Ebook Cancer epidemiology and prevention (4/E): Part 2

(BQ) Part 2 book “Cancer epidemiology and prevention” has contents: Stomach cancer, small intestine cancer, biliary tract cancer, liver cancer, hodgkin lymphoma, multiple myeloma, bone cancers, soft tissue sarcoma, endometrial cancer, ovarian cancer, bladder cancer,… and other contents.

31 Stomach Cancer

CATHERINE DE MARTEL AND JULIE PARSONNET

OVERVIEW

Stomach cancer is the fifth most common incident cancer worldwide

and the third leading cause of cancer death Almost half of the world’s

cases occur in Asia, with 42% in China alone Although the incidence

and mortality from stomach cancer are decreasing, global disease

burden remains high Moreover, the absolute number of cases

contin-ues to rise because of population aging Adenocarcinomas comprise

over 90% of gastric malignancies The adenocarcinomas are further

classified according to anatomic location (cardia vs non- cardia),

his-tology (e.g., intestinal or diffuse, signet ring or non- signet ring) and

most recently by molecular classification Adenocarcinomas in the

stomach’s body and antrum are usually caused by chronic infection

with Helicobacter pylori (H pylori); the incidence of these tumors

is decreasing worldwide Cardia tumors have epidemiological

char-acteristics more similar to those of esophageal adenocarcinoma; the

incidence of these tumors is increasing, particularly in high- income,

Western countries New molecular classification systems have been

proposed based on investigations of tumors in high- income countries

The Cancer Genome Atlas Program has identified four molecular

sub-types: (1) tumors positive for Epstein- Barr virus; (2) those marked

by microsatellite instability; (3) genomically stable tumors; and (4)

tumors with chromosomal instability and extensive somatic copy-

number aberrations These subtypes have not yet been integrated into

etiologic and descriptive studies The most promising public health

strategy for preventing gastric cancer is the eradication of H pylori

with antibiotics This approach is currently being tested in randomized

clinical trials

INTRODUCTION

At the dawn of the twentieth century, stomach cancer represented an

astonishing one- third of all cancers, approximately 1% of hospital

admissions, and 2% of all deaths investigated by necropsy (Fenwick

and Fenwick, 1903) Although stomach cancer remained the

lead-ing cause of cancer death in the world until the mid- twentieth

cen-tury, overall, the rapid decline in stomach cancer throughout the last

100 years has been touted as an “unplanned triumph” (Howson et al.,

1986) The term “unplanned” highlights the lack of directed

inter-vention in preventing this disease; rather, stomach cancer’s decline

parallels the improvements in nutrition, sanitation, and hygiene in

the twentieth century This steady decrease in incidence over time

provides insights into stomach cancer etiology, as well as directions

for the ultimate elimination of stomach cancer as an important health

problem worldwide

Over 90% of stomach cancer cases are adenocarcinomas

aris-ing from the gastric glands (Coleman et  al., 1993; World Health

Organization [WHO], 2010) Other histologic types of epithelial

stomach cancer include adenosquamous carcinoma, carcinoma with

lymphoid stroma (i.e., medullary carcinoma), hepatoid carcinoma,

squamous cell carcinoma, and the small group of neuroendocrine

neoplasms (WHO, 2010) Gastric cancers of non- epithelial origin

include lymphomas and mesenchymal tumors, such as leiomyoma,

schwannoma, and Kaposi sarcoma in immunosuppressed patients

(WHO, 2010) Secondary tumors are rare, the stomach being one of

the five least common metastatic sites (Disibio and French, 2008)

Because they represent the vast majority of gastric tumors, this ter focuses on adenocarcinomas of the stomach, including cancers

chap-of both the gastric cardia (ICD- O code 16.0) and non- cardia (ICD- O codes C16.1– C16.6) (WHO, 2013)

DISEASE BURDEN

According to GLOBOCAN, an estimated 952,000 new cases of gastric cancer occurred worldwide in 2012 (Ferlay et al., 2013), making it the fifth most common cause of cancer worldwide (6.8% of all cancers) More than 70% of cases (677,000) occurred in less developed coun-tries, with 553,000 cases occurring in Eastern Asia, and nearly half of the total number (405,000 cases or 42.5% of the total) in China alone Global incidence and death rates are shown in Figures 31– 1a and 31– 1b Europe contributed nearly 15% of the global burden (140,000 cases), and Latin America contributed a further 6% (61,000 cases) (Ferlay et al., 2013) Eastern Europe and the Andes are areas with par-ticularly high risk In the United States, the American Cancer Society predicts that 26,370 new cases of stomach cancer (16,480 in men and

9890 in women) will be diagnosed in 2016 with 10,730 deaths (6540 men and 4190 women) (American Cancer Society, 2016) This places stomach cancer as the 15th most common malignancy, in terms of both incidence and mortality, in the United States

Survival rates for stomach cancer are generally poor A  recent international comparison of survival in 279 population- based can-cer registries in 67 countries (Allemani et  al., 2015)  shows that the 5- year age standardized net survival from stomach cancer (i.e., the proportion of cancer patients who survive 5  years, after eliminating the background mortality due to other causes) ranges from 15% to 35% for adults Survival is considerably higher in Japan and South Korea (50%– 60%), where systematic screening allows the early detection and surgical treatment of tumors It is not known how much of the improvement in survival in these countries represents lead- time bias, however Some studies suggest a 30%– 60% reduction in mortality with screening (Hamashima et al., 2013; Hamashima et al., 2015) The slope of improvement in mortality has remained relatively constant since the early 1970s, prior to the onset

of widespread endoscopic and radiographic screening programs (Whitlock, 2012)

Because of its high case- fatality, gastric cancer accounts for a larger fraction of cancer deaths than incident cases globally (8.8% versus 6.8%) Despite its declining incidence, gastric cancer remains the third leading cause of cancer death worldwide (723,000 deaths estimated per annum), after lung and liver cancers (American Cancer Society, 2016; Ferlay et  al., 2013) The preceding estimates represent the figures for adenocarcinoma, the most common histological type of gastric cancer The worldwide total number of lymphomas of gastric origin in 2012 was estimated to be 18,000, or less than 2% the number

of adenocarcinomas (Plummer et al., 2016) Other gastric histologic types are even less common

CLASSIFICATION

Epidemiologic studies over the last 50  years have classified gastric cancers according to several systems, beginning with histopathology

in the mid- 1960s, followed by anatomic location in the early 1990s

Molecular classification systems have recently been proposed but

have not yet been integrated into epidemiologic studies, nor are

molecular markers available at this point from population- based

tumor registries

Gastric Anatomy and Function

Grossly, the stomach has four anatomical regions: the cardia, the

fun-dus, the body, and the pylorus, as shown in Figure 31– 2 The gastric

cardia (also known as the gastroesophageal [GE] or esophagogastric

junction [WHO,  2010]) is a narrow circular band, 1.5– 3  cm wide,

located at the junction where the tubular esophagus joins the ach The cardia’s proximal border can usually be seen on endoscopy, but the distal boundary is poorly defined (WHO, 2010) The fundus and body, the sections of the stomach that secrete acid, comprise the majority of the stomach The pylorus is the section of the stomach that transitions from stomach to the small bowel; the proximal portion of the pylorus located within the stomach and before the pyloric sphincter

stom-is known as the antrum

All areas of the stomach are covered with mucus- secreting foveolar and columnar epithelial cells lining the luminal surfaces and invagina-tions called gastric pits Midway down the pits are the gastric stem cells; at the base of the pits are the glands Secretory glands in the

> 15.4 9.7–15.4 6.6–9.7 4–6.6

< 4

Figure 31– 1a Estimated worldwide stomach cancer incidence rates per 100,000 in men for the year 2012 Source: GLOBOCAN 2012, IARC, WHO.

> 13 7.2–13 5.5–7.2 3.6–5.5

< 3.6

Figure 31– 1b Estimated worldwide stomach cancer mortality rates per 100,000 in men for the year 2012 Source: GLOBOCAN 2012, IARC, WHO.

Stomach Cancer 595

body and fundus of the stomach produce hydrochloric acid and

intrin-sic factor (parietal cells), serotonin (enteroendocrine or Kulchitsky

cells), and pepsinogen, leptin, and lipase (chief cells) Enteroendocrine

cells in the gastric antrum secrete gastrin (G cells) and somatostatin (D

cells) Cardiac and pyloric glands have neither parietal nor chief cells

and largely secrete mucus The length of the cardiac mucosa increases

with age and with central obesity, often developing histopathological

signs of moderate to severe inflammation (Derakhshan et  al., 2015;

Miao et al., 2014) When H. pylori infection occurs, it typically starts

in the pylorus, extending to the body and fundus over time, causing

infiltration of the mucosa with inflammatory cells— a condition known

as chronic, active gastritis (Correa and Piazuelo, 2008)

Anatomic Subtypes

In the United States, overall rates of stomach cancer have continued

to decline over the last three decades from an age- standardized rate of

11.7 per 100,000 in 1975 to 6.7 per 100,000 in 2013 However, in the

1990s, investigators noted strikingly different epidemiologic trends in

the incidence rate of cardia versus non- cardia tumors: gastric cardia

cancer increased continuously beginning in the mid- 1970s, whereas

the incidence of non- cardia cancers sharply declined (Blot et al., 1993;

Henson et al., 2004; Levi et al., 1990; Powell and McConkey, 1992; Wu

et al., 2009) Since 1990, however, the age- standardized incidence rate

of cardia cancer— unlike adenocarcinoma of the esophagus— seems to

have stabilized in the United States at approximately 2.1 per 100,000

(Figure 31– 3; Wu et al., 2009) In contrast, the incidence rate of non-

car-dia cancer in the overall population has decreased progressively to 4.1

per 100,000, although incidence remains higher among Asians, blacks,

and Hispanic whites (Devesa et al., 1998) Similar trends have been

reported worldwide In some countries of Europe and the Americas,

the incidence of cardia cancers now approximates or exceeds that of

non- cardia cancer among men (Derakhshan et al., 2016; WHO, 2010)

Although the differentiation of anatomical subsite of gastric cers has improved over time (Camargo et  al., 2011a), it can be dif-ficult to determine the origin of large tumors that bridge the lower esophagus and upper stomach (Misumi et  al., 1989)  or that arise in the poorly defined distal boundary of the gastric cardia As of 2016, tumors registered with overlapping (ICD- O C16.8) or unspecified (C16.9) anatomical subsites constitute nearly half (40%– 50%) of all stomach cancers in Japan, Korea, and the United States; 50%– 60%

can-in the United Kcan-ingdom, Italy, and Australia; and up to 90% can-in Brazil (Machii and Saika, 2016)

Histologic Subtypes

As noted earlier, over 90% of stomach cancer cases are nomas arising from the gastric glands (Coleman et al., 1993; WHO, 2010) Several histologic classification systems have been proposed for gastric adenocarcinoma Among these, the Laurén system has been the most influential

adenocarci-Laurén Classification

Described by Pekka Laurén in 1965, the Laurén classification divides gastric cancer into two histological types:  intestinal and diffuse (Laurén, 1965; Nevalainen, 2013) Intestinal- type cancers form rec-ognizable glands, whereas diffuse adenocarcinomas consist of poorly cohesive cells that infiltrate diffusely into the gastric wall with little or

no gland formation (WHO, 2010) These two types are usually easy

to distinguish under the microscope Moreover, since intestinal and diffuse tumors differ in their clinical, epidemiologic, and etiologic characteristics, the Laurén classification provides a useful framework for understanding pathology, epidemiology, and clinical treatment Because many of the seminal epidemiologic and clinical investigations

of gastric adenocarcinoma have employed the Laurén classification

GE Junction Cardia

Figure 31– 2 Cancer Genome Atlas: key features of the four tumor types by gastric site MSI = microsatellite instability; EBV = Epstein Barr virus;

CIN = chro-mosomal instability; GS = genomically stable Source: The Cancer Genome Atlas Research Network, Comprehensive molecular characterization of gastric carcinoma Adapted from Nature 2014;513(7517):202– 209.

adeno-system, the terms “intestinal” and “diffuse” will be used frequently

throughout this chapter

The intestinal subtype historically represented the majority of

gastric adenocarcinomas worldwide, and the decrease in this

sub-type has been the major contributor to the overall decline in stomach

cancer (Figure 31– 3) In contrast to the decreasing incidence rate

of the intestinal subtype, the incidence of diffuse gastric cancers

has increased by over 400% (Henson et  al., 2004) Consequently,

the proportion of gastric cancers contributed by the two subtypes

has changed dramatically over time In the early 1970s, intestinal-

type cancers represented over 90% of gastric adenocarcinomas

in the United States; by 2000, this proportion had decreased by

50% Similar trends have been reported in Europe (Laurén and

Nevalainen, 1993), Latin America (Rampazzo et  al., 2012), and

Asia (Hajmanoochehri et al., 2013; Kaneko and Yoshimura, 2001),

although one study from Sweden showed decreases in both tumor

types (Ekstrom et al., 2000a)

The Laurén classification has also proven useful in evaluating the

natural history of gastric cancer and temporal trends in cancer

inci-dence and precursor lesions, as discussed in the following (Lewin and

Appelman, 1995; WHO, 2010)

WHO Histologic Classification

In 2010, the WHO devised a more granular classification for gastric

cancer (Hu et al., 2012; WHO, 2010) (Table 31– 1) The intent of this

new classification was to harmonize the histological typing of gastric

cancers with that used for cancers of the small and large intestine

Perhaps because of its increased complexity, the WHO classification has been used less frequently in epidemiologic studies than the Laurén system

Among the intestinal- type tumors, tubular carcinomas are the most common, accounting for 56% of stomach tumors in some studies (Terada, 2016) Papillary carcinomas (13% of tumors) tend to occur in the proximal part of the stomach and to be more aggressive than other histologic types, with higher likelihood of metastasis and mortality

1990 20000.1

110

0.010.1110

0.010.11

Tubular adenocarcinomaMucinous adenocarcinoma

Other poorly cohesive carcinoma

Hepatoid adenocarcinomaOncocytic adenocarcinomaPaneth cell carcinomaGastric carcinoma with lymphoid stromaand others

(Hu et al., 2012)

Stomach Cancer 597

Mucinous adenocarcinomas comprise approximately 10% of gastric

adenocarcinomas, and can have features of both intestinal and diffuse-

type malignancies Signet ring cell and other poorly cohesive

carci-nomas comprise the majority of diffuse types and tend to be highly

invasive In their early stage, hereditary gastric malignancies usually

manifest themselves as patchy intramucosal distributions of signet

ring cells (Hu et al., 2012)

Molecular Classifications

In 2014, the Cancer Genome Atlas program (TCGA) tested 295 gastric

cancer samples with array- based analyses of somatic copy number,

whole- exome sequencing, array- based DNA methylation profiling,

messenger RNA sequencing, microRNA sequencing, reverse- phase

protein array, microsatellite instability testing, and, on a subset, low-

pass whole genome sequencing (The Cancer Genome Atlas [TCGA]

Research Network, 2014) These methods identified four molecular

subtypes of gastric cancer: (1) tumors positive for Epstein- Barr virus

(EBV); (2) tumors marked by microsatellite instability (MSI); (3)

genomically stable tumors (GS); and (4) tumors with chromosomal

instability (CIN), characterized by the presence of extensive somatic

copy- number aberrations (The Cancer Genome Atlas Research

Network, 2014) (Figure 31– 2) The last of these subtypes comprised

approximately half of the tested tumors, whereas the EBV tumors

rep-resented less than 10% Although TCGA found no prognostic

signifi-cance associated with these categories in their small data set, they did

identify statistical associations with gender (men were more likely to

have EBV- related tumors), anatomic subsite (CIN more likely in the

cardia and EBV more likely in the body), Laurén classification (75%

of GS were diffuse- type), and age (GS tumors occurred at younger

ages and MSI, older)

Four subtypes with different nomenclature were also identified by

the Asian Cancer Research Group (ACRG), based on gene

expres-sion profiles, and later were confirmed using the three other cohorts,

including TCGA (Cristescu et al., 2015) These subtypes were labeled

microsatellite instability (MSI), microsatellite stable- epithelial to

mes-enchymal transition (MSS/ EMT), MSS/ tumor protein 53 (TP53)

pos-itive, and MSS/ TP53 negative Unlike the TCGA subtypes, the ACRG

subtypes were clearly linked to prognosis Recurrence and survival

were worst for the MSS/ EMT tumors, which typically displayed loss

of e- cadherin (CDH1) expression and diffuse, signet ring

histopathol-ogy Prognosis was best for MSI tumors

Other studies have categorized tumors based on mutations of

sin-gle genes For example, germline mutations in CDH1 account for

most familial gastric cancers More recently, tumors of the cardia and

the gastroesophageal junction have been classified based on

overex-pression of the human epidermal growth factor 2 (HER- 2) receptor

(Gravalos and Jimeno, 2008) Because these more aggressive tumors

may respond favorably to monoclonal antibody treatments,

expres-sion of HER- 2 may prove to be useful as a clinical marker (Bang

et al., 2010)

TUMOR GRADE AND STAGE

The grading of gastric adenocarcinoma is based on the degree of

differentiation (WHO, 2010) Well- differentiated tumors have well-

formed glands that often resemble metaplastic intestinal epithelium,

whereas poorly differentiated tumors are composed of highly irregular

glands that are recognized with difficulty, or single cells that remain

isolated or are arranged in small or large clusters Moderately

differ-entiated tumors display patterns intermediate between well and poorly

differentiated

TNM Staging

The system most often used to stage stomach cancers in the United

States is the American Joint Commission on Cancer (AJCC) TNM

sys-tem The TNM staging considers the size and the extent of the primary

tumor (T), the involvement of regional lymph nodes (N), and presence

or absence of distant metastasis (M) (Washington, 2010) The staging system used by the SEER program classifies cancer cases into three categories: localized, regional, and distant In the United States, 27%

of stomach cancer cases reported to SEER registries between 2006 and

2012 were diagnosed when localized, 28% were regional, and 35% were distant The remaining 10% were classified as unknown stage at diagnosis (Howlader et al., 2015)

Other classification systems for stomach cancer make the tion between early disease (i.e., invasive cancer not extending beyond the submucosa) and advanced disease (i.e., invasion of the muscularis

distinc-or beyond) This approach is commonly used in countries that employ routine endoscopic screening for gastric cancer

PRECANCEROUS OR PRECURSOR LESIONS

For nearly a century, it has been known that gastric cancers with the intestinal- type histology typically originate from mucosa affected by chronic gastritis (Hartfall, 1936) It was not until the advent of mod-ern endoscopy, however, that serial inspection of unautolyzed speci-mens clarified the temporal sequence of preneoplastic histopathology For intestinal- type tumors, the neoplastic cascade begins with chronic gastritis, an inflammatory condition in which both acute and chronic inflammatory cells invade the mucosa, and mucosal cell turnover is accelerated In a subset of people, the chronic gastritis progresses to chronic atrophic gastritis with loss of gastric glands and diminished ability to produce gastric acid As the atrophic gastritis progresses to involve large areas of the stomach, the mucosa may develop islands

of intestinal metaplasia in which the gastric mucosa recapitulates the morphologic appearance of intestinal mucosa These metaplastic zones can have different histologic morphologies termed “complete” (type I) or “incomplete” (types II and III), as discussed later, and can eventually extend to involve much of the stomach After many years, the mucosa in a subset of individuals progresses to low- and then high- grade gastric epithelial dysplasia and ultimately to early and then inva-sive intestinal- type cancer

Preneoplasia

Correa and colleagues (Correa, 2002; Correa et al., 1976; Correa and Yardley, 1992) have proposed a multistep premalignant process for the intestinal- type gastric carcinoma, involving a sequence of histopath-ological changes in the mucosa from normal to non- atrophic gastri-tis, multifocal atrophic gastritis, intestinal metaplasia, and dysplasia

In high- risk populations, chronic gastritis predominantly affects the antrum and is associated with multifocal atrophic gastritis and gland loss (Correa, 2002) The metaplastic process tends to begin at the antrum- corpus junction, then enlarge and extend to the antrum and/

or the corpus A patchy distribution of dysplastic foci may eventually appear within the area of intestinal metaplasia

As mentioned, the two main types of intestinal metaplasia are

“complete” (also called small intestinal type or type I) and plete” (also called colonic type or types II and III) The epithelium

“incom-in the complete type resembles the small “incom-intest“incom-inal phenotype, small intestine digestive enzymes are present, and only sialomu-cins are expressed In incomplete metaplasia, small intestine diges-tive enzymes are absent or only partially expressed, the epithelium resembles the colonic phenotype, and sulfomucins are expressed, either in combination with (type II) or without (type III) sialomu-cins (Camargo et  al., 2011a) Incomplete metaplasia is frequently associated with frank dysplasia and early carcinoma Controversies still exist regarding the utility of subtyping intestinal metaplasia as

a marker of stomach cancer risk However, the results from tive cohorts suggest that complete intestinal metaplasia occurs first and may transform over time into incomplete metaplasia, which has a much higher risk of malignant transformation (Camargo et al., 2014) Furthermore, the great majority of individuals with intestinal- type gastric cancer have some evidence of metaplasia in surrounding tis-sue (Lauwers and Srivastava, 2007)

prospec-Dysplasia (also called intraepithelial neoplasia), which arises in

either native gastric or intestinalized gastric epithelia, is characterized

by partial or complete loss of differentiation (IARC, 2010)

The magnitude of risk of cancer for each of these precursor stages is

difficult to measure, due to their patchy distribution, inconsistent

termi-nology among pathologists, and the challenge of performing repeated

endoscopy in large cohort studies To rectify some of these

prob-lems, several groups (e.g., Padova International Classification, Vienna

classification, Paris endoscopic classification) (Rugge et al., 2000;

Schlemper et al., 2000; The Paris Classification, 2003) have tried to

eliminate inconsistencies and semantic misunderstandings across

coun-tries— especially between Japan and Europe/ North America—

regard-ing the terminology for preneoplastic lesions and early invasive cancers

(IARC, 2010; Stolte, 2003) Despite differences in nomenclature,

how-ever, it is clear that each step increases risk, even though only a

minor-ity of cases ultimately progress from preneoplasia to invasive cancer In

a nationwide cohort study in the Netherlands, the annual incidence of

gastric cancer was 0.10% for patients whose most severe premalignant

lesion was chronic atrophic gastritis, 0.25% for those with intestinal

metaplasia, 0.60% for those with mild- to- moderate dysplasia, and 6.0%

for those with severe dysplasia (de Vries et al., 2008) A large study

from China demonstrated a similar magnitude of increasing risk with

advancing precancerous conditions: < 0.1% patients with atrophy

gas-tritis, 2.7% with deep intestinal metaplasia, and 7% with moderate or

severe dysplasia developed cancer over 5 years (Camargo et al., 2011a)

More recently, data from a low- risk Swedish cohort of over 400,000

people indicated that, over 20 years, one of 256 individuals with

nor-mal gastric mucosa would develop cancer (standardized incidence ratio

[SIR] = 1.0) compared to 1 in 85 for gastritis (SIR = 1.8), 1 in 50 for

atrophic gastritis (SIR = 2.8), 1 in 39 for intestinal metaplasia (SIR =

3.4), and 1 in 19 for dysplasia (SIR = 6.5) (Song et al., 2015)

Precursor lesions can be identified via endoscopy or by using serum

markers for chronic atrophic gastritis Because gastric glands, and the

chief cells in the glands, can be destroyed in atrophic gastritis,

biomark-ers for proteins produced by chief cells have been used to detect this

damage The most studied of these proteins are pepsinogens I and II

Both low pepsinogen I and a low ratio of pepsinogen I to pepsinogen II

have been used clinically (Charvat et al., 2016; Yamaguchi et al., 2016),

although the sensitivity and specificity for the detection of atrophic

gas-tritis are modest (approximately 65% and 85%, respectively) (Huang et

al., 2015) Given the low incidence of gastric cancer and its precursors,

pepsinogen screening is not recommended in the United States or other

low- incidence countries due to the low predictive power of positive

tests (PDQ® Screening and Prevention Editorial Board, 2016)

No precursor lesions have yet been identified for diffuse cancer or

the rarer subtypes of adenocarcinoma

TUMOR PROGRESSION MODELS

In 1975, Pelayo Correa proposed a model for intestinal- type gastric carcinogenesis positing that gastric cancer arose from chronic gastritis and that subsequent steps in tumor progression were caused by other exposures or cofactors (Correa et  al., 1975) With the discovery of

H. pylori in the 1980s and subsequent research confirming its central

role in causing chronic gastritis, Correa revised his model to

incorpo-rate H. pylori infection as the initiating step in carcinogenesis (Correa,

2004) In this model, dietary factors— including salt- preserved foods, high fats, high nitrates, and decreased fruits and vegetables— fostered subsequent tumor progression Bacterial overgrowth in the hypo-chlorhydric stomach was also thought to contribute to mutation and mucosal changes by inducing the formation of carcinogenic N- nitroso compounds and free radicals Some of the stages of the Correa model have been confirmed experimentally, notably the role for salts in mag-

nifying damage in H. pylori– infected stomachs (D’Elia et al., 2014;

Gaddy et  al., 2013) Molecular aspects of this model of tumor gression are also being elucidated For example, metaplasia associated with gastric overproduction of a trefoil protein normally found in the intestine (spasmolytic polypeptide) is a particularly strong marker for cancer risk

pro-A different model of tumor progression was proposed by Houghton

et al (2004) These researchers showed experimentally that H. pylori–

related tumors of the stomach derive from bone marrow– derived stem cells, rather than resident peripheral stem cells or mucosal cells The bone marrow– derived stem cells are recruited to the stomach by the chronic inflammatory process, fuse with gastric epithelial cells, and then replicate, regenerating the mucosal surface and replacing the original stem cells that were destroyed by chronic gastritis (Bessede

et al., 2015) (Figure 31– 4) As with the Correa model, chronic

inflam-mation due to H. pylori infection is the dominant factor in the

carcin-ogenic process

Some studies suggest that precursor lesions can be reversed or their

progression aborted by eradication of Helicobacter pylori infection

(Fukase et  al., 2008) Thus, several international guidelines

recom-mend H.  pylori eradication for those with precancerous conditions

and even early gastric cancer The latter are thought to be true nancies that have not yet invaded the muscularis They can be cured

malig-by the combination of surgical excision and H.  pylori eradication

(Malfertheiner et al., 2012; Rollan et al., 2014)

No analogous tumor progression model has yet been identified for sporadic diffuse gastric cancers However, genome- wide anal-ysis studies (GWAS) have identified several germline mutations associated with hereditary diffuse gastric cancer (HDGC) Here, the initial stages of carcinogenesis are postulated to begin with germline

4 Homing anddifferentiation

EMT affectingdifferentiationduringregenerativehyperplasia

EMT conferingCSC properties

5 Altereddifferentiation/metaplasia

6 Dysplasia

Composed at 22% of BMDCComposed of CD44+ cellswith CSC properties

Figure 31– 4 Stem cell model of gastric carcinogenesis EMT = epithelial- mesenchymal transition; CSC = cancer stem cells; BMDC = bone marrow- derived cells

Source : Bessede E et al., Helicobacter pylori infection and stem cells at the origin of gastric cancer Oncogene 2015;34(20):2547– 2555.

Stomach Cancer 599

mutations— frameshifts, point mutations, or deletions— that typically

affect one of two genes: CDH1 or, less commonly, CTNNA1 (alpha

e- catenin) (Caldas et  al., 1999; Majewski et  al., 2013) The loss of

function of the second allele in HDGC patients— either due to loss of

heterozygosity or promoter hypermethylation— may then lead to

gas-tric cancer (Grady et al., 2000; Majewski et al., 2013)

SURVIVAL

Survival among patients with stomach cancer is still among the lowest

of all cancer sites in most regions of the world (Allemani et al., 2015)

Only cancers of the lung, liver, and pancreas have worse survival

Five- year net survival from stomach cancer diagnosed between 2005

and 2009 is generally in the range of 25% to 30% Five- year relative

survival is slightly better in some Western countries (i.e., up to 33%

in Belgium and Austria) and in China (31%) and substantially better

in Japan and Korea (58% and 54%, respectively) The success in these

latter countries is due to the larger proportion of early- stage, curable

cancers diagnosed through the intensive screening for stomach cancer

in these countries Between the periods 1995– 1999 and 2005– 2009,

survival statistics have shown very large increases in South Korea

(33% to 58%) and China (15% to 31%), but relative survival rose by

less than 10% in most other locations (Allemani et al., 2015)

Diagnosis at an early stage is critical for survival In stomach

can-cer patients in the United States diagnosed between 1992 and 1998,

the overall relative survival at 5 years for all disease stages combined

was 22% (Jemal et al., 2003) Stage- specific survival was the highest

for localized disease (59%), followed by regional disease (22%), and

lowest for distant disease (2%) (Jemal et al., 2003) Between 2006 and

2012, the overall relative survival improved slightly to 30% Stage-

specific survival had also improved since 1992– 1998, with 67%, 31%,

and 5% observed for localized, regional, and distant disease,

respec-tively (American Cancer Society, 2016) Survival is lower for patients

with cancer of the cardia than for non- cardia cancers, even in early

stage disease (Amini et al., 2015)

DESCRIPTIVE EPIDEMIOLOGY

Traditionally, descriptive analyses of stomach cancer have classified

gastric tumors as either a single entity or as cardia versus non- cardia

cancers Mortality data do not capture either the anatomic location

or the histologic characteristics of the tumor Even when tumors are

subclassified as cardia or non- cardia, tumor registration does not

cur-rently capture the heterogeneity that particularly affects cardia

can-cers These are thought to represent a mix of at least two etiologies: the

first involves severe atrophic gastritis due to H. pylori, similar to the

pathway for non- cardia cancers; the second is thought to result from a

transformation of squamous to columnar metaplasia due to the reflux

of gastric contents, similar to the development of esophageal

adeno-carcinoma (Derakhshan et al., 2008; Miao et al., 2014) The former

predominates in high- risk populations where chronic infection is still

common, such as Northwest Iran and central China The latter

pre-dominates in Western countries where central obesity and esophageal

adenocarcinoma are common (Derakhshan et al., 2015)

Cardia and non- cardia cancers differ in their demographic,

geo-graphical, and temporal distributions The extent of these differences

depends on the mix of the two etiologies of cardia cancer (H. pylori–

related versus reflux- related) in a given area We describe in the

fol-lowing the epidemiological features of the most frequent non- cardia

subsite, and highlight the differences with the cardia subsite at the end

of each paragraph, when necessary

Demographic Characteristics

Stomach cancer is extremely rare before the age of 30 After this age,

the age- specific incidence rate of all subsites combined increases

slowly until the age of 50 and then increases more sharply, with

the highest incidence seen in the oldest age groups In the United

States, the median age at diagnosis is 69 for both sexes (Derakhshan

et al., 2009; Sipponen and Correa, 2002) At all ages, men are twice

as likely as women to develop non- cardia stomach cancer and four times as likely to develop cardia cancer Interestingly, the male- to- female ratio varies across countries, and is reported to be only 1.5

in Sub- Saharan Africa This variation probably reflects differences

in the prevalence of the main etiologic factors for cardia cancer,

as mentioned earlier (Colquhoun et  al., 2015) Within each der, African Americans and Asian Pacific Islanders have twice the incidence rate of overall stomach cancer than non- Hispanic whites (Colquhoun et al., 2015)

gen-Within any country or population, non- cardia gastric cancer is most often seen in lower socioeconomic groups and has been associated with risk factors that correlate with lower socioeconomic status (SES; i.e., lower income, educational level, and occupational standing, greater number of siblings, and crowding) These factors also correlate with

H. pylori infection A large European multicenter study reported that adjustment for H. pylori infection eliminated the association between

lower SES and non- cardia gastric cancer (Nagel et al., 2007) Other factors, such as fruit and vegetable consumption, cigarette smoking, and physical activity, may also confound any observed association with SES Notably, although higher SES is inversely associated with non- cardia gastric cancer, it is strongly associated with cardia gastric cancer and esophageal adenocarcinoma, especially in the middle- and high- income countries

Geographic Variation

A global map of male age- standardized incidence rates from GLOBOCAN 2012 shows that the highest rates of stomach cancers occur in Eastern and Southeastern Asia, Eastern Europe, and parts

of Central and South America The map of female incidence rates is nearly identical except that, in any given country, the rates in women are approximately half those in men

Internationally, the incidence of stomach cancer varies mately 10- fold among men, from more than 60 per 100,000 in Korea

approxi-to less than 6 per 100,000 in the United States, Sweden, or Kuwait (Forman et al., 2013) The extremely high incidence rates among men

in Korea (62 per 100,000) and Japan (46 per 100,000) may partly reflect the intensive endoscopic surveillance conducted in these coun-tries Screening detects even very small lesions that might not oth-erwise progress or be diagnosed Even excluding Korea and Japan, however, a four- fold variation in male incidence persists, with rates over 20 per 100,000 in Belarus, Lithuania, the Russian Federation, and Costa Rica Within Europe, there is considerable variation between the countries at highest risk (generally in Eastern Europe) and those with the lowest risk (in Scandinavia, Switzerland, and the United Kingdom)

In the United States, the National Cancer Institute’s Surveillance, Epidemiology, and End Results Program (SEER), estimates a national standardized incidence rate for stomach cancers using nine population- based cancer registries (San Francisco– Oakland, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle– Puget Sound, Utah, and Atlanta) For the period 2008– 2012, annual incidence rates per 100,000 were 10.1 for men and 5.3 for women, for all races combined Age- standardized incidence rates per 100,000 were higher in blacks than in whites for both men (14.6 vs 9.2) and women (8.4 vs 4.5) (Howlader

et al., 2015) In individual registries, age- standardized incidence rates for white men range from 5.4 per 100,000 in Utah to 11.7 in Los Angeles, and those for white women range from 2.4 per 100,000 in Hawaii to 6.9 in Los Angeles Incidence rates of stomach cancer for black men range from 11.5 per 100,000 in Greater California to 19.1

in Louisiana, and for black women from 6.2 per 100,000 in Kentucky

to 10.5 in Connecticut Male rates are approximately two- fold higher than female rates across the registries for both blacks and whites On average, among the 18 SEER registries during the period 2008– 2012, black males had the highest incidence (14.6 per 100,000) followed by Asian or Pacific Islanders (14.5), Hispanic (14.2), American Indian/ Alaska Native (12.3), and white men (9.2) Female rates showed a similar pattern, with the highest incidence among Asian or Pacific

Islanders (8.8), followed by blacks and Hispanics (both 8.4), American

Indians/ Alaska Natives (7.5), and whites (4.5) (Howlader et al., 2015)

Migrant Studies

Comparisons of cancer risk among people who migrate from a low-

risk to a high- risk population, or vice versa, can provide insights into

the relative contributions of inherited versus acquired risk factors In a

landmark study of Japanese migrants to Hawaii and their descendants

(Kolonel et al., 1985), age- adjusted incidence rates of stomach cancer

in the 1970s were lower in Japanese- born migrants to Hawaii (“Issei”

or first generation) than among the Japanese in Japan The rates were

even lower in the Hawaiian- born Japanese (“Nisei” or second

gen-eration), although still higher than the Caucasian rates Migrants to

Australia from seven European countries with higher stomach cancer

rates than Australia (England, Scotland, Ireland, Poland, Yugoslavia,

Greece, and Italy) showed a risk reduction with increased duration of

residence in Australia (McMichael et al., 1980) A study of migrant

populations within Italy (Fascioli et  al., 1995)  suggested that place

of birth was a stronger predictor of stomach cancer risk than current

place of residence Mortality data from several European countries

also showed a closer relation of stomach cancer risk to county of birth

than county of death (Coggon et al., 1990; Spallek et al., 2012),

indi-cating the persistent significance of environmental factors in earlier

life Collectively, these data suggest that environmental factors acting

early in life have a crucial role in gastric carcinogenesis

Temporal Trends

Over the last half century, the overall incidence and death rates from

stomach cancer have decreased steadily almost everywhere The

data-base Cancer Incidence in Five Continents (CI5plus) contains updated

annual incidence rates for 118 selected populations from 102 cancer

registries through 2007 (Ferlay et al., 2013) A comparison over a 30-

year period (1977– 2007) shows that, apart from year- to- year sporadic

fluctuations, the trends of the decline of stomach cancer incidence are

remarkably similar in men and in women, and in high- risk countries

such as Japan and low- risk countries such as Denmark In the United

States, incidence rates have decreased by more than 80% since 1950

(Siegel et al., 2016) The magnitude and consistency of this downward

trend worldwide parallels the decrease in the prevalence of H. pylori.

