Alleviating the Burden of Cancer: A Perspective on Advances, Challenges, and Future Directions

The 30th Annual Meeting of the American Society of Preventive Oncology provides an appropriate setting for reflecting on the advances, challenges, and future directions in cancer epidemiology and cancer prevention and control. Whereas the universality of cancer incidence and mortality is established, the burden of cancer by type or organ site is represented by contrasting patterns between developing countries in Africa, Asia, and Latin America, and more industrialized countries in North America, Europe, Australia, and New Zealand. The age-period-cohort patterns observed in individual countries may be quite heterogeneous, as in urban compared with rural regions, or in relationship to racial, cultural, ethnic, and socioeconomic characteristics.

Populations in developing countries are disproportionately susceptible to cancers in which infectious agents are causal (1). These organ sites include the uterine cervix, penis, and oropharynx (human papillomavirus), liver (hepatitis viruses B and C), stomach (Helicobacter pylori, non-Hodgkin lymphoma, including endemic Burkitt lymphoma (EBV)), nasopharynx (EBV), urinary bladder (Schistosoma haematobium), biliary tract (opisthorchis viverrini, clonorchis sinensis, and Kaposi's sarcoma (human herpesvirus type 8)). The public health effect of these infections is substantial. The proportion of total cancer deaths attributable to infectious agents is estimated to be about 20% to 25% in developing countries and 7% to 10% in more industrialized countries.

In a worldwide population exceeding six billion in the year 2002, there were ∼10.9 million new cancer cases, 6.7 million cancer deaths, and 22.4 million persons surviving from cancer diagnosed in the previous 5 years (2, 3). In 2020, the world's population is projected to increase to 7.5 billion, which will experience 15 million new cancer cases and 12 million cancer deaths (4). In 2002, the most common incident cancers worldwide, excluding keratinocyte skin cancers, were lung (12.3%), breast (10.4%), and colorectum (9.4%). The three cancer sites contributing to the highest proportions of global cancer deaths were lung (17.8%), stomach (10.4%), and liver (8.8%). Of ∼55 million worldwide deaths in 2002, 12% to 13% were attributed to cancer and 31% to heart disease.

In the United States in 1900, cancer mortality accounted for <4% of all deaths and ranked as the sixth leading cause; by 1940, cancer mortality ranked second and accounted for 11% of all deaths (5). In 2003, there were 556,902 cancer deaths or 22.7% of total deaths. From 1992 to 2002, cancer mortality decreased by 1.5% per year in men and by 0.8% per year in women. Since 1999, in the age-gender combined subgroups younger than 85 years, or in 98% of the population at risk, cancer mortality has surpassed heart disease as the underlying cause of death; in the subgroup older than 85, heart disease deaths were three times more frequent than cancer deaths. In the projected statistics for 2006, among men, cancers of the prostate, lung and bronchus, and colon and rectum will account for at least 56% of all newly diagnosed cases; among women, cancers of the breast, lung and bronchus, and colon and rectum will comprise about 54% of the estimated total cases. In 1987, lung cancer mortality superseded breast cancer mortality in women and, in 2006, will account for 26% of all cancer deaths in women (6).

Approximately 200,000 incident cancers registered in 2003 in the United States, or about 16% of the estimated total annual incidence, were diagnosed in patients with one or more prior independent primary cancers (7). Epidemiologic studies in cohorts of patients who have exhibited multiple primary cancers have contributed to our understanding of the complex etiology of human cancers and to the generating and testing of hypotheses on mechanisms of carcinogenesis (8). A salient research question is whether, or to what extent, risk patterns exist that predict metachronous primary cancers in patients with an index primary cancer of a particular organ and histopathology? The identification of specific predictive patterns of multiple primary cancers would facilitate cost-effective targeting of early detection methods or other preventive interventions. At issue is whether apparent increases in risk result from:

• common environmental and/or genetic factors in the pathogenesis of index and metachronous primary cancers;

• adverse carcinogenic effects of chemotherapy and/or radiotherapy used in the treatment of the index cancer;

• random or chance associations; and

• spurious associations that are the result of biased ascertainment of multiple primary cancers as a result of more careful medical surveillance in patients with a history of cancer when compared with a referent population.

