Pancreatic Cancer, Cystic Pancreatic Neoplasms, and Other Nonendocrine Pancreatic Tumors

Incidence

In the period between 2003 and 2014, the incidence rate of pancreatic cancer (PC) remained stable in the USA, with a slight elevation in the white population. PC is a disease of aging, with the median age at diagnosis of 71 years, with many cases diagnosed before the age of 65. It is rarely seen before the age of 45, with its incidence rising sharply after the age of 50. There is a slight male predominance with a higher incidence in blacks compared to whites.

Populations at Risk

There are a number of risk factors associated with the development of PC. Certain risk factors are immutable: advanced age, male gender, African American race and family history, whereas other risk factors are accrued over time or are environmental (obesity, smoking, alcohol use, chronic pancreatitis). The vast majority of PCs are sporadic in nature. Approximately 5% to 10% of cases are associated with familial PC syndromes, defined as those with 2 or more first-degree family members with PC. Patients with a familial PC syndrome tend to be younger at diagnosis (median age 64 to 65 years). Table 60.1 summarizes some of the genetic syndromes associated with an increased risk of PC.

Increased risk of PC has been associated with a number of inherited gene mutations: BRCA2, CDKN2A, ATM , mismatch repair enzyme instability as present in Lynch syndrome, PALB2, STK11, PRSS1, and SPINK2 . CDKN2A and BRCA2 are the most prominent inherited gene mutations associated with inherited PC, with BRCA2 mutations representing the most common gene mutation. ^,

Familial atypical multiple mole melanoma (FAMMM) is an autosomal dominant condition defined by a mutation of tumor suppressor gene CDKN2A. Normally, CDNKN2A encodes the p16 protein preventing phosphorylation and activation of the retinoblastoma gene. Patients with FAMMM are at risk for melanoma and PC.

Hereditary pancreatitis (see Chapter 57 ), commonly caused by an autosomal dominant mutation in the PRSS gene, is associated with a near 50% incidence of PC by age 74. Patients with other, non-hereditary forms of chronic pancreatitis also have a higher likelihood of developing PC. Smoking and alcohol use are risk factors that may synergize with chronic pancreatitis to influence tumor development. Individuals with Peutz-Jeghers syndrome with STK11 gene mutations, discussed in Chapter 126 , have a cumulative risk of PC by age 65 to 70 of 11% to 36%.

Screening for PC has represented a difficult proposition owing to the low incidence of disease, cost and expertise necessary for imaging, and relatively low sensitivity of imaging modalities. The most efficient mechanism to increase the positive predictive value of screening would be to identify high-risk populations. Currently, family history and the presence of defined gene mutations represent the most effective mechanisms to identify high-risk populations. The number of first-degree relatives (FDRs) with PC strongly influences an individual’s cancer risk. The risk of PC is increased 6.4-fold with 2 affected FDRs and 32-fold with 3 FDRs.

The 2013 International Cancer of the Pancreas Screening (CAPS) consortium has agreed that individuals with 3 or more blood relatives with PC, with at least 1 affected FDR, should be considered for screening. Those with at least 2 affected FDRs should also be considered for screening. Individuals with 2 affected blood relatives with PC, with at least 1 FDR, can be considered for screening. Patients with specific mutations should be considered for screening: Peutz-Jeghers regardless of family history and BRCA2/PALB2/p16/ patients with Lynch syndrome who have an FDR with PC should be considered for screening. Agreed-upon screening modalities include MRI/MRCP and EUS. There is no consensus age to initiate screening, although the AGA suggests screening begins at the age of 35 in patients with hereditary pancreatitis or 10 years before the age of the index case in the setting of familial PC.

One registry of 309 asymptomatic at-risk relatives over age 35 offered participants MRCP followed by EUS if lesions were found. Out of the 109 relatives who had completed at least initial screening, 9 had a significant abnormality by EUS confirmation, with an overall diagnostic yield of 8.3% in at-risk relatives ( Fig. 60.1 ). A different strategy used mRNA biomarkers in saliva to identify specific profiles of molecular changes. A logistic regression model using a combination of 4 mRNA biomarkers could differentiate resectable PC from non-cancer with 90% sensitivity and 95% specificity.