Analysis of cancer registry data over time has provided an average

estimated annual percentage change (EAPC) in gastric cancer

inci-dence of  – 2.5% per annum (Bray et  al., 2012) The decrease in the

incidence rate is outweighed, however, by the increase in the world’s

population Thus, the absolute number of gastric cancer cases is

expected to increase from 952,000 new cases in 2012 to 1,520,000

new cases by 2030

Mortality trends from GLOBOCAN (Ferlay et al., 2013) show that

the decline in mortality is slightly greater and shows more variability

across countries than incidence, suggesting that some improvement in

survival has accompanied the reduction in the incidence rate A recent

global overview of gastric mortality for the period 1980– 2010

demon-strated an average EAPC of – 3.7% in men in the European Union, –

2.7% in the United States, and – 4.1% in Korea (Ferro et al., 2014)

The rate of decrease has slowed in the United States, France, and some

other European countries In the United States, the EAPC decreased

to  – 1.6 during the interval 2006– 2010 A  European study of birth

cohort trends shows that the rate of decrease has diminished among

cohorts born after the 1940s, particularly among women (Malvezzi

et al., 2010)

Incidence trends also vary by histological subtype and anatomic

site, with widespread decreases in non- cardia and intestinal cancers

but increases in cardia and, less commonly, diffuse cancers (Kaneko

and Yoshimura, 2001; Wu et  al., 2009) The SEER Program in the

United States identified a strong decrease of the intestinal type in both

sexes and all age groups, whereas the diffuse type— particularly the

signet ring cell type— progressively increased (Henson et al., 2004)

The increase in signet ring cell type has not been found in other

populations, however (Chapelle et al., 2016; Kaneko and Yoshimura, 2001) A SEER study by Wu et al showed a 23% increase in cardia cancer (ICD code 16.0) from 1978– 1983 to 1996– 2000, whereas non- cardia cancer (ICD code 16.1– 6) and overlapping/ unspecified cancer

of the stomach (ICD code 16.8– 9) decreased (Wu et al., 2009) (Figure 31– 2) Although the increase in cardia cancer may partly reflect improvements in gastric cancer subsite recording during the period,

a true increase is likely and has been reported in other studies in the United States and Northern Europe, including the United Kingdom (Steevens et  al., 2010) Data from the nine oldest US SEER cancer registries (covering 10% of the population) showed that the incidence rates for cardia cancer significantly increased among whites from 1976

to 2007, but did not change among blacks or other races (Camargo

et al., 2011a) This is consistent with white males having higher rates

of cardia cancer than blacks, in contrast to the opposite pattern for tric cancer overall

gas-ETIOLOGIC FACTORS

The strongest risk factor for stomach cancer identified to date is

chronic infection with the bacterium Helicobacter pylori (H. pylori) Because H. pylori infection was established as a risk factor for stom-

ach cancer relatively recently (IARC, 1994), studies conducted earlier

did not have information on H. pylori infection Some risk factors and

protective factors, such as smoking and certain dietary components,

may be correlated with H.  pylori infection, leading to confounding Alternatively, non– H.  pylori risk factors and protective factors may modify the risk due to H. pylori infection, which is likely since only

a small minority of people infected with H. pylori ever develop

stom-ach cancer Hence, the most informative studies are those that collect

information on both H. pylori and other putative risk factors.

Helicobacter pylori

Historical Aspects

Long before the discovery of H. pylori, gastric adenocarcinoma was

known to arise within areas of gastritis The type of gastritis associated with cancer— termed “chronic active gastritis” or “chronic type B gas-tritis” because of the presence of both lymphocytes and neutrophils— was extremely common in elderly populations and was thought to

be a natural consequence of aging In detailed studies from Northern Europe and Latin America, the majority of individuals over the age

of 50 years were found to have chronic type B gastritis (Correa et al., 1976; Siurala et al., 1985)

H. pylori had been described almost a century (Kreinitz, 1906) before

Marshall and Warren’s groundbreaking work (Warren and Marshall,

1983) However, with the rediscovery of H. pylori in the early 1980s,

the idea that gastric inflammation preceded cancer took on new

mean-ing Experimental ingestions and clinical trials of H. pylori eradication all demonstrated that H. pylori was the dominant cause of type B gas-

tritis (Dixon et al., 1996), prompting reconsideration of traditional

die-tary and genetic theories of gastric carcinogenesis If H. pylori caused

gastritis and gastritis was a precursor to malignancy in the majority

of cases, H. pylori was likely to be a critical factor in carcinogenesis.

In 1994, an expert working group convened by the International Agency for Research on Cancer (IARC) classified infection with

H pylori as carcinogenic to humans, based on its association with

gas-tric adenocarcinoma and MALT lymphoma (IARC, 1994) This clusion was upheld in 2009 by a second IARC working group (IARC,

con-2012a), with the added precision that H pylori was designated a cause

of non- cardia gastric carcinoma, the most common subtype globally It was recently estimated that 89% of non- cardia gastric cancers world-

wide, or 730,000 incident cases in 2012, were attributable to H pylori

(Plummer et al., 2015)

Evidence for Carcinogenicity

Helicobacter pylori is a spiral Gram- negative bacterium that colonizes the stomach Although most infections are asymptomatic, H.  pylori

Stomach Cancer 601

is associated with chronic gastritis, peptic ulcer disease, gastric B-

cell mucosa- associated lymphoid tissue (MALT) lymphoma, and

gastric adenocarcinoma It is believed that H. pylori was once

ubiq-uitous in humans, but that its prevalence has declined in successive

birth cohorts, especially in Western Europe, North America, Oceania,

and Japan, so that infection is now rare among children (Herrera and

Parsonnet, 2009) The risk of H. pylori infection is associated with low

socioeconomic status, particularly with overcrowding and poor

sani-tation (Palli et al., 1994; Rothenbacher et al., 1999) Thus the gradual

disappearance of H. pylori in these regions may be largely a

byprod-uct of economic development The widespread use of antibiotics and

improvements in diet may also have played a role It is noteworthy that

the reduction in H.  pylori prevalence matches the decline in gastric

cancer incidence and mortality

Almost all of the epidemiological evidence on the relationship

between H. pylori and gastric cancer comes from serological

assess-ment of anti– H. pylori immunoglobin G (IgG) antibodies It is now

widely accepted that serological assessment of H. pylori infection has

poor sensitivity in retrospective studies, so that case control studies

systematically underestimate the strength of the association This is

because atrophic gastritis, a precancerous lesion, reduces the burden

of H. pylori infection This in turn decreases titers of IgG antibody,

potentially causing the H.  pylori infection to become serologically

undetectable For this reason, the best evidence of H. pylori prevalence

in relation to cancer comes from prospective studies

The most comprehensive relative risk estimates for H pylori and

gastric cancer come from a pooled analysis of 12 prospective

stud-ies, which included 762 cases of non- cardia gastric cancer and 2250

controls The pooled odds ratio was 2.97 (95% CI = 2.34– 3.77) for H

pylori infection (Forman and Helicobacter and Cancer Collaborative

Group, 2001) The same study included 274 cases of cardia gastric

cancer and 827 controls with an odds ratio of 0.99 (95% CI = 0.40–

1.77) for H pylori infection When the pooled analysis was restricted

to cases occurring at least 10 years after the blood draw used for H

pylori diagnosis, the odds ratio increased to 5.93 (95% CI = 3.41–

10.3) for non- cardia cancer but decreased to 0.46 (95% CI = 0.23–

0.90) for cardia cancer This subgroup analysis underscores the much

stronger relationship with non-cardia than cardia cancer and the need

to account for the effect of premalignant disease on the detection of H

pylori infection, even in prospective studies Further follow- up of the

individual studies contributing to this pooled analysis was reviewed in

the IARC Monographs volume 100 part B, but did not substantively

change the conclusions (IARC, 2012a)

Although H.  pylori infection clearly increases risk for cancers of

the gastric body and antrum, its relationship to cardia cancers has not

been firmly established Some data suggest that these topographical

subtypes may represent a mixture of two different tumors, resembling

either non- cardia adenocarcinoma (H. pylori induced) or esophageal

adenocarcinoma (reflux induced) in their etiology In Western

coun-tries, tumors of the cardia occur more frequently in white males (Yang

and Davis, 1988) and a large proportion occur in the setting of

gastro-esophageal reflux disease (GERD) Conversely, in high- risk areas for

H. pylori and non- cardia gastric cancers, most cancers of the cardia

occur in H. pylori– infected people, and may conceivably be related to

H. pylori infection (Shakeri et al., 2015).

Virulence factors

H.  pylori readily loses and acquires DNA fragments and

under-goes various structural genetic changes such as point mutations and

chromosomal rearrangements (Blaser and Berg, 2001) As a

conse-quence, H.  pylori isolates have an extraordinary degree of genetic

variability between and even within infected hosts, and this diversity

may contribute to the clinical outcome of the infection (Aras et  al.,

2002; Patra et al., 2012) A number of genetic factors associated with

H.  pylori colonization (babA, sabA, alphAB, hopZ, and OipA) and

virulence (cagA, vacA) have been identified (Keilberg and Ottemann,

2016)  (Figure 31– 5) The genetic marker that has attracted most

attention in epidemiological studies is the presence of the cag

path-ogenicity island (PAI), a DNA sequence of 40 kbp present in 70%

of H. pylori strains in Europe and North America, but ubiquitous in

Asia and most of Africa (Peek and Crabtree, 2006) PAI- containing organisms cause greater inflammation and are more closely associated with intestinal- type dysplasia and malignancy than are strains without the PAI (Parsonnet et al., 1997) In contrast, diffuse- type cancers have similar associations with both PAI- positive and PAI- negative isolates Other polymorphic genes— such as the vacuolating cytotoxin and babA- 2 adherence gene— have been linked to pathogenicity, but none

CagA- positive H.  pylori by the detection of anti- cagA antibodies

CagA- positive strains are associated with higher risk of gastric cer than cagA- negative strains A  meta- analysis of 16 cohort and case- control studies including 778 cases of non- cardia gastric cancer and 1409 matched controls found an elevated risk of cagA- positive

can-H. pylori infections, with an odds ratio of 2.01 (95% CI = 1.21– 3.32) for cagA- positivity among all H. pylori– infected individuals (Shiota

et al., 2010)

The cag pathogenicity island is also associated with ous gastric lesions Plummer et al (2007) analyzed a cross- sectional endoscopic survey of 2145 individuals from Venezuela, in which both

precancer-the presence of H. pylori DNA and presence of precancer-the cagA gene were

determined by polymerase chain reaction (PCR) on gastric biopsies

Infection with cagA- positive H. pylori strains but not cagA- negative

strains was associated with the severity of precancerous lesions Using individuals with normal gastric mucosa or superficial gastritis

as controls, the OR for dysplasia was 15.5 (95% CI = 6.4– 37.2) for

cagA- positive H.  pylori compared with 0.90 (95% CI  =  0.37– 2.17) for cagA- negative H. pylori González et al (2011) analyzed a follow-

up study of 312 individuals from Spain with an average of 12.8 years

of follow- up between two endoscopies, also using PCR detection and

genotyping of H. pylori The relative risk for progression of

precancer-ous lesions was 2.28 (95% CI = 1.13– 4.58) for cagA- positive strains compared with cagA- negative strains

Host Response and Other Interacting factors

The host also plays an important role in H. pylori outcome El- Omar

et al (2000) reported that H. pylori– infected subjects who developed

gastric cancer were more likely to have specific genotypes of kin (IL)- 1β or the IL- 1β receptor antagonist Interestingly, the higher risk genotypes of IL- 1β not only induce more inflammation than lower risk genotypes but also increase suppression of gastric acid secretion, supporting the pathogenic model devised by Correa Moreover, the IL- 1β genotype is not associated with cancer risk in the absence of infection A  subsequent study identified similar, though less strong,

interleu-interactions between H.  pylori and TNF- α (Machado et  al., 2003)

Other putative host factors that are being explored include p53 morphisms and variants of the HLA genotype.

poly-Environmental factors such as tobacco smoking, diet, and

medica-tions may also enhance or diminish H. pylori’s deleterious effects,

although relatively few of these studies have measured the effects

of combined exposures Studies of dietary factors are particularly sparse A  prospective Scandinavian cohort demonstrated a protec-tive association with ascorbic acid (vitamin C) and beta- carotene in

H. pylori– infected subjects but not in uninfected subjects (Ekstrom

et al., 2000b) A study by Correa and colleagues on gastric sia and other preneoplastic conditions supported this finding They

dyspla-observed that both H. pylori eradication therapy and dietary

antioxi-dants (ascorbic acid and beta- carotene) prevented preneoplastic

pro-gression Combining antioxidants with H. pylori eradication therapy

did not provide added benefit, however, suggesting that the benefit

of antioxidants may be limited to infected hosts In animals,

die-tary salt magnifies H. pylori– associated gastric carcinogenesis (Fox

et al., 2003); this finding has not been substantiated in humans Non-

dietary exposures that may alter H. pylori outcome include aspirin

and non- steroidal anti- inflammatory drugs (NSAIDs), which appear

to protect against gastric cancer only in infected subjects (Zaridze

et al., 1999) Smoking was associated with increased risk of stomach

cancer in H. pylori– infected subjects (OR = 2.2; 95% CI = 1.2– 4.2)

in a nested case control study from Sweden, where cases and

con-trols were tested by enzyme- linked immunosorbent assay (ELISA)

(Siman et al., 2001) Compared to nonsmokers, active smokers had

significantly higher risk of colonization by cagA- positive virulent

strains and a non- significant increase in bacterial load (Santibanez

et al., 2015) In a population- based case- control study in Germany,

the relative risk of gastric cancer was 2.6 (95% CI  =  1.2– 5.7) for

nonsmoking subjects with CagA- positive H.  pylori infections and

7.2 (95% CI = 2.2– 23.6) for smoking subjects with CagA- positive

H. pylori infections, compared with subjects without these risk

fac-tors (Brenner et al., 2002) Hence, smoking could act indirectly on

cancer risk by promoting the emergence of the more virulent strains

in the gastric mucosa

The age at which H. pylori infection is acquired may also modulate

risk, since infection in high- risk areas is usually acquired in childhood

Only indirect data actually support this hypothesis, however (Blaser

et al., 1995) Nevertheless, the theory remains popular, both because of

its biological plausibility (increased duration of chronic inflammation

increases the risk for genetic mutations to accumulate in gastric

epithe-lium) and the epidemiological patterns of disease (cancer occurs more

frequently in regions where childhood infection is common) Even

among infected children, however, differences in gastric response to

infection exist between high- risk and low- risk regions (Bedoya et al.,

2003), suggesting that age at acquisition cannot explain important

dif-ferences in clinical progression

Epstein- Barr Virus

In 2014, The Cancer Genome Atlas Research Network recognized

EBV- associated gastric cancer (EBVaGC) as one of the four subtypes

of a new molecular classification of gastric adenocarcinoma (The

Cancer Genome Atlas Research Network, 2014) EBVaCG tumors

are distinguished from the other subtypes by higher levels of DNA

methylation in CpG islands of promoter regions and by distinctive

genetic alterations The latter include a high frequency of mutations in

PIK3CA and ARID1A, a mutation in BCOR, and amplification of PD-

L1 and PD- L2 (The Cancer Genome Atlas Research Network, 2014)

In epidemiological studies, the established way to identify EBV

in gastric tumors has been through detection of EBV- encoded small

RNAs (EBERs) or EBV DNA using in situ hybridization (ISH)

(Fukayama and Ushiku, 2011; Hamilton- Dutoit and Pallesen, 1994) Two meta- analyses have now been published of gastric cancer stud-ies that used ISH detection methods (Lee et al., 2009; Murphy et al., 2009) One of these was followed by a pooled analysis of over 5000 cancer cases from 15 populations (Camargo et al., 2011b) The pooled analysis detected EBV in malignant epithelial cells in 9% of all gastric cancers, although with high heterogeneity among the studies

EBVaGC displays distinctive epidemiological and clinical features The proportion of EBV- associated gastric cancer is higher at younger than at older cases, higher in men than in women (11% versus 5%, respectively), and slightly higher in American or Caucasian patients than in Asians (10% versus 8%, respectively) (Lee et al., 2009; Murphy

et al., 2009) EBV- related tumors also typically occur in the gastric body or cardia of the stomach, rather than in the antrum (Takada, 2000), and are common in gastric stumps following gastric resection

A large multicenter case series examined the association of EBV status with survival after gastric cancer diagnosis, accounting for tumor stage and other prognostic factors, and found a lower mortality rate (Hazard Ratio [HR] = 0.72; 95% CI = 0.61– 0.86) (Camargo et al., 2014)

A small percentage (probably less than 1%) of all gastric cancers are categorized histopathologically as lymphoepithelioma- like car-cinomas (LELC) These are epithelial tumors with intense lymphoid infiltration in the stroma, similar in appearance to nasopharyngeal car-cinomas Between 80% and 100% of gastric LELC contain monoclon-ally integrated Epstein- Barr virus (Burke et al., 1990; Herrmann and Niedobitek, 2003; Wu et al., 2000) Apart from these histopathological features and the consistent association with EBV, LELC resembles the conventional form of EBVaGC (Cheng et al., 2015)

How and when EBV acts to induce malignant transformation remains largely unknown In the inflammatory mucosa, EBV is prob-ably transferred from EBV- infected B lymphocytes to gastric epithe-lial cells The viral genome is not integrated into the host genome but becomes circular and episomal within the cytoplasm of infected cells EBV uses the cellular machinery of the host cell to propagate its monoclonal viral genome, epigenetically silence viral and host genes, and control the behavior and microenvironment of the infected cell (Abe et al., 2015) EBV does not replicate in gastric tumors but does express latent genes according to specific latency patterns EBVaGC

Shp2 Abi

IL8

Inflammation

Cell integrity

Proliferationcell polarity

BabA S AlpA/B HopZ/OipA

Figure 31– 5 Adherence and virulence factors used by H pylori to promote direct interactions with epithelial cells H pylori possesses multiple adherence

factors to attach to epithelial cells, including BabA, SabA, AlphA/ B, HopZ, and OipA Adherence is important for CagA delivery via the T4SS CagA is phosphorylated inside the host cell, and alters signaling pathways, leading to loss in cell integrity and alteration of cell proliferation and cell polarity, and

induces inflammation Independent of attachment, VacA is secreted by H pylori and can enter the cells via T5SS VacA leads to vacuolation and apoptosis

of its host cells Source: Keilberg D, Ottemann KM How Helicobacter pylori senses, targets and interacts with the gastric epithelium Environ Microbiol

2016;18(3):791– 806

Stomach Cancer 603

belongs to latency type I or II, and typically expresses EBERs, EBNA-

1, BARTs, and BART miRNAs Approximately half of the cases also

express LMP- 2A (Shinozaki- Ushiku et  al., 2015) The clonality of

the viral genome in tumor cells suggests that infection with EBV is

an early event of gastric carcinogenesis (Fukayama et al., 1994; Imai

et al., 1994)

The relationship between EBV and H pylori is still unclear In vitro

studies have shown that the H pylori– associated oxidant NH2Cl may

convert the latent epithelial EBV infection into lytic infection with

activation of the early gene (Minoura- Etoh et al., 2006) Recently,

seroprevalence of antibodies to 15 specific H pylori proteins using

a multiplex assay showed that the serological profile of H pylori is

equivalent in EBV- positive and in EBV- negative gastric biopsies

obtained from 169 non- cardia gastric cancer patients in five countries

These results suggest that H pylori is the main etiological factor in

EBV- positive gastric cancers, and that EBV might act as a cofactor by

increasing the likelihood of malignant transformation In 2011, EBV

was re- evaluated by an IARC Working Group as part of Monograph

100B The group concluded that there was as yet insufficient

epide-miological evidence to implicate EBV as a cause of gastric cancer

The expert group added the following sentence to their evaluation:

“However, the fact that the EBV genome is present in the tumor cell in

a monoclonal form, and that transforming EBV proteins are expressed

in the tumor cell provides a mechanistic explanation for how EBV

might cause a proportion of gastric cancer” (IARC, 2012a)

Smoking

IARC first evaluated the relationship between tobacco smoking and

gastric cancer in 1985 (IARC, 1986) At that time, the available data

were considered insufficient to conclude that the association between

tobacco smoking and stomach cancer was causal Based on additional

data that had accumulated by 2002, IARC concluded that

confound-ing by other factors, such as alcohol consumption, H pylori

infec-tion, and dietary factors, could be reasonably ruled out (IARC, 2004)

Supporting evidence included dose– response relationships with the

duration of smoking and number of cigarettes smoked daily and a

progressive decrease in the association following cessation The

find-ings from the American Cancer Society’s Cancer Prevention Study

II (Chao et al., 2002), published after the 2002 IARC review, further

supported the conclusion of the IARC Working Group In a 14- year

follow- up of 467,788 men and 588,053 women observed from 1982

through 1996, cigarette smoking and use of other tobacco products

were significantly associated with stomach cancer mortality The

rel-ative risk (RR) of dying from stomach cancer among male smokers

versus never smokers, after adjusting for age, race, education,

fam-ily history of stomach cancer, dietary habits, and aspirin intake— but

not H pylori infection— was 2.16 (95% CI = 1.75– 2.67) for cigarette

smokers and 2.29 (95% CI = 1.49– 3.51) for cigar smokers, with larger

risks observed with increasing smoking duration The magnitude of

association between cigarette smoking and stomach cancer mortality

was smaller but still statistically significant in women

A meta- analysis of prospective studies (cohorts, case cohorts, and

nested case- control studies) reported a summary estimate of relative

risk of 1.62 (1.50, 1.75) in male smokers and 1.20 (1.01, 1.43) in

female smokers, compared to never smokers (Ladeiras- Lopes et  al.,

2008) The same meta- analysis suggested that the association may

vary by subsite, although this is not a consistent finding in the

liter-ature A  systematic review and meta- analysis of nine cohort studies

reported a summary RR of 1.87 (95% CI = 1.31– 2.67) for cardia

can-cers and 1.60 (95% CI = 1.41– 1.80) for non- cardia cancan-cers (Ladeiras-

Lopes et al., 2008)

As noted earlier, smoking was associated with increased risk of

stomach cancer in H pylori– infected subjects in a nested case-

con-trol study in Sweden (OR = 2.2; 95% CI = 1.2– 4.2) (Siman et al.,

2001) and higher risk of colonization by cagA- positive virulent strains

(Santibanez et al., 2015) A population- based case- control study in

Germany reported a relative risk of gastric cancer of 2.6 (95% CI = 1.2–

5.7) for nonsmoking subjects with cagA- positive H pylori infections

and 7.2 (95% CI = 2.2– 23.6) for smoking subjects with cagA- positive

H pylori infections, compared with subjects without these risk factors (Brenner et al., 2002) Smoking and H pylori strains, and especially

the cagA- positive, appear to act synergistically in increasing the risk

of non- cardia cancer In 2004 (IARC, 2004) and again in 2009 (IARC, 2012b), IARC concluded that active tobacco smoking increases the risk of gastric cancer but that the evidence for involuntary exposure to tobacco smoke was inconclusive

Food and Nutrition

In June 2015, an expert panel was convened by the World Cancer Research Fund (WCRF) and the American Institute for Cancer Research (AICR) to update a previous report from 2007 and to grade nutritional risk factors for stomach cancer on a five- level scale, accord-ing to the strength of the evidence (1  =  convincing; 2  =  probable;

3 = suggestive; 4 = inconclusive; and 5 = unlikely) (Wiseman, 2008; World Cancer Research Fund International/ American Institute for Cancer Research [WCRF/ AICR], 2016) In this report, only prospec-tive studies (cohort, nested case control, and randomized controlled trials) were considered, and the two anatomic subsites of the stom-ach were analyzed separately when possible The authors reported on

89 new studies conducted around the world, comprising 77,000 cases

of stomach cancers; they noted, however, that whereas most studies

adjusted for tobacco smoking, few did so for H. pylori infection.

According to the WCRF/ AICR panel, no dietary risk factor for tric cancer met the criteria for level one (convincing evidence), but several qualified for level two The evidence was considered probable that consuming more than three drinks (45 g) of alcohol per day and eating processed meat or foods preserved by salting increased the risk

gas-of stomach cancer A high body mass was considered to be a ble risk factor for cardia cancer Eating grilled/ barbecued meat or fish and low consumption of fruit were considered suggestive risk factors (level 3) for both cardia and non- cardia cancers Citrus fruit consump-tion was designated a suggestive protective factor for cardia cancer (WCRF/ AICR, 2016)

proba-fruits and Vegetables

In a meta- analysis based on 22 cohort studies looking at all ach cancers irrespective of anatomic location, Wang et al (2014) reported a small but statistically significant 10% decreased risk of stomach cancer in the high versus low consumers of fruits (RR = 0.90; 96% CI = 0.83– 0.98) (Wang et al., 2014) Similarly, a dose response analysis by the WCRF panel reported a 5% reduction in cancer per 100 g of fruit consumed daily (RR = 0.95; 95% CI = 0.91– 0.99) The association was nonlinear, with most of the risk reduction observed in the low to middle categories of intake The association was strongest when considering only European studies (RR = 0.81; 95% CI = 0.68– 0.96) or subgroups of high- quality stud-ies conducted elsewhere No inverse association was found between vegetable intake and gastric cancer risk The WCRF expert panel concluded that these data and others published up to 2014 were rea-sonably consistent to classify the evidence for low fruit consump-tion and increased risk of both cardia and non- cardia cancers as level 3 (suggestive)

stom-Three studies conducted in Europe and in the United States reported that citrus fruit consumption was associated with decreased risk of cardia gastric cancer when comparing the highest to lowest catego-ries of intake (Freedman et al., 2008; Gonzalez et al., 2012; Steevens

et al., 2011) A meta- analysis of the dose response in these three ies showed a significant decrease of 24% per 100 g of citrus fruit consumed per day (RR = 0.76; 95% CI = 0.58– 0.99) (WCRF/ AICR, 2016) Given the paucity of published studies, however, the evidence regarding citrus consumption and cardia gastric cancer was considered only suggestive

stud-Salt and stud-Salted Preserved food

According to WCRF/ AICR, a relationship between gastric cer incidence and consumption of salt- preserved foods is strongly supported by the literature (WCRF/ AICR, 2016) Two meta- analy-ses showed that high intake of salt- preserved vegetables (“pickled

food”) was associated with summary relative risks for gastric

can-cer of 1.27 (95% CI = 1.09– 1.49) and 1.32 (95% CI = 1.10– 1.59),

respectively (D’Elia et al., 2012; Ren et al., 2012) A subsequent

meta- analysis of nine studies by the WCRF/ AICR expert panel

identified a dose response with no heterogeneity Risk increased by

9% per 20 g increase in daily consumption of salt- preserved

veg-etables (RR = 1.09; 95% CI = 1.05– 1.13) (WCRF/ AICR, 2016)

Salt consumption was a particularly strong predictor of gastric

can-cer in studies in Japan In contrast, a population study in Norway

found no association between high consumption of dietary salt and

risk of gastric cancer (Sjodahl et al., 2008b) In the few studies that

controlled for H pylori infection, the effect of salt appeared to be

independent of cagA status but was stronger in the presence of H

pylori– associated atrophic gastritis (Peleteiro et al., 2011; Shikata

et al., 2006) Animal models have not shown that increased dietary

salt increases the risk of H pylori– associated cancer (Rogers et al.,

2005) Cardia and non- cardia tumors were not investigated

sepa-rately in any of these studies

Salt itself is not carcinogenic, but salted foods contain lower

amounts of micronutrients and may undergo fermentation during

pres-ervation Salt causes mucosal damage in the stomach, induces

inflam-mation, and increases DNA synthesis and cell proliferation, all of

which can facilitate carcinogenesis (Ames and Gold, 1990; Charnley

and Tannenbaum, 1985)

Other Dietary factors

Studies of processed meat consumption (i.e., sausages, bacon,

meat-balls, ham, or pieces of meat having undergone smoking,

fermen-tation, or salt preservation) have been recently reviewed by both

WCRF/ AICR and IARC (Bouvard et al., 2015; WCRF/ AICR, 2016)

Processed meat was classified as carcinogenic to humans (Group 1)

by IARC, based on sufficient evidence in humans regarding

colorec-tal cancer For stomach cancer, however, the evidence was considered

limited Based on the results of three cohort studies, the WCRF

des-ignated consumption of processed meat as a probable cause of non-

cardia cancer (Cross et al., 2011; Gonzalez et al., 2006; Keszei et al.,

2012)  with the mechanism possibly relating to nitrites and nitrate

(Takahashi et  al., 1994) A  dose– response meta- analysis showed an

18% increased risk for non- cardia cancer per 50 g of processed meat

consumed per day (RR = 1.18; 95% CI = 1.01– 1.38) No association

was found with cardia cancer

Diets high in grilled or barbecued meat or fish were also deemed

to possibly increase stomach cancer risk Despite a plausible

bio-logical mechanism (i.e., the formation of carcinogenic

heterocy-clic amines when meat and fish are cooked at high temperature),

data specifically linking heterocyclic amines to gastric cancer are

lacking Similarly, although laboratory studies support a causal link

between N- nitroso compounds and stomach cancer, the WCRF/

AICR panel designated the evidence for carcinogenicity of N-

nitroso compounds in humans to be limited/ suggestive (WCRF/

AICR, 2016)

Few studies have been able to accurately evaluate the effect of

vitamin C on the risk of gastric cancer Within the EPIC cohort,

higher plasma vitamin C levels were associated with a lower risk of

cancer, and did not appear to be limited to a particular anatomical

subsite or histological subtype (Jenab et al., 2006) In contrast,

die-tary vitamin C showed no significant association with cancer risk

The WCRF expert panel was unable to draw any conclusion about

vitamin C based on available evidence (Wiseman, 2008; WCRF/

AICR, 2016)

Alcohol

A possible relationship between alcoholic beverage consumption

and risk for stomach cancer has long been hypothesized, but in

most prospective studies, no overall association has been observed

(IARC, 2010, 2012c) A 2010 meta- analysis by anatomic subtype

found no statistically significant association between heavy

drink-ing and either non- cardia cancer (pooled RR =1.17; 95% CI = 0.78–

1.75) or cardia cancer (RR = 0.99; 95% CI = 0.67– 1.47) (Tramacere

et al., 2012) The strongest evidence for a causal association comes from a meta- analysis of 30 studies (12,000 cases) conducted by WCRF/ AICR in 2015 (WCRF/ AICR, 2016) A dose– response trend

in risk was observed with increasing alcohol consumption above 45

g of ethanol daily The relative risk estimates were 1.06 (95% CI = 1.01– 1.11), 1.15 (95% CI = 1.06– 1.26), and 1.28 (95% CI = 1.08– 1.52) for alcohol intake of 45 g, 80 g, and 120 g per day However, alcohol consumption is difficult to evaluate as an independent vari-able in observational studies due to underreporting, and its associa-tion with smoking and poor nutrition

Ionizing Radiation

Ionizing radiation increases risk of gastric carcinoma (IARC, 2012c) The best evidence comes from the longitudinal study of 38,576 atomic- bomb survivors in Hiroshima and Nagasaki, Japan, followed between 1980 and 1999 (Sauvaget et al., 2005) For a person- years weighted mean dose of 1.6 Gy, the RR was 1.71 (95% CI = 1.27– 2.30) compared to the lowest dose, with a significant dose response trend

(p = 0.009).

Body Mass Index and Physical Activity

Overweight and obesity have been associated with increased risk of many cancers (Lukanova et  al., 2006) A  meta- analysis has shown

an elevated risk for cardia gastric cancer with a summary relative risk estimate of 1.4 (95% CI  =  1.16– 1.68) for overweight (body mass index [BMI] 25– 30), and 2.06 (95% CI = 1.63– 2.61) for obe-sity (BMI ≥ 30) (Yang et al., 2009) A dose response meta- analysis

by WCRF stratified by geographic regions confirmed the positive association in Europe and North America, but not in Asia (WCRF/ AICR, 2016) In Western countries, cardia cancer occurs predom-inantly in obese white men, presumably with similar pathogenesis

as Barrett’s esophagus and esophageal adenocarcinoma It is ble that in healthy volunteers without reflux symptoms, the cardiac mucosa lengthens with the age of the subjects and increasing abdom-inal obesity and histologically resembles Barrett’s esophagus The mechanism by which BMI is associated with cardia cancer could therefore involve not only chronic inflammatory states and hormo-nal disruptions associated with obesity, but also acidic damage at

nota-or around the esophagogastric junction, caused by subclinical acidic reflux (Derakhshan et al., 2015)

No association has been observed between BMI and non- cardia gastric cancer (Corley et al., 2008; Kuriyama et al., 2005; Sjodahl

et al., 2008a) An inverse relationship has been reported between ular physical activity and non- cardia gastric cancer Among many studies that have looked at physical activity in cancer patients, two recent prospective studies specifically examined gastric cancer, and both found a protective association (Leitzmann et al., 2009; Sjodahl

reg-et al., 2008a)

HOST FACTORS

Familial Risk

Increased risk of stomach cancer has long been observed in those with

a family history of stomach cancer (Terry et al., 2002) This moderate

to strong increase in risk does not necessarily indicate heritable ceptibility, however, since many of the established and suspected risk factors for stomach cancer tend to aggregate in families These include

sus-H. pylori infection, smoking and dietary habits.