The multicentric origin of epithelial cancers is demonstrable in patients with oral cancer who manifest an increased incidence of synchronous and metachronous carcinomas of the oropharynx, glottis, supraglottis, esophagus, and lower respiratory tract. The cumulative incidence of metachronous primary cancers in the aerodigestive tract, 6 months after diagnosis of the index primary oral cavity cancer, has been estimated as 3% to 4% per year. These patients exhibit mucosal dysplasias and intraepithelial carcinomas in areas separate from the index primary cancer. The concept of mucosal “field cancerization” would be consistent with protracted exposures to one or more carcinogenic agents (9, 10).

Worldwide, the established causal factors for head and neck cancers include tobacco smoking, heavy alcohol drinking, chewing of betel nut and tobacco, and infection with human papillomavirus. Epidemiologic studies conducted in many countries have established that the risks of lung, upper aerodigestive, stomach, pancreatic, liver, kidney, urinary bladder, and uterine cervical carcinomas; colorectal adenoma and carcinoma; and myelogenous leukemia are significantly increased among cigarette smokers. Approximately 30% of the cancer deaths in the United States are attributable to cigarette smoking, and ethyl alcohol interacts with tobacco smoking in the pathogenesis of >75% of carcinomas of the upper aerodigestive tract (11).

Epidemiologic, experimental, and molecular biological studies have provided a conceptual foundation for multistep carcinogenesis. Experimental models of chemical carcinogenesis have shown distinctive phases of initiation, promotion, neoplastic cell transformation or conversion, and progression resulting from a sequential accumulation of adverse genetic and epigenetic events (12). Chemical agents or mixtures that are tumor promoters are generally not mutagenic and facilitate clonal expansion of initiated cells that are at risk of further genetic mutational events. Notable agents with tumor-promoting properties are tobacco smoke condensate, ethanol, sex steroid hormones, bile acids, dioxin, and agents that are chronic irritants or evoke an inflammatory response (13). Epigenetic events are comprised of stable and heritable, or potentially heritable, alterations in gene expression that do not entail a change in DNA sequence. Patterns of DNA methylation in CpG dinucleotide “island” sequences in the vicinity of a promoter region of a tumor-suppressor gene provide a mechanism by which the gene may be silenced; conversely, genomic demethylation or hypomethylation of a proto-oncogene may result in hyperactivity. Environmental and stochastic factors can affect epigenetic events (14, 15).

As summarized by Hanahan and Weinberg (16), the molecular, genetic, and phenotypic hallmarks in the evolution of a population of cancer cells include unrestrained proliferation, subversion of genetic mechanisms that control apoptosis, interaction of a reactive stroma in the microenvironment of neoplastic tissue that stimulates angiogenesis, the capacity to invade through the basement membrane of the extracellular matrix, and dissemination and proliferation at metastatic sites.

The complex molecular pathogenesis of sporadic colorectal cancer may proceed by means of at least three clinical and morphologic pathways. The three pathways include the predominant pattern of the adenoma precursor progressing to cancer (about 80-85%), defective mismatch repair with microsatellite instability (about 15%), and inflammatory bowel disease (17).

Overall, about 10% of adenomas (tubular, villous, and tubulovillous) will progress to invasive cancer over an interval of about 10 years. Histopathologic features that increase the likelihood of progression to cancer include degree of dysplasia and epithelial atypia, predominance of villous histology, and surface diameter exceeding 1 cm (18). The progression from adenoma of increasing size to dysplasia to carcinoma is anticipated by a cascade of molecular genetic events, including mutations of KRAS oncogene on chromosome 12p and mutations of tumor suppressor genes on 5q [adenomatous polyposis coli (APC)], 17p (TP53), and 18q (DCC; ref. 19). An early mutational event resides at the APC gene locus at 5q, which is correlated with an aberrant proliferation rate in the upper one third of epithelial crypts in otherwise normal-appearing mucosa. Mutation of the APC gene alters critical functions of cell adhesion, cell migration, and apoptosis (20, 21). The truncated APC protein results in increased concentrations of β-catenin, which after adhering to T-cell factor, augments transcription of specific oncogenes. In murine as well as human colon cancer, APC mutations seem earlier in carcinogenesis than mutational and epigenetic events involving KRAS and TP53 (22, 23).

