on the use of radiolabeled
tumor-specific antibodies.
A recent investigation uses PAM4 which is an antibody targeting the tumor-associated MUC-1 antigen . The basis for the study is the success
of initial clinical trials utilizing 131I- and 99m Tc-lebeled PAM4 whole IgG . Cardillo et al. used a bispecific chimeric antibody consisting of
PAM4 Fab' and murine anti-indium-diethylenetriaminepentaacetic acid Fab' fragments to target subcutaneously implanted pancreatic cancer
employing a pretargeting enhancement strategy. The pretargeting strategy exploits the use of a primary targeting monoclonal antibody
carrying a secondary recognition moiety, which is targeted later with a radiolabeled hapten. This system achieves high tumor:nontumor signal
ratios after injection of the hapten, and rapid tumor penetration and clearance from the circulation (Figure 1).
In addition to the traditional methods for pancreatic cancer imaging by PET involving a radiolabeled glucose analogue, different groups have
reported the use of radiolabeled nucleotide or amino-acid tracers for PET or SPECT imaging exploiting the increased metabolic rate of
tumors . These investigations show good tumor localization and kinetics. Furthermore, tumor uptake is directly related to proliferative rate,
unlike imaging with radiolabeled glucose analogues.
bsPAM4-Pretargeted 111In-cPAM4 IgG bsRIT-Pretargeted 111In-pepetide
111
In-pepetide Direct-Labeled 111In-pepetide Alone
Figure 1. Immunoscintigraphy of pancreatic tumor xenografts. Athymic nude mice bearing CaPan1 tumor xenografts were injected with bispecific PAM4 (bsPAM4; 1.5 x 10-10 mol)
followed by administration of 111In-peptide (35  Ci; 1.5 x 10-11 mol) 40 h later. A comparison of the images taken at 48 h after injection of radiolabeled peptide was made between these
mice and mice that were injected with 111In-cPAM4 whole IgG, or pretargeted with control bsRIT, or given radiolabeled peptide alone. Reproduced from Cardillo et al, 2004 with kind
permission from Clinical Cancer Research.
2. Optical molecular imaging of pancreatic cancer
Optical imaging is still an experimental modality for small animal imaging. Optical imaging draws its charm from the fact that it is easy,
relatively cheap to perform, and image acquisition times are short. Optical imaging is also relatively sensitive, ranging between 10-9 and 10-17
mole/L, depending on the precise strategy used. Nevertheless, the depth penetration of optical imaging is only 1-2cm into the tissue due to the
low efficiency of light transmission through an opaque object and the significant tissue scattering effects on image acquisition. As a general
rule, muscle and skin have a high transmission index, whereas highly vascularized tissues have low transmission. Furthermore, the images
obtained by optical imaging lack tomographic resolution which is another drawback of this imaging modality.
Optical imaging is defined by two main imaging strategies, fluorescence imaging and bioluminescence imaging. Bioluminescence imaging
detects photons released from cells genetically engineered to express luciferases- photoroteins which upon contact with their substrate
(luciferin or coelentrazine), induce the release of a photon which can then be detected using a charged coupled device (CCD) .
In fluorescence imaging, the contrast agent is illuminated with light of a certain wavelength which results in a shifted-wavelength emission of
light from the contrast agent. The most promising strategy to date involves imaging in the near-infrared spectrum (700-900nm). Tissue
scattering at that wavelength range is minimal and tissue penetrance is highest, thus partially overcoming the difficulties associated with other
fluorescence imaging molecules, such as GFP. Targets located deeper into the tissue can be imaged using fluorescence mediated tomography
(FMT), which delivers tomographic reconstruction of the image by mathematical modeling of diffusion and scattering . This method has
achieved resolution of 1-2mm and nanomolar sensitivity .
A study published by the group of Robert Hoffman , generated orthotopic pancreatic tumor models in immunocompromised mice using GFPtransformed human pancreatic cancer cell lines. The contribution of this type of model for the study of pancreatic tumor progression and
metastasis is significant in that it allows the real-time tracking of tumor growth and dissemination in a living animal.