Hereditary cancer syndromes involving familial clustering of ach cancer have been recognized for more than 50 years and account for up to 3% of cases (McLean and El- Omar, 2014) In a study of Maori kindred in New Zealand, the familial pattern was consistent with dominant inheritance of a susceptibility gene with incomplete pene-trance (Guilford et al., 1998) A linkage analysis of this and two other Maori pedigrees with early- onset, diffuse- type stomach cancer and

Stomach Cancer 605

subsequent molecular genetic studies led to a discovery that these

pedi-grees had germline mutations in the E- cadherin/ CDH1 gene (Guilford

et  al., 1998) Germline mutations in the E- cadherin gene have also

been found in stomach cancer pedigree studies of families of European

(Gayther et al., 1998; Guilford et al., 1999), Japanese (Shinmura et al.,

1999; Yabuta et  al., 2002), and Korean (Yoon et  al., 1999)  descent

These familial cases led to the recognition of a new syndrome defined

by Guilford et al (1998) as hereditary diffuse gastric cancer (HDGC)

Male and female carriers of germline E- cadherin gene mutations

included in studies of the International Gastric Cancer Linkage

Consortium have a greater than 80% cumulative risk of stomach

cancer by age 80 (Fitzgerald et al., 2010) Somatic mutations in the

E- cadherin gene are frequently detected in the diffuse but not the

intes-tinal type of stomach cancer (Becker et al., 1994) The critical role of

E- cadherin in cell- cell adhesion is notable in this regard Inactivating

mutations in CDH1 are also common in sporadic gastric tumors as

well as HDGC (Pinheiro et al., 2014) Two mutations induce complete

inactivation of the CDH1 gene, which initiates the carcinogenic

proc-ess of diffuse gastric cancer

Until very recently, CDH1 was the only identified gene related to

HDGC However, approximately 60% of HDGC tumors do not

har-bor a defective CDH1 allele In 2013, a CTNNA1 germline truncating

mutation was found in a large HDGC pedigree, prompting an active

search for other susceptibility genes (Majewski et al., 2013; Pinheiro

et al., 2014)

Stomach cancer has also been observed as part of other

heredi-tary cancer predisposition syndromes, as shown in Table 31– 2 These

include hereditary breast/ ovarian cancer due to germline BRCA1 and

BRCA2 mutations (Brose et al., 2002); Lynch syndrome (Chen et al.,

2006; Watson et  al., 2008); Li- Fraumeni syndrome due to germline

p53 mutations (Gonzalez et  al., 2009); and the much rarer familial

adenomatous polyposis (Garrean et al., 2008), Peutz- Jeghers (Howe

et  al., 2004), and juvenile polyposis syndromes (Giardiello et  al.,

2000) associated with germline mutations in the APC gene, SMAD4

and BMPR1A genes, and STK11 gene, respectively (reviewed in Chun

and Ford, 2012)

Lynch syndrome carries a lifetime risk of around 10% for gastric

cancer, considerably lower than the risk of colorectal or endometrial

cancer associated with this syndrome (Chen et al., 2006; Watson et al.,

2008) Approximately 90% of cases have intestinal type histology

A  significantly increased risk of stomach cancer (SIR  =  2.78; 95%

CI = 1.59– 4.52) was observed in Swedish hereditary prostate cancer

families (Gronberg et  al., 2000), but germline E- cadherin mutations

do not seem to contribute to the elevated stomach cancer risk in these

families (Jonsson et al., 2002)

Other Genetic Factors

The role of host susceptibility genes in stomach cancer has also

been investigated in populations without strong familial

aggrega-tion For instance, a higher incidence of stomach cancer in blood

type A  individuals than in those with blood type O was noticed as

early as the 1950s (Aird et al., 1953) Prevalence of chronic atrophic gastritis, intestinal metaplasia, and dysplasia is also higher in sub-jects with blood type A than type O (Haenszel et al., 1976; Kneller

et al., 1992) Numerous studies including a large Swedish and Danish cohort of blood donors (Edgren et  al., 2010)  and a meta- analysis published in 2012 (Wang et al., 2012) confirmed these observations Individuals with blood type A  demonstrated a higher risk of gas-tric cancer (OR = 1.20; 95% CI = 1.02– 1.42; and OR = 1.11; 95%

CI = 1.05– 1.15) than donors with other blood types in the Swedish and Danish cohorts, respectively Further studies found that adher-

ence of H. pylori to human gastric epithelium can be mediated by the blood- group antigen- binding adhesin (BabA) produced by H. pylori

that targets human fucosylated blood group antigens H type I (type O substance) and Lewis b (Leb) (Prinz et al., 2001) The presence of the

babA2 gene, encoding for BabA, in the H. pylori genome is crucial for H. pylori– related pathogenesis and correlates with the activity of

gastritis in the infected stomach

The establishment of H. pylori infection as a risk factor, along with

advances in molecular genetic techniques, has facilitated studies of

interaction between H.  pylori and host genetic factors with regard

to stomach cancer risk Despite the consistent association between

H. pylori infection and stomach cancer risk, only a small fraction of

the infected individuals develop stomach cancer (Parsonnet, 1999)

It is plausible to hypothesize that some individuals are more

suscep-tible to acquiring persistent infection when exposed to H. pylori and

to develop preneoplastic lesions and eventually cancer once infection persists Such variation in susceptibility may be due to interindividual

variability regarding response to and interaction with H.  pylori As mentioned earlier, El- Omar et al (2000) reported that IL- 1 gene clus-

ter polymorphisms were associated with gastric cancer Polymorphisms

in these clusters are suspected of enhancing IL- 1β, an important pro- inflammatory cytokine and a powerful inhibitor of gastric acid secretion.Many studies have investigated a possible association between selected candidate genes and gastric cancer risk (see comprehen-sive review by McLean and El- Omar, 2014) Genes coding for the inflammatory proteins IL- 1β, IL- 1ra, IL- 8, IL- 10, and TNF- α were recently examined in a systematic review and meta- analysis by the Human Genome Epidemiology (HuGE) project (Persson et  al., 2011) The analyses revealed risk differences among histologic

types, anatomic sites, geographic locations, and H. pylori infection

status An increased risk for gastric cancer was clearly observed for IL- 1RN2; however, this risk appeared to be confined to non- Asian populations and was particularly observed for non- cardia cancer,

of both intestinal and diffuse types Conversely, in Asian

popula-tions, IL- 1B- 31C carrier status was associated with a reduced risk

of gastric cancer (Persson et al., 2011) In recent years, IL- 17 has been identified as another candidate gene The IL- 17 family of pro-

inflammatory cytokines has been implicated in many inflammatory- driven diseases, including autoimmune diseases and colorectum or breast cancers Several studies and meta- analyses have evaluated the

187G > A  polymorphism in the IL- 17A gene (rs2275913), which

seems to be associated with an increased risk of gastric cancer, cially in Asian populations (Dai et al., 2016)

espe-Table 31– 2 Inherited Predisposition Syndromes for Gastric Cancer

Hereditary diffuse gastric cancer

PMS2, Epcam

et al., 2008

Genome- wide association studies (GWAS) have also identified

novel susceptibility loci that may offer new insights about gastric

car-cinogenesis For example, gene variant rs2294008, a polymorphic

var-iation in the gene encoding for the prostate stem cell antigen (PSCA),

was identified in 2008 as statistically significantly associated with

dif-fuse gastric cancer in a Japanese population (Sakamoto et al., 2008)

This association was later confirmed in other populations in China

and in Europe, with a similar strength of association with diffuse and

intestinal cancer types (Lu et al., 2010; Sala et al., 2012) Recently

the rs2294008T gene variant has been associated with a worse overall

survival in patients with diffuse type cancers (HR = 1.85; 95% CI =

1.12– 3.06) (Garcia- Gonzalez et al., 2015)

Other Predisposing Conditions

Pernicious anemia, an autoimmune disorder characterized by

atrophic damage restricted to the gastric body mucosa (gastric

atro-phy type A), predisposes to gastric adenocarcinoma, with an

inci-dence rate similar to that seen in gastric atrophy due to H.  pylori

(gastric atrophy type B) (Annibale et  al., 2000) This excess risk

seems to be independent of H. pylori infection, although the

poten-tial interaction between infection and pernicious anemia has not yet

been thoroughly studied

Prior gastric surgery for benign disorders (mainly gastric ulcer) was

linked to increased risk of adenocarcinoma long before the discovery

of H. pylori However, it is not clear if prior gastric surgery is merely a

surrogate for long- term H. pylori infection with more aggressive cagA

strains or if it causes increased risk in the remnant stomach, perhaps

acting synergistically with H. pylori through adverse consequences of

the surgery, such as bile reflux) (Leivonen et al., 1997; Mezhir et al.,

2011) To date, bariatric surgery has not been linked to gastric cancer

risk (Orlando et al., 2014), but as obesity prevalence increases and the

procedures correspondingly become more common, further research

will be needed

FUTURE RESEARCH

The history of gastric cancer is one of impressive, although

inadvert-ent, success Despite the decreasing incidence of this disease, gastric

cancer remains common in many areas of the world, and the absolute

number of cases is increasing Several areas of epidemiologic

investi-gation need attention First, we need improved surveillance for gastric

cancer, particularly in low- income countries where this malignancy is

frequently missed or misdiagnosed Hand in hand with this

surveil-lance, research is needed to dissect the epidemiologic significance of

the various molecular subtypes of gastric cancer that have recently

been identified This research will, in turn, stimulate development

of new screening methodologies as well as development of targeted

treatments Studies indicate that endoscopic screening for gastric

can-cer in high- risk areas with resection of early tumors results in longer

survival, but the extent to which this favorable trend reflects lead- time

bias has not been established What also remains to be established

is the best public health strategy for managing H.  pylori infection,

both in high and low gastric cancer risk countries; although H. pylori

treatment appears to reduce gastric cancer incidence, the broader

consequences of large- scale antibiotic use, the optimal therapy in an

age of increasing antibiotic resistance, and the best implementation

strategies need to be defined on local levels Finally, much more work

needs to be focused on understanding the epidemiology and

preven-tion of gastric cardia cancer that, in the future, could very well surpass

non- cardia cancer in incidence worldwide

References

Abe H, Kaneda A, and Fukayama M 2015 Epstein- Barr virus- associated

gas-tric carcinoma: use of host cell machineries and somatic gene mutations

Pathobiology, 82(5), 212– 223 PMID: 26337667

Aird I, Bentall HH, and Roberts JA 1953 A relationship between cancer

of stomach and the ABO blood groups Br Med J, 1(4814), 799– 801

PMCID: PMC2015995

Allemani C, Weir HK, Carreira H, et al 2015 Global surveillance of cancer survival 1995– 2009: analysis of individual data for 25,676,887 patients from 279 population- based registries in 67 countries (CONCORD- 2)

Lancet, 385(9972), 977– 1010 PMID: 25467588

American Cancer Society 2016 Cancers facts & figures:  stomach cancer

Atlanta: American Cancer Society

Ames BN, and Gold LS 1990 Too many rodent carcinogens:  mitogenesis

increases mutagenesis Science, 249(4972), 970– 971 PMID: 2136249.

Amini N, Spolverato G, Kim Y, et al 2015 Clinicopathological features and prognosis of gastric cardia adenocarcinoma:  a multi- institutional US

study J Surg Oncol, 111(3), 285– 292 PMID: 25308915.

Annibale B, Lahner E, Bordi C, et al 2000 Role of Helicobacter pylori tion in pernicious anaemia Dig Liver Dis, 32(9), 756– 762.

infec-Aras RA, Small AJ, Ando T, and Blaser MJ 2002 Helicobacter pylori strain restriction- modification diversity prevents genome subversion by

inter-chromosomal DNA from competing strains Nucleic Acids Res, 30(24),

5391– 5397 PMCID: PMC140068

Bang YJ, Van Cutsem E, Feyereislova A et al 2010 Trastuzumab in nation with chemotherapy versus chemotherapy alone for treatment of HER2- positive advanced gastric or gastro- oesophageal junction can-

combi-cer (ToGA): a phase 3, open- label, randomised controlled trial Lancet,

376(9742), 687– 697 PMID: 20728210

Becker KF, Atkinson MJ, Reich U, et al 1994 E- cadherin gene mutations

pro-vide clues to diffuse type gastric carcinomas Cancer Res, 54(14), 3845–

3852 PMID: 8033105

Bedoya A, Garay J, Sanzon F, et  al 2003 Histopathology of gastritis in Helicobacter pylori- infected children from populations at high and low

gastric cancer risk Hum Pathol, 34(3), 206– 213 PMID: 12673553.

Bessede E, Dubus P, Megraud F, and Varon C 2015 Helicobacter pylori

infec-tion and stem cells at the origin of gastric cancer Oncogene, 34(20),

2547– 2555 PMID: 25043305

Blaser MJ, and Berg DE 2001 Helicobacter pylori genetic diversity and risk

of human disease J Clin Invest, 107(7), 767– 773 PMCID: PMC199587.

Blaser MJ, Chyou PH, and Nomura A 1995 Age at establishment of Helicobacter pylori infection and gastric carcinoma, gastric ulcer, and

duodenal ulcer risk Cancer Res, 55(3), 562– 565.

Blot WJ, Devesa SS, and Fraumeni JF, Jr 1993 Continuing climb in rates

of esophageal adenocarcinoma:  an update JAMA, 270(11), 1320

PMID: 8360967

Bouvard V, Loomis D, Guyton KZ, et al 2015 Carcinogenicity of

consump-tion of red and processed meat Lancet Oncol, 16(16), 1599– 1600

PMID: 26514947

Bray F, Jemal A, Grey N, Ferlay J, and Forman D 2012 Global cancer sitions according to the Human Development Index (2008– 2030):  a

tran-population- based study Lancet Oncol, 13(8), 790– 801 PMID: 22658655.

Brenner H, Arndt V, Bode G, et al 2002 Risk of gastric cancer among

smok-ers infected with Helicobacter pylori Int J Cancer, 98(3), 446– 449

PMID: 11920598

Brose MS, Rebbeck TR, Calzone KA, et al 2002 Cancer risk estimates for

BRCA1 mutation carriers identified in a risk evaluation program J Natl Cancer Inst, 94(18), 1365– 1372 PMID: 12237282

Burke AP, Yen TS, Shekitka KM, and Sobin LH 1990 Lymphoepithelial noma of the stomach with Epstein- Barr virus demonstrated by polymer-

carci-ase chain reaction Mod Pathol, 3(3), 377– 380 PMID: 2163534.

Caldas C, Carneiro F, Lynch HT, et  al 1999 Familial gastric cancer: 

over-view and guidelines for management J Med Genet, 36(12), 873– 880

PMID: 10593993

Camargo MC, Anderson WF, King JB, et al 2011a Divergent trends for gastric

cancer incidence by anatomical subsite in US adults Gut, 60(12), 1644–

1649 PMCID: PMC3202077

Camargo MC, Kim WH, Chiaravalli AM, et  al 2014 Improved survival of gastric cancer with tumour Epstein- Barr virus positivity: an international

pooled analysis Gut, 63(2), 236– 243 PMCID: PMC4384434.

Camargo MC, Murphy G, Koriyama C, et al 2011b Determinants of Epstein-

Barr virus- positive gastric cancer: an international pooled analysis Br J Cancer, 105(1), 38– 43 PMCID: PMC3137422

Chao A, Thun MJ, Henley SJ, et  al 2002 Cigarette smoking, use of other tobacco products and stomach cancer mortality in US adults: The Cancer

Prevention Study II Int J Cancer, 101(4), 380– 389 PMID: 12209964.

Chapelle N, Manfredi S, Lepage C, et al 2016 Trends in gastric cancer dence: a period and birth cohort analysis in a well- defined French popula-

inci-tion Gastric Cancer, 19(2), 508– 514 PMID: 26118904.

Charnley G, and Tannenbaum SR 1985 Flow cytometric analysis of the effect

of sodium chloride on gastric cancer risk in the rat Cancer Res, 45(11 Pt

2), 5608– 5616 PMID: 4053034

Charvat H, Sasazuki S, Inoue M et al 2016 Prediction of the 10- year probability

of gastric cancer occurrence in the Japanese population: the JPHC study

cohort II Int J Cancer, 138(2), 320– 331 PMID: WOS:000369161200007.

Stomach Cancer 607

Chen S, Wang W, Lee S, et  al 2006 Prediction of germline mutations and

PMCID: PMC2538673

Cheng N, Hui DY, Liu Y, et al 2015 Is gastric lymphoepithelioma- like

car-cinoma a special subtype of EBV- associated gastric carcar-cinoma? New insight based on clinicopathological features and EBV genome polymor-

phisms Gastric Cancer, 18(2), 246– 255 PMID: 24771002.

Chun N, and Ford JM 2012 Genetic testing by cancer site: stomach Cancer J,

18(4), 355– 363 PMID: 22846738

Coggon D, Osmond C, and Barker DJ 1990 Stomach cancer and

PMCID: PMC1971373

Coleman MP, Esteve J, Damiecki P, Arslan A, and Renard H 1993 Trends

PMID: 8258476

Colquhoun A, Arnold M, Ferlay J, et al 2015 Global patterns of cardia and

non- cardia gastric cancer incidence in 2012 Gut, 64(12), 1881– 1888

PMID: 25748648

Corley DA, Kubo A, and Zhao W 2008 Abdominal obesity and the risk of

esophageal and gastric cardia carcinomas Cancer Epidemiol Biomarkers

Correa P 2002 Gastric neoplasia Curr Gastroenterol Rep, 4(6), 463– 470

PMID: 12441036

Correa P 2004 The biological model of gastric carcinogenesis In: Buffler P,

Rice J, Baan R, Bird M, and Boffetta P (Eds.), Mechanisms of ogenesis:  contributions of molecular epidemiology (Vol 157, pp 301–

carcin-310) Lyon: International Agency for Research on Cancer

Correa P, Cuello C, Duque E, et  al 1976 Gastric cancer in Colombia III

Natural history of precursor lesions J Natl Cancer Inst, 57(5), 1027–

1035 PMID: 1003539

Correa P, Haenszel W, Cuello C, Tannenbaum S, and Archer M 1975 A model

for gastric cancer epidemiology Lancet, 2, 58 PMID: 49653.

Correa P, and Piazuelo MB 2008 Natural history of Helicobacter pylori

infec-tion Dig Liver Dis, 40(7), 490– 496 PMCID: PMC3142999.

Correa P, and Yardley JH 1992 Grading and classification of chronic

gas-tritis:  one American response to the Sydney system Gastroenterology,

102(1), 355– 359 PMID: 1727769

Cristescu R, Lee J, Nebozhyn M, et  al 2015 Molecular analysis of gastric

cancer identifies subtypes associated with distinct clinical outcomes Nat

Cross AJ, Freedman ND, Ren J, et  al 2011 Meat consumption and risk

of esophageal and gastric cancer in a large prospective study Am J Gastroenterol, 106(3), 432– 442 PMCID: PMC3039705

D’Elia L, Galletti F, and Strazzullo P 2014 Dietary salt intake and risk of

gas-tric cancer Cancer Treat Res, 159, 83– 95 PMID: 24114476.

D’Elia L, Rossi G, Ippolito R, Cappuccio FP, and Strazzullo P 2012 Habitual

salt intake and risk of gastric cancer: a meta- analysis of prospective

stud-ies Clin Nutr, 31(4), 489– 498.

Dai ZM, Zhang TS, Lin S, et  al 2016 Role of IL- 17A rs2275913 and IL-

17F rs763780 polymorphisms in risk of cancer development: an updated

meta- analysis Sci Rep, 6, 20439 PMCID: PMC4740815.

de Vries AC, van Grieken NC, Looman CW, et al 2008 Gastric cancer risk in

patients with premalignant gastric lesions: a nationwide cohort study in

the Netherlands Gastroenterology, 134(4), 945– 952 PMID: 18395075.

Derakhshan MH, Arnold M, Brewster DH, et  al 2016 Worldwide inverse

association between gastric cancer and esophageal adenocarcinoma

sug-gesting a common environmental factor exerting opposing effects Am J Gastroenterol, 111(2), 228– 239 PMID: 26753891

Derakhshan MH, Liptrot S, Paul J, et  al 2009 Oesophageal and gastric

due to a 17  year delayed development in females Gut, 58(1), 16– 23

PMID: 18838486

Derakhshan MH, Malekzadeh R, Watabe H, et  al 2008 Combination of

gastric atrophy, reflux symptoms and histological subtype indicates

two distinct aetiologies of gastric cardia cancer Gut, 57(3), 298– 305

PMID: 17965056

Derakhshan MH, Robertson EV, Lee YY, et  al 2015 In healthy volunteers,

immunohistochemistry supports squamous to columnar metaplasia as

mechanism of expansion of cardia, aggravated by central obesity Gut,

64(11), 1705– 1714 PMID: 25753030

Devesa SS, Blot WJ, and Fraumeni JF, Jr 1998 Changing patterns in the

inci-dence of esophageal and gastric carcinoma in the United States Cancer,

83(10), 2049– 2053 PMID: 9827707

Disibio G, and French SW 2008 Metastatic patterns of cancers:  results

from a large autopsy study Arch Pathol Lab Med, 132(6), 931– 939

PMID: 18517275

Dixon MF, Genta RM, Yardley JH, and Correa P 1996 Classification and

grad-ing of gastritis: the updated Sydney System International Workshop on

the Histopathology of Gastritis, Houston 1994 Am J Surg Pathol, 20(10),

1161– 1181 PMID: 8827022

Edgren G, Hjalgrim H, Rostgaard K, et al 2010 Risk of gastric cancer and

pep-tic ulcers in relation to ABO blood type: a cohort study Am J Epidemiol,

172(11), 1280– 1285 PMID: 20937632

Ekstrom AM, Hansson LE, Signorello LB, et al 2000a Decreasing incidence of both major histologic subtypes of gastric adenocarcinoma: a population-

based study in Sweden Br J Cancer, 83(3), 391– 396 PMCID: 2374560.

Ekstrom AM, Serafini M, Nyren O, et al 2000b Dietary antioxidant intake and the risk of cardia cancer and noncardia cancer of the intestinal and diffuse

types:  a population- based case- control study in Sweden Int J Cancer,

87(1), 133– 140 PMID: 10861464

El- Omar EM, Carrington M, Chow WH, et  al 2000 Interleukin- 1

poly-morphisms associated with increased risk of gastric cancer Nature,

404(6776), 398– 402 PMID: 10746728

Fascioli S, Capocaccia R, and Mariotti S 1995 Cancer mortality in migrant

populations within Italy Int J Epidemiol, 24(1), 8– 18 PMID: 7797360 Fenwick S, and Fenwick WS 1903 Cancer and other tumours of the stomach

Philadelphia: P Blakiston’s Son

Ferlay J, Soerjomataram I, Ervik M, et  al 2013 GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide:  IARC CancerBase No 11 [Internet] Available from: http:// globocan.iarc.fr Accessed May 2015.Ferro A, Peleteiro B, Malvezzi M, et al 2014 Worldwide trends in gastric can-cer mortality (1980– 2011), with predictions to 2015, and incidence by

subtype Eur J Cancer, 50(7), 1330– 1344 PMID: 24650579.

Fitzgerald RC, Hardwick R, Huntsman D, et  al 2010 Hereditary diffuse gastric cancer:  updated consensus guidelines for clinical manage-

ment and directions for future research J Med Genet, 47(7), 436– 444

PMCID: PMC2991043

Forman D, Bray F, Brewster DH, et al 2013 Cancer incidence in five nents, Vol X Available from: http:// ci5.iarc.fr Accessed May 2015.Forman D, and Helicobacter and Cancer Collaborative Group 2001 Gastric cancer and Helicobacter pylori:  a combined analysis of 12 case con-

conti-trol studies nested within prospective cohorts Gut, 49(3), 347– 353

PMCID: PMC1728434

Fox JG, Rogers AB, Ihrig M, et al 2003 Helicobacter pylori- associated gastric

cancer in INS- GAS mice is gender specific Cancer Res, 63(5), 942– 950

PMID: 12615707

Freedman ND, Subar AF, Hollenbeck AR, et  al 2008 Fruit and vegetable intake and gastric cancer risk in a large United States prospective cohort

study Cancer Causes Control, 19(5), 459– 467 PMID: 18166992.

Fukase K, Kato M, Kikuchi S, et al 2008 Effect of eradication of Helicobacter pylori on incidence of metachronous gastric carcinoma after endoscopic resection of early gastric cancer:  an open- label, randomised controlled

trial Lancet, 372(9636), 392– 397 PMID: 18675689.

Fukayama M, Hayashi Y, Iwasaki Y, et al 1994 Epstein- Barr virus- associated

gastric carcinoma and Epstein- Barr virus infection of the stomach Lab Invest, 71(1), 73– 81 PMID: 8041121

Fukayama M, and Ushiku T 2011 Epstein- Barr virus- associated gastric

carci-noma Pathol Res Pract, 207(9), 529– 537 PMID: 21944426.

Gaddy JA, Radin JN, Loh JT, et al 2013 High dietary salt intake exacerbates

Helicobacter pylori- induced gastric carcinogenesis Infect Immun, 81(6),

adenocarci-matous polyposis syndrome Am Surg, 74(1), 79– 83 PMID: 18274437.

Gayther SA, Gorringe KL, Ramus SJ, et al 1998 Identification of germ- line E-

cadherin mutations in gastric cancer families of European origin Cancer

Giardiello FM, Brensinger JD, Tersmette AC, et  al 2000 Very high risk of

cancer in familial Peutz- Jeghers syndrome Gastroenterology, 119(6),

Into Cancer and Nutrition (EPIC) J Natl Cancer Inst, 98(5), 345– 354.

Gonzalez CA, Lujan- Barroso L, Bueno- de- Mesquita HB, et al 2012 Fruit and vegetable intake and the risk of gastric adenocarcinoma: a reanalysis of the European Prospective Investigation into Cancer and Nutrition (EPIC-

EURGAST) study after a longer follow- up Int J Cancer, 131(12), 2910–

2919 PMID: 22473701

Gonzalez KD, Buzin CH, Noltner KA, et al 2009 High frequency of de novo

mutations in Li- Fraumeni syndrome J Med Genet, 46(10), 689– 693

PMID: 19556618

Grady WM, Willis J, Guilford PJ, et al 2000 Methylation of the CDH1

pro-moter as the second genetic hit in hereditary diffuse gastric cancer Nat

Genet, 26(1), 16– 17 PMID: 10973239

Gravalos C, and Jimeno A 2008 HER2 in gastric cancer:  a new

prognos-tic factor and a novel therapeuprognos-tic target Ann Oncol, 19(9), 1523– 1529

PMID: 18441328

Gronberg H, Bergh A, Damber JE, and Emanuelsson M 2000 Cancer risk in

families with hereditary prostate carcinoma Cancer, 89(6), 1315– 1321

PMID: 11002228

Guilford P, Hopkins J, Harraway J, et al 1998 E- cadherin germline mutations

in familial gastric cancer Nature, 392(6674), 402– 405 PMID: 9537325.

Guilford PJ, Hopkins JB, Grady WM, et al 1999 E- cadherin germline

muta-tions define an inherited cancer syndrome dominated by diffuse gastric

cancer Hum Mutat, 14(3), 249– 255 PMID: 10477433.

Haenszel W, Correa P, Cuello C, et al 1976 Gastric cancer in Colombia II

Case- control epidemiologic study of precursor lesions J Natl Cancer

Hajmanoochehri F, Mohammadi N, Nasirian N, and Hosseinkhani M 2013

Patho- epidemiological features of esophageal and gastric cancers in an

endemic region: a 20- year retrospective study Asian Pac J Cancer Prev,

14(6), 3491– 3497 PMID: 23886134

Hamashima C, Ogoshi K, Okamoto M, et al 2013 A community- based, case-

control study evaluating mortality reduction from gastric cancer by

endo-scopic screening in Japan PLoS One, 8(11), e79088 PMCID: 3827316.

Hamashima C, Shabana M, Okada K, Okamoto M, and Osaki Y 2015 Mortality

reduction from gastric cancer by endoscopic and radiographic screening

Cancer Sci, 106(12), 1744– 1749 PMCID: 4714659

Hamilton- Dutoit SJ, and Pallesen G 1994 Detection of Epstein- Barr virus

small RNAs in routine paraffin sections using non- isotopic RNA/ RNA

in situ hybridization Histopathology, 25(2), 101– 111 PMID: 7982672.

Hartfall SJ 1936 Gastritis in theory and practice Br Med J, 1(3936), 1200–

1204 PMCID: PMC2458877

Henson DE, Dittus C, Younes M, Nguyen H, and Albores- Saavedra J 2004

Differential trends in the intestinal and diffuse types of gastric carcinoma

in the United States, 1973– 2000: increase in the signet ring cell type Arch

Pathol Lab Med, 128(7), 765– 770 PMID: 15214826

Herrera V, and Parsonnet J 2009 Helicobacter pylori and gastric

adenocarci-noma Clin Microbiol Infect, 15(11), 971– 976.

Herrmann K, and Niedobitek G 2003 Epstein- Barr virus- associated

carcino-mas: facts and fiction J Pathol, 199(2), 140– 145 PMID: 12533825.

Houghton J, Stoicov C, Nomura S, et  al 2004 Gastric cancer

originat-ing from bone marrow- derived cells Science, 306(5701), 1568– 1571

PMID: 15567866

Howe JR, Sayed MG, Ahmed AF, et  al 2004 The prevalence of MADH4

and BMPR1A mutations in juvenile polyposis and absence of BMPR2,

PMCID: PMC1735829

Howlader N, Noone AM, Krapcho M, et  al 2015 SEER Cancer Statistics

Review, 1975– 2012, National Cancer Institute Bethesda, MD, http://

seer.cancer.gov/ csr/ 1975_ 2012/ , based on November 2014 SEER data

submission

Howson CP, Hiyama T, and Wynder EL 1986 The decline in gastric

can-cer:  epidemiology of an unplanned triumph Epidemiol Rev, 8, 1– 27

PMID: 3533579

Hu B, El Hajj N, Sittler S, et al 2012 Gastric cancer: classification, histology

and application of molecular pathology J Gastrointest Oncol, 3(3), 251–

261 PMCID: PMC3418539

Huang YK, Yu JC, Kang WM, et  al 2015 Significance of serum

pepsino-gens as a biomarker for gastric cancer and atrophic gastritis

screen-ing:  a systematic review and meta- analysis PLoS One, 10(11), 23

PMID: WOS:000364430700050

IARC 1986 Tobacco smoking Vol 38 Lyon, France February 12– 20, 1985

IARC Monogr Eval Carcinog Risks Hum, 38

IARC 1994 Schistosomes, liver flukes and Helicobacter pylori Vol 61 Lyon,

France, June 7– 14, 1994 IARC Monogr Eval Carcinog Risks Hum,

61, 1– 241

IARC 2004 Tobacco smoke and involuntary smoking Vol 83 Lyon, France,

June 11– 18, 2002 IARC Monogr Eval Carcinog Risks Hum, 83, 1– 1438

PMID: 15285078

IARC 2010 Alcohol consumption and ethyl carbamate Vol 96 Lyon, France,

February 6– 13, 2007 IARC Monogr Eval Carcinog Risks Hum, 96,

3– 1383

IARC 2012a Biological agents Vol 100B A review of human carcinogens

Lyon, France, February 24– March 3, 2009 IARC Monogr Eval Carcinog

Risks Hum, 100B

IARC 2012b Personal habits and indoor combustions Vol 100E A review

of human carcinogens IARC Monogr Eval Carcinog Risks Hum, 100(Pt

E), 1– 538

IARC 2012c Radiation Vol 100D A review of human carcinogens Lyon,

France, June 2– 9, 2009 IARC Monogr Eval Carcinog Risks Hum, 100(Pt

Kaneko S, and Yoshimura T 2001 Time trend analysis of gastric cancer

inci-dence in Japan by histological types, 1975– 1989 Br J Cancer, 84(3),

Kneller RW, You WC, Chang YS, et al 1992 Cigarette smoking and other risk

factors for progression of precancerous stomach lesions J Natl Cancer

Kolonel LN, Hankin JH, and Nomura AM 1985 Multiethnic studies of diet,

nutrition, and cancer in Hawaii Princess Takamatsu Symp, 16, 29– 40

PMID: 3916200

Kreinitz W 1906 Ueber das Auftreten von Spirochaeten verschiedener Form im

Mageninhalt bei Carcinoma ventriculi Deut Med Wochenschr, 32, 872.

Kuriyama S, Tsubono Y, Hozawa A, et al 2005 Obesity and risk of cancer in

Japan Int J Cancer, 113(1), 148– 157.

Ladeiras- Lopes R, Pereira AK, Nogueira A, et al 2008 Smoking and gastric

cancer:  systematic review and meta- analysis of cohort studies Cancer Causes Control, 19(7), 689– 701

Laurén P 1965 The two histological main types of gastric cancer: diffuse and

so- called intestinal type carcinoma Acta Pathol Microbiol Scand, 64, 31

Lauwers GY, and Srivastava A 2007 Gastric preneoplastic lesions and

epi-thelial dysplasia Gastroenterol Clin North Am, 36(4), 813– 829, vi

PMID: 17996792

Lee JH, Kim SH, Han SH, et al 2009 Clinicopathological and molecular acteristics of Epstein- Barr virus- associated gastric carcinoma:  a meta-

char-analysis J Gastroenterol Hepatol, 24(3), 354– 365 PMID: 19335785.

Leitzmann MF, Koebnick C, Freedman ND, et al 2009 Physical activity and

esophageal and gastric carcinoma in a large prospective study Am J Prev

Leivonen M, Nordling S, and Haglund C 1997 Does Helicobacter pylori in the

gastric stump increase the cancer risk after certain reconstruction types?

Anticancer Res, 17(5B), 3893– 3896

Levi F, La Vecchia C, and Te VC 1990 Descriptive epidemiology of carcinomas of the cardia and distal stomach in the Swiss Canton of Vaud

adeno-Tumori, 76(2), 167– 171 PMID: 2330608

Lewin KJ, and Appelman HD 1995 Stomach cancer In: Atlas of tumor

Lu Y, Chen J, Ding Y, et al 2010 Genetic variation of PSCA gene is ated with the risk of both diffuse- and intestinal- type gastric cancer in a

associ-Chinese population Int J Cancer, 127(9), 2183– 2189 PMID: 20131315.

Lukanova A, Bjor O, Kaaks R, et al 2006 Body mass index and cancer: results

from the Northern Sweden Health and Disease Cohort Int J Cancer,

118(2), 458– 466

Machado JC, Figueiredo C, Canedo P, et al 2003 A proinflammatory genetic profile increases the risk for chronic atrophic gastritis and gastric carci-

noma Gastroenterology, 125(2), 364– 371 PMID: 12891537.

Machii R, and Saika K 2016 Subsite distribution of stomach cancer from

Cancer Incidence in Five Continents Vol X Jpn J Clin Oncol, 46(1), 98

PMID: 26709380

Stomach Cancer 609

Majewski IJ, Kluijt I, Cats A, et  al 2013 An alpha- E- catenin (CTNNA1)

mutation in hereditary diffuse gastric cancer J Pathol, 229(4), 621– 629

PMID: 23208944

Malfertheiner P, Megraud F, O’Morain CA et  al 2012 Management of

Helicobacter pylori infection:  the Maastricht IV/ Florence Consensus

Report Gut, 61(5), 646– 664 PMID: 22491499.

