Pancreatic cancer imaging tumor markers diagnosis

Despite increasing interest and research on pancreatic cancer, patients with this devastating disease have a very poor prognosis. The principal reason for this grim prognosis is the inability to diagnose the disease at an early, localized, and curable stage. This review is intended to present an overview of our current status on the diagnosis of pancreatic cancer and emphasize recent advances in imaging and tumor markers that are attempting to overcome the challenges of this dreaded disease. Clinical presentation and risk factors for pancreatic cancer are reviewed. Various imaging moralities—including endoscopic and transabdominal ultrasound, ERCP, CT, MRI and PET—are described, with emphasis on their limitations for diagnosing pancreatic cancer at a size that has a significant impact on survival rates. The most promising approach for the early diagnosis of pancreatic adenocarcinoma utilizes tumor markers. Many of the tumor markers investigated over the past 20 years as well as some promising markers under current investigation are reviewed. The combination of serum markers may improve sensitivity and specificity in making the diagnosis of pancreatic adenocarcinoma. We conclude that the current status of imaging and tumor markers does not permit the early diagnosis of an adenocarcinoma of the pancreas. The development of improved early detection methods for pancreatic cancer is essential; however, at the present time it should be limited to high-risk individuals to allow for a better opportunity for success.

Pancreatic adenocarcinoma is the fifth leading cause of cancer mortality in the United States, with 28,900 deaths anticipated in the year 2001. The primary reason for this grim prognosis is our inability to diagnose this disease at an early stage. The importance of detecting a pancreatic cancer when it is small is best illustrated by a report from Japan that found a postoperative 5-year cumulative survival rate of 100% for patients with tumors < 1 cm. However, irrespective of tumor size, when a carcinoma was greater than 1.1 cm, there was no difference statistically in the survival rate. About 10% of the patients in this highly selected report had a tumor < 1 cm. In most series, it is quite rare to detect a tumor at such a small size. Unfortunately, by the time of diagnosis, an adenocarcinoma of the pancreas will have likely spread due to local infiltration and/or metastases. The ideal histologic stage that warrants aggressive intervention, such as a total pancreatectomy, and offers the best chance of cure would be an advanced ductal precursor lesion (severe atypia or carcinoma-in-situ).

One approach to improve the dismal prognosis for an individual affected with pancreatic adenocarcinoma consists of diagnosing the disease at an earlier—and hopefully more curable—stage. Some of the challenges presented by pancreatic adenocarcinoma include the retroperitoneal location of the pancreas, difficulties in differentiating between focal pancreatitis and carcinoma, and the identification of high-risk groups. This review is intended to present an overview of our current status on the diagnosis of pancreatic cancer and emphasize recent advances in imaging and tumor markers that are attempting to overcome the challenges of this dreaded disease.
CLINICAL PRESENTATION AND RISK FACTORS

Unfortunately it is not possible to reliably diagnose a patient with pancreatic cancer based on symptoms alone. Early symptoms include nonspecific abdominal discomfort, nausea, vomiting, sleeping difficulties, anorexia, and generalized malaise. Despite the perception that a mass in the head of the pancreas presents with “painless jaundice,” the more common presenting symptoms are epigastric pain, weight loss, and obstructive jaundice. Although these symptoms will prompt evaluation of the pancreas and biliary tree, because they occur late in the development of disease, the tumor is usually more advanced, and it is unlikely that the patient will be successfully cured. The lack of specificity for the diagnosis of pancreatic cancer when based on symptoms that are highly suggestive and sensitive for pancreatic cancer has been reported in the landmark study of DiMagno et al. In this study they found that 57% of patients have other problems: 13% had non-pancreatic intra-abdominal cancers, 12% pancreatitis, and 32% had non-pancreatic, non-cancerous disorders including 23% irritable syndrome and 10% miscellaneous problems.