After orthotopic implantation, tumor development was followed by whole-body optical imaging. Furthermore, consecutive whole body
simultaneous images of the primary tumor as well as spleen, bowel, and omentum metastases were obtained and quantitated for up to 64 days
(Figure 2).
The described model is particularly useful for the assessment of therapeutic progress and the study of tumor progression. However, it has no
direct clinical applications due to its reliance on artificial, transgenic strategies for tumor cell labeling.
Figure 2. A, consecutive external whole-body images of internally-growing BxPC-3-GFP tumors. A series of external fluorescence images of the BxPC-3-GFP pancreatic tumor in a single
animal was obtained from days 46 to 64 after SOI of BxPC-3-GFP in a nude mouse. B, growth curves for primary pancreatic tumor (P), splenic metastasis (S), omental metastases (O), and
bowel metastasis (B) as determined by whole-body imaging. Reproduced from Bouvet et al, 2002 with kind permission from Cancer Research.
3. Multimodal NIRF/MR molecular imaging of pancreatic cancer
As mentioned earlier, each of the currently available imaging modalities suffers specific drawbacks. Whereas MRI has unlimited depth
penetration and high spatial resolution (25-100  m), its sensitivity is low (10-3—10-5 M). Optical imaging, on the other hand has a high
sensitivity (10-9 -10-12 M), but limited depth penetration (<1 cm) and low resolution (2-3 mm) (Massoud and Gambhir, 2003). Therefore, it
would be highly beneficial to synthesize imaging probes that would combine these two modalities and take advantage of their best features.
Studies by our group utilized a novel multimodal optical/MR approach for detection of subcutaneous and orthotopically implanted human
pancreatic tumors in the mouse model. We developed a multimodal imaging probe, specific for various epithelial adenocarcinomas, including
pancreatic cancer. The molecular target for our probe is epithelial cell mucin, the product of the MUC-1 gene, which becomes overexpressed
and underglycosylated in a variety of malignancies. As a result of these properties, epitopes on underglycosylated MUC-1 (uMUC-1), which
are cryptic in the nontransformed state, become exposed and available for targeting by imaging or therapeutic probes.
With the recent development of new crosslinked superparamagnetic dextran coated iron oxide nanoparticles (CLIO) for MR imaging and
near-infrared probes (Cy5.5 dye) for optical imaging , it became possible to design multi-modal imaging probes that would combine the
advantages of both methods. In order to target the uMUC-1 antigen on tumor cells in vivo we synthesized an imaging probe that consists of
CLIO nanoparticles, modified with Cy5.5 fluorochrome and carrying EPPT peptides , specifically recognizing uMUC-1, attached to its
dextran coat (Figure 3).
The resultant target specific probe designated CLIO-EPPT was tested in mice, bilaterally implanted with a uMUC-1-positive and a uMUC-1negative tumor. The CLIO-EPPT probe demonstrated selective and specific accumulation in the uMUC-1 positive tumors and produced
signal on MR and optical images. Quantitation of MR imaging-derived signal intensities indicated a reduction in T2 following injection of the
contrast agent (as a measure of probe accumulation) of 53% for pancreatic tumors compared to 13-18% for MUC-1 negative control tumors .
Following these encouraging results, we attempted to apply CLIO-EPPT to the imaging of orthotopically-implanted pancreatic cancer. MR
imaging of the mouse pancreas, however, represents a challenge not found in the clinical imaging of human pancreas since it is not a whole
solid organ but rather a thin, membrane-like tissue spread under the liver, and extending downward and outward under the stomach and
intestines. In order to identify the pancreas on an MR image, one has to first localize it within the abdomen and optimize imaging parameters.