Malvezzi M, Bonifazi M, Bertuccio P, et  al 2010 An age- period- cohort

analysis of gastric cancer mortality from 1950 to 2007 in Europe Ann Epidemiol, 20(12), 898– 905 PMID: 21074104

McLean MH, and El- Omar EM 2014 Genetics of gastric cancer Nat Rev

Gastroenterol Hepatol, 11(11), 664– 674 PMID: 25134511

McMichael AJ, McCall MG, Hartshorne JM, and Woodings TL 1980 Patterns

of gastro- intestinal cancer in European migrants to Australia: the role of

dietary change Int J Cancer, 25(4), 431– 437 PMID: 7372370.

Mezhir JJ, Gonen M, Ammori JB, et  al 2011 Treatment and outcome of

patients with gastric remnant cancer after resection for peptic ulcer

dis-ease Ann Surg Oncol, 18(3), 670– 676.

Miao Q, Ma YZ, Cai GH, and Chen XY 2014 Etiology and morphology of

car-ditis: experiences from a single center in China J Dig Dis, 15(2), 71– 77

PMID: 24151905

Minoura- Etoh J, Gotoh K, Sato R, et al 2006 Helicobacter pylori- associated

(EBV) in gastric epithelial cells latently infected with EBV J Med Microbiol, 55(Pt 7), 905– 911 PMID: 16772418

Misumi A, Murakami A, Harada K, Baba K, and Akagi M 1989 Definition

of carcinoma of the gastric cardia Langenbecks Arch Chir, 374(4), 221–

226 PMID: 2761324

Murphy G, Pfeiffer R, Camargo MC, and Rabkin CS 2009 Meta- analysis

shows that prevalence of Epstein- Barr virus- positive gastric cancer differs

based on sex and anatomic location Gastroenterology, 137(3), 824– 833

PMCID: PMC3513767

Nagel G, Linseisen J, Boshuizen HC, et al 2007 Socioeconomic position and

the risk of gastric and oesophageal cancer in the European Prospective

Epidemiol, 36(1), 66– 76

Nevalainen TJ 2013 Pekka Laurén and histological classification of gastric

carcinoma Gut, 62(8), 1230– 1231.

Orlando G, Pilone V, Vitiello A, et al 2014 Gastric cancer following bariatric

surgery: a review Surg Laparosc Endosc Percutan Tech, 24(5), 400– 405

PMID: 25238176

Palli D, Galli M, Caporaso NE, et al 1994 Family history and risk of stomach

cancer in Italy Cancer Epidemiol Biomarkers Prev, 3(1), 15– 18.

Parsonnet J 1999 Helicobacter pylori and gastric adenocarcinoma

New York: Oxford University Press

Parsonnet J, Friedman GD, Orentreich N, and Vogelman H 1997 Risk for

gas-tric cancer in people with CagA positive or CagA negative Helicobacter

pylori infection Gut, 40(3), 297– 301 PMCID: PMC1027076.

Patra R, Chattopadhyay S, De R, et al 2012 Multiple infection and

microdi-versity among Helicobacter pylori isolates in a single host in India PLoS

PDQ® Screening and Prevention Editorial Board 2016 PDQ Stomach

(Gastric) Cancer Screening Bethesda, MD: National Cancer Institute

Available at: http:// www.cancer.gov/types/stomach/hp/ stomach- ing- pdq Accessed April 25, 2016 [PMID: 26389174]

screen-Peek RM, Jr, and Crabtree JE 2006 Helicobacter infection and gastric

neopla-sia J Pathol, 208(2), 233– 248 PMID: 16362989.

Peleteiro B, Lopes C, Figueiredo C, and Lunet N 2011 Salt intake and

gastric cancer risk according to Helicobacter pylori infection, ing, tumour site and histological type Br J Cancer, 104(1), 198– 207

smok-PMCID: PMC3039805

Persson C, Canedo P, Machado JC, El- Omar EM, and Forman D 2011

Polymorphisms in inflammatory response genes and their association

with gastric cancer: a HuGE systematic review and meta- analyses Am J Epidemiol, 173(3), 259– 270 PMCID: PMC3105271

Pinheiro H, Oliveira C, Seruca R, and Carneiro F 2014 Hereditary diffuse

gastric cancer: pathophysiology and clinical management Best Pract Res Clin Gastroenterol, 28(6), 1055– 1068 PMID: 25439071

Plummer M, De Martel C, Vignat J, et al 2016 Global burden of cancer

attrib-utable to infections in 2012 Lancet Public Health, 4(9), e609– e616

PMID: 27470177

Plummer M, Franceschi S, Vignat J, Forman D, and de Martel C 2015 Global

burden of gastric cancer attributable to Helicobacter pylori Int J Cancer,

136(2), 487– 490 PMID: 24889903

Plummer M, van Doorn LJ, Franceschi S, et  al 2007 Helicobacter pylori

cytotoxin- associated genotype and gastric precancerous lesions J Natl Cancer Inst, 99(17), 1328– 1334

Powell J, and McConkey CC 1992 The rising trend in oesophageal

ade-nocarcinoma and gastric cardia Eur J Cancer Prev, 1(3), 265– 269

PMID: 1467772

Prinz C, Schoniger M, Rad R, et al 2001 Key importance of the Helicobacter pylori adherence factor blood group antigen binding adhesin during chronic

gastric inflammation Cancer Res, 61(5), 1903– 1909 PMID: 11280745.

Rampazzo A, Mott GL, Fontana K, and Fagundes RB 2012 Gastric carcinoma trends in the central region of Rio Grande do Sul (Southern

adeno-Brazil): what has changed in 25 years? Arq Gastroenterol, 49(3), 178–

183 PMID: 23011238

Ren JS, Kamangar F, Forman D, and Islami F 2012 Pickled food and risk

of gastric cancer: a systematic review and meta- analysis of English and

Chinese literature Cancer Epidemiol Biomarkers Prev, 21(6), 905– 915

World J Gastroenterol, 20(31), 10969– 10983 PMCID: PMC4138478

Rothenbacher D, Bode G, Berg G, et al 1999 Helicobacter pylori among school children and their parents: evidence of parent- child transmission J Infect Dis, 179(2), 398– 402

pre-Rugge M, Correa P, Dixon MF, et  al 2000 Gastric dysplasia:  the Padova

international classification Am J Surg Pathol, 24(2), 167– 176 PMID:

10680883

Sakamoto H, Yoshimura K, Saeki N, et al 2008 Genetic variation in PSCA is

associated with susceptibility to diffuse- type gastric cancer Nat Genet,

Helicobacter pylori PLoS One, 10(3), e0120444 PMCID: PMC4368826.

Sauvaget C, Lagarde F, Nagano J, et al 2005 Lifestyle factors, radiation and

gastric cancer in atomic- bomb survivors (Japan) Cancer Causes Control,

16(7), 773– 780

Schlemper RJ, Riddell RH, Kato Y, et  al 2000 The Vienna

PMCID: PMC1728018

Segal ED, Lange C, Covacci A, Tompkins LS, and Falkow S 1997 Induction

of host signal transduction pathways by Helicobacter pylori Proc Natl Acad Sci U S A, 94(14), 7595– 7599 PMCID: PMC23867

Shakeri R, Malekzadeh R, Nasrollahzadeh D, et al 2015 Multiplex H. pylori

serology and risk of gastric cardia and noncardia adenocarcinomas

Cancer Res, 75(22), 4876– 4883 PMCID: PMC4792189

Shikata K, Kiyohara Y, Kubo M, et al 2006 A prospective study of dietary salt intake and gastric cancer incidence in a defined Japanese population: the

Hisayama study Int J Cancer, 119(1), 196– 201.

Shinmura K, Kohno T, Takahashi M, et  al 1999 Familial gastric cer:  clinicopathological characteristics, RER phenotype and germline

PMID: 10357799

Shinozaki- Ushiku A, Kunita A, and Fukayama M 2015 Update on Epstein-

Barr virus and gastric cancer (review) Int J Oncol, 46(4), 1421– 1434

Siman JH, Forsgren A, Berglund G, and Floren CH 2001 Tobacco

smok-ing increases the risk for gastric adenocarcinoma among Helicobacter pylori - infected individuals Scand J Gastroenterol, 36(2), 208– 213 PMID: 11252415

Sipponen P, and Correa P 2002 Delayed rise in incidence of gastric cancer in females results in unique sex ratio (M/ F) pattern:  etiologic hypothesis

Gastric Cancer, 5(4), 213– 219

Siurala M, Sipponen P, and Kekki M 1985 Chronic gastritis: dynamic and

clin-ical aspects Scand J Gastroenterol Suppl, 109, 69– 76 PMID: 3895391.

Sjodahl K, Jia C, Vatten L, et al 2008a Body mass and physical activity and risk of gastric cancer in a population- based cohort study in Norway

Cancer Epidemiol Biomarkers Prev, 17(1), 135– 140

Sjodahl K, Jia C, Vatten L, et al 2008b Salt and gastric adenocarcinoma: a

population- based cohort study in Norway Cancer Epidemiol Biomarkers

Song H, Ekheden IG, Zheng Z, et al 2015 Incidence of gastric cancer among

patients with gastric precancerous lesions: observational cohort study in a low

risk Western population Brit Med J, 351, h3867 PMCID: PMC4516137.

Spallek J, Arnold M, Razum O, et al 2012 Cancer mortality patterns among

Turkish immigrants in four European countries and in Turkey Eur J

Epidemiol, 27(12), 915– 921 PMID: 23179631

Steevens J, Botterweck AA, Dirx MJ, van den Brandt PA, and Schouten LJ

2010 Trends in incidence of oesophageal and stomach cancer subtypes in

Europe Eur J Gastroenterol Hepatol, 22(6), 669– 678 PMID: 19474750.

Steevens J, Schouten LJ, Goldbohm RA, and van den Brandt PA 2011

Vegetables and fruits consumption and risk of esophageal and gastric

cancer subtypes in the Netherlands Cohort Study Int J Cancer, 129(11),

2681– 2693 PMID: 21960262

Stein M, Bagnoli F, Halenbeck R et  al 2002 c- Src/ Lyn kinases activate

Helicobacter pylori CagA through tyrosine phosphorylation of the

EPIYA motifs Mol Microbiol, 43(4), 971– 980 PMID: 11929545.

Stolte M 2003 The new Vienna classification of epithelial neoplasia of the

gas-trointestinal tract: advantages and disadvantages Virchows Arch, 442(2),

99– 106 PMID: 12596058

Takada K 2000 Epstein- Barr virus and gastric carcinoma Mol Pathol, 53(5),

255– 261 PMCID: PMC1186978

Takahashi M, Nishikawa A, Furukawa F, et al 1994 Dose- dependent promoting

effects of sodium chloride (NaCl) on rat glandular stomach

carcinogene-sis initiated with N- methyl- N’- nitro- N- nitrosoguanidine Carcinogenecarcinogene-sis,

15(7), 1429– 1432 PMID: 8033321

Terada T 2016 Histopathological study using computer database of 10 000

consecutive gastric specimens: (2) malignant lesions Gastroenterol Rep

Terry MB, Gaudet MM, and Gammon MD 2002 The epidemiology of gastric

cancer Semin Radiat Oncol, 12(2), 111– 127 PMID: 11979413.

The Cancer Genome Atlas Research Network 2014 The Cancer Genome Atlas

Research Network: comprehensive molecular characterization of gastric

adenocarcinoma Nature, 513(7517), 202– 209 PMCID: PMC4170219.

The Paris Classification 2003 The Paris endoscopic classification of superficial

neoplastic lesions: esophagus, stomach, and colon: November 30 to December

1, 2002 Gastrointest Endosc, 58(6 Suppl), S3– 43 PMID: 14652541.

Tramacere I, Negri E, Pelucchi C, et al 2012 A meta- analysis on alcohol

drink-ing and gastric cancer risk Ann Oncol, 23(1), 28– 36.

Wang Q, Chen Y, Wang X, et al 2014 Consumption of fruit, but not vegetables,

may reduce risk of gastric cancer: results from a meta- analysis of cohort

studies Eur J Cancer, 50(8), 1498– 1509 PMID: 24613128.

Wang Z, Liu L, Ji J, et  al 2012 ABO blood group system and gastric

can-cer: a case- control study and meta- analysis Int J Mol Sci, 13(10), 13308–

13321 PMCID: PMC3497328

Warren JR, and Marshall BJ 1983 Unidentified curved bacilli on gastric

epi-thelium in active chronic gastritis Lancet, 1(8336), 1273– 1275.

Washington K 2010 7th edition of the AJCC cancer staging manual: stomach

Ann Surg Oncol, 17(12), 3077– 3079 PMID: 20882416

Watson P, Vasen HF, Mecklin JP, et al 2008 The risk of extra- colonic, extra-

endometrial cancer in the Lynch syndrome Int J Cancer, 123(2), 444–

WHO 2010 WHO Classifications of tumours of the digestive system Lyon,

France: IARC Press

WHO 2013 International classification of diseases for oncology, 3rd ed

Geneva: World Health Organization

Wiseman M 2008 The second World Cancer Research Fund/ American Institute for Cancer Research expert report Food, nutrition, physical activity, and

the prevention of cancer:  a global perspective Proc Nutr Soc, 67(3),

253– 256

World Cancer Research Fund International/ American Institute for Cancer Research (WCRF/ AICR) 2016 Continuous update project report: diet, nutrition, physical activity and stomach cancer Available at:  wcrf.org/ stomach- cancer- 2016

Wu H, Rusiecki JA, Zhu K, Potter J, and Devesa SS 2009 Stomach noma incidence patterns in the United States by histologic type and

carci-anatomic site Cancer Epidemiol Biomarkers Prev, 18(7), 1945– 1952

PMCID: PMC2786772

Wu MS, Shun CT, Wu CC, et al 2000 Epstein- Barr virus- associated

gas-tric carcinomas: relation to H pylori infection and genetic alterations Gastroenterology, 118(6), 1031– 1038 PMID: 10833477

Yabuta T, Shinmura K, Tani M, et al 2002 E- cadherin gene variants in gastric cancer families whose probands are diagnosed with diffuse gastric cancer

Int J Cancer, 101(5), 434– 441 PMID: 12216071

Yamaguchi Y, Nagata Y, Hiratsuka R, et al 2016 Gastric cancer screening by combined assay for serum anti- Helicobacter pylori IgG anbitody and

serum pepsinogen levels: the ABC method Digestion, 93, 13– 18.

Yamazaki S, Yamakawa A, Ito Y, et al 2003 The CagA protein of Helicobacter pylori is translocated into epithelial cells and binds to SHP- 2 in human

gastric mucosa J Infect Dis, 187(2), 334– 337 PMID: 12552462.

Yang P, Zhou Y, Chen B, et al 2009 Overweight, obesity and gastric cancer

risk: results from a meta- analysis of cohort studies Eur J Cancer, 45(16),

2867– 2873

Yang PC, and Davis S 1988 Epidemiological characteristics of noma of the gastric cardia and distal stomach in the United States, 1973–

adenocarci-1982 Int J Epidemiol, 17(2), 293– 297 PMID: 3403124.

Yoon KA, Ku JL, Yang HK, et al 1999 Germline mutations of E- cadherin gene

in Korean familial gastric cancer patients J Hum Genet, 44(3), 177– 180

SAMUEL O ANTWI, RICK J JANSEN, AND GLORIA M PETERSEN

OVERVIEW

Pancreatic cancer (PC) is an uncommon but often rapidly lethal

malignancy Worldwide, PC is the twelfth most commonly

diag-nosed cancer and the seventh most common cause of cancer- related

death Globally, the estimated incidence and death rates from PC

were almost identical in 2012, resulting in 338,000 new cases and

330,400 deaths In high- resource countries, the recorded incidence

rates are higher than in developing countries, at least partly reflecting

more available means of diagnosis, but the prognosis is only slightly

better In the United States, PC ranks twelfth in cancer incidence

among men and ninth among women; it ranks fourth in cancer

mor-tality in both sexes, and third when men and women are combined

Etiologic research on PC is complicated by the relatively

inacces-sible location of the pancreas, the obstacles to early diagnosis, the

aggressiveness and resistance to therapy of these malignancies, and

the tendency of PC to progress rapidly from diagnosis to death Until

recently, the only etiologic factors considered definitive causes of PC

were tobacco use, chronic pancreatitis, and several rare high-

pene-trance genetic disorders In the past 10– 15 years, the evidence for

other causal relationships has strengthened, especially for metabolic

risk factors (obesity, type 2 diabetes mellitus, insulin, and insulin- like

growth factor), chronic local inflammation, heavy alcohol

consump-tion, dietary consumption of grilled meat, and non- O ABO blood

type Among US whites, the incidence and mortality rates from PC,

which were decreasing in the late twentieth century following large

reductions in smoking, are now increasing, paralleling large increases

in the prevalence of obesity and type 2 diabetes Looking at SEER 13

area trends from 2005- 2014, increasing incidence and mortality were

observed for all racial groups, except Blacks and American Indians/

Alaska Natives who show a decreasing trend in both incidence and

mortality Research on PC is focused on factors along the entire

chain connecting structural changes in DNA (germline and somatic),

through intermediate epigenetic and other influences on transcription

and translation, to the phenotypic abnormalities of neoplasia

INTRODUCTION

Research on PC has long been impeded by the relatively

inaccessi-ble location of the pancreas, the lack of access to tissue during the

preclinical development of tumors, the difficulties of early diagnosis,

aggressiveness of the disease, resistance to therapy, and the typically

rapid progression from diagnosis to death (Anderson et  al., 2006)

Partly because of these barriers, etiologic studies in the twentieth

century identified few factors that were considered definite causes of

PC Multiple epidemiologic studies established that tobacco use was

an important and potentially modifiable cause Clinical studies

con-firmed strong relationships between PC and chronic pancreatitis, and

smaller increases in risk associated with several rare high- penetrance

genetic disorders and a history of gastric surgery Despite numerous

studies of other exposures, the evidence remained inconsistent and/

or difficult to interpret into the early twenty- first century, as reviewed

by Anderson et  al in the third edition of this text (Anderson et  al.,

2006) This chapter is divided into sections on the two major types

of pancreatic carcinoma (ductal adenocarcinoma and neuroendocrine/

islet cell tumors), and discusses classification, diagnosis, demographic

patterns, environmental and host risk factors, molecular pathogenesis, and preventive measures

CLINICAL AND PATHOLOGICAL FEATURES

Anatomy, Presenting Symptoms, and Diagnosis

The pancreas is a glandular organ of the digestive system, located behind the stomach and lying against and connected to the duodenum

by a main duct that joins to the common bile duct The head of the pancreas is closest to the duodenum, and distally from it are the body and tail The pancreas has two functions Its exocrine function is to produce digestive enzymes (proteases, lipase, and amylase) that are secreted through a network of smaller pancreatic ducts that join the main pancreatic duct Its endocrine function is to produce hormones (glucagon and insulin) in the islet cells located throughout the pan-creas, which help maintain glucose homeostasis Cancer can occur either in the cells lining the ducts or in the islet cells

Most symptoms of PC do not appear until the tumor is at a late stage Approximately 80% of patients have unresectable disease at the time of diagnosis due to metastatic spread or locally advanced disease Typically, individuals diagnosed with PC report non- specific abdomi-nal pain, jaundice, and/ or unintended weight loss Jaundice is usu-ally due to ductal tumors located in the pancreatic head, obstructing the bile duct (Ryan et al., 2014; Yamada et al., 2009) Approximately 70% of ductal tumors are located in the head of the pancreas, 5%– 10% in the body, and 10%– 15% in the tail Diagnosis of pancreatic cancer is usually confirmed by imaging that includes a combination of computed tomography (CT) scanning, magnetic resonance imaging (MRI), and positron emission tomography (PET), and/ or endoscopic ultrasound (EUS) and endoscopic retrograde cholangiopancreatogra-phy (ERCP), pancreatic biopsy, and serum levels of CA 19- 9

Tumor Subtypes

The most common form of pancreatic cancer is ductal noma (in the exocrine portion of the gland), accounting for > 95% of all pancreatic cancers Ductal cells constitute 10%– 15% of the volume but give rise to 90% of all tumors, as discussed in the following section

adenocarci-of the chapter Acinar cell tumors are rare (1% adenocarci-of exocrine atic cancer), but these cells comprise 80% of the volume of the gland (Antonello et al., 2009) The other main group of pancreatic tumors is neuroendocrine in origin (islet cells), as discussed in the subsequent section It is proposed that ductal adenocarcinoma progresses in a step-wise fashion through acquisition of molecular alterations and histolog-ically well- defined non- invasive precursor lesions (Maitra and Hruban, 2008) Precursor lesions progress from a benign intraductal tumor through increasing grades of dysplasia to invasive adenocarcinoma, providing models of neoplastic pancreatic progression (Grutzmann

pancre-et al., 2010; Mettu and Abbruzzese, 2016) The classic morphologic progression is suggested to occur through pancreatic intraepithelial neoplasias (PanINs) PanINs are microscopic lesions in small (less than 5 mm) pancreatic ducts and are classified into two grades (low- grade PanIN- 1 and PanIN- 2), and high- grade (PanIN- 3, formerly known as carcinoma in situ) (Basturk et al., 2015; Hruban et al., 2001; Ying et al., 2016) Another precursor of invasive pancreatic carcinomas

is pancreatic intraductal papillary mucinous neoplasia (IPMN), which

are cystic lesions With advances in endoscopic imaging, more patients

are being diagnosed with IPMN and other cystic tumors, although the

vast majority are asymptomatic and do not progress to malignancy

An estimated 2% of adults, and up to 10% of individuals aged 70 and

older, have IPMNs (Ryan et al., 2014)

Molecular Pathogenesis/ Somatic Mutations

Following a tumor progression model delineated by genetic events in

pancreatic tumorigenesis (Bardeesy and DePinho, 2002), there has

been consistency and confirmation in the body of research over many

studies since 2002 The categories of molecular events include the

following:

(a) Activation by mutations in the KRAS oncogene, among the

earli-est events; 95% of pancreatic ductal adenocarcinomas possess a

KRAS mutation, most commonly in codon 12.

(b) Inactivation of tumor suppressor genes, TP53 in 60% of tumors,

SMAD4 in half of tumors, and CDKN2A in 95% of tumors, occur

later in the progression pathway

(c) Mutations of mismatch repair genes, such as MLH1 and MSH2,

have been found in 4% of pancreatic tumors

It is thought that these molecular events, whether gradual or rapid,

collec-tively have an impact in several core pathways (Notch, Hedgehog, beta-

catenin, axon guidance, chromatin remodeling, and DNA repair) that have

been proposed by Jones et al (2008) and further refined and expanded by

Biankin and colleagues (2012) Understanding these molecular events

will shed light on future treatment and early detection strategies

DUCTAL ADENOCARCINOMA

Descriptive Epidemiology

The diagnosis of PC is challenging, even in high- resource countries,

and tends to be underdiagnosed, particularly in countries that lack

resources for sophisticated diagnostic testing (Wolfgang et al., 2013)

Thus, the interpretation of incidence and mortality data is more

trac-table when comparisons involve countries and populations at a similar

level of economic development

Demographic Characteristics

Age

PC occurs almost exclusively in adults, ages 20 years or older (American

Cancer Society 2016) It is rarely diagnosed in persons younger than

age 40; disease incidence rises sharply by age 50 and continues to rise

until about age 80 (http:// seer.cancer.gov/ ) The median age at diagnosis

of PC in the United States is 71 years (American Cancer Society 2016)

The majority (~80%) of PC patients are diagnosed between ages 60

and 80 years (Ahlgren, 1996; Gold and Goldin, 1998) Persons

diag-nosed before age 50 may be more likely to have a positive family

his-tory or inherited genetic predisposition (Lowenfels and Maisonneuve,

2004; Raimondi et  al., 2007; Wolfgang et  al., 2013) The age

distri-bution at diagnosis in a high- resource country is illustrated by data

from the National Cancer Institute’s Surveillance, Epidemiology, and

End Results (SEER- 18) Registries between 2008 and 2012 The

per-centage of incident cases under age 20 years was approximately 0%

This increased to 3% at ages 20– 44  years, 9% at ages 45– 54  years,

22% at ages 55– 64 years, 27% at ages 55– 64 years, 26% at ages 75–

84 years, and 13% among those older than age 85 (The Surveillance

Epidemiology and End Results [SEER] Program Fact Sheets: Pancreas

Cancer [http:// seer.cancer.gov/ canques/ incidence.html])

Gender

The incidence rate of PC is about 36% higher in men than women

(American Cancer Society, 2013) Men are usually diagnosed at earlier

ages than women (American Cancer Society 2013; Siegel et al., 2015), and have a higher death rate from PC (American Cancer Society 2016; Lowenfels and Maisonneuve, 2006; Torre et al., 2015) The age- adjusted incidence rate in the United States in 2012 was also higher in men than

in women (14.5 vs 11.7 per 100,000), with correspondingly higher mortality rates (12.6 vs 9.6 per 100,000) (http:// seer.cancer.gov/ ) In contrast to the age- standardized rates, the lifetime risk of developing

PC is similar among men and women (1.5% for both), due to the greater longevity of women and the decreasing gap in exposure to tobacco use between men and women (American Cancer Society, 2013)

Race/ Ethnicity

Racial disparities in PC incidence and mortality rates exist both within and among countries (American Cancer Society, 2016; Arnold et al., 2009; Center et  al., 2011; Khawja et  al., 2015; Parkin et  al., 2002; Silverman et  al., 2003) In the United States, African Americans have the highest incidence and mortality rates from PC, while Asian Americans and Pacific Islanders have the lowest rates (Table 32– 1) The recorded incidence and mortality are actually higher among African Americans than among Native Africans (Curado et al., 2007; Kovi and Heshmat, 1972; Parkin et  al., 2002; Walker et  al., 1993), although this at least partly reflects differences in detection Between

2000 and 2012, African Americans were diagnosed with PC at a rate of 15.7 per 100,000 versus 11.9 per 100,000 among whites (Table 32–1).Whereas Hispanic women and white women in the United States have similar incidence rates, white men have about a 15% higher inci-dence rate than Hispanic men (Table 32– 1) PC incidence rates among American Indian/ Alaska Natives are the second lowest of the racial and ethnic groups in the United States

While reasons for the increased incidence of PC among African Americans are not entirely clear, it has been suggested that a higher prevalence of risk factors (such as cigarette smoking, diabetes, fam-ily history of PC, and high body mass index [BMI]) among African Americans may explain the excess risk in this population (Khawja et al., 2015; Parkin et  al., 2002) However, studies that attempt to explain the racial disparity in PC between African Americans and whites have reached varying conclusions (Arnold et  al., 2009; Silverman et  al., 2003) In a population- based, case- control study of African Americans and whites recruited from metropolitan regions of Atlanta, Georgia; Detroit, Michigan; and 10 counties in New Jersey between 1986 and

1989, Silverman et al (2003) examined racial differences in PC risk among 526 individuals with PC and 2153 cancer- free controls They observed that the determinants of higher incidence of PC among African Americans compared to whites varied by sex Among men, the excess risk of PC in African Americans was explained largely

by cigarette smoking, long- standing history of diabetes, and a tive family history of PC Among women, the higher risk in African Americans was explained mainly by heavy alcohol use and elevated BMI (Silverman et al., 2003) Arnold et al (2009) performed a similar but much larger study that included 6243 cases and 1,054,146 controls using data from the American Cancer Society’s Cancer Prevention Study II (CPS II) These investigators used PC mortality as a surrogate for incidence, and observed that variation in smoking habits, diabe-tes status, BMI, family history of PC, and cholecystectomy did not explain the higher incidence of PC in African Americans compared to whites, even in sex- specific analyses (Arnold et al., 2009) These stud-ies suggest that other unexplored factors or attributes of exposure not captured by the available metrics may contribute to the observed racial differences in PC Even though rates of PC incidence and mortality are higher in economically developed than in developing regions of the world, the male: female ratios are identical (1.4:1) for incidence and similar (1.5:1 and 1.4:1) for mortality in developed and developing regions, respectively (Torre et al., 2015)

posi-Geographic Variation

The recorded incidence and mortality rates of PC vary substantially across the globe due, in part, to differences in diagnosis and surveil-lance patterns In general, higher rates are reported in countries with

Cancer of the Pancreas 613

better diagnostic imaging (high- income countries) and lower rates in

countries lacking high- level diagnostic imaging (low- income

coun-tries) The reported PC incidence and mortality rates are high in North

America, Western Europe, and Central and Eastern Europe and low

in South- Central Asia and Central Africa (Figure 32– 1) Worldwide,

the United States ranked 20th in PC incidence and 23rd in PC-

related deaths (with 7.5 and 7.0 per 100,000, respectively) in 2012

(Ferlay et  al., 2013) The highest age- standardized incidence rates

in 2012 were recorded in the Czech Republic, Slovakia, Armenia, and Hungary (with 9.7, 9.4, 9.3, and 9.3 per 100,000, respectively) (Ferlay et  al., 2013) Countries with a high incidence of PC also include Canada (6.4 per 100,000), Australia (6.6 per 100,000), and many European countries (Figure 32– 2) In South America, PC inci-dence rates are high in French Guyana, Uruguay, and Argentina (with 8.1, 7.6, and 6.7 per 100,000, respectively), moderately high

in Peru, Brazil, and Paraguay (with 4.6 per 100,000 in each), mediate in Colombia, Venezuela, and Ecuador (with 3.8, 3.6, and 3.1 per 100,000, respectively), and lowest in Bolivia (2.3 per 100,000)

inter-In Asia, incidence rates are high in Japan, Kazakhstan, and South Korea (8.5, 6.8, and 6.7 per 100,000, respectively), while countries such as India, Pakistan, Nepal, Bangladesh, Laos, and Vietnam have some of the world’s lowest rates of PC (ranging from ~0.5 to 1.2 per 100,000) (Ferlay et al., 2013) In Africa, rates are moderately high in South African and Libya (4.7 per 100,000 for both) (Figure 32– 2) The lowest recorded rates in Africa are observed mainly in Central African countries such as Cameroon, the Central African Republic, the two Republics of Congo, Gabon, and Angola, with rates that range from 0.7 to 1.1 per 100,000, although these estimates are based on very limited data (Ferlay et al., 2013) As mentioned, the low inci-dence in these countries may also reflect the limited availability

Table 32– 1 Incidence of Pancreatic Cancer in the United States per

100,000 Population, 2000–2012

African

American Indian/ Alaska Native

Asian or Pacific Islander

Program SEER*Stat Database: Incidence - SEER 18 Registries Research Data National

Cancer Institute Bethesda, MD, http://seer.cancer.gov/canques/incidence.html All

reported materials are publicly available.