Awareness of risk factors for pancreatic cancer may lead to an earlier or more aggressive evaluation for pancreatic cancer. The greatest risk factor for the development of a pancreatic cancer is genetic predisposition. It is estimated that approximately 5%–10% of patients with pancreatic cancer will have one or more first- or second-degree relatives with pancreatic cancer. The recognition of a familial aggregation first occurred in the setting of individual reports of pancreatic cancer-prone families. Further evidence for this finding comes from multiple case-control studies. Falk et al, in a case-control study in Louisiana, found an increased risk for pancreatic cancer among persons reporting the occurrence of any cancer in a first-degree relative (OR = 1.86; 95% CI = 1.42–2.44). The highest risk was in those with a history of pancreatic cancer in a close relative (OR = 5.25; 95% CI = 2.08–13.21). In another case-control study of pancreatic cancer, Fernandez et al studied the relationship between family history in the first-degree relatives of 362 patients with pancreatic cancer and 1,408 controls in northern Italy. The authors reported a significant association between the family history of pancreatic cancer in the cases as compared with the controls (relative risk = 3.3; 95% CI = 1.4–6.6). Furthermore, the authors found that the risk for pancreatic cancer did not significantly change after allowance for dietary factors, use of alcohol and tobacco, and medical history of pancreatitis and diabetes mellitus. Pancreatic cancer may also be an integral tumor of recognized inherited syndromes. These hereditary syndromes include hereditary pancreatitis, familial breast cancer, hereditary nonpolyposis colorectal cancer, a subset of kindreds with familial atypical multiple molemelanoma syndrome affected with a p16 germline mutation, ataxia-telangiectasia, and Peutz-Jeghers polyposis.

The most significant and reproducible environmental etiologic risk factor appears to be cigarette smoking. Most case-control studies have found an odds ratio for current smokers ranging from 1.3- to 5.5-fold. It appears that cigarette smoking, in concert with a family history of pancreatic cancer, may be synergistic in its etiology. Schenk et al reported that a family history of PC or history of smoking approximately doubled the risk for developing pancreatic cancer. Furthermore, the risk increased to 8-fold for developing pancreatic cancer in those individuals who ever smoked and had a positive family history of pancreatic cancer with a first-degree relative diagnosed before age 60. This synergistic effect has also been recognized in established pancreatic cancer-prone registries. Lowenfels et al found that the risk for developing pancreatic cancer in their hereditary pancreatitis registry increased from 54-fold to 154-fold as compared to the general population. In our pancreatic cancer registry, we have found an almost 2-fold increase risk of developing pancreatic cancer in pancreatic cancer-prone family members who have ever smoked as compared to non-smoking family members (unpublished data).
IMAGING STUDIES

This section will review the currently available imaging modalities employed in the diagnosis of pancreatic adenocarcinoma. Transabdominal ultrasound (US) and computed tomography (CT) are the two most commonly utilized imaging modalities for the evaluation of pancreatic disease, especially carcinoma. Endoscopic retrograde cholangiopancreatography (ERCP) is considered the gold standard for the visualization of the pancreatic duct and biliary system. The clinical role of newer imaging studies, such as endoscopic ultrasound (EUS), positron emission tomography (PET), and magnetic resonance imaging (MRI), is still evolving. All of these techniques base the diagnosis cancer on at least one of the following imaging findings in the pancreas: morphologic changes (i.e., focal deformity or enlargement), density changes, pancreatic and biliary duct changes (i.e., dilatation or stricture) or signs of local extension into adjacent structures.
Transabdominal Ultrasound

Most of the anatomy of pancreatic cancer can be demonstrated by transabdominal ultrasound. The results of this imaging modality are dependent on multiple variables such as patient body habitus, the skill of the examiner, and the presence or absence of overlying loops of gas-filled bowel. The entire gland can be seen by transabdominal ultrasound in only about 25% of cases. On ultrasound, a pancreatic carcinoma typically appears as a focal hypoechoic, hypovascular solid mass with irregular margins Main pancreatic duct dilatation proximal to the tumor is often seen. Biliary dilatation may occur when the tumor is located in the head of the pancreas. The sensitivity of transabdominal ultrasound imaging in the diagnosis of pancreatic cancer has varied based on the experience of several investigators and has ranged from as low as 44% to as high as 94%.

