Pancreatic Cancer Risk Factors

Over the past 20 years, the incidence of pancreatic carcinoma in Europe and North America has remained unchanged, with an estimated 9-10 cases per 100 000 and slightly increased male:female and black:white ratios. Pancreatic cancer currently ranks as the fifth most common cause of cancer-related deaths in western countries. Various risk factors have been implicated and are shown in the panel. Other observations include familial aggregation of pancreatic carcinoma and a 13-fold increased risk of this cancer in certain families with malignant melanoma. Diabetes mellitus, although not an independent risk factor is more prevalent in patients with pancreatic cancer; this increase is largely due to recent-onset diabetes, presumably as a result of the tumour.
Pathology and molecular abnormalities

About 90% of pancreatic tumours are adenocarcinomas with a ductal phenotype. Neuroendocrine tumours and acinar cell carcinomas represent about 2-5% of all pancreatic tumours. The assumption is that tumorigenesis starts from pluripotent stem cells. However, the presence of such protodifferentiated cells with the capability of transdifferentiation in the adult human pancreas remains elusive. Pancreatic ductal adenocarcinomas are characterised histologically by atypical glands embedded in a dense fibrotic stroma. Although histologically very similar, adenocarcinomas of the distal bile ducts and the ampulla of Vater should be considered separately because they usually have a better prognosis owing to a higher rate of resectability. Most tumour types other than ductal adenocarcinomas tend to be more amenable to therapeutic interventions such as resection, chemotherapy, or irradiation. Thus, histological diagnosis is essential if a pancreatic tumour other than an adenocarcinoma is suspected. For example, the very rare mucinous cystic and solid-pseudopapillary (solid and papillary) tumours, which occur predominantly in women under 40, and intraductal papillary-mucinous tumours (also known as mucin-producing carcinoma) have a more favourable prognosis after resection. This generally is especially true for functional neuroendocrine carcinoma (eg, gastrinoma, glucagonoma, and VIPoma) and for hormone-inactive, non-functional neuroendocrine pancreatic cancer, lymphomas, and the extremely rare mesenchymal tumours (liomyosarcoma, fibrosarcoma, haemangiopericytoma). Although acinar cell carcinomas are usually well differentiated, the prognosis of this rare tumour type is almost as poor as for ductal adenocarcinoma.

Ductal adenocarcinomas are characterised by expression of mucin-related carbohydrate antigens such as CEA, CA19-9, DU-PAN-2, and SPan 1, the mucin MUC 1, and cytokeratins 7, 8, 18, and 19; acinar cell carcinomas express pancreatic digestive enzymes and neuroendocrine tumours express synaptophysin and chromogranin A. 70% of ductal adenocarcinomas are located in the head, 20% in the body, and 10% in the tail of the organ.

Local tumour extension almost invariably involves the peripancreatic fat tissue through direct invasion of lymphatic channels and perineural spaces. Duodenum, stomach, gallbladder, and peritoneum are infiltrated by tumours located in the pancreatic head; body and tail tumours can invade liver, spleen, and left adrenal gland. Lymphatic spread to adjacent and distant lymph nodes seems to precede haematogenous spread, which affects, in descending order, liver, peritoneum, lungs, adrenals, kidneys, bones, and brain.

Both tumour-suppressor genes and oncogenes contribute to tumorigenesis in pancreatic adenocarcinoma. K-RAS oncogene activation has been studied most extensively because it is found in 80-90% of ductal pancreatic carcinomas but is almost absent in acinar cell carcinomas and neuroendocrine tumours. Whereas the expression of these altered gene products is increasingly applied for diagnosis, no correlation with tumour stage or patient survival has been observed so far. Similarly, p53, a nuclear phosphoprotein, is overexpressed mainly due to point mutations in 50-70% of pancreatic adenocarcinomas but not in acinar and neuroendocrine carcinomas; this observation suggests different pathways of tumorigenesis. Somatic mutations, homozygous deletions, or loss of expression of the tumour-suppressor genes DCC (deleted in colon carcinoma) and MTS1 are observed in about 50 and 70% of pancreatic adenocarcinomas, respectively. DPC 4 (deleted in pancreatic carcinoma), a novel candidate tumour suppressor gene, has also been identified in human pancreatic carcinoma.
Diagnosis and staging

Presenting signs

Early symptoms of pancreatic carcinoma, including weight loss, anorexia, epigastric discomfort, and back pain, are often non-specific and vague, so diagnosis may be considerably delayed. Weight loss (usually greater than 10% of body weight) and jaundice due to biliary obstruction are most common in cancer of the pancreatic head, whereas epigastric and back pain are more common in tumours of the body and tail.[ 8] Pain is probably caused by invasion of the tumour into the splanchnic plexus and retroperitoneum. Nausea and vomiting due to duodenal obstruction are late manifestations of the disease. New onset of diabetes is observed in 15% of cases and about 3% of pancreatic cancer patients present with acute pancreatitis.