In a study done by our group, we delineated the mouse pancreas in a live animal within the abdomen and identified crucial "landmarks"
which allowed
Figure 3. A, the core protein of the MUC-1 tumor antigen. The immunodominant region of the tandem repeat is recognized by the EPPT1 peptide derived from an ASM2 monoclonal
antibody (25). B, synthesis (left) and scheme of the probe (right). C, the absorption spectrum of CLIO-EPPT showed the presence of three peaks corresponding to FITC, Cy5.5, and iron
oxide nanoparticles. Reproduced from Moore et al, 2004 with kind permission from Cancer Research.
Figure 4. (a) Anatomic positioning of the pancreas (P, black frame) within the mouse abdomen. The sacrificed animal was opened with a median incision and the pancreas was exposed
carefully. D, duodenum; L, liver; V, stomach; S, spleen. (b-f) Coronal MRI of the pancreas on T1-weighted spin echo sequence without (b and c) and with (d) Gd-DTPA and on T2weighted sequence (e and f). The pancreas is framed red, delineating the splenic (c and f) and the duodenal (b) parts of the pancreas. Reproduced from Grimm et al, 2003 with kind
permission from International Journal of Cancer
differentiating the pancreatic tissue from adjacent gut structures (Figure 4). The tail of the pancreas is closely adjacent to the spleen and the
stomach. The spleen appears as a triangular shaped structure on coronal slices; the stomach presents as an oval with a low (or mixed) signal
lumen, continuing to the right into the duodenum adjacent to the right lobe of the liver. The pancreatic tail is situated distal of the stomach,
partly overlaying the spleen and the upper pole of the left kidney, filling a triangular-shaped space between the stomach, the kidney and the
spleen. Following the tail of the pancreas to the right allows easy delineation of the body of the pancreas, located immediately adjacent to the
duodenum.
Having resolved some of the challenges involved in localization of the structure of the pancreas by MR imaging, we utilized our probe as a
tumor targeting tool. For that purpose, we utilized an orthotopic model of pancreatic cancer that we had previously established (Grimm et al,
2003) by delivering CAPAN-2 human pancreatic adenocarcinoma cells directly into the tail of the pancreas in nude mice. Twenty one days
following tumor implantation, mice were injected i.v. with CLIO-EPPT (10mg Fe/kg) and subjected to MR and NIRF imaging pre- and 24 h
post-injection
As evidenced in Figure 5A, both on coronal and transverse T2-weighted MR images, there is a clearly defined area of signal reduction
medial to the spleen and superior to the kidney, a region corresponding to the coordinates of the mouse pancreas. Based on the observed MR
contrast effects of CLIO-EPPT accumulation in ectopic tumors, the highly localized nature of the observed signal, and the good spatial
correlation between coronal and transverse views of the lesion, we hypothesize that this discrete T2 shortening is due to specific
accumulation of the probe. Further support for our hypothesis came from the detection of a high-intensity NIRF signal associated with the
area of the pancreas in tumor-implanted animals (Figure 5B).
The ability to detect orthotopic tumors opens the possibility not only for recognition of lesions growing in a natural physiologic environment
but also for monitoring of tumor progression and the collection of time-course data defining tumor response to therapy. To that end, we
performed time course imaging of mice orthotopically implanted with human pancreatic adenocarcinoma.
Figure 5. A. Coronal and transverse T2 weighted MR images of a CAPAN-2 pancreatic adenocarcinoma bearing mouse 24h after administration of CLIO-EPPT. Arrowheads point to
regions of T2 shortening associated with accumulation of the probe. B. Coronal and saggital light (left), NIRF (middle), and color-coded NIRF (right) images of the same mouse. The visible
signal enhancement in the region of the pancreas is representative of probe accumulation.
MR and NIRF imaging utilizing CLIO-EPPT as a tumor contrast agent were performed on 19 th and 26th days after tumor implantation. MR
imaging was performed before and 24 hours after injection of the CLIO-EPPT probe on both days. As shown in Figure 6A, the implanted
growing tumors were identified on MR images as discrete areas of T2 reduction at both time points in the transverse and coronal planes.
NIRF imaging confirmed the presence of CLIO-EPPT-mediated signal associated with the area of the pancreas suggesting the consistent
localization of uMUC-1-expressing tumor cells in that region throughout the study (Figure 6B).
