At the University of Texas M.D. Anderson Cancer Center (MDACC), the surgeon and pathologist evaluate each specimen first by frozen-section examination of the common bile duct transection margin and the pancreatic transection margin. The retroperitoneal margin is defined as the soft-tissue margin directly adjacent to the proximal 3 to 4 cm of the superior mesenteric artery (SMA) (Fig. 32.4-1) (Figure Not Available) , and is evaluated by permanent-section examination of a 2- to 3-mm full-face (en-face) section of the margin. The retroperitoneal margin must be taken at the time of tumor resection by the pathologist and surgeon. Identification of this margin of resection is not possible once the gross examination of the specimen has been completed. A positive bile duct or pancreatic transection margin is treated with re-resection; this is not possible in the retroperitoneum, where the aorta and SMA origin limit the extent of surgical resection. Samples of multiple areas of each tumor, including the interface between tumor and adjacent uninvolved tissue, are submitted for paraffin-embedded histologic examination (5 to 10 blocks). Sections 4 mum thick are cut and stained with hematoxylin and eosin. Final pathologic evaluation of permanent sections includes a description of tumor histology and differentiation; gross and microscopic evaluation of the tissue of origin (pancreas, bile duct, ampulla of Vater (or duodenum); and assessments of maximal transverse tumor diameter, lymph node status, and the presence or absence of perineural, lymphatic, and vascular invasion (Table 32.4-6) (Table Not Available) . When segmental resection of the superior mesenteric vein (SMV) is required, the area of presumed tumor invasion of the vein wall is serially sectioned and examined in an attempt to discriminate benign fibrous attachment from direct tumor invasion. In patients who received preoperative chemoradiation, the grade of treatment effect is assessed on permanent sections using the grading schema developed by Cleary and reported by Evans and coworkers.

The method for classifying subsets of regional lymph nodes in pancreaticoduodenectomy specimens is based on the work of Cubilla and colleagues. The soft fibrofatty tissue containing regional lymph nodes is divided into six regions as outlined on the anatomic pathology dissection board (Fig. 32.4-2) (Figure Not Available) . If lymph nodes are not identified, fat or other potentially neoplastic tissue is submitted for microscopic examination. Staley and colleagues have demonstrated that the number of lymph nodes identified in the surgical specimen is increased by the use of a standardized system of specimen analysis. The dissection board illustrated in Figure 32.4-2 (Figure Not Available) provides a simple means of improving lymph node identification and documenting the location of histologically confirmed lymph node metastases. In contrast, the Japanese staging system, which involves extremely detailed analysis of margins and lymph node groups, is not a practical system for widespread application.
As the use of multimodality treatment strategies for pancreatic cancer becomes more common, it will be even more important to standardize pathologic assessment of tumor specimens.

CLINICAL AND PATHOLOGIC (SURGICAL) STAGING
A standardized system for the clinical and pathologic staging of pancreatic cancer does not currently exist in the United States. The system modified from the American Joint Committee on Cancer and the TNM Committee of the International Union Against Cancer appears in Table 32.4-4 (Table Not Available) . However, this staging system is rarely used because it is difficult to apply, the stages do not directly correlate with treatment or prognosis, and lymph node status cannot be determined without surgical treatment. Pathologic staging can be applied only to patients who undergo pancreatectomy; in all other patients, only clinical staging, based on radiographic examinations, can be done. Treatment and prognosis are based on whether the tumor is potentially resectable, locally advanced, or metastatic, definitions that do not directly correlate with TNM status. For example, both potentially resectable and locally advanced tumors may be stage T3. Further, because lymph node metastases are often small (2 to 4 mm) and not accurately assessed by preoperative imaging, many patients who undergo complete negative-margin pancreaticoduodenectomy are found to have positive regional lymph nodes on permanent-section pathologic analysis of the resected specimen. Such patients would then be classified as having stage III disease. However, a patient with unresectable, locally advanced disease is usually classified as having stage II (T3, N0) disease because operative (pathologic) staging is not performed.

Tumors of the pancreas are unlike other solid tumors of the gastrointestinal tract in that accurate diagnosis, clinical staging, and pathologic evaluation of resected specimens require extensive interaction and cooperation between physicians of different specialties. Accurate clinical staging requires high-quality computed tomography (CT) to accurately define the relationship of the tumor to the celiac axis and superior mesenteric vessels. The development of objective radiographic criteria for preoperative tumor staging allows physicians to develop detailed treatment plans for their patients, avoid unnecessary laparotomy in patients with locally advanced or metastatic disease, and improve resectability rates.
Similar standardized criteria are needed for the pathologic analysis of pancreaticoduodenectomy specimens to allow accurate interpretation of survival statistics.