Imaging procedures

There is no ideal diagnostic procedure for evaluation of pancreatic carcinoma. Transabdominal ultrasound is used as the first step. Tumours of more than 2 cm, dilated biliary and pancreatic ducts, and extrapancreatic spread are detectable by this means. Contrast enhanced computed tomography (CT) is the single most useful imaging procedure. Local spread, distant metastasis, and invasion of adjacent organs or vessels can be detected. Although the accuracy of CT for predicting unresectability is more than 90%, prediction of resectability ranges beween 50 and 90%. Limitatons of CT and ultrasound are small tumours and peritoneal spread. If ultrasound and CT scanning remain equivocal, the site of biliary or pancreatic-duct obstruction can be further defined by endoscopic retrograde cholangiopancreatography (ERCP), shown in the figure. This procedure permits delineation of discrete alterations of the biliary tree and the pancreatic duct that have been missed by other imaging methods, and biopsy and brush cytology can also be carried out. Some centres prefer the percutaneous transhepatic approach for cholangiography (PTC), although this procedure is more invasive. Percutaneous, transabdominal, or endoscopic ultrasound-guided fine-needle aspiration of tumour lesions can be done in selected patients. Whether this procedure is required for patients whose lesions are amenable to surgery is questionable for two reasons: (a) small, potentially curable tumours are easily missed by this approach; and (b) metastatic spread has been observed as a consequence of diagnostic puncture. However, in patients with unresectable lesions, cytological or even histological analysis allows diagnosis of tumours such as neuroendocrine neoplasias or lymphoma that are responsive to specific therapeutic approaches. Selective coeliac and mesenteric angiography, combined with evaluation of the portal venous anatomy, may be helpful for delineation of major arterial and venous anatomical abnormalities as well as for staging of resectability. Consequently, this imaging procedure is often used before surgery. Nevertheless, a recent study showed no advantage over high-resolution dynamic CT in terms of preoperative staging. Except for identification of liver metastases, CT arterial portography has no advantage in tumour staging. Endoscopic ultrasound is a relatively new imaging procedure that can display small lesions undetectable by methods such as CT, ultrasound (figure), or angiography. It can also localise lymph node metastases and/or vascular tumour infiltration with high sensitivity. A major limitation of this very sensitive technique is its operator dependence. Localisation of distant metastases is crucial for the patient's management. Since metastases in pancreatic cancer involve liver, peritoneum, and omentum as multifocal and small (diameter <5 mm) lesions, the imaging procedures often give false-negative results. Thus, before laparotomy, diagnostic laparoscopy has proved valuable in combination with peritoneal washings and cytological analysis.[ 15]

New developments for pancreatic tumour imaging have been reported. Among these, new routes for the local applicaton of ultrasound devices--eg, via a laparoscope, the portal vein, or the biliary and pancreatic ducts--have been studied. Scintigraphic visualisation of G-protein-coupled receptors in pancreatic tumour tissue has also been achieved with radiolabelled ligands such as somatostatin or vasoactive intestinal polypeptide (VIP). Whereas somatostatin receptors were only observed in pancreatic neuroendocrine carcinomas, VIP receptors were additionally present in adenocarcinomas.[ 16] Several studies on neuroendocrine tumours suggest that receptor imaging in vivo can sometimes be better than conventional imaging procedures. The diagnostic value of positron emission tomography (PET) awaits stringent evaluation. Magnetic resonance imaging with ultrarapid echoplanar imaging shows promise.

Many serological tumour markers, the most common being CEA and CA 19-9, have been proposed for diagnosis and follow-up of patients with pancreatic cancer. None of the markers studied so far is suitable for screening. Moreover, even with a probable diagnosis of pancreatic cancer based on imaging procedures, many false-positives (eg, due to biliary obstruction) and false-negatives occur. And for postoperative surveillance, there are no convincing data that tumour recurrence detected by tumour markers, leading to prompt therapeutic intervention, prolongs survival.

Distinguishing between pancreatic carcinoma and chronic pancreatitis is notoriously difficult because chronic pancreatitis may present with localised, inflammatory intrapancreatic lesions, pancreatic duct irregularities, or even stenoses. No imaging procedure can reliably differentiate between the two entities. A valuable tool in the differential diagnosis is cytological analysis of fine-needle biopsy specimens, pure pancreatic juice, or ductal brushings. The sensitivity of cytology, which largely depends on the quality of the material, reaches 70-80% in most studies. Sensitivity might be further enhanced by use of flow cytometry and molecular analysis of the aspirates; specificity of cytology is consistently greater than 95%. Molecular techniques have also been investigated for diagnosis. For example, K-RAS mutations can be detected, in order of decreasing sensitivity, in pancreatic duct brushings, fine-needle aspirates, pure pancreatic and duodenal juice, faeces, and even peripheral blood of pancreatic cancer patients.[ 17] Despite this, non-mutatedK-RAS is observed in about 15% of pancreatic carcinomas.[ 7] Furthermore, mutated K-RAS has recently been detected, albeit in a lower percentage, in hyperplastic duct cells of disease-free human pancreas.
Therapy

Surgery is the only curative treatment for pancreatic carcinoma, and outcome is related mainly to tumour stage and histology (eg, neuroendocrine tumour vs adenocarcinoma). Based on the preoperative staging, resectability depends largely on the following factors:

    * size (`magic' 2-3 cm) and location of the primary tumour
    * invasion into the peripancreatic tissue and lymph node involvement
    * evidence of vascular infiltration
    * presence of distant metastasis.