Due to its lower cost and wide availability, transabdominal ultrasound is commonly utilized as an initial screening technique for biliary-pancreatic disease. Support for this approach comes from a large prospective cohort study of 900 patients who underwent transabdominal ultrasound of the pancreatic area for routine clinical indications. The sensitivity for the detection of exocrine pancreatic cancer was 90% (79 of 88 patients). Nine exocrine tumors were missed from 779 cases, which had reportedly not shown any tumor in the pancreatic area for a specificity of 98.8%. The diagnosis of no tumor in the pancreatic area was determined by follow-up in the Swedish Cancer Registry for at least 2 years after the ultrasound or death from other causes. Results from this study suggest that the use of transabdominal ultrasound as the initial imaging modality is appropriate.
Computed Axial Tomography

Some authorities have recommended CT as the preferred initial imaging modality to evaluate for a suspected pancreatic carcinoma. Significant technical advances in CT imaging, including the development of a helical or spiral CT and electron beam scanning, have allowed for better visualization and staging of pancreatic carcinoma. Conventional CT involves obtaining images with incremental repositioning of the patient with repetitive breath-hold acquisitions; thereby, having the risk of not imaging part of the pancreas. The use of helical CT allows for continuous imaging of a large volume of tissue in a single breath-hold period. Moreover, this technique permits retrospective reconstruction of images, which increases its diagnostic accuracy.

A mass at the head of the pancreas is the most common CT finding. Enlargement of the whole gland is a less common pattern. The use of intravenous contrast permits differentiation of normal pancreas from carcinoma since the normal pancreas increase in density while a carcinoma is usually a low-density lesion. This is best seen with a bolus administration of contrast material When compared to other imaging modalities such as ultrasound, reported sensitivities for detecting a pancreatic neoplasm by CT range from 20% to 89%.
Endoscopic Retrograde Cholangiopancreatography

Endoscopic Retrograde Cholangiopancreatography (ERCP) is considered the gold standard for the visualization of the pancreatic duct and biliary system. The procedure is performed by passing a side-viewing duodenoscope into the second portion of the duodenum and positioning a catheter into the ampullary orifice to obtain retrograde opacification of the pancreatic as well as the intrahepatic and extrahepatic biliary ductal systems. Limitations of the procedure include the need for conscious sedation and a complication rate of 5%–10%. Reported complications include hemorrhage, pancreatitis, cholangitis, retroduodenal perforation, and mortality. ERCP also has potential for the performance of concurrent diagnostic interventions such as the collection of pancreatic juice (with or without secretin stimulation) for cytology or analysis for different biological markers, and tissue sampling of the pancreatic duct by brush or biopsy.

It is reasonable to expect that direct visualization of the pancreatic duct would be a sensitive technique for detecting malignancy because the majority of adenocarinomas of the pancreas (> 90%) are of ductal origin. An important paper by Niederau and Grendell summarized data from 16 studies and calculated a sensitivity of 92% and specificity of 96% for diagnosing pancreatic cancer by ERCP. The presence of chronic pancreatitis often contributes to diagnostic errors on ERCP, either by mimicking carcinoma or causing a carcinoma to be missed due to the underlying ductal changes. A study by Brentnall et al on the surveillance of pancreatic cancer families suggests that pancreatography can detect ductal changes such as very mild and focal side branch duct irregularities, small sacculations, and main duct strictures that correspond to lesions in a dysplastic stage.

The development of an ultrathin peroral pancreatoscope presents an interesting approach for the early detection of pancreatic cancer. Uehara et al have used this modality in combination with pancreatoscopic cytology to identify 11 patients with carcinoma in situ. The main duct showed changes of irregular mucosa, papillary mucosa, or nodular mucosa while changes of chronic pancreatitis revealed a smooth stenosis with protein plugs, erythema and scarring. The use of pancreatoscopic guidance increased the number of cancer cells as compared to pancreatic juice.
Magnetic Resonance Imaging

Magnetic resonance creates images using a static magnetic field and evaluates nuclei for absorption or emission of electromagnetic energy. Recent technical improvements in MRI techniques such as fat suppression, breath holding, intravenous contrast administration, and the use of glucagon for suppression of intestinal motility have markedly improved imaging of the normal and diseased pancreas. The use of T1-weighted spin echo sequences, which is performed during one breath hold, has been shown to be feasible for the detection of pancreatic neoplasms. The use of intravenous contrast administration optimizes image quality, since an increased signal is seen in normal pancreatic tissue, but not within a pancreatic cancer (or after contrast administration). Although gadolinium chelate is the most commonly used contrast agent, a recent study suggests that the use of mangafodipir may improve the detection of pancreatic lesions. The sensitivity and specificity of MRI in the diagnosis of pancreatic carcinoma has ranged from 83%–87% and 81%–100% respectively when compared to other imaging modalities such as endoscopic ultrasound and CT.