Figure 6. A. T2 weighted coronal and transverse MR images of a CAPAN-2 pancreatic adenocarcinoma bearing mouse 24h after CLIO-EPPT administration. MR imaging of the same
mouse was performed on days 19 and 26 after tumor implantation. Arrowheads point to areas of T2 shortening associated with probe accumulation. B. Light (left), NIRF (middle), and
color-coded NIRF (right) images of the same mouse. The signal intensity in the region of the pancreas is evident on both day 19 and day 26 after tumor implantation.
Currently, we are undertaking studies to evaluate the CLIO-EPPT probe as a diagnostic tool for monitoring the response of orthotopicallyimplanted pancreatic tumors to various forms of chemotherapy. Our progress on multimodal imaging of pancreatic cancer combines the
sensitivity of optical imaging with the high resolution of MR imaging and allows not only the detection of primary tumor and metastases
(MUC-1 is overexpressed and underglycosylated on metastatic lesions) but also, potentially, the tracking of tumor progression and response
to therapy non-invasively. The development of new therapeutics for pancreatic cancer can be very costly. Therefore, there has been an
increased demand for accurate and non-invasive assessment of their effectiveness that can be achieved using our approach.
IV. Future outlook
Carcinogenesis of pancreatic cancer is a multistep signal transduction process, in which the most likely early events involve activation of
oncogenes, such as k-ras , overexpression of growth signal receptors, the EGFR in particular, and mutations in various tumor suppressor
genes, including DPC4 , p53 , and p16 . Based on the currently available information regarding tumor origin and progression, there is an
abundance of potential strategies for targeting malignancy, whether for the purpose of imaging or therapy. Monoclonal antibodies to the
EGFR, probes/inhibitors of tumor-upregulated proteases, apoptosis-related molecules, such as caspase-3 and tumor necrosis factor family
members, all constitute prospective targets and are actively being investigated.
Detailed knowledge about these targets and their interactions represents the basis upon which one can design various diagnostic approaches
by combining the specificity and selectivity of molecular interrogation of malignancy with the capacity of the presently available imaging
modalities to provide a real-time global view of anatomical characteristics and physiologic processes in a living subject. The potential
implications of these methodologies for the management of pancreatic cancer become obvious in view of the aggressiveness of the disease
and its high mortality rate.
References
Ahmad NA, Lewis JD, Siegelman ES, Rosato EF, Ginsberg GG, and Kochman ML (2000) Role of endoscopic ultrasound and magnetic resonance imaging in the preoperative
staging of pancreatic adenocarcinoma. Am J Gastroenterol 95, 1926-1931.
Almoguera C, Shibata D, Forrester K, Martin J, Arnheim N, and Perucho M (1988) Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 53,
549-554.
Bardeesy N, and DePinho RA (2002) Pancreatic cancer biology and genetics. Nat Rev Cancer 2, 897-909.
Blanchard JA, 2nd, Barve S, Joshi-Barve S, Talwalker R, and Gates LK, Jr. (2000) Cytokine production by CAPAN-1 and CAPAN-2 cell lines. Dig Dis Sci 45, 927-932.
Bluemke DA, Cameron JL, Hruban RH, Pitt HA, Siegelman SS, Soyer P, and Fishman EK (1995) Potentially resectable pancreatic adenocarcinoma: spiral CT assessment with
surgical and pathologic correlation. Radiology 197, 381-385.
Bouvet M, Wang J, Nardin SR, Nassirpour R, Yang M, Baranov E, Jiang P, Moossa AR, and Hoffman RM (2002) Real-time optical imaging of primary tumor growth and
multiple metastatic events in a pancreatic cancer orthotopic model. Cancer Res 62, 1534-1540.
Caldas C, Hahn SA, da Costa LT, Redston MS, Schutte M, Seymour AB, Weinstein CL, Hruban RH, Yeo CJ, and Kern SE (1994) Frequent somatic mutations and homozygous
deletions of the p16 (MTS1) gene in pancreatic adenocarcinoma. Nat Genet 8, 27-32.