As a result of better surgical technique, mean operating time, morbidity, and mortality, as well as the need for perioperative transfusion, have been greatly reduced over the past few decades. Nevertheless, most (about 80-90%) tumours are not resectable. Resectable tumours are generally located in the pancreatic head. Consequently, a standard pancreatoduodenectomy (Whipple's operation) or a modification of this approach (eg, pylorus-preserving technique) is performed. About a third of patients with primary tumours smaller than2 cm but with localised spread survive 5 years. When there is no lymph node or vascular invasion, more than half the patients survive 5 years. To improve the therapeutic outcome in patients with locally advanced disease, several strategies have been pursued: extended pancreaticoduodenectomy; total pancreatectomy; `super-radical' resection procedures; and pre-operative or intraopertive radiotherapy alone or in combination with chemotherapy. Irrespective of the therapy used, local recurrence in the retroperitoneum, followed by liver metastases and peritoneal dissemination, are common and thereby determine survival. Comparison of the various surgical approaches does not show significant differences in terms of overall survival, but does reveal higher mortality and morbidity with the more radical procedures. However, a comparison of patients with negative and positive resection margins showed 5-year survival rates of 22% and 0%, respectively. This result has prompted several studies of surgery in combination with radiotherapy and/or chemotherapy. In combination with surgery, radiotherapy alone (ie, intraoperative) or together with chemotherapy ( 5-fluorouracil, postoperative) showed improved local control with an extended survival for some patients. In addition, pain control is an important benefit of this approach. For localised, unresectable pancreatic carcinoma, only the combination of radiation therapy and chemotherapy (especially 5-fluorouracil) can increase survival. By use of neoadjuvant radiochemotherapy, applicaton of more aggressive chemotherapeutic agents alone or in combination does not result in a significant 5-year survival benefit. For tumours with distant metastases, various systemic strategies have been studied. Of the chemotherapeutic agents, 5-fluorouracil has been most extensively investigated and has been shown to have a response rate in the range of 10-20%. Multidrug regimens have not been reproducibly shown to be more effective than 5-fluoruoracil alone. In addition to chemotherapy, biotherapies alone or in combination, including antioestrogens, interferons, or peptide hormones have been used. However, an effect on 5-year or even 2-year survival rates has not been documented for any ofthese agents.

When evaluating new treatments, researchers must bear in mind that most patients with metastatic cancer are already debilitated and will be unable to tolerate agents with severe side effects. Moreover, in the absence of a standard therapeutic approach, every patient with advanced pancreatic adenocarcinoma should be treated only in randomised trials with an obligatory untreated control group. All studies should also incorporate quality-of-life assessment. In this context, gemcitabine seems to be of some value.

Palliation

Palliation remains a cornerstone of therapy, the main focus being management of jaundice, duodenal outlet obstruction, and pain. For pain relief, about 50% of the patients respond satisfactorily to conventional narcotic analgesics. Percutaneous or intraoperative chemical splanchnectomy with alcohol or phenol by either fluoroscopic or CT guidance results in effective analgesia in 80-90% of patients. External-beam radiation is likewise effective for pain relief but sometimes takes several weeks to act--a definite drawback in patients with a short life expectancy. Surgical bypass (choledochojejunostomy or cholecystojejunostomy) and biliary stenting (transhepatic or endoscopic) are equally effective for the relief of jaundice due to malignant biliary obstruction. Neither procedure improves survival, but the gain in quality of life obtained by elimination of biliary obstruction is considerable. In patients with a limited life expectancy (<6 months), which essentially holds true for almost every pancreatic cancer patient, endoscopic stenting is generally preferred to bypass surgery, owing to lower early procedure-related morbidity and mortality. However, bypass surgery is associated with a lower incidence of late complications (recurrent jaundice, cholangitis) and there is the possibility of intraoperative splanchnectomy and gastrojejunostomy in case of existing or imminent duodenal obstruction. Gastrojejunostomy, either retrocolic or via a Roux-en-Y limb, effectively relieves duodenal outlet obstruction but often results in delayed gastric emptying, bleeding from stromal ulceration, or biliary gastritis.

Risk factors for pancreatic cancer

Increased dietary fat intake
Reduced consumption of fruit and vegetables
Occupational exposure to petroleum products,
  b-naphthylamine and benzidine
Tobacco smoking
Tropical calcifying pancreatitis
Chronic pancreatitis

Last updated Jan 2/07