Eastern Asia

Polynesia

Southeast AsiaMelanesiaMicronesia

Middle Africa

Western Africa

South-Central Asia

Less developed regions

More developed regions

World

Eastern Africa

Southern Africa

CaribbeanSouth America

Central America

Northern Africa

Northern EuropeSouthern Europe

Western Europe

Australia/New Zealand

6.3 5.8 5.9 5.5 5 5.9 5.6 5.4 4.9 5.3 4.9

3.6 3.4 3.5 3.4 3.6 3.4 3.1 3 3.6 3.6 3.7 3.5 1.9 1.9 2.4 2.3 1.8 1.8 2 1.9 1.1 1.3 1.2 1.7 1.6 0.8 0.3 1 0.9 0.8 0.7

4.9

8.3 8.6 8.3 8.9 9 7.3 7.1 7.5 6.3 7.6 7.4

5 5.1 5.5 5.2 5.3 5.2 4.9 4.7 4.2 4.1 3.7 4.4 4.4 3.3 3.3 3.2 3.2 2.5 2.4 2.3 2 1.9 1.5 1.5 2.2 3.6 1.3 1.2 1.4 1.3

8 8

Figure 32– 1 Age- standardized incidence and mortality rates (per 100,000) for pancreatic cancer among men and women, by geographic region, 2012

Source: International Agency for Research on Cancer (IARC) (2015)

of sophisticated diagnostic procedures Overall, the global patterns of

PC incidence are nearly identical to the mortality rates (Figure 32– 3)

While African Americans have traditionally had the highest incidence

of PC among all other racial groups in the world, current data shows

that incidence rates in Eastern European countries, such as the Czech

Republic and Slovakia, are rising sharply (Figure 32– 4) As of 2005,

incidence of PC among men in the Czech Republic had surpassed the

incidence among African- American men (Figure 32– 4)

Time Trends in the United States

Among US whites, the incidence and mortality rates from PC decreased

in the late twentieth century, following reductions in smoking, but have increased since the year 2000, following substantial increases in the prevalence of obesity and type 2 diabetes Data from nine cancer regis-tries of the SEER program show that between 1973 and 2002, PC inci-dence decreased among men in the United States by 0.62% each year

6.3+

4.1–6.3 2.4–4.1 1.2–2.4

<1.2

No Data Pancreatic cancer

Figure 32– 2 Global variation in pancreatic cancer incidence rates (per 100,000), estimated for 2012 Source: International Agency for Research on Cancer

(IARC) (2015)

6.2+

4.0–6.2 2.4–4.0 1.2–2.4

<1.2

No Data Pancreatic cancer

Figure 32– 3 Global variation in pancreatic cancer mortality rates (per 100,000), estimated for 2012 Source: International Agency for Research on Cancer

(IARC) (2015)

Cancer of the Pancreas 615

Among women, incidence rates increased by 0.94% annually from

1973 to 1984 before beginning to decline slowly (Zhang et al., 2007)

Overall, the incidence of PC decreased in both African Americans and

whites between 1973 and 2002, except during a brief period in the early

to mid- 1980s, when PC incidence increased in whites (Zhang et  al.,

2007) A recent analysis of data from 42 states, representing 87% of the

US population, observed that between 2000 and 2009, PC incidence

increased by 0.9% each year among white men and women and among

African- American men, while incidence remained relatively stable

among African- American women (Jemal et al., 2013) The patterns of

PC incidence in the United States did not change significantly among

Asians/ Pacific Islanders or Alaska Natives/ American Indians during

this period (2000– 2009) (Jemal et al., 2013) Mortality from PC in the

United Sates increased by about 3- fold between 1920 and 1968, from

3 cases per 100,000 in 1920 to ~8.5 cases per 100,000 in 1968 (Krain,

1970) Between 1970 and 1995, the death rate from PC decreased by

0.7% annually among white men before increasing by 0.4% each year

between 1995 and 2009 (Ma et al., 2013) Among white women, annual

incidence of PC increased by 0.4% between 1970 and 1984, remained

stable between 1984 and 1998, and then increased by 0.5% annually

between 1998 and 2009 Among African Americans, the incidence rate

increased between 1970 and 1989 by 0.5% annually in men and 1.3%

in women, but decreased from 1989 to 2009 by 0.9% among men and

by 0.5% among women (Ma et al., 2013) Overall, PC- related deaths

increased by 0.4% annually between 2000 and 2009 in US men and

women, but this increase was mainly confined to whites (by 0.5% per

year in each sex) and Asian/ Pacific Islander men in the United States

(by 1% per year) (Jemal et al., 2013) Additional details on racial/ ethnic

trends in PC incidence by gender are presented Figure 32– 5

Risk Factors

Age is the strongest risk factor for PC, as discussed earlier The other

established risk factors are cigarette smoking, preexisting diabetes,

ABO blood type, chronic pancreatitis, and hereditary syndromes

caused by germline mutations in the BRCA1 or BRCA2 gene tary breast- ovarian syndrome), p16/ CKDN2A (familial atypical mul- tiple mole melanoma syndrome), PRSS1, PRSS2, SPINK1, or CTRC (hereditary pancreatitis), STK11/ LKB1 (Peutz- Jeghers syndrome),

(heredi-CFTR (cystic fibrosis), APC (familial adenomatous polyposis), TP53 (Li- Fraumeni syndrome), and MSH2 or MLH1 (Lynch syndrome)

(Klein, 2012; Petersen, 2015) Risk factors are discussed starting with modifiable lifestyle and environmental factors, medical conditions, and ending with genetic and molecular factors, and other suspected risk factors

Cigarette Smoking

Tobacco smoking was the first firmly established risk factor for PC (Lowenfels and Maisonneuve, 2006) Cigarette smoking was desig-nated as causally related to PC by the US Surgeon General in 1982 (see Chapter 11) The associations have been somewhat stronger in recent studies from Europe than those from the United States A pooled anal-ysis of large US cohort studies in 2000– 2010 reported relative risk estimates for current cigarette smoking in relation to death from PC of 1.6 in men and 1.9 in women (Carter et al., 2015) Based on this, the

2014 US Surgeon General Report estimated that 10%– 14% of all PC deaths in the United States are currently attributable to tobacco smok-ing (Reports of the Surgeon General, 2014) In contrast, European cohort studies have estimated that ~20%– 25% of all PC cases are attributable to tobacco smoking (Lowenfels and Maisonneuve, 2006) Studies have shown that PC risk increases with number of cigarettes smoked per day, with smoking duration and with pack- years of smok-ing (Bosetti et  al., 2012a; Hassan et  al., 2007; Lynch et  al., 2009; Nilsen and Vatten, 2000; Tranah et al., 2011; Vrieling et al., 2010; Yun

et al., 2006) In a meta- analysis of 82 published studies, Iodice et al reported a significantly higher risk of PC among current and former

Thailand (3 registries)

United Kingdom (9 registries) United Kingdom(9 registries)

USA, SEER (9 registries):

Figure 32– 4 Trends in age- standardized pancreatic cancer incidence rates in men and women, comparing United States (black, white) to other countries

Source: International Agency for Research on Cancer (IARC) (2015)

cigarette smokers, compared to never smokers, with odds ratios (ORs)

and 95% confidence intervals (CIs) of 1.74 (95% CI: 1.61, 1.87) and

1.2 (95% CI: 1.11, 1.29), respectively (Iodice et al., 2008) In a nested

case- control study of the International Pancreatic Cancer Cohort

Consortium, the association between cigarette smoking and PC risk

was examined by pooling primary data from eight prospective studies

involving 1481 PC cases and 1539 controls (Lynch et al., 2009) The

study reported a higher risk of PC among current cigarette smokers

(odds ratio [OR = 1.77; 95% CI: 1.38, 2.26) and a non- significant risk

among former smokers (OR = 1.09; 9% CI: 0.91, 1.30) When duration

of smoking cessation was examined, a substantially higher risk of PC

was observed among former smokers with less than 10 years of

smok-ing cessation (OR = 2.19; 95% CI: 1.25, 3.83), whereas a positive, but

non- statistically significant, association was observed among former

smokers with 10– 14 years of smoking cessation (OR = 1.24; 95% CI:

0.78, 1.98) The study further showed that risks of PC among former

smokers with more than 15 years of smoking cessation were similar

to that of never smokers (ORs and 95% CI for 15– 19 years, 20– 29

years, and ≥ 30 years of smoking cessation were 0.91 [0.58, 1.44], 0.91

[0.66, 1.26], and 0.93 [0.73, 1.20], respectively) (Lynch et al., 2009)

A pooled analysis from the international Pancreatic Cancer Case-

Control Consortium (PanC4) also examined the association between

cigarette smoking and PC using primary data from 12 case- control

studies consisting of 6507 PC cases and 12,890 controls (Bosetti et al.,

2012a) These investigators observed that the PC risk was more than

twice as high among current cigarette smokers compared with never

smokers (OR = 2.2; 95% CI: 1.7, 2.8), and 20% higher among former

smokers (OR =1.20; 95% CI: 1.0, 1.3) (Bosetti et al., 2012a) They

also observed decreasing risk of PC with increasing duration of

smok-ing cessation, such that the risk among former smokers approached

that of never smokers after 15 years of smoking cessation The ORs

and 95% CIs for 1– 9 years, 10– 15 years, 16– 19 years, 20– 29 years,

and ≥ 30 years duration of smoking cessation were 1.64 (1.36, 1.97),

1.42 (1.11, 1.82), 1.12 (0.86, 1.44), 0.98 (0.77, 1.23), and 0.98 (0.83,

1.16), respectively (Bosetti et al., 2012a) Another pooled analysis of

11 case- control studies in PanC4, involving 6058 cases and 11,338 controls, observed that cigar- only and cigarette- only smokers had similar increased risk of PC when compared to individuals who had never used any tobacco product (OR and 95% CI for cigar- only and cigarette- only smokers were 1.62 [1.15, 2.29] and 1.50 [1.39, 1.62], respectively) (Bertuccio et  al., 2011) Thus, while tobacco smoking

is strongly associated with an increase in risk of PC, the risk tends

to decrease with longer duration of smoking cessation Although two European studies have suggested that the risk of PC among former smokers approaches that of never smokers after 5 years of smoking cessation (Nilsen and Vatten, 2000; Vrieling et al., 2010), the majority

of studies show that the PC risk among former smokers only becomes similar to that of never smokers after ~15 years of smoking cessation (Bosetti et al., 2012a; Fuchs et al., 1996; Iodice et al., 2008; Lynch

et al., 2009) It should be noted that survival bias likely has a cant effect on OR estimates related to longer cessation periods.Despite the consistent evidence that smoking increases the risk of PC, the specific components of tobacco smoke that alter risk have not been clearly identified (Edderkaoui and Thrower, 2013) Cigarette smoke contains at least 5300 toxic chemicals (and as many as 7000 different chemicals, as estimated by some), including 69 known carcinogens (Hoffmann et al., 2001; Reports of the Surgeon General, 2014) Animal studies have shown that PC can be induced through the administration

signifi-of tobacco- derived N- nitrosamines in drinking water or through

paren-teral administration (Risch, 2003; Rivenson et al., 1988) Metabolites

of tobacco- derived N- nitrosamines are known to cause damage to

DNA (e.g., adduct formation and strand breaks), to inactivate tumor suppressor genes, and to promote oncogene activity, all of which can lead to pancreatic tumorigenesis (Hang, 2010; Hecht, 2003; Wogan

et al., 2004) Such tobacco- derived carcinogens can reach the exocrine pancreas via the bloodstream or through refluxed bile (Lowenfels and Maisonneuve, 2006; Wogan et  al., 2004) Higher levels of tobacco- derived carcinogen adducts have been detected in human pancreatic tissues of smokers than non- smokers (Thompson et al., 1999; Wang

et al., 1998), as have higher levels of tobacco- derived carcinogens in

201816141210

Rate per 100,000 8

642

Hispanic femaleWhite femaleWhite male

Black maleBlack female

Hispanic male

Figure  32– 5 Racial/ ethnic trends in age- standardized pancreatic cancer incidence rates by gender in the United States, 1992– 2013 Source:  The

Surveillance, Epidemiology, and End Results (SEER) Program SEER *Stat Database: Incidence - SEER 13 Registries Research Data National Cancer Institute Bethesda, MD, http:// seer.cancer.gov/ canques/ incidence.html

Cancer of the Pancreas 617

human pancreatic juice (Prokopczyk et al., 2002) In addition,

stud-ies have shown that mutations of the K- ras oncogene, which occur in

more than 90% of PC cases (Bryant et al., 2014), are more common

in pancreatic tumors and plasma DNA of smokers than non- smokers

(Hruban et al., 1993; Li J et al., 2007) Despite the strong relationship

with smoking, the risk of PC varies widely among individuals with

similar smoking histories Data from molecular epidemiologic studies

suggest that these differences may be due to inter- individual variation

in the genetic factors responsible for detoxifying tobacco- derived

car-cinogens (Duell et al., 2002; Li et al., 2005; Suzuki et al., 2008) or

dif-ferences in the ability to repair smoking- related DNA damage (Duell

et al., 2002; McWilliams et al., 2008)

Smokeless tobacco has long been suspected as a cause of PC The

International Agency for Research on Cancer (IARC) has classified

both cigarette smoking and use of smokeless tobacco products as

car-cinogenic and causally related to PC in humans (Cogliano et al., 2011)

In a meta- analysis of six studies from the United States, Sweden, and

Norway, use of smokeless tobacco (chew or snus) was associated

with an increased risk of PC, similar to the association with active

smoking (OR = 1.6; 95% CI: 1.1, 2.2, users vs non- users) (Boffetta

et al., 2008) A population- based case- control study of 526 PC cases

and 2153 controls in the United States also observed that, compared

to non- users of tobacco products, users of more than 2.5 ounces of

smokeless tobacco per week had increased risk of PC, although with

wide confidence intervals (OR  =  3.5; 95% CI:  1.1, 10.6) (Alguacil

and Silverman, 2004) At least one case- control study (Muscat et al.,

1997)  and one cohort study (Zheng et  al., 1993)  have reported a

non- statistically significant increase in risk of PC among smokeless

tobacco users compared to non- users, with risk estimates ranging from

1.7 to 3.6 Other studies have reported null associations, which that

may reflect differences in the smokeless tobacco products used or

dif-ferences in study methods (Bertuccio et  al., 2011; Sponsiello- Wang

et al., 2008)

Few studies have examined the association between exposure to

environmental tobacco smoke (ETS) and the risk of PC Among non-

smokers, exposure to ETS increases the risk of PC in a dose- dependent

manner, with childhood exposure doubling the risk of PC (Vrieling

et al., 2010) A population- based case- control study in Canada

involv-ing 583 incident PC cases and 4813 controls reported a positive

asso-ciation between exposure to ETS during childhood or adulthood and

PC risk (OR = 1.21; 95% CI: 0.60, 2.44, exposed vs never exposed)

(Villeneuve et al., 2004) A hospital- based, case- control study of 808

newly diagnosed PC patients and 808 healthy controls at the M. D

Anderson Cancer Center also reported a non- significant increased PC

risk among individuals with self- reported exposure to ETS (Hassan

et  al., 2007), as did a population- based cohort study of 93,921

par-ticipants that included 148 PC cases in Washington County, Maryland

(Gallicchio et  al., 2006) European studies have reported stronger

association between ETS exposure and PC risk (Chuang et al., 2011;

Vrieling et al., 2010), and while the overall data suggest that ETS may

contribute to PC risk, no consensus group has yet designated this

evi-dence as definite

Overweight/ Obesity and Physical Activity

The relationship between increased body mass index (BMI) and

PC has been designated as causal by the World Cancer Research

Fund/ American Institute for Cancer Research (WCRF/ AICR) (see

Chapter  20) At least 22 prospective studies in independent cohorts

have examined this relationship, with 19 studies reporting a positive

association (summarized in Aune et  al., 2012) Four meta- analyses

confirm the association between excess body weight and increased risk

of PC (Aune et al., 2012; Berrington de Gonzalez et al., 2003; Larsson

et al., 2007; Renehan et al., 2008) The first covered studies were

pub-lished between 1966 and 2003, and included 14 studies with a total

of 6391 PC cases (Berrington de Gonzalez et al., 2003) This meta-

analysis reported that the risk of PC increased by 2% for every 1 kg/ m2

increase in BMI (relative risk [RR] = 1.02; 95% CI: 1.01, 1.03) overall,

with nearly identical findings in men (RR = 1.03; 95% CI: 1.01, 1.06)

and women (RR = 1.02; 95% CI: 1.00, 1.03) (Berrington de Gonzalez

et al., 2003) The second meta- analysis included 21 prospective ies published between 1966 and 2006 involving 3,495,981 participants with 8062 PC cases It reported the risk of PC per 5 kg/ m2 increase

stud-in BMI of 16% among men (RR = 1.16; 95% CI: 1.05, 1.28), 10% among women (RR = 1.10; 95% CI: 1.02, 1.19), and 12% in both sexes combined (RR  =  1.12; 95% CI:  1.06, 1.17) (Larsson et  al., 2007) The third meta- analysis reviewed studies published up to 2007 and included 16 studies consisting of 4443 PC cases and 3,303,073 partici-pants, and reported RR estimates per 5 kg/ m2 increase in BMI of 1.07 (95% CI: 0.93, 1.23) in men and 1.12 (95% CI: 1.03, 1.23) in women (Renehan et al., 2008) The most recent meta- analysis considered all studies published up to 2011 and included 29 publications in an anal-ysis that involved 5,037, 555 participants and 9504 PC cases (Aune

et al., 2012) It estimated an overall 10% increase in risk of PC per

5 kg/ m2 increase in BMI (RR = 1.10; 95% CI: 1.07, 1.14), with similar risks in men (RR = 1.13; 95% CI: 1.04, 1.22) and women (RR = 1.10; 95% CI: 1.04, 1.16) A 2004 study estimated that about 27% of all PC cases in the United States were attributable to being overweight and obese (Calle and Kaaks, 2004) In the United Kingdom, it was esti-mated in 2010 that about 12% of PC cases were attributable to being overweight or obese (Parkin et al., 2011)

The relationship between excess body weight and PC is thought to be mediated by insulin resistance (Calle and Kaaks, 2004) Insulin resis-tance generally occurs from metabolic adaptation to excessive release of free circulating fatty acids from adipocytes, particularly from adipocytes

in intra- abdominal adipose tissue, and compensated through tion of insulin from the pancreas (Calle and Kaaks, 2004; Kahn and Flier, 2000) There is mounting evidence that insulin resistance and its com-pensatory hyperinsulinemia are associated with increased risk of many cancer types, including colon (Giovannucci, 2007; McKeown- Eyssen, 1994), endometrial (Hernandez et  al., 2015; Kaaks et  al., 2002), and pancreatic (Stolzenberg- Solomon et al., 2005; Weiderpass et al., 1998; Wolpin et al., 2013) cancers There also are suggestions that excess body weight might promote pancreatic tumor growth through the activation of insulin- like growth factor pathways, leading to increased cell prolifera-tion and inhibition of apoptosis (Min et al., 2003)

oversecre-Few studies have examined the association between abdominal ness or central adiposity and risk of PC In a study of the American Cancer Society Cancer Prevention Study II (CPS- II) Nutrition Cohort consisting of 145,627 participants, of whom 242 had developed PC, participants with central weight gain (i.e., weight gain in the “waist” or

fat-“chest and shoulders”) were compared to those with peripheral weight gain (i.e., weight gain in “hips and thighs” or “equally all over”) Men and women with central weight gain were found to have a higher risk

of PC (RR = 1.45; 95% CI: 1.02, 2.07) (Patel et al., 2005) A European- wide cohort study involving 438,405 participants that included 324

PC cases found that larger waist circumference and higher waist- to- hip ratios were both associated with greater risk of PC: RR per 10 cm increase in waist circumference = 1.13 (95% CI: 1.01, 1.26), and RR per 0.1 increment in waist- to- hip ratio = 1.24 (95% CI: 1.04, 1.48), respectively (Berrington de Gonzalez et al., 2006) In the analysis of pooled data from 14 cohort studies described earlier, higher waist- to- hip ratio was found to be associated with an increased risk of PC (RR = 1.35; 95% CI: 1.03, 1.78, highest vs lowest quartile) (Genkinger

et al., 2011) A 2008 analysis of 5- year follow- up data from the NIH- AARP Diet and Health Study cohort involving 495,035 participants and including 212 cases among men and 100 cases among women found that larger waist circumference was associated with a higher risk of PC in women (RR = 2.53; 95% CI: 1.13, 5.65, highest vs low-est quartile), but not in men (Stolzenberg- Solomon et al., 2008) Luo

et al also observed a positive association between waist- to- hip ratio and PC in the Women’s Health Initiative based on analysis of 251 cases accrued from 138,503 participants after an average of 7.7 years

of follow- up (hazard ratio [HR] = 1.70; 95% CI: 1.1, 2.6, highest vs lowest quintile; and HR = 1.27; 95% CI: 1.1, 1.5, per 0.1 increase in waist- to- hip ratio) (Luo et al., 2008) The most recent meta- analysis

on PC also showed that the disease risk increases by 11% for every

10 cm increase in waist circumference and by 19% for every 0.1- unit increment in waist- to- hip ratio (Aune et al., 2012)

Regular physical activity has been known to reduce body fat mass

and improve glucose intolerance and insulin sensitivity; therefore, it is

hypothesized that physical activity may modulate PC risk through these

mechanisms (Forman et al., 2007) However, study results have varied

Michaud et  al reported that moderate recreational physical activity

(activity performed solely for pleasure or enjoyment) was associated

with reduced risk of PC (RR = 0.45; 95% CI: 0.29, 0.70), but no

asso-ciation was found with total physical activity (Michaud et al., 2001)

Brenner et  al also reported an inverse association between regular

leisure- time physical activity and PC risk (OR = 0.65; 95 % CI: 0.52,

0.87) (Brenner et al., 2014), as did Kollarova et al (OR = 0.63; 95%

CI:  0.43, 0.92) (Kollarova et  al., 2014)  and Stolzenberg- Solomon

et al (HR = 0.42, 95% CI: 0.22, 0.83) for moderate or heavy

activ-ity (Stolzenberg- Solomon et al., 2002a), but other studies did not find

an association (Patel et al., 2005; Stolzenberg- Solomon et al., 2008)

Occupational physical activity also has been associated with reduced

risk of PC in some but not all studies (as reviewed in Bao and Michaud,

2008; O’Rorke et al., 2010) Measurement error in the assessment of

physical activity is a major limiting factor in almost all studies that

may have obscured the true association

Diet

Based on an extensive literature review for PC, the WCRF/ AICR

reported that evidence was limited that foods containing folate or fruits

provide protection, inconsistent regarding vegetables, and suggestive

but limited for an increased risk associated with red and processed

meat, food and beverages containing fructose, and saturated fatty acids

(Wiseman, 2008) Since that report was published, four cohort and

five case- control studies have investigated various dietary components

and PC The focus of the more recent literature has largely been on

nutrient components (e.g., specific fatty acids) rather than food

cat-egories (e.g., citrus fruit) Most studies have generally suggested that

numerous components of fruits and vegetables (including vitamin B6,

choline, β- carotene, zeaxanthin, α- tocopherol, and flavonoids) and

whole grains provide a protective effect (Heinen et al., 2012; Huang

et al., 2016; Jansen et al., 2011) Unsaturated fatty acids appear to also

provide a protective effect, while fats and vitamin D found in dairy

may increase association with risk (Jansen et al., 2014; Thiebaut et al.,

2009; Waterhouse et al., 2016) Compared to cohort studies, the case-

control studies tended to report a higher number of significant results

For meat mutagens and meat preparation/ doneness preferences, the

evidence from two cohort and two case- control studies has generally

shown positive associations between PC and increasing intake of well-

done grilled/ barbecued meat, heterocyclic amines, and a

mutagenic-ity activmutagenic-ity index (revertants / grams of daily meat intake) based on

mutagenicity in the Salmonella- based Ames Assay (Anderson et al.,

2012; Li D et al., 2007; Stolzenberg- Solomon et al., 2007) A newer

case- control study has reported no association (Jansen et al., 2013a)

Proposed explanations for the inconsistencies between case- control

and cohort studies include information and reporting bias with respect

to dietary ascertainment, variability in histologically verified tumor

types, and heterogeneous measures of intake (Vrieling et  al., 2009)

Case- control study participant selection (i.e., healthy survivor bias [Hu

Z- H et al., 2016]) could explain the observed inverse associations since

the majority of cohort studies report null results, and cohort studies are

not affected by possible diet changes after a cancer diagnosis Another

suggested explanation for these diet- cancer inconsistencies is variation

in underlying gene polymorphisms involved in metabolizing

compo-nents of the diet (Huang et al., 2016; Jansen et al., 2013b) The main

underlying potential pathways for how dietary intake could lead to or

prevent PC include chronic inflammation, insulin sensitivity, and

oxi-dative stress Specific foods and/ or food components have been

experi-mentally shown to alter the cellular state and to lead to PC development

in model organisms

Alcohol

The WCRF/ AICR reported that there is suggestive evidence of

increased risk associated with heavy alcohol use (Wiseman, 2008)

Across epidemiologic studies, there are often variations in ing and reporting alcohol exposure, leading to difficulty in comparing study results For pooled analyses, when comparing the highest versus lowest intake categories, the RR ranged from 1.22 to 1.38 Since 2009, two pooled data analyses and one meta- analysis have been performed (Genkinger et al., 2009; Michaud et al., 2010; Tramacere et al., 2010) There were several different control groups for these studies (0 g etha-nol/ day or > 0– 4.99 g ethanol/ day or < 1 drink/ day) with a wide range

measur-of definitions measur-of heavy drinking (> 30 gram/ day or > 45 grams/ day or

> 9 drinks/ day or > 3 drinks/ day) All studies showed a significantly increased risk for PC among heavy drinkers Among three recent case- control studies, one in Japan estimated that those who drank > 80 grams/ day were diagnosed with PC on average 3.1 years earlier than those who drank < 50 grams/ day The other two studies, one in China (Zheng et al., 2016) and one in Canada (Rahman et al., 2015), showed

no significant increased risk when comparing any alcohol tion to no consumption or heavy to low categories, and type consumed

consump-In individual epidemiological studies, this association is difficult to detect since they typically are limited by sample size, potential recall bias, or possible selection bias Additionally, power issues arise when alcohol is split based on the type of alcohol consumed (i.e., beer, wine,

conse-or where PC patients were screened fconse-or diabetes using fasting glucose values or glucose tolerance tests, it has been observed that ~80% of PC patients have a diagnosis of diabetes or glucose intolerance that meets the American Diabetes Association’s diagnostic criteria for diabetes (Pannala et al., 2009) However, this includes both preexisting diabetes and situations where the diagnosis of PC preceded the diagnosis of diabetes (referred to as “PC- associated diabetes”) This reverse causa-tion is based on the observation that diabetes generally appears within

3 years before the diagnosis of PC and is thus an early manifestation

of PC (Chari et al., 2008; Pannala et al., 2009) Studies have reported

a wide range of frequencies of new- onset diabetes attributable to PC, from as high as 52% to as low as 16% (Chari et al., 2008; Gullo et al., 1994) In patients with PC- associated diabetes, it is believed that the tumor produces “diabetogenic” substances that impair glucose metab-olism (Li, 2012; Pannala et al., 2009) This is supported by observa-tions that PC- associated diabetes generally resolves after pancreatic tumor resection, whereas removal of a pancreatic tumor does not resolve preexisting diabetes (Chari, 2007) Thus, while long- standing diabetes is a risk factor for PC, new- onset diabetes in later adulthood could be an early sign of undiagnosed PC

At least five overlapping meta- analyses (Batabyal et al., 2014; Ben

et  al., 2011; Everhart and Wright, 1995; Huxley et  al., 2005; Song

et al., 2015) and three pooled analyses (Bosetti et al., 2014; Elena et al., 2013; Li et al., 2011) attempted to clarify the bidirectional relationship between diabetes and PC by examining the duration of diabetes rela-tive to PC diagnosis (Batabyal et al., 2014; Ben et al., 2011; Everhart and Wright, 1995; Huxley et al., 2005; Song et al., 2015) Although all of the studies reported ≥ 50% increase in risk of PC among indi-viduals with preexisting diabetes, there were suggestions of decreas-ing PC risk as the duration of diabetes increases A meta- analysis by Huxley and colleagues (2005) that included 19 prospective studies and

17 case- control studies reported smaller magnitudes of association as the duration of diabetes increased, with RRs for 1– 4 years, 5– 9 years, and ≥ 10 years duration of diabetes of 2.05 (95% CI: 1.87, 2.25), 1.54 (95% CI:  1.31, 1.81), and 1.51 (95% CI:  1.16, 1.96), respectively The majority of diabetic cases in the 1– 4  year group are most cer-tainly PC- associated diabetes In another meta- analysis involving 30

Cancer of the Pancreas 619

cohort studies, Ben et al observed a significant decline in the

associa-tion between preexisting diabetes and PC as the duraassocia-tion of diabetes

increased (Ben et al., 2011) The RRs for the subgroups of 1– 4 years,

5– 9  years, and ≥ 10  years duration of diabetes were 1.95 (95%

CI: 1.65, 2.31), 1.49 (95% CI: 1.05, 2.12), and 1.47 (95% CI: 0.94,

2.31), respectively (Ben et al., 2011) In a pooled analysis of 15 case-

control studies involving 8305 PC cases and 13,987 controls, Bosetti

and colleagues (2014) also observed decreasing PC risk with

increas-ing duration of diabetes, although the risk persisted in those with more

than 20 years duration of diabetes The RRs were 2.92 (95% CI: 2.44,

3.50), 1.84 (1.54, 2.20), 1.69 (1.36, 2.09), 1.54 (1.17, 2.03) and 1.30

(1.03, 1.63) for diabetes duration of 2– 4 years, 5– 9 years, 10– 14 years,

15– 19 years, and ≥ 20 years, respectively (Bosetti et al., 2014) Elena

et al examined the association between diabetes duration and PC by

pooling primary data from 12 cohort studies that included 1621

inci-dent PC cases and 1719 controls They observed a strong association

among individuals with diabetes duration of 2– 8  years (OR  =  1.79;

95 % CI: 1.25, 2.55) and no association among those who had been

diagnosed with diabetes for ≥ 9 years (OR = 1.02; 95 % CI: 0.68, 1.52)

(Elena et al., 2013) It is possible that the decreasing PC risk associated

with a longer duration of diabetes could be related to lifestyle

modi-fications (e.g., healthy diet, physical activity) following the diabetes

diagnosis or the use of some diabetic medications, such as metformin,

which has been associated with reduced risk of PC in some studies

(Bosetti et al., 2014; Li et al., 2009)

Very few studies have investigated the association between type

1 diabetes and PC The large majority of these have been null, but

individual studies have included only small numbers of both type 1

diabetes and PC cases One cohort study of 109,581 participants

iden-tified 11 individuals with type 1 diabetes, among whom the risk of PC

was greater than that of non- diabetics; it reported a standardized PC

incidence ratio of 3.50 (95% CI: 1.80, 6.30) (Wideroff et al., 1997)

In a meta- analysis of three cohort studies and six case- control studies,

individuals with type 1 diabetes or onset before age 40 had a 2- fold

increase in PC risk (RR = 2.0; 95% CI: 1.37, 3.01) This observation

was based on 39 PC cases with diabetes (Stevens et al., 2007) It has

been suggested that the use of exogenous insulin in type 1 diabetes

may partly account for the increased PC risk (Li, 2012) Given the

small number of diabetic cases included in the existing studies, further

research with larger numbers of both conditions is needed to verify

the findings

Reasons for the association between preexisting type 2 diabetes and

PC include the fact that they share many risk factors, such as obesity,

physical inactivity, cigarette smoking, and excessive alcohol

consump-tion (Herman, 2007; Lowenfels and Maisonneuve, 2006) Proposed

mechanisms for the link between diabetes and PC include insulin

resistance, chronic hyperinsulinemia, increased circulating levels of

insulin- like growth factors (IGFs), and chronic inflammation (Bao

et al., 2011; Li, 2012) In the setting of insulin resistance, excessive

amounts of insulin are secreted by the pancreatic β-cells as a

com-pensatory mechanism to regulate blood glucose levels This exposes

the exocrine pancreas to high concentrations of endogenous insulin

for many years (Bao et al., 2011) Experimental evidence shows that

chronic hyperinsulinemia can promote cell proliferation and decrease

cellular apoptosis, which are conditions that favor carcinogenesis (Li,

2012; Magruder et al., 2011)

Other studies have investigated the hypothesis that impaired

gly-cemic control, insulin resistance, and its compensatory

hyperinsu-linemia may promote pancreatic tumorigenesis In 2005, a study by

Stolzenberg- Solomon et  al examined the association between pre-

diagnostic fasting glucose level, blood insulin concentration, the

extent of insulin resistance, and risk of PC in a case- cohort study of

169 incident PC cases and 400 non- cancer controls (Stolzenberg-

Solomon et al., 2005) The study participants were randomly selected

from the Alpha- Tocopherol Beta- Carotene Cancer Prevention (ATBC)

trial of 29,133 Finnish male smokers of age 50– 65 years, who were

followed for up to 17  years The study found that increased blood

glucose, insulin levels, and insulin resistance were positively

associ-ated with PC risk The associations were particularly strong for PC

cases that were diagnosed after 10 years of follow- up (HRs and 95%

CIs for highest vs lowest quartiles: glucose 2.16 [1.05, 4.42], insulin concentration 2.90 [1.22, 6.92], insulin resistance 2.71 [1.19, 6.18]) (Stolzenberg- Solomon et al., 2005) Maisonneuve et al also reported

on the use of insulin injection versus oral hypoglycemic medication for control of diabetes and risk of PC in a population- based study of

823 newly diagnosed PC cases and 1679 controls (Maisonneuve et al., 2010) The study showed over 2- fold increase in PC risk among dia-betes patients compared to non- diabetics (OR = 2.16; 95% CI: 1.60, 2.91) The management of diabetes with insulin injection was associ-ated with a greater risk (OR = 3.54; 95% CI: 1.64, 7.61) than the use

of oral hypoglycemic medication (OR  =  1.78; 95% CI:  1.16, 2.75) (Maisonneuve et al., 2010) These epidemiological reports are corrob-orated by an earlier experimental study that demonstrated that physi-ologic levels of insulin stimulate pancreatic tumor cell proliferation and glucose utilization through the activation of mitogen- activated protein (MAP) kinase and phosphatidylinositol 3- kinase (PI3- kinase),

as well as increased expression of glucose transporter 1 (GLUT1) (Ding et al., 2000)

Insulin’s effect on pancreatic tumor development and progression has been shown in animal models to occur also through promoting insulin- like growth factor 1 (IGF- 1) activity (Li, 2012) High levels of insulin can stimulate synthesis of IGF- 1 and suppress IGF- 1 binding activity in the liver at the same time, which increases circulating levels

of IGF- 1 (Bao et al., 2011; Li, 2012) Increased levels of free lating IGF- 1 have been shown to promote cell proliferation, inhibit apoptosis, and enhance angiogenesis in the tumor microenvironment (Bao et al., 2011) Moreover, high IGF- 1 levels have been associated directly with increased risk of PC (McCarty, 2001) Thus, preexisting diabetes may contribute to PC risk through diabetes- associated insulin resistance, its impact on insulin secretion from the pancreas or admin-istration of exogenous insulin, and the subsequent effect of insulin on IGF- 1 activity Diabetes is also associated with chronic inflammation (Garcia et al., 2010), and chronic inflammation is known to promote pancreas tumorigenesis, possibly by high levels of inflammatory cyto-kines in the setting of diabetes (Li, 2012)

circu-Chronic Pancreatitis

The exocrine portion of the pancreas is susceptible to two main eases: pancreatitis and PC (Raimondi et al., 2010) Pancreatitis is gen-erally classified as acute or chronic, based on presenting symptoms, past medical history, and clinical course While acute pancreatitis, which is typically caused by gallstones or heavy alcohol consump-tion, is characterized by inflammation that resolves after removal of the offending agent, chronic pancreatitis is a continuing inflamma-tory condition marked by irreversible morphological changes to the pancreas that include calcification, fibrosis, and pancreatic ductal inflammation (Forsmark, 2007; Raimondi et  al., 2010) Although chronic pancreatitis has been linked to habitual alcohol abuse (~70%

dis-of cases) and cigarette smoking (< 10% dis-of cases), in a significant proportion of patients (~20%), the etiology is unknown (Forsmark, 2007) There is strong epidemiological and clinical evidence linking long- standing history of chronic pancreatitis with increased risk of PC (Duell et al., 2012; Howes and Neoptolemos, 2002; Lowenfels et al., 1993; Raimondi et al., 2010) However, because pancreatic tumors can obstruct the flow of enzymes from the pancreatic duct into the blood-stream, pancreatitis can result from tumor- associated ductal obstruc-tion (Forsmark, 2007) Therefore, as with diabetes and PC, there is a two- way relationship between chronic pancreatitis and PC

In studies that excluded from analyses cases of PC diagnosed temporally close to diagnosis of pancreatitis in order to reduce confounding by reverse causation, results showed a strong associa-tion between antecedent chronic pancreatitis and PC (Duell et al., 2012; Karlson et  al., 1997; Lowenfels et  al., 1993; Olson, 2012) For example, in 1993, Lowenfels et al conducted a cohort study of

2015 chronic pancreatitis patients who were followed at major pitals in Europe and in the United States for an average of 7.4 years (Lowenfels et al., 1993) After excluding PC cases that occurred in the first two years of follow- up, they observed a 16 times higher incidence of PC in the chronic pancreatitis patients than expected in the general population (standardized incidence ratio [SIR] = 16.5;

hos-95% CI: 11.1, 23.7) The excess risk of PC remained when analyses

were performed separately for patients who had lived with chronic

pancreatitis for 10  years or more, suggesting an etiological link

between preexisting chronic pancreatitis and PC (Lowenfels et al.,

1993) In 1999, Talamini et  al also conducted a cohort study of

715 chronic pancreatitis patients followed for a median of 10 years

(Talamini et al., 1999) After eliminating PCs that occurred within

4 years of diagnosis of pancreatitis, they observed a 13 times greater

risk of PC in the chronic pancreatitis patients than expected in the

general population (SIR = 13.3; 95% CI: 6.4, 24.5) However, after

excluding the early- onset PCs, no cases of PC had occurred among

non- smokers The investigators, therefore, concluded that cigarette

smoking and perhaps other factors associated with chronic

inflam-mation (e.g., alcohol abuse) may play a key role in the association

between chronic pancreatitis and PC (Talamini et  al., 1999) In a

Swedish hospital- based cohort study of 4546 chronic pancreatitis

patients conducted in 1997, Karlson et al reported a nearly 8- fold

excess risk of PC in chronic pancreatitis patients than expected

(SIR  =  7.6; 95% CI:  6.0, 9.7) (Karlson et  al., 1997) They also

observed a generally declining, but persistently higher magnitude of

PC risk for patients living with chronic pancreatitis for 1– 4 years,

5– 9 years, and 10– 24 years, with SIRs of 22.2 (95% CI: 16.2, 29.6),

4.6 (95% CI: 2.6, 7.5), and 2.2 (95% CI: 0.9, 4.4), respectively The

study further showed that the risk of PC remained significantly high

among patients living with alcohol- related chronic pancreatitis for

10 years or more (SIR = 3.8; 95% CI: 1.5, 7.9), suggesting that

alco-hol abuse may contribute to reported associations between chronic

pancreatitis and PC risk (Karlson et al., 1997)