 Retrospective pathologic analysis of archival material does not allow accurate assessment of margins of resection or number of lymph nodes retrieved. However, these are the most accurate predictors of outcome. In the recent study by Yeo and colleagues, resection margin, lymph node status, and tumor size and DNA content were the tumor characteristics that most strongly predicted survival by multivariate analysis. To determine which patient subsets may benefit from the most aggressive treatment strategies, accurate pathologic staging and histologic assessment of response are mandatory.

CELLULAR PATHOLOGY

The normal pancreatic architecture is characteristic of a secretory gland: a background of acinar cells accounts for approximately 80% of the cell number and volume of the gland; 1% to 2% are clusters of islet cells; 10% to 15% are single-layered, cuboidal ductal cells; and there is a sparse interlacing network of blood vessels, lymphatics, nerves, and collagenous stroma. This architecture is markedly altered in carcinoma, in which the predominant histologic feature is a dense collagenous stroma with atrophic acini, remarkably preserved islet cell clusters, and a slight to moderate increase in the number of ducts, both normal-appearing and cancerous. The diagnosis of ductal adenocarcinoma rests on the identification of mitoses, nuclear and cellular pleomorphism, discontinuity of ductal epithelium, and evidence of perineural, vascular, or lymphatic invasion.

CLINICAL SIGNS AND SYMPTOMS

The lack of obvious clinical signs and symptoms delays diagnosis in most patients. Jaundice, due to extrahepatic biliary obstruction, is present in approximately 50% of patients at diagnosis and is associated with a less advanced stage of disease than are other signs or symptoms. Small tumors of the pancreatic head may obstruct the intrapancreatic portion of the bile duct and cause the patient to seek medical attention when the tumor is still localized and potentially resectable. In the absence of extrahepatic biliary obstruction, few patients present with potentially resectable disease.

The pain typical of locally advanced pancreatic cancer is a dull, fairly constant pain of visceral origin localized to the region of the middle and upper back. The pain is due to tumor invasion of the celiac and mesenteric plexus. Vague, intermittent epigastric pain occurs in some patients; its etiology is less clear. Fatigue, weight loss, and anorexia are common even in the absence of mechanical gastric outlet obstruction. Pancreatic exocrine insufficiency due to obstruction of the pancreatic duct may result in malabsorption and steatorrhea. Although malabsorption and mild changes in stool frequency are common, diarrhea occurs infrequently.

Glucose intolerance is present in most patients with pancreatic cancer. Although the exact mechanism of hyperglycemia remains unclear, both altered beta-cell function and impaired tissue insulin sensitivity are present. The importance of islet cell function to the development of exocrine cancer is suggested by the work of Bell and Stayer, who demonstrated that pretreatment of hamsters with streptozocin and the resulting destruction of islet cells prevented the induction of pancreatic cancer in these animals by the carcinogen N-nitrosobis-(2-oxopropyl)amine (BOP). This work was substantiated by studies in Chinese hamsters demonstrating that only genetically diabetic animals did not develop cancers in response to N-nitroso-(2-oxopropyl)amine.

In the absence of jaundice, patient complaints are nonspecific, as are clinical signs on physical examination. However, important staging information with direct implications for therapy can be obtained from the physical examination. This information includes performance status, cardiopulmonary function, and the presence or absence of left supraclavicular adenopathy and ascites.

The recent emergence of the importance of inherited genetic abnormalities in gastrointestinal tract neoplasia has led to closer investigation of the potential role for heritable factors in pancreatic cancer. Several rare hereditary disorders predispose persons to both endocrine and exocrine pancreatic cancer. These include the multiple endocrine neoplasia type I syndrome, hereditary pancreatitis, Lynch syndrome II, von Hippel-Lindau syndrome, ataxia-telangiectasia, and possibly the familial atypical multiple mole melanoma syndrome. In addition, case reports and formal epidemiologic studies have suggested the possibility of familial aggregations of pancreatic cancer outside the context of these rare familial syndromes. One case-control study estimated that 3% of pancreatic cancers had a hereditary origin. Evaluation of approximately 30 extended families with presumed familial pancreatic cancer has suggested that transmission is consistent with an autosomal dominant pattern. The age at onset, tumor histopathology, male preponderance, and overall survival of persons affected by pancreatic cancer in these families are reported to be similar to those of persons with pancreatic cancer in the general population. Continued study of these patients and their families may provide insight into the critical molecular genetic abnormalities leading to familial pancreatic cancer. Familial genetic abnormalities may then provide insight into the process of pancreatic carcinogenesis for patients with sporadic pancreatic cancer and provide opportunities for early detection and chemoprevention.