An alternative carcinogenic pathway involves mutations in DNA mismatch repair genes manifested by high-level microsatellite instability and the appearance of mucin-producing poorly differentiated cancers more frequently located in the proximal colon (24). Microsatellites represent a limited number of DNA sequences repeated in tandem that show polymorphisms in length throughout the human genome. Because they are repetitive, microsatellites are prone to strand slippage and errors in replication. Microsatellite instability occurs when there is dysfunction of the DNA mismatch-repair enzyme system. The alterations in microsatellite length are shown as band shifts on gel electrophoresis (25). In the majority of sporadic colorectal cancers associated with microsatellite instability, repression of transcription by hypermethylation of the promoter region most commonly of the hMLH1 mismatch repair gene has been described. In Lynch syndrome patients with HNPCC, 90% of germ line mutations reside at hMSH2 and hMLH1 loci (26, 27).

Inflammatory bowel disease provides an illustrative mechanism of increased risk of neoplasia experienced by patients with chronic inflammation in gastrointestinal tissues. The pathogenesis of inflammatory bowel disease suggests that there has been an aberrant immune response to luminal indigenous microorganisms or ingested foreign antigens. The inflammatory response mediated by CD4⁺ T cells and expressed in the intestinal epithelium is exaggerated, unregulated, and cytotoxic. The mucosal ulcerations would allow for microbial flora to gain access to submucosal lymphoid tissue and thus trigger an immune response (28-30). The risk of colorectal cancer in ulcerative colitis patients increases with longer duration of disease and with extent of involvement of the large intestine. In addition, the risk increases with the appearance of higher-grade dysplastic lesions. Dysplastic areas may seem flat or polypoid, localized, multifocal, or diffuse (31).

Colorectal cancer chemopreventive trials are focusing on targeting segments of the population at increased risk (e.g., history of adenomas or family history) by prescribing nutritional supplements of calcium, vitamin D, and/or folic acid; or treating with nonsteroidal anti-inflammatory agents; or assessing benefits of postmenopausal estrogen replacement therapy (32-36). Combinations of concurrent administration of chemopreventive agents may be more effective than one agent acting independently. As reported by Grau et al., in an analysis of trials in patients with recurrent adenomas, the combination of supplemental calcium (1,200 mg daily of elemental calcium) and “frequent use” of nonsteroidal anti-inflammatory drugs resulted in a reduction of risk of advanced adenomas (i.e., severe dysplasia, villous elements, and 1 cm or larger in diameter) of 65% (P_interaction = 0.01). Similarly, in a trial of daily aspirin (81 mg) and supplemental calcium, there was an 80% risk reduction (37). If corroborated in future randomized clinical trials, investigation of potential synergistic mechanisms in relationship to the pathologic characteristics and anatomic location of colorectal adenomas would be appropriate.

Optimal dietary and supplemental sources of folate over a protracted period of time have been reported to be associated with a 20% to 50% reduction in risk of colorectal adenomas and carcinomas (38-40). In a pooled analysis of nine cohort studies conducted in North America and Europe, the top quintile of folate intake, when compared with the lowest quintile, was associated with a 21% lower risk of colorectal cancer (multivariate-adjusted relative risk, 0.79; 95% confidence interval, 0.70-0.89; ref. 41). The results in various case-control and cohort studies were in particular compelling when folate and ethanol intakes were considered jointly. Among individuals with a low intake of folate and a high intake of ethanol, the risks of colorectal adenoma and cancer were increased >2-fold (42, 43). In patients with chronic excessive consumption of ethanol, folate deficiency may occur because of inadequate dietary intake, malabsorption, or defects in folate-binding proteins or in the enterohepatic absorption and recirculation of folate.