In more recent studies where both alcohol intake and cigarette

smok-ing were accounted for in the analyses, havsmok-ing a history of chronic

pancreatitis remained significantly associated with an increase in risk

of PC (Duell et al., 2012; Olson, 2012; Raimondi et al., 2010) Duell

et al reported a higher risk of PC among individuals with more than

2  years duration of chronic pancreatitis (OR  =  2.71; 95% CI:  1.96,

3.74), after adjusting for known risk factors of PC, including alcohol

intake and cigarette smoking in a pooled analysis of data from 10 case-

control studies consisting of 5048 cases and 10,947 controls (Duell

et  al., 2012) They also reported a stronger association for chronic

pancreatitis cases diagnosed less than 2 years before diagnosis of PC

(OR = 13.56; 95% CI: 8.72, 21.90), which likely reflects reverse

cau-sation or initial misdiagnosis of PC as pancreatitis (Duell et al., 2012)

In addition, they observed that the association remained statistically

significant for individuals with 10 years or more duration of chronic

pancreatitis Raimondi et al also reported a 5.8- fold higher risk of PC

among individuals with chronic pancreatitis from a meta- analysis of

22 studies after excluding PC cases diagnosed within 2 years of the

diagnosis of pancreatitis (RR  =  5.8; 95% CI:  2.1, 15.9) (Raimondi

et al., 2010) This pair of reports is consistent with findings from an

earlier review of 25 epidemiological studies that included eight studies

that examined the time interval between diagnosis of pancreatitis and

PC (Olson, 2012) The review showed that although the association

generally declined with increasing duration of pancreatitis, all eight

studies reported elevated risk of PC among individuals with

pancreati-tis, even when pancreatitis was diagnosed many years before the

diag-nosis of PC (RR ranging from 1.8 to 5.1) (Olson, 2012) Moreover,

PC has been associated with tropical chronic pancreatitis (Chari et al.,

1994) and hereditary pancreatitis (a germline autosomal- dominant

dis-order) (Lowenfels et al., 2000) Thus, it appears that patients with any

form of chronic pancreatitis have a high risk of PC However, chronic

pancreatitis is a relatively rare disease (annual incidence is ~5– 10 per

100,000), and it is estimated to account for only 1% of the overall

bur-den of PC (Duell et al., 2012; Raimondi et al., 2010)

Drugs

Non- Steroidal Anti- Inflammatory Drugs (NSAIDs)

Experimental studies suggest that NSAIDs can inhibit pancreatic

tumor growth (Kokawa et al., 2001; Molina et al., 1999; Perugini et al.,

2000); however, the majority of epidemiologic studies do not show an association between NSAID use and PC risk (Anderson et al., 2002; Capurso et al., 2007; Coogan et al., 2000; Larsson et al., 2006; Tan

et  al., 2011) Several studies suggest that regular use of aspirin, the most frequently studied NSAID, is associated with reduced risk of PC, but the overall evidence is inconsistent In a prospective cohort study

of 28,283 postmenopausal women in the Iowa Women’s Health Study,

80 of whom had PC, Anderson et al observed that current use of rin was associated with lower risk of PC (OR = 0.57; 95% CI: 0.36, 0.90, users vs non- users) (Anderson et  al., 2002) They found also that higher frequency of aspirin use was associated with lower risk

aspi-of PC (OR = 0.40; 95% CI: 0.20, 0.82; ≥ 6 times per week vs never

users, P trend = 0.005) (Anderson et al., 2002) In a clinic- based, case-

control study of 904 rapidly ascertained incident PC cases and 1224 controls, Tan et al observed a lower risk of PC among aspirin users (OR = 0.74; 95% CI: 0.60,0.91; ≥ 1 days/ month vs < 1 day/ month) and among those with greater frequency of aspirin use (OR = 0.63;

95% CI: 0.47, 0.85; ≥ 6 day/ week vs < 1 day/ month, P trend = 0.007)

(Tan et  al., 2011) A  similar finding was reported by Streicher and colleagues (2014) In contrast, Schernhanner et  al observed a sug-gestive increased PC risk among aspirin users in the Nurses’ Health Study among 88,378 women, of whom 161 had PC (OR = 1.20; 95% CI:  0.87, 1.65; current users vs never users) (Schernhammer et  al., 2004) The study further showed a pattern of increasing risk of PC with increasing duration of aspirin use (OR  =  1.58; 95% CI:  1.03, 2.43;

≥ 20 years of regular use vs non- use, P trend = 0.01) (Schernhammer

et al., 2004) Results from meta- analyses do not support the tion that aspirin or NSAID use reduce PC risk (Bosetti et al., 2012b; Capurso et al., 2007; Larsson et al., 2006)

sugges-Metformin

Epidemiologic reports indicate that the use of metformin, a commonly prescribed oral antidiabetic medication, is associated with reduced risk

of PC (Li, 2012) A hospital- based case- control study of 973 cases and

863 controls observed a 62% lower risk of PC among diabetic patients who had taken metformin when compared to non- diabetic control patients and after controlling for duration of diabetes, use of insulin, and other risk factors of PC (OR = 0.38; 95% CI: 0.21, 0.67) (Li et al., 2009) A further slight reduction in risk was observed among diabetics with longer duration of metformin use (OR = 0.30; 95% CI: 0.13, 0.69;

≥ 5 years vs never use) (Li et al., 2009) This finding is corroborated

by a retrospective cohort study of 62,809 diabetics, of whom 89 had

PC (HR = 0.20; 95% CI: 0.011, 0.36; metformin users vs non- users) (Currie et al., 2009), and by a recent meta- analysis involving 10 cohort studies and three case- control studies (HR = 0.63; 95% CI: 0.46, 0.86; users vs non- users) (Wang et al., 2014) Potential mechanisms for the effect of metformin on PC include lowering circulating levels of insu-lin, improving peripheral insulin resistance, and blocking the mito-genic effects of IGF- 1 (Li et al., 2009; Li, 2012)

Statins

In vitro (Gbelcova et  al., 2008; Kusama et  al., 2002; Wong et  al., 2001) and in vivo (Gbelcova et al., 2008; Kusama et al., 2002; Sumi

et  al., 1992; Wong et  al., 2001)  studies suggest that statins, a class

of cholesterol- lowering drugs, have anti- pancreatic tumor ties Postulated mechanisms of these 3- hydroxy- 3- methylglutaryl- coenzyme A  (HMG- CoA) reductase inhibiting agents include pro- apoptotic activity and blockade of the mevalonate metabolic pathway (Kusama et  al., 2002; Wong et  al., 2001) However, in a

proper-2008 meta- analysis that included three randomized clinical trials, four cohort studies, and five case- control studies, no association was observed between statin use and PC risk (Bonovas et al., 2008) It is, however, possible that the antitumor effectiveness of statins depends

on their combination with other anticancer agents (Bocci et al., 2005; Yao et al., 2006)

Insulin/ Insulin Secretagogues

While there is mounting evidence linking the use of insulin or lin secretagogues (drugs that increase insulin secretion by pancreatic β- cells, e.g., sulfonylureas) with increased risk of PC (Bodmer et al.,

insu-Cancer of the Pancreas 621

2012; Currie et al., 2009; Li et al., 2009; Stolzenberg- Solomon et al.,

2005), it remains unclear whether long- term insulin use is associated

with an increase in PC risk In a pooled analysis of three case- control

studies involving 2192 cases and 5113 controls, insulin use for 3 years

or less was associated with higher risk of PC, whereas use of

insu-lin for more than 10 years was associated with lower risk (Li et al.,

2011) The study reported ORs for ≤ 3  years, >3– ≤ 10  years, and

> 10 years of 2.4 (95% CI: 1.6, 3.7), 1.2 (95% CI: 0.7, 1.9), and 0.50

(95% CI: 0.30, 0.90), respectively, for insulin- using diabetes patients

compared to non- insulin- using diabetics (Li et al., 2011) Other studies

that reported an overall increase in PC risk among insulin users also

reported null associations between PC and insulin use for more than

5 years (Hassan et al., 2007; Wang et al., 2006) or more than 10 years

(Silverman et al., 1999) Thus, although considerable evidence links

insulin use to increased PC risk, the association may be confounded by

duration of insulin use and, by extension, duration of diabetes because

of its complex relationship with PC, especially in studies where

analy-ses were not restricted to diabetic patients or where duration of

diabe-tes was not considered in the analyses

Occupation and Physical Environment

PC incidence has been associated with various occupational

expo-sures, including chlorinated hydrocarbons (CHC), polycyclic aromatic

hydrocarbons (PAHs, mainly derived from diesel exhaust and

alumi-num production), organochlorine pesticides, and heavy metals such

as nickel, chromium, and cadmium (Andreotti and Silverman, 2012;

Ojajarvi et al., 2000) Results from a meta- analysis of 92 occupational

studies published between 1969 and 1998 showed excess risk of PC

among individuals exposed to chlorinated hydrocarbons and nickel,

with meta- risk- ratios of 1.4 (95% CI: 1.0, 1.8) and 1.9 (95% CI: 1.2,

3.2), respectively (Ojajarvi et al., 2000) In 2012, a review of

occu-pational studies published since this meta- analysis concluded that

work- related exposure to CHC, pesticides, PAHs, nitrosamines, and

radiation have been consistently associated with increased PC risk,

with stronger evidence for CHC and PAHs (Andreotti and Silverman,

2012) Non- occupational studies have also shown that increased

expo-sure to environmental chemicals and heavy metals is associated with

increased risk A population- based case- control study in southeastern

Michigan reported excess risk of PC among individuals with self-

reported exposure to organochlorine pesticides, including ethylan,

chloropropylate, and dichlorodiphenyltrichloroethane (DDT) (Fryzek

et al., 1997) A hospital- based case- control study conducted in the East

Nile Delta region of Egypt observed an elevated risk of PC among

patients who self- reported exposure to pesticides (OR  =  2.6; 95%

CI: 0.97, 7.2) (Lo et al., 2007) In a clinic- based, case- control study

in Rochester, Minnesota, excess risk of PC was observed among

indi-viduals with self- reported exposure to pesticides, CHC, benzene, and

asbestos (Antwi et al., 2015) However, these hospital or clinic- based

case- control studies are limited by their retrospective nature,

particu-larly the reliance on self- reported exposure after cancer diagnosis It

has been suggested that environmental chemicals and heavy metals

may reach the pancreas via the bloodstream or through bile reflux into

the pancreatic duct and, in the process, these substances exert their

carcinogenic effects, such as activation of oncogenes, formation of

DNA adducts, and impairment of DNA damage repair mechanisms

(Wogan et al., 2004) It remains to be demonstrated whether reducing

exposures to these substances in occupational and non- occupational

settings would have a significant effect on the incidence of PC

Genetic Susceptibility

family History and familial PC

Family history of PC is a consistent and important risk factor of

PC Early familial clustering studies in which families with

mul-tiple siblings or over mulmul-tiple generations were affected (Dat and

Sontag, 1982; Ehrenthal et  al., 1987; Friedman and Fialkow, 1976;

Ghadirian et al., 1987; MacDermott and Kramer, 1973) were followed

by familial aggregation studies and analysis of families using formal study designs, finding increased risk A  comprehensive summary of these studies and estimated risks are listed in Table 32– 2 Seven case- control studies, two cohort studies, one population- based genealogic analysis, and one case series that estimated incidence of PC in relatives have found that first- degree relatives have at least a 2- fold increased risk of developing PC These findings are remarkably consistent, given that case ascertainment and data collection spanned 30 or more years, multiple countries and cultures, and different methods for estimat-ing risk A systematic review and meta- analysis by Permuth- Wey and Egan (2009) of a cohort study and seven case- control studies totaling

6568 PC cases calculated an overall relative risk of 1.80 (95% CI: 1.48, 2.12) They also found that 1.3% of PC in the population is attributable

to family history The risk was consistent for both males and females, and did not differ by early or late age at diagnosis With respect to risk for second- degree relatives (aunts, uncles, grandparents, grand-children), both Hassan et al (2007) and Shirts et al (2010) reported risks comparable to those of first- degree relatives (relatives risks of 2.9 (95% CI: 1.3, 6.3) and 1.59 (95% CI: 1.31, 2.91), respectively In addition, a large multicenter cohort study examined risk by number of affected individuals and showed even high risk associated with hav-ing two or more first- degree relatives with PC with an odds ratio of 4.26 (95% CI: 0.48, 37) (Jacobs et al., 2010) A population- based twin study of cancer in Sweden by Lichtenstein et al (2000) estimated PC heritability to be 36%, similar to colorectal cancer (35%), higher than breast cancer (27%), and slightly lower than prostate cancer (42%)

In the first published segregation analysis of PC, Klein et al (2002) analyzed family histories in 287 PC patients seen during 1994– 1999

at the Johns Hopkins Hospital in Baltimore, Maryland The analysis rejected non- genetic transmission models The data best fit a major gene model that was predicted to follow an autosomal dominant pat-tern of a rare allele; 0.7% of the population would carry a high risk of developing PC due to this putative gene A smaller study of 70 families drew a similar conclusion (Banke et al., 2000)

Familial PC (FPC) was defined to advance consistency in research, and was defined as kindreds containing at least a pair of individuals who were affected with pancreatic adenocarcinoma and who were first- degree relatives (Hruban et  al., 1998) Numerous studies using this definition have focused on genetic epidemiology (Petersen et al., 2006) and susceptibility gene discovery (Amundadottir, 2016; Klein, 2012; Petersen, 2015) These studies have shown that FPC is geneti-cally heterogeneous, and that increased risk of PC is seen in hereditary cancer syndromes Using increasingly sophisticated genomic analysis

technologies, susceptibility genes include most commonly BRCA1/

2, CDKN2A, PALB2, ATM, and mutations associated with hereditary

colorectal cancer syndromes These are summarized in Table 32– 3

It remains unclear whether or how lifestyle factors influence

PC risk among FPC family members Cigarette smoking is a well- established risk factor for sporadic PC A pooled case- control study that included 6507 cases with PC found that the prevalence of never smokers was 36.5% and 60.5% for ever smokers, with 2.6% miss-ing or other (Bosetti et al., 2012a) In the PACGENE study, 37% of affected members of FPC kindreds were never smokers, 47.1% were ever smokers, but smoking status was unknown in 14.9% (Petersen

et al., 2006) In comparison, a regional hospital- based Australian study found that 60.3% of 68 FPC patients were never smokers compared to

45.6% never smokers in 698 non- FPC patients (p = .0315) However,

there were no differences in alcohol intake in this study (Humphris

et al., 2014) It has been suggested that smoking may potentiate PC risk among predisposition gene mutation carriers (McWilliams et al., 2011), but to date, this has not been possible to evaluate more compre-hensively Data on other PC risk factors, such as diet or obesity, are too sparse among FPC to draw meaningful inferences Similarly, disease associations, such as diabetes or pancreatic cyst disease, have not been systematically studied in the context of FPC

Genetic Predisposition Among Sporadic PC Patients

Two distinct approaches have been taken to uncover genetic disposition in patients unselected for family history These include

pre-Table 32– 2 Family History and Estimated Risks of PC in Case- Control and Cohort Studies

Risk of PC in Family Members

Reference

International, PanScan Cohort Consortium

(1 case- control and 10 cohort studies), 1985– 2001

2005)

Case- control study designs reported unless otherwise specified.

Abbreviations: CI = confidence interval; RR = relative risk; SIR = standardized incidence ratio.

Table 32– 3 Associated Genes and Syndromes and Estimates of Risk of Developing Pancreatic Adenocarcinoma

Predisposition Syndrome

(Chaffee et al., 2016; Grant et al., 2015; Roberts et al., 2012)

BRCA1 17q21.31 Hereditary breast and

ovary, male breast, prostate

CDKN2A 9p21.3 Familial atypical mole

RR = 80.8 (95%

CI: 44.7, 146)

Potjer et al., 2015; Zhen et al., 2015)Mismatch

Hereditary non- polyposis colorectal cancer (Lynch syndrome)

Colorectum, endometrial, ovary, stomach, small bowel, urinary tract (ureter, renal pelvis) biliary, glioblastoma, skin (sebaceous)

No effect up to SIR = 8.6 (95% CI: 4.7, 15.7)

Kastrinos et al., 2009)

PALB2 16p12.2 Familial breast cancer Fanconi anemia, breast,

esophagus, prostate, stomach

Increased risk: not well defined

3/ 963/ 521

3.10.6

(Jones et al., 2009; Zhen

et al., 2015)

PRSS1

SPINK1

7q345q32

The probabilities of detecting a deleterious mutation in the predisposition genes shown were based upon studies that sequenced the entire gene in series of familial PC (FPC) patients Abbreviations: FPC = familial PC; OR = odds ratio; CI = confidence interval; RR = relative risk; SIR = standardized incidence ratio.

Cancer of the Pancreas 623

genome- wide association studies (GWAS) and next- generation

sequencing of patients

Genome- Wide Association Studies (GWAS) In the search

for common variants with low penetrance, the GWAS approach has

been the most- used strategy A series of these large case- control

stud-ies performed by the PanScan Consortium and the PC Case- Control

Consortium (PanC4) have been completed, with ongoing meta-

analysis By applying agnostic genetic scans of single nucleotide

poly-morphisms (SNPs), highly stringent thresholds of significance (i.e.,

P < 1 × 10– 7), and independent validation, these consortia have

identi-fied a roster of at least 13 novel SNPs in genes, as well as

confirm-ing known associations For example, the initial published GWAS

(PanScan I  in 2009), using 2000 case- control pairs, identified the

most consistent SNP association on chromosome 9q34.2 (in the ABO

blood group gene) (Amundadottir et al., 2009) In the following year,

PanScan II reported from an analysis of 3850 pairs associations of

SNPs on chromosome 1q32.1 (in NR5A2), 5p15.33 (in the CLPTM1L-

TERT gene region) and 13q22.1 (in a non- genic region between KLF5

and KLF12) (Petersen et al., 2010) PanScan III included 7683 cases

and 14,397 controls and found more risk loci: on chromosome 5p15.33

(an association that independently identified the CLPTM1L- TERT

gene region), on 7q23.2 (LINC- PINT), 16q23.1 (BCAR1), 13q12.2

(PDX1), 22q12.1 (ZNRF3), and a non- genic SNP on 8q24.1 (Wolpin

et al., 2014) In 2015, the PanC4 studied 4164 cases and 3792

con-trols, combined with PanScan data on 9925 cases and 11,569 concon-trols,

with replication and reported four new loci on chromosomes 17q24.3

(LINC00673), 2p14 (ETAA1), 7p14.1 (SUGCT), and 3q28 (TP63)

(Childs et al., 2015)

Genetic Analysis and Sequencing of  PC Patient

Series Among the first large unselected series reported was

sequencing for 39 known mutations in the cystic fibrosis

transmem-brane regulator (CFTR) gene of 949 PC patients by McWilliams et al

(2010), who reasoned from established data that the CFTR gene,

known to be associated with chronic pancreatitis, may increase risk

of PC Compared with data on 13,340 controls from a clinical

labora-tory database, they found that 5.3% carried a common CFTR mutation

versus 3.8% of controls, giving an OR = 1.40 (95% CI: 1.04, 1.89)

Among patients who were younger when their disease was diagnosed

(< 60  years), the carrier frequency was higher than in controls, and

the resulting OR increased to 1.82 (95% CI: 1.14, 2.94) McWilliams

et al (2011) also estimated the prevalence of deleterious mutations in

CDKN2A in a series of 1537 sporadic, unselected patients, and found

a prevalence of 0.6%

Whole exome sequencing of patients with and without family

his-tory of PC led to the identification of PALB2 and ATM (Jones et al.,

2009; Roberts et al., 2012), which had not been previously known to

increase risk of PC Moreover, the agnostic gene discovery efforts

(such as whole exome or whole genome sequencing) are

identify-ing novel genes (Roberts et  al., 2016), but the prevalence, gene by

gene, is quite low, and in many cases is unique to specific individuals

(Hu C et al., 2016) In summary, research to date has clearly revealed

the extensive genetic heterogeneity of the FPC phenotype In

addi-tion to expanding the catalog of genes, they provide an opportunity to

study the potential effect of genetic mutations on age at diagnosis and

the risk of developing other cancers This promises to be an area for

researchers to fill in the gaps of our knowledge about the

epidemiol-ogy, functional, and clinical implications

Interactive Effects Between Lifestyle and Genetic

Variants on PC Risk

Smoking and Genes

Genes from several pathways have been suggested to interact with

smoking to influence PC risk Hypothesized interactions include those

with genes that affect carcinogen metabolism, DNA repair, nicotine

dependence, oxidative stress, hormone metabolism, inflammation,

insulin secretion, and chromatin- remodeling (Ayaz et  al., 2008; Bartsch et al., 1998; Duell et al., 2002; Jiao et al., 2007; Liu et al., 2000; Vrana et  al., 2009) Laboratory evidence suggests that the major cigarette smoke carcinogen NNK can activate Cox2, EGFR, and Erk in pancreatic cancer cells and ductal cells (Askari et al., 2005; Blackford et al., 2009), thereby modifying proliferation and cell death (Edderkaoui and Thrower, 2013; Park et al., 2013) From an obser-vational epidemiologic standpoint, 14 PC case- control studies have investigated the potential interaction between smoking and polymor-phisms using a candidate gene approach In individual studies, an increased risk of PC has been reported between smoking status and

those with a minor allele for SNPs in XRCC2 (p = 0.02; involved in DNA repair) (Jiao et al., 2008); XRCC4 (p = 0.02) (Shen et al., 2015); CAPN10 (involved in insulin uptake) (Fong et al., 2010); adiponectin gene (p = 0.022; involved in metabolic and hormone processes) (Yang

et al., 2015); EPHX1 (p = 0.04; involved with metabolism); and NAT2 (p = 0.03; involved with detoxification of drugs and bioactivation of carcinogens) (Jang et al., 2012) Variants in the CYP1A2 (p = < 0.001; involved with metabolism) and NAT1 (p = 0.012; involved with detox-

ification of drugs and bioactivation of carcinogens) genes have also been observed to interact with heavy smoking among women (Suzuki

et al., 2008b) Both genes are involved in detoxifying and

bioactiva-tion of aromatic amines, and NAT1 rapid acetylator genotypes have

been identified (Hein et  al., 2000; Hirvonen, 1999) Genes in the

IGF axis regulate cell differentiation, proliferation, and migration,

and play an important part in initiating carcinogenesis (Gukovskaya

et al., 2002; Haber et al., 2004; Verma, 2005; Zatonski et al., 1993)

Two genes that encode components of the IGF- axis, IGF2R and IRS1,

interact with smoking, but the mechanisms are unknown (Dong et al.,

2012) There is an observed interaction between XPD (involved in DNA repair) and smoking where having a polymorphism in XPD

Asn312Asn and being an ever smoker (current and former) reduced

the risk of PC (OR = 0.42 [0.21, 083]; p = 0.01) (Li J et al., 2007)

Functionally, the Asn312Asn polymorphism may change the folding pattern of the resulting protein and corresponding function (Affatato

et al., 2004) There is a significant interaction between smoking and cytotoxic T lymphocyte- associated protein (CTLA- 4; involved in immune response) and the risk of PC where smokers with at least one

A allele have an increased risk of PC (p for interaction = 0.037) (Yang

et al., 2012) Contrary to these individual study results, no significant smoking– gene interactions were observed after multiple comparison correction in a discovery study using existing GWAS and smoking datasets and a combined sample of 2028 cases and 2109 controls (Tang et al., 2014a)

Obesity and Genes

Four studies have investigated the potential interaction for PC risk between obesity and genes responsible for regulating the balance of energy and tumor development and progression Nakao et al (2011)

studied the interaction with the IGF- 1 gene in a Japanese hospital-

based case- control study Weight was self- reported at baseline and recalled for 20 years of age Those with a minor allele for rs574214 and BMI ≥ 25 were at an increased risk of PC In a previous study (Lin et al., 2011), this polymorphism was found to be associated with risk of PC and diabetes mellitus, but not BMI Genetic variation in

FTO has been associated with obesity (Hinney et al., 2007; Scott et al.,

2007)  and is regulated by fasting and feeding status (Gerken et  al., 2007) and negatively regulates lipid metabolism (Klöting et al., 2008)

Those with the FTO polymorphism and BMI < 25 have a reduced risk

of PC, and those with BMI ≥ 25 have an increased risk (Tang et al., 2011) The mechanistic relationship with BMI is currently not known

ADIPOQ codes for adipocyte- secreted hormone and has a low

fre-quency of the homozygous variant in the study population However, a

significant interaction with BMI < 25 was observed (p = 0.005) (Tang

et al., 2011) Among 172 cases and 181 controls in a Chinese tion, 6 SNPs were analyzed in the adiponectin gene (Yang et al., 2015) One was significantly associated with a decreased risk (OR  =  0.66; 95% CI:  0.47, 0.93) of PC, while another was associated with an increased risk (OR = 1.42; 95% CI: 1.04, 1.94) Pooling information

from multiple sites into a discovery study using existing GWAS and

smoking data sets for 2028 cases and 2109 controls, no significant

obesity– gene interactions were observed after multiple comparison

correction (Tang et al., 2014b)

Diet and Genes

Dietary intake has been proposed to interact with genes involved with

metabolism, antioxidant defense, and DNA repair (Cullen et al., 2003;

Hauptmann and Cadenas, 1997; Jansen et al., 2013b; Ough et al., 2004;

Suzuki et al., 2008b; Tang et al., 2010; Weydert et al., 2003; Yamauchi,

2007; Zhang et al., 2011) Alcohol and its major metabolite,

acetalde-hyde, are categorized as carcinogens (Baan et  al., 2009), which has

been hypothesized to interact with genes that play a role in tumor

development and progression (Appleman et al., 2000; Mohelnikova-

Duchonova et al., 2010; Scheipers and Reiser, 1998; Yang et al., 2015,

2012) However, most of the reports based on potential interactions

between diet or alcohol intake and candidate genes have not been

rep-licated sufficiently to warrant conclusions about clear interaction

Molecular Factors

Telomere Length

Telomeres, the complex chromosomal end- capping structures

con-sisting of repetitive nucleotide sequences (TTAGGG) and the

telo-mere binding- protein complex called shelterin protect chromosomal

DNA from degradation, and prevent end- to- end joining and aberrant

recombination during cell division (Blackburn et al., 2015) Telomeres

thereby preserve the integrity and stability of genomic material

Peripheral blood leukocyte telomere length (a marker of host telomere

status) is approximately 10−16 kilobase pairs long at birth (Drury

et al., 2015; Okuda et al., 2002), but progressively shortens by about

50– 100 base pairs with each cell division, due to incomplete

repli-cation (Collins and Mitchell, 2002) Because telomere length

short-ens with each cell division, telomere length correlates inversely with

age and may serve as a marker of biological aging (Blackburn et al.,

2015) Factors associated with chronic inflammation and oxidative

stress, such as cigarette smoking, diabetes, and excess body weight

are known contributors to the telomere length- shortening process

(Blackburn et al., 2015; Harley, 1997; Jennings et al., 2000)

Telomere attrition has been linked to increased risk of many

can-cer types Five studies examined the association between telomere

length and PC with mixed findings (Bao et  al., 2016; Campa et  al.,

2014; Lynch et al., 2013; Skinner et al., 2012; Zhang et al., 2016) The

first was a clinic- based, case- control study of 499 incident PC cases

and 963 non- cancer control patients (Skinner et al., 2012) This study

reported a “U- shaped” association between peripheral blood leukocyte

telomere length and PC, such that individuals with extremely short

(lower 1%) or extremely long (top 10%) telomere length had increased

PC risk (Skinner et al., 2012) Among those within the 1st and 90th

percentile distribution of telomere length, short telomere length was

associated with an increased risk of PC (Skinner et  al., 2012) The

second was a nested case- control study among Finnish male

smok-ers that consisted of 193 cases and 660 controls (Lynch et al., 2013)

This prospective study reported that long leukocyte telomere length

was associated with increased risk of PC in a dose– response manner

(OR = 1.57; 95% CI: 1.01, 2.43; longest vs shortest telomere length

quartile, P trend = 0.0007) (Lynch et al., 2013) The third report was

from a nested case- control study of 331 PC cases and 331 controls

(Campa et al., 2014) It reported a modest increase in PC risk among

individuals with long leukocyte telomere lengths (OR  =  1.13; 95%

CI: 1.01, 1.27, continuous variable) Further analysis by cubic spline

regression found that the association was non- linear (P for non-

linearity  =  0.02) (Campa et  al., 2014) One of the two most recent

studies was a pooled analysis of five prospective cohorts involving

386 cases and 896 controls (Bao et al., 2016) The results from this

study showed that short pre- diagnostic leukocyte telomere length was

associated with increased PC in a dose- dependent manner (OR = 1.72;

95% CI:  1.07, 2.78; shortest vs longest telomere length quintile, P

trend = 0.048; P for non- linearity > 0.05) (Bao et al., 2016) The final

report was from a population- based, nested case- control study of 900 cases and 900 controls in Liaoning, China, and showed a “U- shaped” association between leukocytes telomere length and PC risk (Zhang

et al., 2016) Using the third quartile as the reference category, they found increased risk for those in the shortest telomere length quartile (OR  =  3.10; 95% CI:  1.84, 5.21) and those in the longest telomere length quartile (OR = 1.49; 95% CI: 1.11, 2.00) (Zhang et al., 2016) These studies differed in many ways, including differences in the stud-ied populations, variations in the measurement of telomere length, and variations in the time between blood collection and diagnosis of PC, which may explain the inconsistent finding Tumor- based studies show that pancreatic tumors have shortened telomere lengths, suggesting that telomere shortening might be central to pancreatic tumorigenesis (Bardeesy and DePinho, 2002; Hashimoto et al., 2008)

Epigenetic Factors

Our knowledge of the molecular structure of DNA beyond the gene sequence and how these mechanisms interact with the genetic struc-ture is continuously expanding, but is currently limited The genome

is generally considered to represent inherited disease ity and has been thought to be fairly stable over time with the help

susceptibil-of efficient DNA repair mechanisms We also know that additional

mechanisms, collectively referred to as epigenetics, can be a means

by which non- inherited effects (i.e., due to environmental insults) can dynamically affect the expression (not coding) of the genetic sequence (Heard and Martienssen, 2014; Nagy and Turecki, 2015) Examples of epigenetic mechanisms include DNA methylation, his-tone modification, and alterations in microRNA and non- coding RNA Epigenetic changes are thought to represent early influences

in carcinogenesis (Verma et  al., 2014), with external factors such

as diet, drugs, or infectious agents generating interactive effects to modify a person’s cancer risk (Costa, 2010; Hou et al., 2011; Verma, 2013) The potential ability to reverse or prevent harmful changes

to these epigenetic mechanisms through behavior modification or other interventions is what makes epigenetic research attractive from

a public health and disease prevention standpoint Epigenetic ers and patterns are tissue- specific, although many markers have been identified as playing an important role in different cancer types

mark-To date, most research relevant to epidemiology of PC has been in DNA methylation of promoter sites in tumor suppressor genes (i.e.,

APC, BRCA1, and CDKN2A) identified as the most common regions

in human pancreatic neoplasms (Guo et  al., 2014) Genome- wide DNA methylation patterns in PDAC tissues suggest methylated genes involved in TGF- beta, WNT, integrin signaling, cell adhesion, and stellate cell activation (Dutruel et al., 2014; Nones et al., 2014; Vincent et al., 2014)

Other Risk Factors

perturba-non referred to as dysbiosis or dysbacteriosis (Zambirinis et al., 2014)

However, the role of shifts in the human microbiome composition as

a causative factor of PC has yet to be thoroughly elucidated It is well recognized that exposure to certain environmental agents can alter the composition of the gut microbiome and disrupt mucosal permeability These bacteria or their products (e.g., lipopolysaccharides and bacte-rial DNA) can then evoke local and systemic inflammatory responses (Yan and Schnabl, 2012; Zambirinis et al., 2014) For example, stud-ies have shown that excess consumption of alcohol can lead to dis-ruption of the intestinal mucosal barrier, accompanied by increased

Cancer of the Pancreas 625

proliferation of pathogenic bacteria in the gut (Purohit et  al., 2008;

Yan and Schnabl, 2012; Zambirinis et al., 2014) Such gut microbes

can migrate to the pancreas via reflux of intestinal contents into the

main pancreatic duct

Studies also indicate that changes in oral microbiota because of

poor hygiene may contribute to PC In this context, periodontal disease

and tooth loss in adults, caused by bacterial infection and periodontal

inflammation, have been associated with increased PC risk (Michaud

et al., 2007; Stolzenberg- Solomon et al., 2003) Moreover, Farrell and

colleagues examined microbiome profiles in human saliva and found

specific oral microbiota profiles associated with increase of PC and

pancreatitis (Farrell et al., 2012) The study found 31 bacterial species

that were significantly higher in saliva pellets of PC patients compared

to healthy controls, and 25 bacteria species that were significantly

lower in saliva of PC patients than controls The study further showed

that the presence of two bacteria (Streptococcus mitis and Neisseria

elongata), when combined, can distinguish PC patients from healthy

controls with 96% sensitivity and 82% specificity Although this was

a small study consisting of 10 cases and 10 controls, the findings were

validated in an independent sample of 28 PC cases and 28 controls

(Farrell et al., 2012) The small samples in these studies limit

gener-alizability, but provide intriguing clues regarding the role of

micro-biomes in PC Larger studies are needed to refine and validate such

biological markers Investigations into the prognostic relevance of the

oral cavity and gut microbiome may provide new insights for

preven-tive interventions

Inflammation

Experimental, clinical, and epidemiologic evidence show a clear link

between inflammation and PC The association appears to be

medi-ated by a “field effect,” such that PC tends to be fostered in a pro-

inflammatory microenvironment with high concentration of growth

factors, cytokines, and chemokines, and activation of inflammatory

pathways such as cyclooxygenase- 2 (COX- 2), and nuclear factor

kappa B (NF- kB) (Garcea et al., 2005; Greer and Whitcomb, 2009)

The epidemiologic data show a strong association between PC and

pro- inflammatory stimuli, such as tobacco smoking, and

inflamma-tory conditions, such as chronic pancreatitis, diabetes, and obesity

(Lowenfels and Maisonneuve, 2006) A  chronic inflammatory state,

particularly chronic inflammation of the pancreas (pancreatitis), favors

pancreatic tumorigenesis in that it causes a cycle of repeated DNA

damage and repair, leading to increased frequency of cell division,

which, in turn, increases the chances of cellular aberrations, malignant

cell formation, and ultimately, frank malignancy (Garcea et al., 2005;

Greer and Whitcomb, 2009) Individuals with hereditary pancreatitis,

a rare form of chronic pancreatitis that is caused by mutations in the

cationic trypsinogen gene (PRSS1), have a 40% risk of developing

PC by age 70 years (Lowenfels et al., 1997), while individuals with

non- genetic forms of chronic pancreatitis have a 13- fold higher risk

of PC compared to healthy populations (Yadav and Lowenfels, 2013)