Cantero D, Friess H, Deflorin J, Zimmermann A, Brundler MA, Riesle E, Korc M, and Buchler MW (1997) Enhanced expression of urokinase plasminogen activator and its
receptor in pancreatic carcinoma. Br J Cancer 75, 388-395.
Cardillo TM, Karacay H, Goldenberg DM, Yeldell D, Chang CH, Modrak DE, Sharkey RM, and Gold DV (2004) Improved targeting of pancreatic cancer: experimental studies
of a new bispecific antibody, pretargeting enhancement system for immunoscintigraphy. Clin Cancer Res 10, 3552-3561.
Clarke DL, Thomson SR, Madiba TE, and Sanyika C (2003) Preoperative imaging of pancreatic cancer: a management-oriented approach. J Am Coll Surg 196, 119-129.
Coussens LM, and Werb Z (1996) Matrix metalloproteinases and the development of cancer. Chem Biol 3, 895-904.
Cowgill SM, and Muscarella P (2003) The genetics of pancreatic cancer. Am J Surg 186, 279-286.
Davis JL, Milligan FD, and Cameron JL (1975) Septic complications following endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 140, 365-367.
Freeny PC, Marks WM, Ryan JA, and Traverso LW (1988) Pancreatic ductal adenocarcinoma: diagnosis and staging with dynamic CT. Radiology 166, 125-133.
Freeny PC, Traverso LW, and Ryan JA (1993) Diagnosis and staging of pancreatic adenocarcinoma with dynamic computed tomography. Am J Surg 165, 600-606.
Furukawa H (2002) Diagnostic clues for early pancreatic cancer. Jpn J Clin Oncol 32, 391-392.
Gerdes B, Ramaswamy A, Ziegler A, Lang SA, Kersting M, Baumann R, Wild A, Moll R, Rothmund M, and Bartsch DK (2002) p16INK4a is a prognostic marker in resected
ductal pancreatic cancer: an analysis of p16INK4a, p53, MDM2, an Rb. Ann Surg 235, 51-59.
Gold DV, Cardillo T, Goldenberg DM, and Sharkey RM (2001) Localization of pancreatic cancer with radiolabeled monoclonal antibody PAM4. Crit Rev Oncol Hematol 39,
147-154.
Gress TM, Muller-Pillasch F, Lerch MM, Friess H, Buchler M, and Adler G (1995) Expression and in-situ localization of genes coding for extracellular matrix proteins and
extracellular matrix degrading proteases in pancreatic cancer. Int J Cancer 62, 407-413.
Grimm J, Potthast A, Wunder A, and Moore A (2003) Magnetic resonance imaging of the pancreas and pancreatic tumors in a mouse orthotopic model of human cancer. Int J
Cancer 106, 806-811.
Hahn SA, Schutte M, Hoque AT, Moskaluk CA, da Costa LT, Rozenblum E, Weinstein CL, Fischer A, Yeo CJ, Hruban RH, and Kern SE (1996) DPC4, a candidate tumor
suppressor gene at human chromosome 18q21.1. Science 271, 350-353.
Hawes RH, Xiong Q, Waxman I, Chang KJ, Evans DB, and Abbruzzese JL (2000) A multispecialty approach to the diagnosis and management of pancreatic cancer. Am J
Gastroenterol 95, 17-31.
Hosten N, Lemke AJ, Wiedenmann B, Bohmig M, and Rosewicz S (2000) Combined imaging techniques for pancreatic cancer. Lancet 356, 909-910.
Hussain R, Courtenay-Luck NS, and Siligardi G (1996) Structure-function correlation and biostability of antibody CDR-derived peptides as tumour imaging agents. Biomed
Pept Proteins Nucleic Acids 2, 67-70.
Jemal A, Tiwari RC, Murray T, Ghafoor A, Samuels A, Ward E, Feuer EJ, and Thun MJ (2004) Cancer statistics, 2004. CA Cancer J Clin 54, 8-29.