Folic acid has multiple metabolic effects, including biosynthesis of purines and pyrimidines and production of S-adenosylmethionine, a requisite metabolite for methylation of DNA cytosines, histones, and phospholipids. Methylene tetrahydrofolate reductase catalyzes the reduction of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate, the principal folate congener supplied to tissues, and a vital methyl donor, in concert with vitamins B¹² and B⁶, and dietary choline upon conversion to betaine. The folate congener (5,10-methylenetetrahydrofolate) is the source of a complementary vital function (i.e., the transfer of a methylene group to deoxyuridylate for the synthesis of thymidylate in the DNA nucleotide sequence). Thus, when in the presence of folate deficiency there is increasing methylene tetrahydrofolate reductase activity, the resulting increase in 5-methyltetrahydrofolate occurs at the “cost” of reduced thymidylate synthesis (44-46).

A thermolabile variant of methylene tetrahydrofolate reductase is associated with reduced enzyme activity. The molecular structure of the methylene tetrahydrofolate reductase variant is a substitution at nucleotide 677 of a cytosine (C) for thymine (T), which converts an alanine to valine amino acid. In relationship to low folate intake, a homozygous TT variant, when compared with CC or CT genotypes, would be associated with markedly elevated levels of homocysteine, a putative risk factor for cardiovascular disease (47). However, with adequate intake of folate and vitamin B¹², TT individuals are reported to be at decreased risk of colorectal cancer when compared with the CC or CT genotype (48). Various studies have suggested that individuals with TT genotype and low intake of folate were more susceptible to the carcinogenic effects of alcohol (49-51). A plausible inference based on the apparent gene-nutrient interactions would be that reduced activity of methylene tetrahydrofolate reductase in TT carriers, with normal folate intake, favored 5,10-methylenetetrahydrofolate, the pathway affecting DNA synthesis and repair.

Recurrent or persistent inflammation, resulting from an exposure to a specific microbial or chemical agent, ionizing or UV radiation, or physical injury, may induce, promote, or influence susceptibility to carcinogenesis by causing DNA damage, inciting aberrant tissue reparative proliferation, and/or by creating a stromal “soil” that is enriched with proinflammatory cytokines and growth factors (e.g., vascular endothelial growth factor). Various pathologic conditions prominently associated with chronic inflammation are linked with cancers in gastrointestinal, respiratory, anogenital, and lymphoid organs and tissues (Table 1; ref. 52).

Inflammation involves a complex reaction to a toxic agent in vascularized tissue resulting in the influx of circulating leukocytes, connective tissue cells, and extracellular matrix constituents. Prominent among the inflammatory cells are macrophages, lymphocytes, granulocytes, mast cells and, dendritic cells, in combination with fibroblasts. Activated macrophages and lymphocytes release inflammatory mediators or cytokines that stimulate the immune system with potential adverse consequences on genomic stability, matrix metabolism, fibroblast activation, and angiogenesis (53). Fibroblasts are primarily responsible for production and remodeling of the extracellular matrix, as well as coparticipating with inflammatory cells in the production of paracrine growth factors that regulate cell proliferation (54).

Phases of cancer promotion and progression are intimately linked to a dysfunctional cytokine network. Cytokines represent a family of biological response modifiers, including interleukins, chemokines, IFNs, growth factors, leukocyte colony-stimulating factors, and tumor necrosis factor-α. Chemokines stimulate leukocyte recruitment and migration as part of the host response to foreign antigens. Dvorak, in contrasting mechanisms involved in physiologic wound healing with those in neoplastic cell proliferation and invasion, referred to tumors as “wounds that do not heal” (55, 56). The dysfunctional network of cytokines, chemokines, growth factors, and extracellular proteases interact with cell surface receptors and target genes that ultimately influence tumor cell proliferation and survival, angiogenesis, and migration of tumor cells into the stromal matrix (57, 58). An additionally important component in the cytokine signaling pathway is activation of the nuclear transcription factor-κB. Activation of nuclear transcription factor-κB by proinflammatory cytokines is expressed in a second messenger system, with potential pathologic effects on tumor promotion (59).

Chronic inflammation and the metabolic products of phagocytosis are often accompanied by the excessive formation of reactive oxygen and nitrogen species that are potentially damaging to DNA, lipoproteins, and cell membranes (60). Inflammatory cells also release metabolites of arachidonic acid, or eicosanoids, including prostanoids or prostaglandins and leukotrienes. The cyclooxygenases are key enzymes that control rate-limiting steps in prostaglandin synthesis. The expression of the isoform cyclooxygenase-2 is induced by inflammatory and neoplastic cells, and metabolites produced by the action of cyclooxygenase-2 on arachidonic acid have been shown to affect carcinogenic pathways (61-63).