Cigarette smoking, diabetes, and obesity contribute to PC by

promot-ing a persistent state of systemic inflammation, which has been

associ-ated with DNA adducts formation, altered gene expression, oncogene

activation, and inhibition of apoptosis (Garcea et al., 2005; Greer and

Whitcomb, 2009) Obesity accounts for ~12% of PCs (Parkin et al.,

2011) and is associated with increased circulating levels of the pro-

inflammatory adipokines, resistin and leptin, and lower levels of the

anti- inflammatory adipokine, adiponectin (Kowalska et  al., 2008)

Diabetes, a well- known risk factor of PC, has been associated with

higher circulating levels of inflammatory cytokines such as C- reactive

protein (CRP) and interleukin 6 (IL- 6) (Pradhan et al., 2001), while

cigarette smoking, which accounts for ~20%– 25% of PCs, is

associ-ated with elevassoci-ated levels of many inflammatory cytokines including

CRP, IL- 6, and intercellular adhesion molecule- 1 (ICAM- 1) (Levitzky

et al., 2008) Thus, inflammation appears to be a common underlying

mechanism through which these factors influence PC risk

Although inconclusive, emerging evidence suggest that diet- derived

inflammation, mainly from consumption of red meat, high- fat and high-

calorie foods, and sugar- sweetened beverages, increases PC risk (Antwi

et al., 2016) However, as with other environmental exposures such as

smoking, the effect of diet- derived inflammation may be dependent on inter- individual variation in metabolism of pro- carcinogens, such that the effect might be less pronounced in persons with efficient metabolic detoxification of carcinogens than in those with inefficient detoxifica-tion of carcinogens Studies also show that COX- 2 inhibitors can delay the development of the PC precursor lesion, pancreatic intraepithelial neoplasia (PanIN) (Hruban et al., 2008), while omega- 3 fatty acids can suppress the growth of pancreatic tumors (Hering et al., 2007), further supporting a role of inflammation in PC

Infectious Agents

Pathogenic infections are estimated to account for ~10%– 20% of all cancers worldwide (Chang and Parsonnet, 2010) Although several studies suggest the possibility of an association with an infectious agent, no agent has been confirmed as causative for PC Infection with hepatitis B and C has been associated with PC in some studies (Fiorino et al., 2013; Majumder et al., 2014) However, there is stron-

ger evidence for a plausible role of Helicobacter pylori (H.  pylori),

an infectious organism that can survive the acidic conditions of the gastric lumen and is associated with increased risk of gastric can-cer (Lowenfels and Maisonneuve, 2006) A  1998 study by Raderer and colleagues was the first to report an epidemiologic association

between H. pylori infection and PC (Raderer et al., 1998) This case-

control study of 27 cases and 27 controls reported a 2- fold increase in

risk among H. pylori positive individuals (OR = 2.1; 95% CI: 1.09,

4.05) (Raderer et  al., 1998) In 2001, Stolzenberg- Solomon et  al

measured seroprevalence of H. pylori among 121 cases and 226

con-trols in a nested case- control study within the Alpha- Tocopherol, Beta- Carotene Cancer Prevention Study, a Prospective Cohort study

of Finnish Male Smokers (Stolzenberg- Solomon et al., 2001) These

investigators observed increased risk among individuals with H. pylori

infection (1.87; 95% CI: 1.05, 3.34), with a much stronger association among those with the cytotoxin- associated gene- A- positive (CagA+)

H.  pylori strains (2.01; 95% CI:  1.09, 3.70) (Stolzenberg- Solomon

et al., 2001) Since these studies, at least four additional studies have investigated this association, and none found a statistically significant association (de Martel et al., 2008b; Lindkvist et al., 2008; Risch et al., 2010; Schulte et al., 2015), although two reported a positive relation-ship (Lindkvist et al., 2008; Risch et al., 2010) A 2011 meta- analysis

of the six studies, including five of the studies cited earlier (de Martel

et al., 2008a; Lindkvist et al., 2008; Raderer et al., 1998; Risch et al., 2010; Stolzenberg- Solomon et  al., 2001)  reported a 38% increased

risk among H. pylori seropositive individuals (OR 1.38; 95% CI: 1.08,

1.75) (Trikudanathan et al., 2011) In a more recent meta- analysis of

nine studies, increased PC risk was observed among H. pylori positive

individuals (1.47; 95% CI: 1.22, 1.77) (Xiao et al., 2013) When ses were stratified by geographic regions, an excess risk of 56% was observed in studies conducted in Europe, 2- fold excess risk among studies conducted in East Asia, and a non- significant 17% excess risk among studies conducted in North America (Xiao et al., 2013) Therefore, while the evidence is not entirely consistent, it is plausible

analy-that H. pylori may be involved in pancreatic carcinogenesis, possibly

through interaction with other host factors such as ABO blood type (Risch et al., 2010)

a 26% and 38% lower PC risk among individuals with self- reported history of hay fever and allergy to animals, respectively, which the investigators suggested indicates a potential role of immunoglob-ulin E (an allergic response antibody) in PC (Olson et  al., 2013) The results further showed a suggestion toward lower risk among those with a history of any allergy (OR = 0.79; 95% CI: 62, 1.00),

excluding asthma (Olson et al., 2013) A thorough review of 11

pub-lications on allergy history and PC risk showed an overall inverse

association, with ORs ranging between 0.39 (95% CI: 0.22, 0.71)

and 0.77 (95% CI: 0.63, 0.95) (Olson, 2012) In general, the

epide-miologic evidence indicates an inverse association between history

of respiratory allergies (particularly hay fever) and PC, but does not

support an association with asthma, possibly because of the

com-plex and varied etiology of asthma (Olson, 2012)

Previous Surgery

Partial gastrectomy and cholecystectomy have been reported to be

associated with PC risk Among 15 studies that examined

associa-tion between gastrectomy and PC risk, 11 reported elevated risk

among post- gastrectomy patients, with five showing statistically

significant increased risk (Olson, 2012) In agreement with these

results, a multicenter case- control study of 4717 cases and 9374

con-trols reported increased PC risk among post- gastrectomy patients

(OR  =  1.53; 95% CI:  1.15, 2.03), but the association was limited

to gastric resections occurring within 2  years before diagnosis of

PC (OR = 6.18; 95% CI: 1.82, 20.96) (Bosetti et al., 2013) Other

studies have suggested that cholecystectomy (surgical removal of

gallbladder) may be associated with PC, but the evidence is

incon-clusive While some studies have reported elevated PC risk after

cholecystectomy (Silverman et al., 1999; Zhang et al., 2014), many

show no association (Goldacre et  al., 2005; Schernhammer et  al.,

2002; Ye et al., 2001) The link between cholecystectomy and PC

is based on the notion that removal of the gallbladder increases

cir-culating levels of cholecystokinin (a digestive hormone thought to

promotes PC growth), and enhances degradation of bile salt into

bile acid and other metabolites that increases cell proliferation (Ura

et al., 1986) However, this is not completely supported by the

exist-ing epidemiologic data

NEUROENDOCRINE/ ISLET CELL TUMORS

Pancreatic neuroendocrine tumors (PNETs) are rare tumors of

neu-roendocrine origin arising in the islet cells of the pancreas PNETs

comprise approximately 3% of PC They are typically more indolent

than pancreatic adenocarcinomas, although some PNETS can

prog-ress more rapidly to malignancy (Amin and Kim, 2016; Halfdanarson

et  al., 2008b) However, the reasons for progression are unclear

Population- based studies have reported wide estimates of 5- year

over-all survival, ranging from 15% to 60%, based on disease stage and

treatment center (Bilimoria et al., 2007; Strosberg et al., 2011) Based

upon SEER data, the US crude annual incidence per 100,000 for all

ages is 0.1 in females and 0.2 in males, and incidence increases with

age (Halfdanarson et al., 2008a) Incidence of PNET has risen sharply

in the United States over the past two decades; this may be related

to improvements in imaging and increased screening of the pancreas

(Yao et  al., 2008) In 1973, the incidence was 0.17/ 100,000; it was

0.47/ 100,000 by 2007 (Halfdanarson et al., 2008a; Hallet et al., 2015;

Lawrence et al., 2011)

PNETs are divided into two main types: functional (the tumors

produce excess hormones of different varieties), often associated

with clinical syndromes; and non- functional (non- secretory),

which are often metastatic at diagnosis (Halfdanarson et  al.,

2008b) Functional PNETs constitute 10%– 40% of all PNETs

and are distinguished by the main hormones they secrete (insulin,

gastrin, glucago- vaso- intestinal peptide, or associations with

clin-ical conditions such as hyperinsulinemia, peptic ulcer disease)

Inherited syndromes such as multiple endocrine neoplasia type

1 (MEN1) (Lakhani et  al., 2007) or von Hippel- Lindau disease

(Lonser et al., 2003) can feature functional PNETs, although most

PNETs are thought to be sporadic, not associated with any known

hereditary or familial syndromes The mean age at diagnosis of

patients with PNETs is 58.5 + 14.9, with slightly older mean age

for those with non- functional tumors (58.8 + 14.7) compared to

those with functional tumors (55.2 + 16.3) (Strosberg et al., 2011)

For unselected PNETs, a family history of other gastrointestinal neuroendocrine tumors and other cancers in first- degree relatives was suggested (Hassan et al., 2008b; Hiripi et al., 2009) One study reported that patients with non- functional PNETs were more likely

to have a family history of pancreatic adenocarcinoma, but no mal comparison was done (Gullo et  al., 2003) In a study of 309 cases and 602 controls by Halfdanarson et al (2014), PNET cases were more likely to report a family member with sarcoma, PNET, cancer of the gallbladder, stomach, and ovary, but in the absence

for-of any association for-of PNET with a personal history for-of pancreatic adenocarcinoma

There are conflicting data on the association between smoking and alcohol use with risk of neuroendocrine tumors (Hassan et al., 2008a; Kaerlev et  al., 2002) In patients with five different types of neuro-endocrine tumors, no association was observed between smoking or alcohol consumption and the development of neuroendocrine tumors (Hassan et al., 2008a) An association was suggested between smoking and the risk of neuroendocrine malignancies in two population- based studies that included a limited sample of patients with small bowel neuroendocrine tumors (carcinoid tumors) (Hassan et  al., 2008a; Kaerlev et  al., 2002) In the study by Halfdanarson et  al., slightly more cases were likely to have a personal history of tobacco use than controls, but this difference was not statistically significant However,

alcohol use was less common among cases (54% vs 67%, P < 0.001)

(Halfdanarson et al., 2014)

Several studies found associations with a personal history of diabetes (Feldman et al., 1975; Halfdanarson et al., 2014; Hassan et al., 2008a; Kaerlev et al., 2002) Interestingly, Hassan et al reported that patients with PNETs were more likely to have been diagnosed with diabetes

in the year leading up to the PNET diagnosis (Hassan et al., 2008a) Halfdanarson et al found that diabetes was more commonly reported

by cases than controls (19% vs 11%, P < 0.001) (Halfdanarson et al.,

2014)  and also suggested that the association between recent- onset diabetes and pancreatic adenocarcinoma that they observed may rep-resent the effect of proteins secreted by the tumor on beta- cell function and insulin resistance

In summary, pancreatic neuroendocrine tumors (functional and non- functional), constitute a small fraction of PCs, but are increasing in incidence, possibly due to improved diagnostic procedures Other than rare genetic syndromes that feature functional PNETs, studies of risk factors are limited and do not conclusively implicate any factors that are associated with pancreatic adenocarcinoma

CURRENT STUDY LIMITATIONS

All epidemiologic study designs are subject to limitations and biases that affect the interpretation and generalizability of reported results For example, differential misclassification and recall of dietary pat-terns between cases and controls could contribute to biased risk esti-mates Comorbidities associated with smoking, obesity, and alcohol intake affect selection of cases with these exposures For many of the exposures discussed here, there are social stigmas associated with high levels of consumption that may influence how a partici-pant completes survey questions In retrospective population- based studies of rapidly fatal disease, bias can occur due to the demise of eligible cases with a higher proportion of later- stage disease, pos-sibly resulting in non- random non- response In prospective studies, the rarity of PC limits the number of potential cases seen during follow- up Both these situations lead to a reduced power to detect associations Moreover, gene– environment interaction studies are often criticized for being underpowered, and it has been suggested that the associations seen are often false positives and cannot be replicated

OPPORTUNITIES FOR PREVENTION

Smoking prevention and cessation, maintaining a healthy body weight, and consuming a well- balanced diet present the best opportunities for

Cancer of the Pancreas 627

the prevention of PC The 2014 US Surgeon General’s Report

esti-mates that ~10%– 14% of PC deaths in the United States are

attrib-utable to tobacco smoking (Reports of the Surgeon General, 2014),

whereas European studies suggest that some 20%– 25% of PC cases

are attributable to smoking (Lowenfels and Maisonneuve, 2006)

While current smoking, smoking duration, and smoking dose are all

strongly associated with increased PC risk, many studies have shown

that the PC risk associated with smoking regresses to null after about

15 years of smoking cessation, a clear benefit of smoking cessation

(Bosetti et al., 2012a; Fuchs et al., 1996; Iodice et al., 2008; Lynch

et al., 2009) The risk fraction attributable to excess adiposity is not

completely clear; Calle and Kass (2004) attribute 27% of all PCs in

the United States to overweight and obesity, while Parkin et al (2011)

attribute 12% of PCs in the United Kingdom to overweight and

obe-sity It is, nonetheless, plausible that at least 25% of PCs in the United

States can be prevented through the elimination of smoking and

long- term maintenance of a healthy body weight Although the role

of diet in PC is not well characterized, there is evidence indicating

that excessive alcohol consumption and increased intake of saturated

fat, red meat, and processed meat are associated with increased risk

(Michaud et  al., 2010; Nöthlings et  al., 2005; Stolzenberg- Solomon

et  al., 2002b) Limiting intake of these foods may therefore reduce

risk Long- standing diabetes has been associated with increased PC

risk For diabetics, better glycemic control involving dietary

modifica-tion and regular physical activity may reduce insulin resistance and its

attendant hyperinsulinemia, which are associated with increased PC

risk (Bao et al., 2011; Li, 2012) While inconclusive, there are

sug-gestions that smokeless tobacco use and secondhand tobacco smoke

may increase risk (Alguacil and Silverman, 2004; Boffetta et al., 2008;

Vrieling et al., 2010); therefore, avoiding these exposures may reduce

individual risk

Future Directions

The development of robust risk prediction models that combine

genomic and risk- factor information is central to the identification

of individuals at high risk of developing PC for timely

preven-tive interventions, including chemoprevention and early detection

Epidemiologic and risk factors data as reviewed in this chapter

pro-vide important clues for risk stratification With the increasing

obe-sity epidemic, especially among youth, and the strong association

between obesity and PC, it can be expected that obesity- related PC

rates will increase over the coming decades Dietary results

regard-ing PC risk have largely been inconsistent, with the potential

excep-tion of certain fatty acids and well- done red meat Dietary data have

been fraught with measurement error, and often a large percentage of

the data is missing for participants Technology provides a potential

solution for a more accurate assessment of dietary habits and a better

understanding of how diet influences PC risk, but not for

reconstruct-ing past dietary exposures As smokreconstruct-ing rates continue to decrease,

cigarette smoking– related PC will also decrease The role that e-

cig-arettes may play in PC has yet to be determined, while the effect of

environmental exposure, especially in early childhood, needs further

exploration Alcohol appears to be a risk for PC only among those

in the heaviest consumption category; however, remaining potential

confounding of smoking behaviors, even after adjustment, is a

con-cern Genetic data can assist in identifying individuals at high risk

of developing PC; new statistical and epidemiological methods or

processes are needed to pinpoint the responsible genetic variants or

regions and their interaction with modifiable risk factors Large and

well- designed prospective studies investigating the potential

chemo-preventive effects of metformin, statins, and aspirin, both

indepen-dently and in combination, may inform strategies for addressing the

rising incidence of PC

Epidemiologic research is also needed to increase precision in

methods for identifying and targeting individuals for earlier detection

Evaluation of biomarkers for early detection of PC will require

partic-ular focus on designs to deliver highly specific and sensitive assays

New research directions involving advanced technology and better

understanding of the biology of pancreatic carcinogenesis in molecular

and cell biology, assays using proteomics, microarray analysis of gene expression profiles, and detection of circulating tumor- secreted DNA and exosomes hold promise

views Cancer Epidemiol Biomarkers Prev, 13(1), 55– 58.

American Cancer Society 2013 Cancer Facts & Figures Atlanta, GA:

American Cancer Society

American Cancer Society 2016 Cancer Facts & Figures Atlanta, GA:

American Cancer Society

Amin S, and Kim MK 2016 Islet cell tumors of the pancreas Gastroenterol Clin North Am, 45(1), 83– 100 PMID: 26895682

Amundadottir LT 2016 Pancreatic cancer genetics Int J Biol Sci, 12(3), 314–

pancreatic cancer JNCI, 94(15), 1168– 1171 PMID: 12165642.

Anderson KE, Mack TM, and Silverman DT 2006 Cancer of the pancreas

In: Schottenfeld D, and Fraumeni FJ (Eds.), Cancer Epidemiology and Prevention (3rd ed., pp 721– 763) New York: Oxford University Press.Anderson KE, Mongin SJ, Sinha R, et al 2012 Pancreatic cancer risk: asso-ciations with meat- derived carcinogen intake in the Prostate, Lung,

Colorectal, and Ovarian Cancer Screening Trial (PLCO) cohort Mol Carcinog, 51(1), 128– 137 PMCID: PMC3516181

Andreotti G, and Silverman DT 2012 Occupational risk factors and

pancre-atic cancer:  a review of recent findings Mol Carcinog, 51(1), 98– 108

PMID: 22162234

Antonello D, Gobbo S, Corbo V, et al 2009 Update on the molecular genesis of pancreatic tumors other than common ductal adenocarcinoma

patho-Pancreatology, 9(1- 2), 25– 33 PMID: 19077452

Antwi SO, Eckert EC, Sabaque CV, et  al 2015 Exposure to

environmen-tal chemicals and heavy meenvironmen-tals, and risk of pancreatic cancer Cancer Causes & Control, 26(11), 1583– 1591 PMCID: PMC4624268

Antwi SO, Oberg AL, Shivappa N, et al 2016 Pancreatic cancer: associations

of inflammatory potential of diet, cigarette smoking and long- standing

diabetes Carcinogenesis PMID: 26905587 PMCID: PMC4843052.

Appleman LJ, Berezovskaya A, Grass I, and Boussiotis VA 2000 CD28 costimulation mediates T cell expansion via IL- 2- independent and IL-

2- dependent regulation of cell cycle progression J Immunol, 164(1),

144– 151

Arnold LD, Patel AV, Yan Y, et  al 2009 Are racial disparities in pancreatic

cancer explained by smoking and overweight/ obesity? Cancer Epidemiol Biomarkers Prev, 18(9), 2397– 2405 PMCID: PMC3630792

Askari MDF, Tsao M- S, and Schuller HM 2005 The tobacco- specific

proliferation of immortalized human pancreatic duct epithelia through

beta- adrenergic transactivation of EGF receptors J Cancer Res Clin Oncol, 131, 639– 648 PMID: 16091975

Aune D, Greenwood DC, Chan DS, et al 2012 Body mass index, abdominal fatness and pancreatic cancer risk:  a systematic review and non- linear

dose- response meta- analysis of prospective studies Ann Oncol, 23(4),

Ayaz L, Ercan B, Dirlik M, Atik U, and Tamer L 2008 The association between

N- acetyltransferase 2 gene polymorphisms and pancreatic cancer Cell Biochem Funct, 26(3), 329– 333 PMID: 18027363

Baan R, Grosse Y, Straif K, et al 2009 A review of human carcinogens: Part

F:  chemical agents and related occupations Lancet Oncol, 10(12),

1143– 1144

Banke MG, Mulvihill JJ, and Aston CE 2000 Inheritance of pancreatic cancer

in pancreatic cancer- prone families Med Clin North Am, 84(3), 677– 690,

x- xi PMID: 10872424

Bao B, Wang Z, Li Y, et al 2011 The complexities of obesity and diabetes with

the development and progression of pancreatic cancer Biochim Biophys

Bao Y, and Michaud DS 2008 Physical activity and pancreatic cancer risk: a

systematic review Cancer Epidemiol Biomarkers Prev, 17(10), 2671–

2682 PMCID: PMC4896495

Bao Y, Prescott J, Yuan C, et al 2016 Leucocyte telomere length, genetic

vari-ants at the TERT gene region and risk of pancreatic cancer Gut, 1– 7

doi:10.1136/ gutjnl- 2016- 312510 PMID: 27797938

Bardeesy N, and DePinho RA 2002 Pancreatic cancer biology and genetics

Nat Rev Cancer, 2(12), 897– 909 PMID: 12459728

Bartsch H, Malaveille C, Lowenfels AB, et al 1998 Genetic polymorphism

H:quinone oxidoreductase in relation to malignant and benign pancreatic

disease risk The International Pancreatic Disease Study Group Eur J

Cancer Prev, 7(3), 215– 223 PMID: 9696930

Basturk O, Hong SM, Wood LD, et al 2015 A revised classification system

and recommendations from the Baltimore Consensus Meeting for

neo-plastic precursor lesions in the Pancreas Am J Surg Pathol, 39(12), 1730–

1741 PMCID: PMC4646710

Batabyal P, Vander Hoorn S, Christophi C, and Nikfarjam M 2014 Association

of diabetes mellitus and pancreatic adenocarcinoma: a meta- analysis of

88 studies Ann Surg Oncol, 21(7), 2453– 2462 PMID: 24609291.

Ben Q, Xu M, Ning X, et al 2011 Diabetes mellitus and risk of pancreatic

can-cer: a meta- analysis of cohort studies Eur J Cancer, 47(13), 1928– 1937

PMID: 21458985

Berrington de Gonzalez A, Spencer EA, Bueno- de- Mesquita HB, et al 2006

Anthropometry, physical activity, and the risk of pancreatic cancer in the

European Prospective Investigation into Cancer and Nutrition Cancer

Epidemiol Biomarkers Prev, 15(5), 879– 885 PMID: 16702364

Berrington de Gonzalez A, Sweetland S, and Spencer E 2003 A meta- analysis

of obesity and the risk of pancreatic cancer Br J Cancer, 89(3), 519– 523

PMCID: PMC2394383

Bertuccio P, La Vecchia C, Silverman DT, et al 2011 Cigar and pipe

smok-ing, smokeless tobacco use and pancreatic cancer: an analysis from the

International Pancreatic Cancer Case- Control Consortium (PanC4) Ann

Oncol, 22(6), 1420– 1426

Biankin AV, Waddell N, Kassahn KS, et al 2012 Pancreatic cancer genomes

reveal aberrations in axon guidance pathway genes Nature, 491(7424),

399– 405 PMCID: PMC3530898

Bilimoria KY, Bentrem DJ, Merkow RP, et al 2007 Application of the

pan-creatic adenocarcinoma staging system to panpan-creatic neuroendocrine

tumors J Am Coll Surg, 205(4), 558– 563 PMID: 17903729.

Blackburn EH, Epel ES, and Lin J 2015 Human telomere biology: a

contribu-tory and interactive factor in aging, disease risks, and protection Science,

350(6265), 1193– 1198 PMID: 26785477

Blackford A, Parmigiani G, Kensler TW, et al 2009 Genetic mutations

associ-ated with Cigarette smoking in pancreatic cancer Cancer Res, 69, 3681–

3688 PMID: 19351817

Bocci G, Fioravanti A, Orlandi P, et al 2005 Fluvastatin synergistically enhances

the antiproliferative effect of gemcitabine in human pancreatic cancer

MIAPaCa- 2 cells Br J Cancer, 93(3), 319– 330 PMCID: PMC2361561.

Bodmer M, Becker C, Meier C, Jick SS, and Meier CR 2012 Use of

antidia-betic agents and the risk of pancreatic cancer: a case- control analysis Am

J Gastroenterol, 107(4), 620– 626 PMID: 22290402

Boffetta P, Hecht S, Gray N, Gupta P, and Straif K 2008 Smokeless tobacco

and cancer Lancet Oncol, 9(7), 667– 675 PMID: 18598931.

Bonovas S, Filioussi K, and Sitaras NM 2008 Statins are not associated with a

reduced risk of pancreatic cancer at the population level, when taken at low

doses for managing hypercholesterolemia: evidence from a meta- analysis

of 12 studies Am J Gastroenterol, 103(10), 2646– 2651 PMID: 18684187.

Bosetti C, Lucenteforte E, Bracci PM, et al 2013 Ulcer, gastric surgery and

pancreatic cancer risk:  an analysis from the International Pancreatic

Cancer Case- Control Consortium (PanC4) Ann Oncol, 24(11), 2903–

2910 PMCID: PMC3811904

Bosetti C, Lucenteforte E, Silverman DT, et al 2012a Cigarette smoking and

pancreatic cancer: an analysis from the International Pancreatic Cancer

Case- Control Consortium (Panc4) Ann Oncol, 23(7), 1880– 1888.

Bosetti C, Rosato V, Gallus S, Cuzick J, and La Vecchia C 2012b Aspirin and

cancer risk: a quantitative review to 2011 Ann Oncol, 23(6), 1403– 1415

PMID: 22517822

Bosetti C, Rosato V, Li D, et  al 2014 Diabetes, antidiabetic medications,

and pancreatic cancer risk: an analysis from the International Pancreatic

PMCID: PMC4176453

Brenner DR, Wozniak MB, Feyt C, et al 2014 Physical activity and risk of pancreatic cancer in a central European multicenter case– control study

Cancer Causes & Control, 25(6), 669– 681

Bryant KL, Mancias JD, Kimmelman AC, and Der CJ 2014 KRAS: feeding

pancreatic cancer proliferation Trends Biochem Sci, 39(2), 91– 100.

Calle EE, and Kaaks R 2004 Overweight, obesity and

cancer: epidemiologi-cal evidence and proposed mechanisms Nat Rev Cancer, 4(8), 579– 591

PMID: 15286738

Campa D, Mergarten B, De Vivo I, et al 2014 Leukocyte telomere length in

relation to pancreatic cancer risk: a prospective study Cancer Epidemiol Biomarkers Prev, 23(11), 2447– 2454 PMID: 25103821

Capurso G, Schunemann HJ, Terrenato I, et al 2007 Meta- analysis: the use of non- steroidal anti- inflammatory drugs and pancreatic cancer risk for dif-

ferent exposure categories Aliment Pharmacol Ther, 26(8), 1089– 1099

PMID: 17894651

Carter BD, Freedman ND, and Jacobs EJ 2015 Smoking and mortality: beyond

established causes N Engl J Med, 372(22), 2170 PMID: 26017836 Center M, Siegel R, and Jemal A 2011 Global Cancer Facts & Figures

Atlanta, GA: American Cancer Society

Chaffee, K.G., et al Genetic heterogeneity and survival among pancreatic nocarcinoma (PDAC) patients with positive family history, in American Society of Clinical Oncology Annual Meeting 2016: Chicago, IL

ade-Chang AH, and Parsonnet J 2010 Role of bacteria in oncogenesis Clin Microbiol Rev, 23(4), 837– 857 PMCID: PMC2952975

Chari ST 2007 Detecting early pancreatic cancer:  problems and prospects

Semin Oncol, 34(4), 284– 294 PMCID: PMC2680914

Chari ST, Leibson CL, Rabe KG, et  al 2008 Pancreatic cancer- associated diabetes mellitus: prevalence and temporal association with diagnosis of

cancer Gastroenterology, 134(1), 95– 101 PMCID: PMC2271041.

Chari ST, Mohan V, Pitchumoni CS, et al 1994 Risk of pancreatic carcinoma

in tropical calcifying pancreatitis: an epidemiologic study Pancreas, 9(1),

62– 66 PMID: 8108373

Childs EJ, Mocci E, Campa D, et al 2015 Common variation at 2p13.3, 3q29,

7p13 and 17q25.1 associated with susceptibility to pancreatic cancer Nat Genet, 47(8), 911– 916 PMCID: PMC4520746

Chuang S- C, Gallo V, Michaud D, et  al 2011 Exposure to environmental tobacco smoke in childhood and incidence of cancer in adulthood in never smokers in the EuropeanProspective Investigation into Cancer and

Nutrition Cancer Causes & Control, 22(3), 487– 494.

Cogliano VJ, Baan R, Straif K, et al 2011 Preventable exposures associated

with human cancers JNCI, 103(24), 1827– 1839 PMCID: PMC3243677.

Collins K, and Mitchell JR 2002 Telomerase in the human organism

Oncogene, 21(4), 564– 579 PMID: 11850781

Coogan PF, Rosenberg L, Palmer JR, et al 2000 Nonsteroidal anti- inflammatory drugs and risk of digestive cancers at sites other than the large bowel

Cancer Epidemiol Biomarkers Prev, 9(1), 119– 123 PMID: 10667472

Costa FF 2010 Epigenomics in cancer management Cancer Manag Res, 2,

Currie CJ, Poole CD, and Gale EA 2009 The influence of glucose- lowering

therapies on cancer risk in type 2 diabetes Diabetologia, 52(9), 1766–

1777 PMID: 19572116

Dat NM, and Sontag SJ 1982 Pancreatic carcinoma in brothers Ann Intern

de Martel C, Llosa AE, Friedman GD, et al 2008a Helicobacter pylori

infec-tion and development of pancreatic cancer Cancer Epidemiol Biomarkers

de Martel C, Llosa AE, Friedman GD, et al 2008b Helicobacter pylori

infec-tion and development of pancreatic cancer Cancer Epidemiol Biomarkers

Ding XZ, Fehsenfeld DM, Murphy LO, Permert J, and Adrian TE 2000 Physiological concentrations of insulin augment pancreatic cancer cell proliferation and glucose utilization by activating MAP kinase, PI3

kinase and enhancing GLUT- 1 expression Pancreas, 21(3), 310– 320

PMID: 11039477

Dong X, Li Y, Tang H, et al 2012 Insulin- like growth factor axis gene

poly-morphisms modify risk of pancreatic cancer Cancer Epidemiol, 36(2),

206– 211 PMID: 21852217

Drury SS, Esteves K, Hatch V, et al 2015 Setting the trajectory: racial

dispari-ties in newborn telomere length J Pediatrics, 166(5), 1181– 1186.

Cancer of the Pancreas 629

Duell EJ, Holly EA, Bracci PM, et al 2002 A population- based, case- control

study of polymorphisms in carcinogen- metabolizing genes, smoking, and

pancreatic adenocarcinoma risk JNCI, 94(4), 297– 306 PMID: 11854392.

Duell EJ, Holly EA, Bracci PM, Wiencke JK, and Kelsey KT 2002 A

population- based study of the Arg399Gln polymorphism in x- ray repair cross- complementing group 1 (XRCC1) and risk of pancreatic adenocar-

cinoma Cancer Res, 62(16), 4630– 4636.

Duell EJ, Lucenteforte E, Olson SH, et al 2012 Pancreatitis and pancreatic

cancer risk:  a pooled analysis in the International Pancreatic Cancer

PMCID: PMC3477881

Dutruel C, Bergmann F, Rooman I, et al 2014 Early epigenetic

downregula-tion of WNK2 kinase during pancreatic ductal adenocarcinoma

develop-ment Oncogene, 33(26), 3401– 3410.

Edderkaoui M, and Thrower E 2013 Smoking and pancreatic disease J

Cancer Ther, 4, 34– 40 PMID: 24660091

Ehrenthal D, Haeger L, Griffin T, and Compton C 1987 Familial pancreatic

adenocarcinoma in three generations: a case report and a review of the

literature Cancer, 59(9), 1661– 1664 PMID: 3828965.

Elena JW, Steplowski E, Yu K, et  al 2013 Diabetes and risk of pancreatic

cancer: a pooled analysis from the pancreatic cancer cohort consortium

Cancer Causes & Control, 24(1), 13– 25 PMCID: PMC3529822

Everhart J, and Wright D 1995 Diabetes mellitus as a risk factor for pancreatic

cancer: a meta- analysis JAMA, 273(20), 1605– 1609 PMID: 7745774.

Falk RT, Pickle LW, Fontham ET, Correa P, and Fraumeni JF, Jr 1988 Life-

style risk factors for pancreatic cancer in Louisiana: a case- control study

Am J Epidemiol, 128(2), 324– 336 PMID: 3394699

Farrell JJ, Zhang L, Zhou H, et al 2012 Variations of oral microbiota are

asso-ciated with pancreatic diseases including pancreatic cancer Gut, 61(4),

582– 588 PMCID: PMC3705763

Feldman JM, Plonk JW, Bivens CH, and Lebovitz HE 1975 Glucose intolerance

in the carcinoid syndrome Diabetes, 24(7), 664– 671 PMID: 125668.

Ferlay J SI, Ervik M, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin

DM, Forman D, Bray F 2013 GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide:  IARC CancerBase No 11 Lyon, France:  International Agency for Research on Cancer; 2013 Available from: http:// globocan.iarc.fr Accessed October 5, 2016

Fernandez E, La Vecchia C, D’Avanzo B, Negri E, and Franceschi S 1994

Family history and the risk of liver, gallbladder, and pancreatic cancer

Cancer Epidemiol Biomarkers Prev, 3(3), 209– 212 PMID: 8019368

Fiorino S, Chili E, Bacchi- Reggiani L, et al 2013 Association between

hepati-tis B or hepatihepati-tis C virus infection and risk of pancreatic adenocarcinoma

development:  a systematic review and meta- analysis Pancreatology,

13(2), 147– 160 PMID: 23561973

Fong PY, Fesinmeyer MD, White E, et al 2010 Association of diabetes

sus-ceptibility gene calpain- 10 with pancreatic cancer among smokers J Gastrointest Cancer, 41(3), 203– 208 PMCID: PMC2952047

Forman D, Burley V, Cade J, et al 2007 The associations between food,

nutri-tion and physical activity and the risk of pancreatic cancer and underlying mechanisms World Cancer Res Fund Food, nutrition, physical activity, and the prevention of cancer: a global perspective Washington, DC: AICR

Forsmark CE 2007 Pancreatitis and Its Complications New York: Springer

Science & Business Media

Friedman JM, and Fialkow PJ 1976 Familial carcinoma of the pancreas Clin

Genet, 9(5), 463– 469 PMID: 1269168

Fryzek JP, Garabrant DH, Harlow SD, et  al 1997 A case- control study of

self- reported exposures to pesticides and pancreas cancer in southeastern

Michigan Int J Cancer, 72(1), 62– 67 PMID: 9212224.

Fuchs CS, Colditz GA, Stampfer MJ, et al 1996 A prospective study of

ciga-rette smoking and the risk of pancreatic cancer Arch Intern Med, 156(19),

2255– 2260 PMID: 8885826

Gallicchio L, Kouzis A, Genkinger JM, et al 2006 Active cigarette smoking,

household passive smoke exposure, and the risk of developing pancreatic

cancer Prev Med, 42(3), 200– 205.