Josephson L, Groman E, and Weissleder R (1991) Contrast agents for magnetic resonance imaging of the liver. Targeted Diagn Ther 4, 163-187.
Josephson L, Tung CH, Moore A, and Weissleder R (1999) High-efficiency intracellular magnetic labeling with novel superparamagnetic-Tat peptide conjugates. Bioconjug
Chem 10, 186-191.
Karayiannakis AJ, Syrigos KN, Polychronidis A, and Simopoulos C (2001) Expression patterns of alpha-, beta- and gamma-catenin in pancreatic cancer: correlation with Ecadherin expression, pathological features and prognosis. Anticancer Res 21, 4127-4134.
Kawesha A, Ghaneh P, Andren-Sandberg A, Ograed D, Skar R, Dawiskiba S, Evans JD, Campbell F, Lemoine N, and Neoptolemos JP (2000) K-ras oncogene subtype
mutations are associated with survival but not expression of p53, p16(INK4A), p21(WAF-1), cyclin D1, erbB-2 and erbB-3 in resected pancreatic ductal adenocarcinoma. Int J
Cancer 89, 469-474.
Keleg S, Buchler P, Ludwig R, Buchler MW, and Friess H (2003) Invasion and metastasis in pancreatic cancer. Mol Cancer 2, 14.
Kleeff J, Ishiwata T, Maruyama H, Friess H, Truong P, Buchler MW, Falb D, and Korc M (1999) The TGF-beta signaling inhibitor Smad7 enhances tumorigenicity in
pancreatic cancer. Oncogene 18, 5363-5372.
Kollmannsberger C, Peters HD, and Fink U (1998) Chemotherapy in advanced pancreatic adenocarcinoma. Cancer Treat Rev 24, 133-156.
Korc M (1998) Role of growth factors in pancreatic cancer. Surg Oncol Clin N Am 7, 25-41.
Li D, Xie K, Wolff R, and Abbruzzese JL (2004) Pancreatic cancer. Lancet 363, 1049-1057.
Mariani G, Molea N, Bacciardi D, Boggi U, Fornaciari G, Campani D, Salvadori PA, Giulianotti PC, Mosca F, Gold DV, and et al. (1995) Initial tumor targeting,
biodistribution, and pharmacokinetic evaluation of the monoclonal antibody PAM4 in patients with pancreatic cancer. Cancer Res 55, 5911s-5915s.
Massoud TF, and Gambhir SS (2003) Molecular imaging in living subjects: seeing fundamental biological processes in a new light. Genes Dev 17, 545-580.
Mertz HR, Sechopoulos P, Delbeke D, and Leach SD (2000) EUS, PET, and CT scanning for evaluation of pancreatic adenocarcinoma. Gastrointest Endosc 52, 367-371.
Moore A, Josephson L, Bhorade RM, Basilion JP, and Weissleder R (2001) Human transferrin receptor gene as a marker gene for MR imaging. Radiology 221, 244-250.
Moore A, Marecos E, Bogdanov A, Jr., and Weissleder R (2000) Tumoral distribution of long-circulating dextran-coated iron oxide nanoparticles in a rodent model. Radiology
214, 568-574.
Moore A, Medarova Z, Potthast A, and Dai G (2004) In vivo targeting of underglycosylated MUC-1 tumor antigen using a multimodal imaging probe. Cancer Res 64, 18211827.
Moore A, Sun PZ, Cory D, Hogemann D, Weissleder R, and Lipes MA (2002) MRI of insulitis in autoimmune diabetes. Magn Reson Med 47, 751-758.
Moore A, Weissleder R, and Bogdanov A, Jr. (1997) Uptake of dextran-coated monocrystalline iron oxides in tumor cells and macrophages. J Magn Reson Imaging 7, 11401145.
Ntziachristos V, Tung CH, Bremer C, and Weissleder R (2002) Fluorescence molecular tomography resolves protease activity in vivo. Nat Med 8, 757-760.