Future studies of the complex cascade of cellular and humoral factors participating in the chronic inflammatory response will further understanding of the pathogenesis of various cancers and provide a rationale for targeted chemopreventive interventions. Preclinical and clinical investigations of agents that target the tumor microenvironment and epithelial-stromal signaling pathways are evaluating, for example, the efficacy of inducers of apoptosis and inhibitors of angiogenesis, collagen synthesis and cyclooxygenase-2-dependent and prostaglandin-independent pathways (64-66). The anticarcinogenic properties of nonsteroidal anti-inflammatory drugs in the colorectum are well documented by experimental and epidemiologic studies and by randomized trials. Regular use of aspirin for at least 10 years, for example, has been associated with a reduced risk of colorectal adenomas in addition to adenocarcinomas, suggesting an early-stage mechanistic pathway in colorectal neoplasia (33, 67-69).

The prevalence of obesity has been increasing over the past two decades in many industrialized countries and in urban areas of developing countries and is incurring substantial public health and economic costs. In 2003 to 2004, among adults in the United States, ages ≥20 years, 32% were classified as obese (body mass index ≥ 30 kg/m²) and 34% as overweight (body mass index = 25-29.9 kg/m²). In the National Health and Nutrition Examination Survey of children and adolescents ages 2 to 19 years, conducted in 2003 to 2004, ∼17% were classified as overweight; this represented, over the past 25 years, more than doubling of prevalence rates among children under 12 years of age and tripling among those 12 to 19 years of age. Among overweight children and adolescents, the younger the age of onset, the higher the correlation with obesity in adulthood (70, 71).

Results from epidemiologic studies have indicated that central or visceral adiposity [waist circumference >40 in. (102 cm) in men and >35 in. (88 cm) in women] is associated with increased incidence or mortality from cancers of the endometrium, breast (postmenopausal women), kidney, gallbladder, pancreas, esophagus and gastric cardia (adenocarcinoma), colon, and prostate (Table 2). Other suggested organ sites at increased risk include thyroid, multiple myeloma, and liver. In the American Cancer Society Cancer Prevention Study II, 15% to 20% of cancer deaths in women and 10% to 14% in men were attributed to being overweight or obese (72, 73).

The pathophysiology of obesity may be viewed as a dysfunctional disparity between energy intake (kilocalories) in relation to mechanisms controlling appetite and satiety and basal and adaptive energy expenditure, expressed as metabolic equivalents per unit of time, or thermogenesis. Body mass is determined by the interaction of genetic, dietary, and other lifestyle and psychosocial factors (74). The marked increase in the global prevalence of obesity during the past two decades underscores the significance of trends in dietary practices, sedentary lifestyle, and corporate production and marketing of foods, rather than genetic determinants (75, 76).

Cardiovascular disease, cerebrovascular disease, type II diabetes, and neoplastic diseases associated with obesity are highly correlated with the presence of total body fat mass and intra-abdominal (visceral) fat. Patients with abdominal obesity tend to have elevated glucose and insulin levels during an oral glucose tolerance test and high rates of lipolysis. Lipolysis is the catabolic process leading to the breakdown of triglycerides into fatty acids and glycerol. The metabolic and endocrine effects of obesity are intimately connected with elevated production of free fatty acids from triglycerides in adipose tissue and impaired insulin-mediated glucose metabolism. Being overweight or obese and sedentary decreases insulin sensitivity in target tissues as in the liver, skeletal muscle, and fat and is accompanied by hyperinsulinemia (77-79).