The recent development of fast imaging techniques and improved abdominal image quality have led to the noninvasive evaluation of the pancreatic duct and biliary tree by magnetic resonance cholangiopancreatography (MRCP). Heavy T2-weighted pulse sequences allow stationary fluids with a long T2 relaxation time, such as pancreatic juice and bile, to be shown as white (high signal intensity) on a black background. Images are similar to those obtained by ERCP, although it appears that the pancreatic tail may be poorly visualized due to the inability of patients to hold their breath causing the tail to shift outside the area of interest. A recent study demonstrated that MRCP with a 84% sensitivity and 97% specificity was as accurate as ERCP, which had a 70% sensitivity and 94% specificity, for the detection of pancreatic carcinomas. It has been reported that MRCP can detect a stenosis of the main pancreatic duct in 38 of 44 (86%) patients, including identification of a small (13 mm) lesion. Like other imaging modalities, MRCP cannot easily differentiate between benign and malignant strictures of the pancreatic duct. MRCP has several reported advantages over ERCP including its noninvasive nature, the lack of dependence on operator's skill, and the capability to visualize the pancreatic duct proximal to the obstruction. At some institutions, MRCP has replaced ERCP as the initial method of choice for imaging studies in the diagnosis of pancreaticobiliary diseases.
Positron Emission Tomography

The technique of positron emission tomography (PET) is based on the use of a positron-emitting, radiolabeled tracer that is incorporated into the normal metabolic processes that occur in the body. Two high-energy photons are produced when the emitted positron combines with an electron. Images are obtained by utilizing a machine that simultaneously detects these photons. Based on the assumption that glucose utilization is higher in malignant cells, the glucose analog 2-(fluorine-18)-fluoro-2-deoxy-D-glucose (FDG) has been shown to have increased uptake in a variety of tumors including those that originate in the brain, lung, breast, colon, rectum, and pancreas. For pancreatic cancer in particular, it has been shown that there is increased expression of the glucose transporter 1 gene, supporting the premise that FDG would have increased uptake.

Several studies have demonstrated excellent sensitivities for PET in the identification of pancreatic carcinoma (> 93%). Specificity and negative predictive value for the detection of malignancy have also been reported to be high, ranging from 82%–93% and from 82%–90% respectively. False positives in these above-mentioned studies were related to a focal inflammatory response in the pancreas, often in the setting of chronic pancreatitis. False negative studies may be related to the size (< 15 mm) of the primary lesion. Although initially it was felt that false negative scans may occur in patients with insulin-dependent diabetes due to competition of glucose with FDG in the transporter system, other reports have not found this phenomenon to contribute to their false negative rate. It has been suggested that delayed (2 hours post-injection) FDG-PET scanning increases the diagnostic accuracy.

The role of PET scanning in the detection of adenocarcinoma of the pancreas is not known. It is difficult to draw meaningful conclusions from currently available studies due to the effects of selection bias, and limitations in most studies that the cancer was already known or suspected, and comparative data with benign conditions having been absent. Additionally, this procedure is limited to tertiary care facilities due to equipment requirements (cyclotron to generate the positron-emitting radionuclides) and expense.
Endoluminal Ultrasound

Endoscopic ultrasound is felt to be the most sensitive technique currently available for the detection and local staging of adenocarcinoma of the pancreas. By imaging the pancreas with the use of an endoscope positioned in the stomach and duodenum, one can overcome many of the variables that limit transabdominal ultrasound, such as body habitus and overlying air in the gastrointestinal tract, and thereby improve the acquisition of images by placement of the transducer in closer proximity to the pancreas. The availability of miniaturized ultrasound probes that can be introduced directly into the ductal system has allowed for even closer evaluation of pancreatic strictures.