Palazzo L, Roseau G, Gayet B, Vilgrain V, Belghiti J, Fekete F, and Paolaggi JA (1993) Endoscopic ultrasonography in the diagnosis and staging of pancreatic adenocarcinoma.
Results of a prospective study with comparison to ultrasonography and CT scan. Endoscopy 25, 143-150.
Pellegata NS, Sessa F, Renault B, Bonato M, Leone BE, Solcia E, and Ranzani GN (1994) K-ras and p53 gene mutations in pancreatic cancer: ductal and nonductal tumors
progress through different genetic lesions. Cancer Res 54, 1556-1560.
Petrovsky A, Schellenberger E, Josephson L, Weissleder R, and Bogdanov A, Jr. (2003) Near-infrared fluorescent imaging of tumor apoptosis. Cancer Res 63, 1936-1942.
Pinto MM, Avila NA, and Criscuolo EM (1988) Fine needle aspiration of the pancreas. A five-year experience. Acta Cytol 32, 39-42.
Ringel J, and Lohr M (2003) The MUC gene family: their role in diagnosis and early detection of pancreatic cancer. Mol Cancer 2, 9.
Rosch T (1995) Staging of pancreatic cancer. Analysis of literature results. Gastrointest Endosc Clin N Am 5, 735-739.
Rosch T, Lorenz R, Braig C, and Classen M (1992) Endoscopic ultrasonography in diagnosis and staging of pancreatic and biliary tumors. Endoscopy 24 Suppl 1, 304-308.
Rosewicz S, and Wiedenmann B (1997) Pancreatic carcinoma. Lancet 349, 485-489.
Rozenblum E, Schutte M, Goggins M, Hahn SA, Panzer S, Zahurak M, Goodman SN, Sohn TA, Hruban RH, Yeo CJ, and Kern SE (1997) Tumor-suppressive pathways in
pancreatic carcinoma. Cancer Res 57, 1731-1734.
Saito K, Ishikura H, Kishimoto T, Kawarada Y, Yano T, Takahashi T, Kato H, and Yoshiki T (1998) Interleukin-6 produced by pancreatic carcinoma cells enhances humoral
immune responses against tumor cells: a possible event in tumor regression. Int J Cancer 75, 284-289.
Samnick S, Romeike BF, Kubuschok B, Hellwig D, Amon M, Feiden W, Menger MD, and Kirsch CM (2004) p-[123I]iodo-L-phenylalanine for detection of pancreatic cancer:
basic investigations of the uptake characteristics in primary human pancreatic tumour cells and evaluation in in vivo models of human pancreatic adenocarcinoma. Eur J Nucl
Med Mol Imaging 31, 532-541.
Santo E (2004) Pancreatic cancer imaging: which method? Jop 5, 253-257.
Scarpa A, Capelli P, Mukai K, Zamboni G, Oda T, Iacono C, and Hirohashi S (1993) Pancreatic adenocarcinomas frequently show p53 gene mutations. Am J Pathol 142, 15341543.
Seitz U, Wagner M, Vogg AT, Glatting G, Neumaier B, Greten FR, Schmid RM, and Reske SN (2001) In vivo evaluation of 5-[(18)F]fluoro-2'-deoxyuridine as tracer for
positron emission tomography in a murine pancreatic cancer model. Cancer Res 61, 3853-3857.
Sheridan MB, Ward J, Guthrie JA, Spencer JA, Craven CM, Wilson D, Guillou PJ, and Robinson PJ (1999) Dynamic contrast-enhanced MR imaging and dual-phase helical CT
in the preoperative assessment of suspected pancreatic cancer: a comparative study with receiver operating characteristic analysis. AJR Am J Roentgenol 173, 583-590.
Shi Q, Abbruzzese JL, Huang S, Fidler IJ, Xiong Q, and Xie K (1999) Constitutive and inducible interleukin 8 expression by hypoxia and acidosis renders human pancreatic
cancer cells more tumorigenic and metastatic. Clin Cancer Res 5, 3711-3721.