In Table 3, the metabolic and endocrine systemic effects of obesity are listed, suggesting possible mechanisms for tumorigenesis in hormone-dependent organs and colon and gallbladder. The adipocyte in secreting free fatty acids and hormones, such as leptin, adiponectin, and resistin, and cytokines, such as interleukin-6 and tumor necrosis factor-α, participates as a central mediator of the insulin-resistance syndrome in obese individuals (80). The physiologic dynamics of insulin resistance and compensatory hyperinsulinemia greatly increases the risk of metabolic sequelae of residual glucose intolerance, high plasma triglyceride, low high-density lipoprotein cholesterol, hypertension, and increased levels of proinflammatory cytokines and of pro-coagulation factors, such as fibrinogen and plasminogen activator inhibitor. It has been estimated that ∼23% or 47 million adults in the United States report symptoms and clinical sequelae of the metabolic or insulin resistance syndrome (81, 82).

In 1996, the American Cancer Society projected the results of targeting various risk factors on cancer incidence and mortality in the United States over a 25-year period. Byers et al. (83) examined the feasibility of achieving the targeted goals of 25% reduction in cancer incidence and 50% reduction in cancer mortality in the year, 2015, by affecting trends in prevalence proportions of various behavioral risk factors. The risk factors included current smoking behavior, dietary patterns of high fat or low fruit and vegetable consumption, frequent or heavy use of alcohol, nonuse of antagonists of endogenous estrogens (e.g., tamoxifen), failure to have screening with sigmoidoscopy or mammography, or failure to receive optimal treatment for cancer. If current trends in the rate of change in prevalence of the risk factors were to continue, by the year 2015, the linear projection cumulatively over 25 years would be a 13% decline in cancer incidence (3.2 million fewer cases) and a 21% decline in cancer mortality (160,000 fewer cancer deaths).

In a recent American Cancer Society report, it was concluded that >50% of the cancer deaths in the United States could be prevented with modification in lifestyle risk factors and greater use of cost-effective cancer screening tests (84). In the National Cancer Institute Cancer Trends Progress Report (2005), it was estimated that 20% to 30% of cancer incidence in the United States would be attributed to obesity and lack of physical activity (85). Risk reduction associated with maintaining moderate or vigorous physical activity, independently of weight reduction in overweight or obese individuals, has been shown consistently in studies of colon cancer and adenomatous polyps and to a somewhat lesser degree, of breast cancer (86-91).

A collaborating group participating in the Comparative Risk Assessment Project reviewed behavioral and environmental risk factors that would offer “the best option for reducing the large and increasing burden of cancer worldwide” (92). Risk factors selected were those with a high likelihood of causality, reasonably complete data on population exposure levels, and a leading potentially modifiable cause of worldwide or regional disease burden. Tobacco was estimated to have caused 21% of cancer deaths worldwide, with a higher attributable fraction of 29% in high-income countries and 18% in low- and middle-income regions. Regular use of moderate to high levels of alcohol (i.e., exceeding 25 g of ethanol per day) accounted for an estimated 5% of global cancer deaths, including upper aerodigestive organ cancers, liver cancer and breast cancer (93, 94). Overweight and obesity and physical inactivity (i.e., <2.5 hours per week of moderate intensity activity or <4,000 kJ/wk) were causally associated with <5% of all cancer deaths worldwide, but of 20% of global colorectal and breast cancer deaths.

In Table 4, priority is assigned to potentially modifiable lifestyle risk factors that affect the global burden of cancer. As suggested in the publication by Byers et al., ideally, the control of the burden of cancer would be achievable through the integration of population-based, health-maintaining lifestyle behavioral practices, and access to affordable and effective periodic screening and early detection examinations in conjunction with comprehensive treatment services (83). In arriving at evidence-based recommendations in cancer screening, clinicians and public health professionals are reviewing alternative methods that are culturally sensitive and in tune with individual preferences (95). Of particular interest is the substantial effect of exposures to microbial agents that potentially may be lessened by the availability and use of effective antimicrobial therapy, immunization to reduce rates of persistent infection, and the application of global surveillance and early detection methods in high-risk populations. Primary preventive measures that effectively control the persistence or transmission of infectious agents causing gastric cancer (H. pylori), hepatocellular cancer (hepatitis B virus), and uterine cervical cancer (human papillomaviruses) are high-priority strategic interventions. Although further understanding of causal factors and pathways as a result of experimental and epidemiologic research will continue to contribute to innovative preventive strategies, there are already at hand compelling opportunities for behavioral and therapeutic interventions that have the potential to reduce substantially the global cancer burden.