Like transcutaneous ultrasound, the majority of pancreatic carcinomas appear on endoscopic ultrasound (EUS) as ill-defined, irregular, hypoechoic masses. There are no pathognomonic features to differentiate between benign and malignant lesions. Despite this limitation, multiple studies have shown that EUS is quite sensitive in detecting pancreatic cancer. The improved sensitivity seen with EUS may be related to examiner bias in its interpretation because of knowledge of the patient's clinical symptoms and the results of their prior diagnostic studies.

After combining data from 16 studies, Rösch[ 80] showed that EUS had a sensitivity of 93% in the diagnosis of pancreatic cancer and was superior to CT (71%), ultrasound (69%), and ERCP (89%). Yasuda et al in a large series of 146 patients with pancreatic cancer obtained similar findings with a reported sensitivity of 98% for EUS compared to 75% by ultrasound, 80% by CT, 86% by ERCP, and 89% by angiography. The use of fine-needle aspiration (FNA) under EUS guidance has been successfully performed and may improve specificity by providing a definitive pathological diagnosis. A recent study by Gress et al of 102 patients with prior negative pancreatic duct cytology by ERCP or CT-guided FNA demonstrated a high diagnostic accuracy for EUS-guided FNA with positive results for 57 of the 61 cases subsequently determined to have pancreatic cancer.

The administration of secretin during the performance of EUS allows for the collection of pancreatic juice in the duodenum. Limitations to the performance of endoscopic ultrasound in the evaluation of pancreatic cancer include the expense, steep learning curve for pancreatic lesions, the need for conscious sedation, and limited availability.

Intraductal ultrasonography is performed with ultrasonic probes introduced through the biopsy channel of a standard duodenoscope. These probes have a limited depth of penetration because of their high frequency, typically 12, 20, or 30 MHz. One potential role for intraductal ultrasound may be in the differentiation between a benign and a malignant localized stenosis of the main pancreatic duct.
TUMOR MARKERS

Current imaging studies are inadequate for the identification of lesions at the severe dysplastic stage (PanIN-3); moreover, these imaging modalities do not reliably detect tumors less than 1–2 cm in size. Therefore, the use of biological markers for the early detection of pancreatic cancer is our most promising approach. It has been suggested that the use of tumor markers could allow for the earlier diagnosis of pancreatic cancers, either as alone or in association with imaging modalities. A tumor marker can be defined as any substance that when measured in abnormal concentration in a body fluid or in tissue, may detect a malignancy or define its site of origin. Attributes for the ideal tumor marker include high sensitivity and specificity, inexpensive, reproducible, rapid, widely available, acceptable to patients and correlation between concentration and tumor mass. For pancreatic disease, the most commonly studied specimen source is serum or plasma. Other potential sources include stool, pancreatic juice, or cells obtained from intraoperative specimens, FNA, cytologic brushings, or large-bore needle biopsy. Tumor markers can be classified as tumor-specific or tumor-associated. Currently, there are no known tumor-specific markers for pancreatic cancer. Tumor-associated markers in pancreatic cancer can be subclassified based on their biological actions or products. Table 1 classifies many of the tumor markers evaluated over the past 20 years as well as some promising markers under current investigation. Some of these markers are discussed in greater detail below.

Most studies regarding tumor markers in pancreatic malignancies have focused on their role in establishing a diagnosis of the primary tumor or in monitoring treatment response. To date, the clinical role of tumor markers has been limited. No tumor marker has been shown to be useful in the screening of an asymptomatic population due to a required specificity of greater than 99% to avoid the consequences of a high rate of false positive results. CA 19–9 is presently considered the gold standard tumor marker for pancreatic disease, and clinically, is the most widely used marker. Although other tumor markers have been reported to have comparable results to CA 19–9 in the diagnosis of pancreatic cancer, none have proven superior.
Selected Tumor Markers

CA 19–9 CA 19–9 is a sialylated Lewis[a] antigen associated with circulating mucins. It was originally defined by a monoclonal antibody raised in mice immunized with a human colon cancer cell line. The antigen is found in the normal epithelial cells of the pancreas, biliary ducts, gallbladder, and stomach. About 5% of individuals, those who are Lewis a-b-negative, lack the enzyme necessary to synthesize the CA 19–9 antigen; thus limiting, the sensitivity of this tumor marker to a maximum of 95%. CA 19–9 is most frequently elevated in pancreatic adenocarcinoma but may also be expressed in other malignancies, particularly those of the bile duct, stomach and colon. Levels can also be elevated in such benign conditions as acute and chronic pancreatitis, hepatitis, and biliary obstruction; however, marked elevations are typically limited to acute cholangitis and cirrhosis.