Environmental Factors

The most significant environmental factor in PC, and possibly the only one that has been firmly established, is cigarette smoking. Meta-analysis data have demonstrated a 1.7 relative risk (RR) for smokers. The risk for PC remains elevated for up to 10 years after cessation of smoking. The risk seems to be dose dependent, with risk rising with every 5 cigarettes smoked daily. Smoking is likely a contributing cause in 20% to 25% of all PCs.

Obesity has been associated with an increased incidence of PC. The 2012 World Cancer Research Fund Panel directly linked increased body mass, abdominal girth and abdominal weight to PC. There is a consistent increase in risk of PC with an increase in BMI.

Dietary constituents appear to have less influence as an environmental risk factor for PC than was initially found on population studies. There is some evidence that a high intake of red or processed meat may slightly increase the risk of development of PC. A protective effect of dietary folate has not been supported by recent analysis. There is limited evidence from case-control studies supporting a protective effect of fruit and vegetable consumption on PC. ^,

Diabetes mellitus is associated with an increased risk of PC. Longstanding diabetes, greater than 10 years, has been associated with an increased risk of PC. Cohort analysis has demonstrated an 8-fold increased risk of PC versus age-matched patients without diabetes mellitus. One fourth of patients are diabetic at the time of diagnosis, with another 40% experiencing prediabetic glucose elevations. The causal relationship between diabetes and tumor induction is not known. Loss of islet cell mass to a small tumor burden should not be sufficient to cause endocrine insufficiency in most patients with PC, and glucose tolerance sometimes improves in patients who have undergone tumor resection. GLP-1 mimetic drugs (e.g., exenatide) and inhibitors of GLP-1 metabolism (DPP4 inhibitors, e.g., sitagliptin) used in type 2 diabetes mellitus (see Chapter 4 ), have been associated with a nearly 3-fold increased risk of PC, unlike other classes of oral antidiabetic agents.

Molecular Pathology and Genetic Alterations

Pancreatic tumorigenesis is the result of a complex series of events, likely combining the effects of multiple intracellular genetic mutations with an altered inflammatory extracellular microenvironment. PC is thought to develop in a stepwise manner from a non-malignant precursor lesion, referred to as pancreatic intraepithelial neoplasia (PanIN), through progressive cellular and nuclear atypia to invasive adenocarcinoma. This histopathologic progression is mediated through a series of potentially inherited and acquired genetic alterations over time ( Table 60.2 ). Recently, gene sequencing of pancreatic adenocarcinomas has implicated changes in the SWI/SNF chromatin remodeling complex, specifically the subunit BRG1. BRG1, when associated with KRAS, has led to intraductal papillary mucinous neoplasm and invasive adenocarcinoma.

KRAS is an oncogene where activating mutations represent the most common gene mutation present in pancreatic adenocarcinomas. An activating mutation in KRAS is detected in approximately 30% of early neoplasms and 95% of advanced malignancy.

Mutations in KRAS allows for constitutive activation and dysregulation within the mitogen-activated protein kinase and AKT pathways leading to uncontrolled cellular proliferation and survival.

KRAS codon 12 mutations have been demonstrated to be an early acquired mutation in the transition from normal tissue through non-malignant neoplastic lesions, Pan-In, to invasive adenocarcinoma. Due to their relatively common prevalence in pancreatic neoplasia, KRAS targets represent a potential early detection test as well as a possible target for treatment.

CDKN2A, a tumor-suppressor gene, is an acquired mutation generally found in advanced rather than early Pan-IN mutations. Loss of the CDKN2A gene product, p16, has been associated with uncontrolled cyclin D1 activation. Loss of the p16 gene product has been associated with progressive histologic dysplasia. Mutations in the tumor-suppressor genes SMAD4 and TP53 are found almost exclusively in high-grade PanINs. The TP53 protein product, p53, is integral in DNA repair and apoptosis. Loss of TP53 leads to uncontrolled cell growth of potentially damaged cells.