OCCUPATIONAL EXPOSURES

Exposure to certain chemicals, usually in a manufacturing setting, has been associated with an increased risk of pancreatic cancer. Previous studies implicated 2-naphthylamine, benzidine, and derivatives of gasoline. More recently, a cohort mortality study of more than 5000 chemical manufacturing workers suggested that heavy and prolonged exposure to DDT and related compounds can cause pancreatic cancer in humans. Finally, specific occupations (e.g., stone miners, cement workers, gardeners, and textile workers) have been associated with an increased risk of pancreatic cancer; however, the specific causative factors or carcinogens accounting for this increase are unclear. Currently, most newly diagnosed patients with pancreatic cancer do not have evidence of a specific chemical exposure or relevant occupational history. Other factors, such as smoking and diet, will probably play a much greater overall role in determining individual risk of pancreatic cancer.

Patients with a history of surgical procedures for peptic ulcer disease appear to be at an increased risk for pancreatic cancer. The physiologic basis for this apparent risk is unknown, but in an experimental rat system, chronic duodenogastric reflux produced sustained hypergastrinemia and promoted pancreatic carcinogenesis. Other researchers have suggested that the increased risk of pancreatic cancer following peptic ulcer surgery may result from the increased formation of N-nitroso compounds by nitrate-reducing bacteria that proliferate in the relatively achlorhydric stomach.

Cholecystectomy has also been associated with an increased risk of pancreatic cancer. This association is controversial, however, and not all series have supported the claim that cholecystectomy increases risk. The possibility that prior cholecystectomy can influence pancreatic carcinogenesis is supported by observations that CCK, which circulates at higher levels after cholecystectomy, is important in experimental pancreatic carcinogenesis and can stimulate the growth of experimental pancreatic tumors.

Diabetes Mellitus

Diabetes mellitus has been long associated with pancreatic cancer, however, the precise relationship has yet to be defined. Diabetes mellitus has been implicated as both an early manifestation of pancreatic carcinoma and a predisposing factor. It is known that pancreatic cancer can induce peripheral insulin resistance, and the argument that long-standing diabetes mellitus is also a risk factor for pancreatic cancer is supported by a recent cohort study showing that after an initial hospitalization for diabetes, patients had an increased risk of developing pancreatic cancer and that this risk persisted for more than a decade. In addition, a metaanalysis of studies published between 1975 and 1994 showed that pancreatic cancer occurred with increased frequency in patients with long-standing diabetes. The specific type of diabetes mellitus is also a factor; the cohort study described above agreed with other studies that suggested that the increased risk was limited to patients with noninsulin-dependent diabetes or patients whose diabetes was diagnosed before age 40. The mechanisms underlying the association between pancreatic cancer and diabetes are obscure; however, the diabetic state seems to enhance the growth of pancreatic cancer in animal models.

Chronic Pancreatitis

An association between pancreatitis and an increased risk of pancreatic cancer has long been suspected, although the magnitude of the risk remains uncertain. Older clinical studies suggested that chronic forms of pancreatitis, particularly those accompanied by pancreatic calcifications, were most closely associated with the subsequent development of pancreatic cancer. A series of recent reports have validated the epidemiologic association between chronic pancreatitis and pancreatic cancer, but the magnitude of the risk of pancreatic cancer attributable to pancreatitis remains unclear. Calculation of a general estimate of population-attributable risk has suggested that chronic pancreatitis may explain as many as 5% of pancreatic cancer cases.

Recent pathologic and molecular biologic studies have begun to explore the relationship between chronic pancreatitis and pancreatic cancer. Chronic inflammatory processes have long been associated with increased cancer risk. Pathologic examination of lesions along the pancreatic duct has revealed the spectrum of mucous cell hyperplasias (papillary and nonpapillary hyperplastic lesions and atypical hyperplastic lesions) in patients with chronic pancreatitis and patients with pancreatic cancer. The recent report of the identification of mutations in the K- ras oncogene, a mutation found almost universally in established pancreatic cancers, in regions of mucous cell hyperplasia in patients with chronic pancreatitis provides the first molecular link between chronic inflammation and the initiation of multistep pancreatic carcinogenesis. However, additional research is needed because other recent reports failed to identify K- ras mutations in pathologically similar lesions.