Two studies which summarized results from multiple investigations utilizing CA 19–9 levels in the diagnosis of pancreatic adenocarcinoma among symptomatic patients, reported the sensitivity and specificity of CA 19–9 (using a cutoff of 37 U/mL) to be 81 to 85% and 81 to 90%, respectively. The sensitivity and specificity for the diagnosis of adenocarcinoma of the pancreas are dependent on the cutoff level selected for CA 19–9. For example, using a cutoff of 15 U/mt yields a sensitivity of 92% but a specificity of only 60%, while a CA 19–9 level of 1,000 U/mL gives a sensitivity of only 40% but a specificity of 99%. The degree of CA 19–9 elevation may be useful in differentiating pancreatic adenocarcinoma from inflammatory conditions of the pancreas and may prove useful in the clinical evaluation of patients with suspected abdominal disorders. Studies have demonstrated that CA 19–9 levels in pancreatic juice are not as specific or sensitive as serum levels in the diagnosis of pancreatic carcinoma.

Carcinoembryonic antigen Carcinoembryonic antigen (CEA) is a protein highly expressed in the colon during embryonic growth. It is found in cord blood and amniotic fluid as well as in the mucus secretions of the biliary tract, stomach and small bowel. CEA is best known as a tumor marker for colorectal carcinoma; however, because it was the only commercially available assay for many years, it was also evaluated as a tumor marker for pancreatic adenocarcinoma. CEA has a limited sensitivity and specificity in the diagnosis of adenocarcinoma of the pancreas. In one investigation, which combined results from 18 studies, a sensitivity of 58% and specificity of 75% at a reference cutoff of 5 ng/mL was reported. In addition, the specificity was even lower at 61%, when one only included patients with conditions that may feature in the differential diagnosis of pancreatic adenocarcinoma, such as benign diseases of the pancreas or biliary system and nonpancreatic malignancies. Following the availability of CA 19–9, serum levels of CEA have not been routinely used as a tumor marker for pancreatic carcinoma. However, CEA levels in the pancreatic juice have been reported to have a diagnostic accuracy of 83% in the diagnosis of early pancreatic cancer (< 20mm, confined to pancreas), and when combined with cytology, which alone had a diagnostic accuracy of 84%, the diagnostic accuracy increased to 93%.

K-ras K-ras mutations occur in more than 90% of pancreatic adenocarcinomas as well as in more than half of intraductal neoplasms. These observations have led to studies that have assessed whether the detection of K-ras mutations in pancreatic juice is useful in the diagnosis of pancreatic adenocarcinoma. There have been reports demonstrating that K-ras mutations can be detected from pancreatic juice obtained during ERCP in 55%–77% of patients with pancreatic carcinoma. It has also been shown that K-ras mutations can be found in duodenal juice collected after secretin stimulation in 63%–79% of patients, thereby potentially decreasing the morbidity and cost of this test.

The potential role for the identification of K-ras mutations from pancreatic juice to serve as a marker for the early detection of pancreatic cancer is suggested by several reports. Wakabayashi et al described a patient who had several ERCPs with concurrent collection of pancreatic juice done three years and six months prior to his diagnosis of pancreatic adenocarcinoma. On retrospective examination of the pancreatic juice and paraffin-embedded tissue from the resected carcinoma, the same K-ras mutation at codon 12 was found in all of the specimens. Berthelemy et al reported on two patients, one with chronic pancreatitis and one with idiopathic pancreatitis, found to have a K-ras mutation in their pancreatic juice, in whom pancreatic cancer became clinically evident many months later.