Shi Q, Le X, Abbruzzese JL, Peng Z, Qian CN, Tang H, Xiong Q, Wang B, Li XC, and Xie K (2001) Constitutive Sp1 activity is essential for differential constitutive expression
of vascular endothelial growth factor in human pancreatic adenocarcinoma. Cancer Res 61, 4143-4154.
Siegelman ES, Outwater EK, Vinitski S, and Mitchell DG (1995) Fat suppression by saturation/opposed-phase hybrid technique: spin echo versus gradient echo imaging. Magn
Reson Imaging 13, 545-548.
Soto JA, Barish MA, Yucel EK, Clarke P, Siegenberg D, Chuttani R, and Ferrucci JT (1995) Pancreatic duct: MR cholangiopancreatography with a three-dimensional fast spinecho technique. Radiology 196, 459-464.
Steiner E, Stark DD, Hahn PF, Saini S, Simeone JF, Mueller PR, Wittenberg J, and Ferrucci JT (1989) Imaging of pancreatic neoplasms: comparison of MR and CT. AJR Am J
Roentgenol 152, 487-491.
Takaku K, Oshima M, Miyoshi H, Matsui M, Seldin MF, and Taketo MM (1998) Intestinal tumorigenesis in compound mutant mice of both Dpc4 (Smad4) and Apc genes. Cell
92, 645-656.
Tamm EP, Silverman PM, Charnsangavej C, and Evans DB (2003) Diagnosis, staging, and surveillance of pancreatic cancer. AJR Am J Roentgenol 180, 1311-1323.
Tempia-Caliera AA, Horvath LZ, Zimmermann A, Tihanyi TT, Korc M, Friess H, and Buchler MW (2002) Adhesion molecules in human pancreatic cancer. J Surg Oncol 79,
93-100.
Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, and Bos JL (1988) Genetic alterations during colorectal-tumor
development. N Engl J Med 319, 525-532.
Warshaw AL, and Fernandez-del Castillo C (1992) Pancreatic carcinoma. N Engl J Med 326, 455-465.
Watanabe N, Tsuji N, Kobayashi D, Yamauchi N, Akiyama S, Sasaki H, Sato T, Okamoto T, and Niitsu Y (1997) Endogenous tumor necrosis factor functions as a resistant
factor against hyperthermic cytotoxicity in pancreatic carcinoma cells via enhancement of the heart shock element-binding activity of heart shock factor 1. Chemotherapy 43,
406-414.
Weissleder R (2002) Scaling down imaging: molecular mapping of cancer in mice. Nat Rev Cancer 2, 11-18.
Weissleder R, Moore A, Mahmood U, Bhorade R, Benveniste H, Chiocca EA, and Basilion JP (2000) In vivo magnetic resonance imaging of transgene expression. Nat Med 6,
351-355.
Winston CB, Mitchell DG, Outwater EK, and Ehrlich SM (1995) Pancreatic signal intensity on T1-weighted fat saturation MR images: clinical correlation. J Magn Reson
Imaging 5, 267-271.
Yamanaka Y, Friess H, Buchler M, Beger HG, Uchida E, Onda M, Kobrin MS, and Korc M (1993) Overexpression of acidic and basic fibroblast growth factors in human
pancreatic cancer correlates with advanced tumor stage. Cancer Res 53, 5289-5296.
Yasuda K, Mukai H, Nakajima M, and Kawai K (1993) Staging of pancreatic carcinoma by endoscopic ultrasonography. Endoscopy 25, 151-155.
Zeman RK, Cooper C, Zeiberg AS, Kladakis A, Silverman PM, Marshall JL, Evans SR, Stahl T, Buras R, Nauta RJ, Sitzmann JV, and al-Kawas F (1997) TNM staging of
pancreatic carcinoma using helical CT. AJR Am J Roentgenol 169, 459-464.
Zimny M, and Schumpelick V (2001) [Fluorodeoxyglucose positron emission tomography (FDG-PET) in the differential diagnosis of pancreatic lesions]. Chirurg 72, 989-994.