Gandini S, Lowenfels AB, Jaffee EM, Armstrong TD, and Maisonneuve P

2005 Allergies and the risk of pancreatic cancer:  a meta- analysis with

review of epidemiology and biological mechanisms Cancer Epidemiol Biomarkers Prev, 14(8), 1908– 1916 PMID: 16103436

Garcea G, Dennison AR, Steward WP, and Berry DP 2005 Role of

inflamma-tion in pancreatic carcinogenesis and the implicainflamma-tions for future therapy

Pancreatology, 5(6), 514– 529 PMID: 16110250

Garcia C, Feve B, Ferre P, et al 2010 Diabetes and

inflammation: fundamen-tal aspects and clinical implications Diabetes Metab, 36(5), 327– 338

PMID: 20851652

Gbelcova H, Lenicek M, Zelenka J, et al 2008 Differences in antitumor effects

of various statins on human pancreatic cancer Int J Cancer, 122(6),

1214– 1221 PMID: 18027870

Genkinger JM, Spiegelman D, Anderson KE, et al 2009 Alcohol intake and

pancreatic cancer risk: a pooled analysis of fourteen cohort studies Cancer Epidemiol Biomarkers Prev, 18(3), 765– 776 PMCID: PMC2715951.Genkinger JM, Spiegelman D, Anderson KE, et al 2011 A pooled analysis of

14 cohort studies of anthropometric factors and pancreatic cancer risk Int

J Cancer, 129(7), 1708– 1717 PMCID: PMC3073156

Gerken T, Girard CA, Tung Y- CL, et al 2007 The obesity- associated FTO gene

encodes a 2- oxoglutarate- dependent nucleic acid demethylase Science,

318(5855), 1469– 1472

Ghadirian P, Simard A, and Baillargeon J 1987 Cancer of the pancreas

in two brothers and one sister Int J Pancreatol, 2(5- 6), 383– 391

PMID: 3693982

Ghadirian P, Liu G, Gallinger S, et al 2002 Risk of pancreatic cancer among

individuals with a family history of cancer of the pancreas Int J Cancer,

97(6), 807– 810 PMID: 11857359

Giardiello FM, Brensinger JD, Tersmette AC, et  al 2000 Very high risk of

cancer in familial Peutz- Jeghers syndrome Gastroenterology, 119(6),

1447– 1453 PMID: 11113065

Giovannucci E 2007 Metabolic syndrome, hyperinsulinemia, and colon

can-cer: a review Am J Clin Nutr, 86(3), s836– s842 PMID: 18265477.

Gold EB, and Goldin SB 1998 Epidemiology of and risk factors for pancreatic

cancer Surg Oncol Clin N Am, 7(1), 67– 91 PMID: 9443987.

Goldacre MJ, Abisgold JD, Seagroatt V, and Yeates D 2005 Cancer after

chole-cystectomy: record- linkage cohort study Br J Cancer, 92(7), 1307– 1309.

Goldstein AM, Struewing JP, Fraser MC, Smith MW, and Tucker MA 2004

Prospective risk of cancer in CDKN2A germline mutation carriers J Med Genet, 41(6), 421– 424 PMCID: PMC1735800

Grant RC, Selander I, Connor AA, et al 2015 Prevalence of germline tions in cancer predisposition genes in patients with pancreatic cancer

muta-Gastroenterology, 148(3), 556– 564 PMCID: PMC4339623

Greer JB, and Whitcomb DC 2009 Inflammation and pancreatic

can-cer:  an evidence- based review Curr Opin Pharmacol, 9(4), 411– 418

PMID: 19589727

Grutzmann R, Niedergethmann M, Pilarsky C, Kloppel G, and Saeger HD 2010 Intraductal papillary mucinous tumors of the pancreas: biology, diagnosis,

and treatment Oncologist, 15(12), 1294– 1309 PMCID: PMC3227924.

Gukovskaya AS, Mouria M, Gukovsky I, et  al 2002 Ethanol metabolism and transcription factor activation in pancreatic acinar cells in rats

Gastroenterology, 122(1), 106– 118 PMID: 11781286

Gullo L, Migliori M, Falconi M, et al 2003 Nonfunctioning pancreatic

endo-crine tumors:  a multicenter clinical study Am J Gastroenterol, 98(11),

2435– 2439 PMID: 14638345

Gullo L, Pezzilli R, and Morselli- Labate AM 1994 Diabetes and the risk of

pancreatic cancer N Engl J Med, 331(2), 81– 84.

Guo M, Jia Y, Yu Z, et al 2014 Epigenetic changes associated with neoplasms

of the exocrine and endocrine pancreas Discovery Med, 17(92), 67.

Haber PS, Apte MV, Moran C, et al 2004 Non- oxidative metabolism of

etha-nol by rat pancreatic acini Pancreatology, 4(2), 82– 89 PMID: 15056978.

Halfdanarson TR, Bamlet WR, McWilliams RR, et al 2014 Risk factors for pancreatic neuroendocrine tumors:  a clinic- based case- control study

Hallet J, Law CH, Cukier M, et  al 2015 Exploring the rising incidence of

epidemiol-ogy, metastatic presentation, and outcomes Cancer, 121(4), 589– 597

PMID: 25312765

Hang B 2010 Formation and repair of tobacco carcinogen- derived bulky DNA

adducts J Nucleic Acids, 2010, 709521 PMCID: PMC3017938 Harley CB 1997 Human ageing and telomeres Ciba Found Symp, 211, 129–

atic cancer: a case- control study Cancer, 109(12), 2547– 2556.

Hassan MM, Bondy ML, Wolff RA, et  al 2007 Risk factors for pancreatic

cancer:  case- control study Am J Gastroenterol, 102(12), 2696– 2707

PMCID: PMC2423805

Hassan MM, Phan A, Li D, et al 2008a Family history of cancer and

asso-ciated risk of developing neuroendocrine tumors:  a case- control study

Cancer Epidemiol Biomarkers Prev, 17(4), 959– 965 PMID: 18398037

Hassan MM, Phan A, Li D, et al 2008b Risk factors associated with

neuroen-docrine tumors: A U.S.- based case- control study Int J Cancer, 123(4),

867– 873 PMID: 18491401

Hauptmann N, and Cadenas E 1997 The oxygen paradox:  biochemistry of

active oxygen Cold Spring Harbor Monogrh Arch, 34, 1– 20.

Heard E, and Martienssen RA 2014 Transgenerational epigenetic

inheri-tance: myths and mechanisms Cell, 157 95– 109.

Hecht SS 2003 Tobacco carcinogens, their biomarkers and tobacco- induced

cancer Nat Rev Cancer, 3(10), 733– 744.

Hein DW, Doll MA, Fretland AJ, et al 2000 Molecular genetics and

epide-miology of the NAT1 and NAT2 acetylation polymorphisms Cancer

Epidemiol Biomarkers Prev, 9(1), 29– 42 PMID: 10667461

Heinen MM, Verhage BA, Goldbohm RA, and van den Brandt PA 2012 Intake

of vegetables, fruits, carotenoids and vitamins C and E and pancreatic

cancer risk in The Netherlands Cohort Study Int J Cancer, 130(1), 147–

158 PMID: 21328344

Hering J, Garrean S, Dekoj TR, et  al 2007 Inhibition of proliferation by

omega- 3 fatty acids in chemoresistant pancreatic cancer cells Ann Surg

Oncol, 14(12), 3620– 3628 PMID: 17896154

Herman WH 2007 Diabetes epidemiology: guiding clinical and public health

practice:  the Kelly West Award Lecture, 2006 Diabetes Care, 30(7),

1912– 1919 PMID: 17496237

Hernandez AV, Pasupuleti V, Benites- Zapata VA, et al 2015 Insulin resistance

and endometrial cancer risk: a systematic review and meta- analysis Eur

J Cancer, 51(18), 2747– 2758 PMID: 26597445

Hinney A, Nguyen TT, Scherag A, et  al 2007 Genome wide association

(GWA) study for early onset extreme obesity supports the role of fat mass

and obesity associated gene (FTO) variants PloS One, 2(12), e1361.

Hiripi E, Bermejo JL, Sundquist J, and Hemminki K 2009 Familial

gastro-intestinal carcinoid tumours and associated cancers Ann Oncol, 20(5),

950– 954 PMID: 19150948

Hirvonen A 1999 Polymorphic NATs and cancer predisposition IARC Sci

Hoffmann D, Hoffmann I, and El- Bayoumy K 2001 The less harmful

ciga-rette:  a controversial issue a tribute to Ernst L.  Wynder Chem Res

Toxicol, 14(7), 767– 790 PMID: 11453723

Hou L, Zhang X, Wang D, and Baccarelli A 2012 Environmental

chemi-cal exposures and human epigenetics Int J Epidemiol, 41(1), 79– 105

PMID: 22253299 PMCID: PMC3304523

Howes N, and Neoptolemos JP 2002 Risk of pancreatic ductal adenocarcinoma

in chronic pancreatitis Gut, 51(6), 765– 766 PMCID: PMC1773485.

Hruban R, Petersen GM, Ha P, and Kern S 1998 Genetics of pancreatic

can-cer: from genes to families Surg Oncol Clinics North Am, 7(1), 1– 23.

Hruban RH, Adsay NV, Albores- Saavedra J, et al 2001 Pancreatic

intraepithe-lial neoplasia: a new nomenclature and classification system for

pancre-atic duct lesions Am J Surg Pathol, 25(5), 579– 586 PMID: 11342768.

Hruban RH, Maitra A, and Goggins M 2008 Update on pancreatic

PMC2480542

Hruban RH, van Mansfeld AD, Offerhaus GJ, et al 1993 K- ras oncogene

acti-vation in adenocarcinoma of the human pancreas: a study of 82

carcino-mas using a combination of mutant- enriched polymerase chain reaction

analysis and allele- specific oligonucleotide hybridization Am J Pathol,

143(2), 545– 554 PMICD: PMC1887038

Hu C, Hart SN, Bamlet WR, et al 2016 Prevalence of pathogenic mutations

in cancer predisposition genes among pancreatic cancer patients Cancer

Epidemiol Biomarkers Prev, 25(1), 207– 211 PMCID: PMC4754121

Hu Z- H, Connett JE, Yuan J- M, and Anderson KE 2016 Role of survivor bias

in pancreatic cancer case- control studies Ann Epidemiol, 26, 50– 56

PMID: 26688282

Huang JY, Butler LM, Wang R, et  al 2016 Dietary intake of one- carbon

metabolism- related nutrients and pancreatic cancer risk: The Singapore

Chinese Health Study Cancer Epidemiol Biomarkers Prev, 25, 417– 424.

Humphris JL, Johns AL, Simpson SH, et  al 2014 Clinical and pathologic

features of familial pancreatic cancer Cancer, 120(23), 3669– 3675

PMID: 25313458

can-cer:  a meta- analysis of 36 studies Br J Cancer, 92(11), 2076– 2083

PMCID: PMC2361795

Inoue M, Tajima K, Takezaki T, et  al 2003 Epidemiology of pancreatic

cancer in Japan:  a nested case- control study from the Hospital- based

Epidemiologic Research Program at Aichi Cancer Center (HERPACC)

Int J Epidemiol, 32(2), 257– 262 PMID: 12714546

Iodice S, Gandini S, Maisonneuve P, and Lowenfels AB 2008 Tobacco and the

risk of pancreatic cancer: a review and meta- analysis Langenbeck’s Arch Surgery, 393(4), 535– 545

Iqbal J, Ragone A, Lubinski J, et al 2012 The incidence of pancreatic cancer in

BRCA1 and BRCA2 mutation carriers Br J Cancer, 107(12), 2005– 2009

control study Cancer Epidemiol Biomarkers Prev, 22(7), 1336– 1339

veg-Cancer Causes Control, 22(12), 1613– 1625 PMCID: PMC3522747

Polymorphisms in metabolism/ antioxidant genes may mediate the effect

of dietary intake on pancreatic cancer risk Pancreas, 42(7), 1043– 1053

PMCID: PMC3779344

Jemal A, Simard EP, Dorell C, et  al 2013 Annual Report to the Nation on the Status of Cancer, 1975– 2009, Featuring the Burden and Trends in Human Papillomavirus (HPV)– Associated Cancers and HPV Vaccination

Coverage Levels JNCI, 105(3), 175– 201.

Jennings BJ, Ozanne SE, and Hales CN 2000 Nutrition, oxidative damage, telomere shortening, and cellular senescence:  individual or connected

agents of aging? Mol Genet Metab, 71(1– 2), 32– 42.

Jiao L, Bondy ML, Hassan MM, et al 2007 Glutathione S- transferase gene

polymorphisms and risk and survival of pancreatic cancer Cancer,

109(5), 840– 848 PMCID: PMC1892189

Jiao L, Hassan MM, Bondy ML, et al 2008 XRCC2 and XRCC3 gene

poly-morphism and risk of pancreatic cancer Am J Gastroenterol, 103(2),

360– 367 PMCID: PMC2268638

Jones S, Hruban RH, Kamiyama M, et al 2009 Exomic sequencing identifies

PALB2 as a pancreatic cancer susceptibility gene Science, 324(5924),

217 PMCID: PMC2684332

Jones S, Zhang X, Parsons DW, et  al 2008 Core signaling pathways in

human pancreatic cancers revealed by global genomic analyses Science,

321(5897), 1801– 1806 PMCID: PMC2848990

Kaaks R, Lukanova A, and Kurzer MS 2002 Obesity, endogenous hormones,

and endometrial cancer risk:  a synthetic review Cancer Epidemiol Biomarkers Prev, 11(12), 1531– 1543 PMID: 12496040

Kaerlev L, Teglbjaerg PS, Sabroe S, et al 2002 The importance of smoking and medical history for development of small bowel carcinoid tumor: a

European population- based case- control study Cancer Causes Control,

Kastrinos F, Mukherjee B, Tayob N, et al 2009 Risk of pancreatic cancer in

families with Lynch syndrome Jama, 302(16), 1790– 1795 PMCID:

PMC4091624

Khawja SN, Mohammed S, Silberfein EJ, et al 2015 Pancreatic cancer

dispari-ties in African Americans Pancreas, 44(4), 522– 527 PMID: 25872128 Klein AP 2012 Genetic susceptibility to pancreatic cancer Mol Carcinog,

51(1), 14– 24 PMCID: PMC3570154

Klein AP, Beaty TH, Bailey- Wilson JE, et al 2002 Evidence for a major gene

influencing risk of pancreatic cancer Genet Epidemiol, 23(2), 133– 149

Cancer of the Pancreas 631

Kollarova H, Azeem K, Tomaskova H, et al 2014 Is physical activity a

protec-tive factor against pancreatic cancer? Bratisl Lek Listy, 115(8), 474– 478

PMID: 25246281

Korsse SE, Harinck F, van Lier MG, et al 2013 Pancreatic cancer risk in Peutz-

Jeghers syndrome patients: a large cohort study and implications for

sur-veillance J Med Genet, 50(1), 59– 64 PMID: 23240097.

Kovi J, and Heshmat MY 1972 Incidence of cancer in negroes in Washington,

D.C.  and selected African cities Am J Epidemiol, 96(6), 401– 413

PMID: 4643672

Kowalska I, Straczkowski M, Nikolajuk A, et al 2008 Insulin resistance, serum

adiponectin, and proinflammatory markers in young subjects with the

metabolic syndrome Metabolism, 57(11), 1539– 1544 PMID: 18940391.

Krain LS 1970 The rising incidence of carcinoma of the pancreas:  real or

apparent? J Surg Oncol, 2(2), 115– 124 PMID: 5520829.

Kusama T, Mukai M, Iwasaki T, et  al 2002 3- hydroxy- 3- methylglutaryl-

coenzyme a reductase inhibitors reduce human pancreatic cancer

PMID: 11832446

Lakhani VT, You YN, and Wells SA 2007 The multiple endocrine neoplasia

syndromes Annu Rev Med, 58, 253– 265 PMID: 17037976.

Larsson SC, Giovannucci E, Bergkvist L, and Wolk A 2006 Aspirin and

non-steroidal anti- inflammatory drug use and risk of pancreatic cancer:  a

meta- analysis Cancer Epidemiol Biomarkers Prev, 15(12), 2561– 2564

PMID: 17164387

Larsson SC, Orsini N, and Wolk A 2007 Body mass index and pancreatic

cancer risk: a meta- analysis of prospective studies Int J Cancer, 120(9),

1993– 1998 PMID: 17266034

Lawrence B, Gustafsson BI, Chan A, et al 2011 The epidemiology of

gastro-enteropancreatic neuroendocrine tumors Endocrinol Metab Clin North

Levitzky YS, Guo CY, Rong J, et al 2008 Relation of smoking status to a panel

of inflammatory markers:  the framingham offspring Atherosclerosis,

201(1), 217– 224 PMCID: PMC2783981

Li D 2012 Diabetes and pancreatic cancer Mol Carcinog, 51(1), 64– 74

PMCID: PMC3238796

Li D, Day RS, Bondy ML, et al 2007 Dietary mutagen exposure and risk of

pancreatic cancer Cancer Epidemiol Biomarkers Prev, 16(4), 655– 661

PMCID: PMC1892159

Li D, Jiao L, Li Y, et al 2005 Polymorphisms of cytochrome P4501A2 and

Carcinogenesis, 27(1), 103– 111

Li D, Tang H, Hassan MM, et al 2011 Diabetes and risk of pancreatic

can-cer: a pooled analysis of three large case- control studies Cancer Causes Control, 22(2), 189– 197 PMID: 21104117

Li D, Yeung SC, Hassan MM, Konopleva M, and Abbruzzese JL 2009

Antidiabetic therapies affect risk of pancreatic cancer Gastroenterology,

137(2), 482– 488 PMCID: PMC2735093

Li J, Zhu J, Hassan MM, et al 2007 K- ras mutation and p16 and

preproen-kephalin promoter hypermethylation in plasma DNA of pancreatic

can-cer patients:  in relation to cigarette smoking Pancreas, 34(1), 55– 62

PMCID: PMC1905887

Lichtenstein P, Holm NV, Verkasalo PK, et al 2000 Environmental and

heri-table factors in the causation of cancer:  analyses of cohorts of twins

from Sweden, Denmark, and Finland N Engl J Med, 343(2), 78– 85

PMID: 10891514

Lin Y, Yagyu K, Egawa N, et  al 2011 An overview of genetic

polymor-phisms and pancreatic cancer risk in molecular epidemiologic studies J Epidemiol, 21(1), 2– 12 PMCID: PMC3899511

Lindkvist B, Johansen D, Borgstrom A, and Manjer J 2008 A prospective

study of Helicobacter pylori in relation to the risk for pancreatic cancer

BMC Cancer, 8, 321 PMCID: PMC2613155

Liu G, Ghadirian P, Vesprini D, et al 2000 Polymorphisms in GSTM1, GSTT1

and CYP1A1 and risk of pancreatic adenocarcinoma Br J Cancer,

82(10), 1646– 1649 PMCID: PMC2374522

Lo AC, Soliman AS, El- Ghawalby N, et al 2007 Lifestyle, occupational, and

reproductive factors in relation to pancreatic cancer risk Pancreas, 35(2),

120– 129 PMID: 17632317

Lonser RR, Glenn GM, Walther M, et  al 2003 von Hippel- Lindau disease

Lancet, 361(9374), 2059– 2067 PMID: 12814730

Luo J, Margolis KL, Adami HO, et  al 2008 Obesity and risk of pancreatic

cancer among postmenopausal women:  the Women’s Health Initiative

(United States) Br J Cancer, 99(3), 527– 531 PMCID:  PMC2527801

PMID: 18628761Lowenfels AB, and Maisonneuve P 2004 Epidemiology and prevention of

pancreatic cancer Jpn J Clin Oncol, 34(5), 238– 244.

Lowenfels AB, and Maisonneuve P 2006 Epidemiology and risk factors for

pancreatic cancer Best Pract Res Clin Gastroenterol, 20(2), 197– 209

PMID: 16549324

Lowenfels AB, Maisonneuve P, and Whitcomb DC 2000 Risk factors for cer in hereditary pancreatitis International Hereditary Pancreatitis Study

can-Group Med Clin North Am, 84(3), 565– 573 PMID: 10872414.

Lowenfels AB, Maisonneuve P, Cavallini G, et al 1993 Pancreatitis and the

risk of pancreatic cancer International Pancreatitis Study Group N Engl

J Med, 328(20), 1433– 1437 PMID: 8479461

Lowenfels AB, Maisonneuve P, DiMagno EP, et al 1997 Hereditary tis and the risk of pancreatic cancer International Hereditary Pancreatitis

pancreati-Study Group JNCI, 89(6), 442– 446 PMID: 9091646.

Lynch SM, Major JM, Cawthon R, et al 2013 A prospective analysis of mere length and pancreatic cancer in the alpha- tocopherol beta- carotene

telo-cancer (ATBC) prevention study Int J Cancer, 133(11), 2672– 2680

PMID: 23674344

Lynch SM, Vrieling A, Lubin JH, et al 2009 Cigarette smoking and pancreatic cancer: a pooled analysis from the pancreatic cancer cohort consortium

Am J Epidemiol, 170(4), 403– 413

Ma J, Siegel R, and Jemal A 2013 Pancreatic cancer death rates by race

among US men and women, 1970– 2009 JNCI, 105(22):1694– 1700

PMID: 24203988

MacDermott RP, and Kramer P 1973 Adenocarcinoma of the pancreas in four

siblings Gastroenterology, 65(1), 137– 139 PMID: 4720820.

Magruder JT, Elahi D, and Andersen DK 2011 Diabetes and pancreatic

can-cer: chicken or egg? Pancreas, 40(3), 339– 351 PMID: 21412116.

Maisonneuve P, Lowenfels AB, Bueno- de- Mesquita HB, et al 2010 Past cal history and pancreatic cancer risk:  results from a multicenter case-

medi-control study Ann Epidemiol, 20(2), 92– 98 PMID: 20123159.

Maitra A, and Hruban RH 2008 Pancreatic cancer Annu Rev Pathol, 3, 157–

188 PMCID: PMC2666336

Majumder S, Bockorny B, Baker WL, and Dasanu CA 2014 Association

between HBsAg positivity and pancreatic cancer:  a meta- analysis J Gastrointest Cancer, 45(3), 347– 352 PMID: 24788082

McCarty MF 2001 Insulin secretion as a determinant of pancreatic cancer risk

Med Hypotheses, 57(2), 146– 150 PMID: 11461162

McKeown- Eyssen G 1994 Epidemiology of colorectal cancer revisited:  are

serum triglycerides and/ or plasma glucose associated with risk? Cancer Epidemiol Biomarkers Prev, 3(8), 687– 695 PMID: 7881343

McWilliams RR, Rabe KG, Olswold C, De Andrade M, and Petersen GM

2005 Risk of malignancy in first- degree relatives of patients with

pancre-atic carcinoma Cancer, 104(2), 388– 394 PMID: 15912495.

McWilliams RR, Bamlet WR, Cunningham JM, et al 2008 Polymorphisms in

DNA repair genes, smoking, and pancreatic adenocarcinoma risk Cancer

McWilliams RR, Petersen GM, Rabe KG, et al 2010 Cystic fibrosis brane conductance regulator (CFTR) gene mutations and risk for pancre-

transmem-atic adenocarcinoma Cancer, 116(1), 203– 209 PMCID: PMC2807917.

McWilliams RR, Wieben ED, Rabe KG, et al 2011 Prevalence of CDKN2A mutations in pancreatic cancer patients: implications for genetic counsel-

ing Eur J Hum Genet, 19(4), 472– 478 PMCID: PMC3060321.

Mettu NB, and Abbruzzese JL 2016 Clinical insights into the biology and

treat-ment of pancreatic cancer J Oncol Pract, 12(1), 17– 23 PMID: 26759461.

Michaud DS, Giovannucci E, Willett WC, et al 2001 Physical activity,

obe-sity, height, and the risk of pancreatic cancer JAMA, 286(8), 921– 929

PMID: 11509056

Michaud DS, Joshipura K, Giovannucci E, and Fuchs CS 2007 A prospective study of periodontal disease and pancreatic cancer in US male health pro-

fessionals JNCI, 99(2), 171– 175 PMID: 17228001.

Michaud DS, Vrieling A, Jiao L, et  al 2010 Alcohol intake and atic cancer:  a pooled analysis from the pancreatic cancer cohort

PMCID: PMC3098295

Min Y, Adachi Y, Yamamoto H, et al 2003 Genetic blockade of the insulin- like growth factor- I receptor: a promising strategy for human pancreatic can-

cer Cancer Res, 63(19), 6432– 6441 PMID: 14559833.

Mohelnikova- Duchonova B, Vrana D, Holcatova I, et  al 2010 CYP2A13, ADH1B, and ADH1C gene polymorphisms and pancreatic cancer risk

Pancreas, 39(2), 144– 148

Molina MA, Sitja- Arnau M, Lemoine MG, Frazier ML, and Sinicrope FA

1999 Increased cyclooxygenase- 2 expression in human pancreatic nomas and cell lines: growth inhibition by nonsteroidal anti- inflammatory

carci-drugs Cancer Res, 59(17), 4356– 4362 PMID: 10485483.

Muscat JE, Stellman SD, Hoffmann D, and Wynder EL 1997 Smoking and

pancreatic cancer in men and women Cancer Epidemiol Biomarkers

Nagy C, and Turecki G 2015 Transgenerational epigenetic inheritance:  an

open discussion Epigenomics, 7, 781– 790 PMID: 26344807.

Nakao M, Hosono S, Ito H, et al 2011 Interaction between IGF- 1

polymor-phisms and overweight for the risk of pancreatic cancer in Japanese Int J Mol Epidemiol Genet, 2(4), 354– 366 PMCID: PMC3243451

Nilsen TIL, and Vatten LJ 2000 A prospective study of lifestyle factors and

the risk of pancreatic cancer in Nord- Trøndelag, Norway Cancer Causes

Control, 11(7), 645– 652

Nones K, Waddell N, Song S, et al 2014 Genome- wide DNA methylation

pat-terns in pancreatic ductal adenocarcinoma reveal epigenetic deregulation of

SLIT- ROBO, ITGA2 and MET signaling Int J Cancer, 135(5), 1110– 1118.

Nöthlings U, Wilkens LR, Murphy SP, et  al 2005 Meat and fat intake as

risk factors for pancreatic cancer: The Multiethnic Cohort Study JNCI,

97(19), 1458– 1465 PMID: 16204695

Ojajarvi IA, Partanen TJ, Ahlbom A, et al 2000 Occupational exposures and

pancreatic cancer: a meta- analysis Occup Environ Med, 57(5), 316– 324

PMCID: PMC1739949

Okuda K, Bardeguez A, Gardner JP, et al 2002 Telomere length in the

new-born Pediatr Res, 52(3), 377– 381 PMID: 12193671

Olson SH 2012 Selected medical conditions and risk of pancreatic cancer Mol

Carcinog, 51(1), 75– 97 PMID: 22162233

Olson SH, Hsu M, Satagopan JM, et  al 2013 Allergies and risk of

pancre-atic cancer: a pooled analysis from the Pancrepancre-atic Cancer Case- Control

Consortium Am J Epidemiol, 178(5), 691– 700 PMCID: PMC3755648.

O’Rorke MA, Cantwell MM, Cardwell CR, Mulholland HG, and Murray LJ 2010

Can physical activity modulate pancreatic cancer risk? a systematic review

and meta- analysis Int J Cancer, 126(12), 2957– 2968 PMID: 19856317.

Ough M, Lewis A, Zhang Y, et al 2004 Inhibition of cell growth by

overex-pression of manganese superoxide dismutase (MnSOD) in human

pancre-atic carcinoma Free Radical Res, 38(11), 1223– 1233.

Pannala R, Basu A, Petersen GM, and Chari ST 2009 New- onset diabetes: a

potential clue to the early diagnosis of pancreatic cancer Lancet Oncol,

10(1), 88– 95 PMCID: PMC2795483

Park C- H, Lee I- S, Grippo P, et al 2013 Akt kinase mediates the pro- survival

effect of smoking compounds in pancreatic ductal cells Pancreas, 42,

655– 662 PMID: 23271397

Parkin D, Whelan S, Ferlay J, Teppo L, and Thomas D 2002 Cancer

Incidence in Five Continents, Vol VIII Lyon, France: IARC Scientific

Publications, 155

Parkin DM, Boyd L, and Walker LC 2011 16 The fraction of cancer

attrib-utable to lifestyle and environmental factors in the UK in 2010 Br J

Cancer, 105 Suppl 2, S77– S81 PMCID: PMC3252065

Patel AV, Rodriguez C, Bernstein L, et al 2005 Obesity, recreational

physi-cal activity, and risk of pancreatic cancer in a large U.S Cohort Cancer

Epidemiol Biomarkers Prev, 14(2), 459– 466 PMID: 15734973

Permuth- Wey J, and Egan KM 2009 Family history is a significant risk factor

for pancreatic cancer: results from a systematic review and meta- analysis

Fam Cancer, 8(2), 109– 117 PMID: 18763055

Perugini RA, McDade TP, Vittimberga FJ, Jr, Duffy AJ, and Callery MP 2000

Sodium salicylate inhibits proliferation and induces G1 cell cycle arrest

in human pancreatic cancer cell lines J Gastrointest Surg, 4(1), 24– 32,

discussion 32– 23 PMID: 10631359

Petersen GM, Amundadottir L, Fuchs CS, et al 2010 A genome- wide

asso-ciation study identifies pancreatic cancer susceptibility loci on

chro-mosomes 13q22.1, 1q32.1 and 5p15.33 Nat Genet, 42(3), 224– 228

PMCID: PMC2853179

Petersen GM, de Andrade M, Goggins M, et al 2006 Pancreatic cancer genetic

epidemiology consortium Cancer Epidemiol Biomarkers Prev, 15(4),

704– 710 PMID: 16614112

Petersen GM 2015 Familial pancreatic adenocarcinoma Hematol Oncol Clin

North Am, 29(4), 641– 653 PMCID: PMC4522044

Potjer TP, Kranenburg HE, Bergman W, et al 2015 Prospective risk of cancer

and the influence of tobacco use in carriers of the p16- Leiden germline

variant Eur J Hum Genet, 23(5), 711– 714 PMCID: PMC4402641.

Pradhan AD, Manson JE, Rifai N, Buring JE, and Ridker PM 2001 C- reactive

protein, interleukin 6, and risk of developing type 2 diabetes mellitus

Prokopczyk B, Hoffmann D, Bologna M, et al 2002 Identification of tobacco-

derived compounds in human pancreatic juice Chem Res Toxicol, 15(5),

677– 685

Purohit V, Bode JC, Bode C, et al 2008 Alcohol, intestinal bacterial growth,

intestinal permeability to endotoxin, and medical consequences: 

sum-mary of a symposium Alcohol, 42(5), 349– 361 PMCID: PMC2614138.

Raderer M, Wrba F, Kornek G, et al 1998 Association between Helicobacter

PMID: 9428370

Rahman F, Cotterchio M, Cleary SP, and Gallinger S 2015 Association

between alcohol consumption and pancreatic cancer risk: a case- control

study PLoS One, 10(4), e0124489 PMCID: PMC4391718.

Raimondi S, Lowenfels AB, Morselli- Labate AM, Maisonneuve P, and Pezzilli

R 2010 Pancreatic cancer in chronic pancreatitis; aetiology, incidence,

and early detection Best Pract Res Clin Gastroenterol, 24(3), 349– 358

ade-tive series Am J Gastroenterol, 103(1), 111– 119 PMID: 18184119.

Renehan AG, Tyson M, Egger M, Heller RF, and Zwahlen M 2008 Body- mass index and incidence of cancer:  a systematic review and meta- analysis

PMID: 18280327

Reports of the US Surgeon General 2014 The Health Consequences of Smoking— 50 Years of Progress: A Report of the Surgeon General Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, Coordinating Center for Health Promotion, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health Available at http:// www.surgeongeneral.gov/ library/ reports/ 50- years- of- progress/ full- report.pdf

Resta N, Pierannunzio D, Lenato GM, et  al 2013 Cancer risk associated

patients:  results of an Italian multicenter study Dig Liver Dis, 45(7),

606– 611 PMID: 23415580

Risch HA 2003 Etiology of pancreatic cancer, with a hypothesis concerning

the role of N- nitroso compounds and excess gastric acidity JNCI, 95(13),

948– 960

Risch HA, Yu H, Lu L, and Kidd MS 2010 ABO blood group, Helicobacter pylori seropositivity, and risk of pancreatic cancer: a case- control study

Rivenson A, Hoffmann D, Prokopczyk B, Amin S, and Hecht SS 1988 Induction

of lung and exocrine pancreas tumors in F344 rats by tobacco- specific and

Areca- derived N- nitrosamines Cancer Res, 48(23), 6912– 6917.

Roberts NJ, Jiao Y, Yu J, et  al 2012 ATM mutations in patients with

hereditary pancreatic cancer Cancer Discov, 2(1), 41– 46 PMCID:

PMC3676748

Roberts NJ, Norris AL, Petersen GM, et al 2016 Whole genome sequencing

defines the genetic heterogeneity of familial pancreatic cancer Cancer Discov, 6(2), 166– 175 PMCID: PMC4744563

Rosato V, Polesel J, Bosetti C, et  al 2015 Population attributable risk

PMID: 25479588

Ryan DP, Hong TS, and Bardeesy N 2014 Pancreatic adenocarcinoma N Engl

J Med, 371(11), 1039– 1049 PMID: 25207767

Scheipers P, and Reiser H 1998 Fas- independent death of activated CD4+ T

lymphocytes induced by CTLA- 4 crosslinking Proc Natl Acad of Sci,

95(17), 10083– 10088

Schenk M, Schwartz AG, O’Neal E, et  al 2001 Familial risk of pancreatic

cancer J Natl Cancer Inst, 93(8), 640– 644 PMID: 11309441.

Schernhammer ES, Kang JH, Chan AT, et al 2004 A prospective study of

aspi-rin use and the risk of pancreatic cancer in women JNCI, 96(1), 22– 28

PMID: 14709735

Schernhammer ES, Michaud DS, Leitzmann MF, et al 2002 Gallstones,

cho-lecystectomy, and the risk for developing pancreatic cancer Br J Cancer,

86(7), 1081– 1084 PMCID: PMC2364180

Schulte A, Pandeya N, Fawcett J, et  al 2015 Association between

helico-bacter pylori and pancreatic cancer risk: a meta- analysis Cancer Causes Control, 26(7), 1027– 1035 PMID: 25951801

Scott LJ, Mohlke KL, Bonnycastle LL, et al 2007 A genome- wide association study of type 2 diabetes in Finns detects multiple susceptibility variants

Science, 316(5829), 1341– 1345

Shen Q, Tian Y, Li K, et al 2015 Association of single nucleotide

polymor-phisms of DNA repair gene and susceptibility to pancreatic cancer Int J Clin Exp Pathol, 8(3), 3180– 3185 PMCID: PMC4440146

population- based description of familial clustering of pancreatic cancer

Clin Gastroenterol Hepatol, 8(9), 812– 816 PMID: 20570637

Siegel RL, Miller KD, and Jemal A 2015 Cancer statistics, 2015 CA Cancer