A potential limitation of this marker is the finding that K-ras mutations can occur in the setting of chronic pancreatitis. Watanabe et al[ 94] evaluated a hybridization protection assay that could be quantitated to determine if they could overcome this limitation. Twenty-nine patients with adenocarcinoma of the pancreas and 26 patients with chronic pancreatitis were studied. With conventional PCR-restriction fragment length polymorphism, 5 of 26 (19%) patients with chronic pancreatitis and 22 of 28 (77%) patients with pancreatic cancer were positive for the K-ras mutation. When using the quantitated analysis at their selected cut-off value, only one of the 26 chronic pancreatitis patients was positive for the mutation, while the majority (19 of 29 [66%]) of pancreatic cancer patients still remained positive for the mutation.

p53 p53 mutations occur in up to 75% of pancreatic cancers. In the setting of benign and malignant pancreatic and biliary diseases, serum antibodies against a mutant p53 demonstrated an excellent specificity for pancreatic malignancy, but only a 20% sensitivity, thus limiting its clinical role as a tumor marker. Immunocytochemical staining for p53 mutations has been evaluated as a tool to increase the sensitivity of cytology samples. Iwao et al reported that this technique increased the sensitivity from ~60% for Papanicolaou staining alone to 90% in combination with staining and p53 immunocytochemistry.

CA 242 CA 242 is a tumor-associated antigen expressed in the same mucin fraction as CA 19–9. Although not completely defined, the antigenic determinant of CA 242 appears to be similar to but different from CA 19–9. It appears to have a slightly greater specificity but lower sensitivity than CA 19–9 in the diagnosis of adenocarcinoma of the pancreas. This improved specificity may be related, in part, to the finding that CA 242 is less influenced by the stagnation of pancreatic juice found in chronic pancreatitis. CA 242 serum levels appear to be a better predictor of prognosis than CA 19–9. In a recent study, only resectability or preoperative CA 242 levels, and not preoperative serum CA 19–9 levels, were found on stepwise multivariate analysis to be independent prognostic factors for survival. The role of this marker remains to be defined, but it may be of particular use in patients with chronic pancreatitis, in which the diagnosis of pancreatic carcinoma is being entertained.

Amylin Amylin (islet amyloid polypeptide), a hormone produced by the beta cells of the islets may cause insulin resistance and may be responsible for the occurrence of diabetes in patients with pancreatic adenocarcinoma, as evidenced by a study showing that amylin levels are elevated in patients with pancreatic adenocarcinoma who have diabetes. From the observation that diabetes occurs in up to 80% of patients with pancreatic adenocarcinoma, it has been suggested that amylin may be a valuable marker for the early detection of pancreatic cancer. Unfortunately, preliminary results have shown suboptimal specificity rates for serum amylin levels.

CA 72–4 The CA 72–4 immunoassay measures tumor-associated glycoprotein 72 (TAG-72), a high molecular weight glycoprotein expressed in malignant epithelium-derived tumors. Due to poor sensitivity when compared to CA 19–9 this marker does not appear to be useful in the management of patients with pancreatic adenocarcinoma. However, CA 72–4 levels in pancreatic cyst fluid appear to be useful in differentiating mucinous cystic neoplasms from benign serous cystadenomas and pseudocysts.

Telomerase Telomerase is the ribonucleoprotein enzyme that prevents chromosome shortening during DNA replication. Because DNA polymerase can not replicate the extreme 5' terminus of a linear DNA molecule, cells must have an alternative mechanism for maintaining the terminal DNA sequences. This usually involves telomerase adding additional DNA to the chromosome terminus, or telomere. Abundant telomerase activity has been observed in immortalized human cells and in the majority of human tumors. Thus, there appears to be a close correlation between the presence of telomerase and human cancer.

Several research groups have examined the level of telomerase in samples from surgically resected pancreatic carcinomas, benign pancreatic tumors and chronic pancreatitis tissue. Although telomerase was detected in 80–95% of the carcinoma samples, no activity was observed in the benign pancreatic lesions. In a follow up study, one group examined telomerase activity in pancreatic duct cells collected by endoscopic retrograde pancreatic duct brushing (ERDB). They found telomerase in 13 out of 15 (87%) of the patients who were later diagnosed as having duct cell carcinoma or carcinoma in situ (a precancerous condition) and in none of the 17 subjects who had benign pancreatic disease (12 with pancreatitis, 2 pancreatic cysts, 1 abscess and 2 benign tumors). Of the 15 samples from patients with duct cell carcinoma, 6 were originally diagnosed as benign on the basis of cytology and the final diagnosis was only made after surgery. This study indicates that telomerase activity is an excellent marker for pancreatic cancer and telomerase assays using pancreatic specimens may not only help with diagnosis, but also provide a tool for early detection of pancreatic carcinomas.