Progression of neoplastic cells to invasive malignancy often occurs via clonal evolution through the acquisition of mutations. As detailed earlier, there are a number of mutations that occur concomitantly with KRAS mutations: CDKN2A/p16, TP53, SMAD4. Genetically engineered mouse models have recapitulated this spectrum of lesions experimentally. For example, progression of PanIN lesions from grade 1 through grade 3 to invasive carcinoma has been linked to a stepwise presence of an increasing number of genetic mutations ( Fig. 60.2 ). Better understanding of the cascade of events requires analysis of not only individual cellular alterations but also cell-cell interactions and microenvironmental forces. Global genomic sequencing of human pancreatic adenocarcinoma revealed the extreme complexity of tumor genetics, but did identify a core set of only 12 cellular signaling pathways and processes that were each altered in up to 100% of the tumors.

Genome sequencing of patients with widespread disease revealed genetic heterogeneity in the metastatic clonal populations distinct from the primary carcinoma. Quantitative analysis of the timing of these mutations showed a decade lag between the initiating mutation and the birth of the primary cancer cell. Another 5 years were required for the metastatic ability, with subsequent patient death in approximately 2 years. The number of the 4 major driver genes was significantly correlated with disease-free survival and overall survival, important information that is independent of traditional clinical staging. These data imply a much longer window of opportunity to diagnose and eventually intervene in a disease with such a dismal outcome, using currently available diagnostic and treatment schemes.

The desmoplastic stroma, long-described histologically in pancreatic adenocarcinoma, may be permissive to tumor growth and block delivery of therapeutic agents. Poor vascularity of the matrix and the various cytokine and other immune modulators suggest areas for future therapeutic approaches, with local inflammation being a key factor. A clinical correlation has been noted in genetic polymorphisms in inflammatory-related genes. Single nucleotide polymorphisms (SNPs) in inflammatory pathway genes MAPK8IP1 and SOCS3 were associated with a 10- or 6-month survival advantage, respectively, in carriers of 1 minor allele, and a 2-year survival advantage if both minor alleles were present.

US and CT

Transabdominal US is frequently the first modality used in many patients with PC who present with jaundice. US is useful to evaluate the presence of gallstones and confirm biliary dilation. CT is the method of choice for diagnosis and staging of PC. Recent improvements in this technology have greatly increased the resolution by faster and thinner section imaging. A multiphase imaging technique obtains pre-contrast, peak pancreatic parenchymal phase, and peak liver portal venous phase images. The early phase delineates the tumor, whereas the late phase enhances the vascular relationships and liver metastases. Overall sensitivity of CT for PC is 86% to 97%, but sensitivity for lesions less than 2 cm is near 77%.

The pancreatic CT protocol consists of dual-phase scanning using IV and oral contrast agents. The first, early arterial phase, scan is obtained at 25 seconds after IV contrast injection and offers visualization of the arterial anatomy for surgery. The second, arterial (pancreatic) phase scan is obtained 40 seconds after administration of IV contrast agent. At this time, maximum enhancement of the normal pancreas is obtained, allowing identification of nonenhancing neoplastic lesions ( Fig. 60.3 A ). The third, portal venous phase scan is obtained 70 seconds after injection of IV contrast agent and allows accurate detection of liver metastases and assessment of tumor involvement of the portal and mesenteric veins (see Fig. 60.3 B ).

Longstanding CT criteria for unresectability of a pancreatic tumor are as follows: (1) distant metastasis (e.g., to liver, peritoneum, or other sites), (2) encasement of the celiac axis or superior mesenteric artery, and/or (3) occlusion of the portal vein or superior mesenteric vein, although venous reconstruction is challenging this criteria. CT has demonstrated a sensitivity of 55% to 97% and specificity of 91% to 100% in detecting vascular invasion. Regarding resectability, CT has demonstrated a sensitivity of 76% to 92% and specificity of 82% to 100%. Some patients (5% to 15%) predicted to have resectable disease according to these CT criteria are found at laparoscopy to have unresectable lesions. Another equal-sized group of patients that appear resectable by CT criteria are found to be unresectable at exploratory surgery, usually by virtue of a T4 lesion. Such patients clearly do not benefit from surgical exploration, and their identification by preoperative imaging remains a challenge.