Unfortunately, ERDB can cause pancreatitis and hence is not suitable as a routine screening procedure. Detection of telomerase in cells from pancreatic juice looks much more promising. In one study where pancreatic juice was obtained by ERP from 34 patients (12 with ductal carcinoma, 12 with pancreatic adenoma, and 12 with pancreatitis), there was a strong correlation between the level of telomerase in the pancreatic juice sample and the presence of a pancreatic carcinoma. The level of telomerase in the pancreatic juice from adenoma and pancreatitis patients was very low while the levels in the samples from carcinoma patients were much higher. This suggests that telomerase activity in pancreatic juice can be used to distinguish between adenocarcinoma, adenoma and pancreatitis. Support for the use of telomerase expression in pancreatic juice as a means for the early detection of pancreatic adenocarcinoma is provided by a case report from Suehara et al. They describe a patient who had pancreatic juice collected during an ERCP for pancreatitis and subsequently was found 19 months later on CT scan to have a tumor for which the patient underwent pancreaticoduodenectomy. This patient developed a detectable pancreatic adenocarcinoma 19 months later. Further studies are warranted to better define the role of telomerase in the early detection of pancreatic cancer.
ADDITIONAL MARKERS

There are many new molecular markers that are under investigation for the early diagnosis of pancreatic cancer. Tissue polypeptide specific antigen has been reported to have better discrimination than CA 19–9 levels for the differentiation between pancreatic carcinoma and chronic pancreatitis. Tissue inhibitor of metalloproteinase type 1 (TIMP-1) was the first marker to be identified by gene expression analysis. This serum tumor marker was not felt to be useful alone for the diagnosis of pancreatic cancer, but when included in a panel with CA 19–9 and CEA, it had a sensitivity of 60% and a specificity of 100%.
SUMMARY

Optimal diagnostic algorithms for pancreatic cancer are dependent on clinical context including patient presentation and availability of imaging modalities. In most instances, a helical CT scan should be the initial study, since more staging information (metastasis, vascular involvement) can be obtained than on an abdominal ultrasound. EUS may detect a smaller lesion and has the advantage of allowing fine needle aspiration for histologic confirmation. ERCP remains a useful diagnostic and therapeutic procedure. It can aid in diagnosis by providing cytologic brushings and in therapy of biliary obstruction by stent placement. PET and MRI may be useful in those selected cases in whom the suspicion of carcinoma remains high and CT scan and EUS are not diagnostic. The ability to differentiate between an inflammatory mass and cancer remains a problem with all imaging modalities. Pancreatoscopy warrants further study, given its promise of diagnosing pancreatic cancer at an earlier stage; however, its lack of availability, technical demand and invasiveness limits its current clinical usefulness.

The role of tumor markers in pancreatic disease, particularly pancreatic adenocarcinoma, needs to be clarified. Presently CA 19–9 is the most widely studied and available serum marker. It may have a complementary role in the diagnosis of pancreatic adenocarcinoma but should not replace a histologic diagnosis. CA 19–9 levels appear to have the potential to serve as a marker for response to therapy following surgery or chemotherapy; thereby, allowing earlier opportunities to explore other therapeutic options. The combination of serum markers may improve sensitivity and specificity in making the diagnosis of pancreatic adenocarcinoma. Our growing understanding of the molecular pathogenesis of pancreatic cancer will likely lead to the development of other tumor markers.

The early detection of pancreatic cancer in high-risk individuals remains challenging. Unfortunately, there is no proven method to achieve this goal. The use of biological markers for the early detection of pancreatic cancer is our most promising approach, since our current imaging technology does not reliably permit the visualization of small (< 1 cm) tumors. Although it may be reasonable to utilize pancreatoscopy in this setting, the risk of repeated manipulation of the pancreatic duct is unknown. Thus, this tool requires additional evaluation and may best be used in those clinical situations in which abnormal biological markers have been identified.

Last updated Jan 2/07