Unprecedented Opportunities and Promise for Cancer Prevention Research

That cancer exacts an enormous toll on the world population has been recognized for a long time. So great is the direct and collateral damage of this disease to the U.S. population that former U.S. President Richard Nixon declared a “war on cancer,” which led to the National Cancer Act of 1971 supporting research to provide a greater understanding of this enemy and better weapons for combating it. The ensuing decades of research have provided astonishing amounts of information on the molecular alterations that underlie different cancers and inform potential new avenues for their control.

Perhaps the most potentially far-reaching information to emerge from these studies points to cancer prevention as the most efficient way of waging this war. Why would prevention be more efficient than therapy? In aggregate, cancer studies have described in great depth and detail the derailment of cellular security systems and signaling gone awry, which is coupled with the daunting ability of tumor cells to mutate and evade therapeutic agents. Again and again, cancer cells can evade chemotherapeutic agents and generate drug-resistant variants that progress to kill the patient. Yet, before cells transform to the point of needing chemotherapy, they likely exist in states that are much more amenable to interventions that control their future directions. New research efforts are identifying the factors that control these future directions and open the opportunity for modulating the context and outcome with respect to cancer development. Thus, these new insights suggest that prevention of cancer is a rational and ultimately effective method for dealing with this set of diseases.

It is useful here to define the range of concepts and approaches encompassed by the term “cancer prevention.” Several steps within carcinogenic initiation and progression are appropriate targets for cancer prevention, which may take place at primary and secondary levels as described in detail below. Primary prevention is the avoidance of exposure to carcinogens or carcinogenic processes and frequently is directed at the population as a whole or at a rather broadly defined population. Although this area of prevention, for example through dietary changes or exercise, is perhaps most associated with the term cancer prevention among people not working in or very familiar with the field, it is important to remember that cancer prevention is not limited to this category. Examples of primary prevention are avoidance of tobacco smoke and psychosocial stress, chemically blocking exposure to pervasive environmental carcinogens such as aflatoxin, and induction of phase 1 or 2 enzymes to alter carcinogen metabolism. Another important area of primary prevention is natural agent chemoprevention in people with no signs of premalignancy (an approach that may be most effective in nutrient-deficient populations). Secondary prevention includes screening for and early detection of premalignancy or early, subclinical cancer and intervention to prevent progression to invasive disease or symptomatic cancer. Secondary prevention aims to eliminate or reduce existing risk in a generally more-defined, more-specified risk population (compared with primary prevention). The broad spectrum of secondary prevention comprises molecular high-risk settings (not necessarily including premalignant lesions) such as germline BRCA mutation carriers (e.g., prophylactic mastectomy) and colorectal adenomas [e.g., polypectomy, nonsteroidal anti-inflammatory drugs (NSAIDs)]. Last, tertiary prevention has been defined as preventing or controlling the symptoms and morbidity of cancer or the morbidity of cancer therapy (e.g., controlling pain with morphine or nausea with Compazine) and further as preventing recurrence or second primary cancer (e.g., with molecular-targeted chemoprevention) in patients after successful definitive treatment of early-stage cancer. These definitions of primary, secondary, and tertiary prevention conform conceptually with a gradient of cancer risk from normal or low, through moderate or high, to risk realized in established cancer.

In general contrast to participation in tertiary prevention or cancer therapy, enthusiasm for primary or secondary cancer prevention can be problematic. Presumptively healthy and disease-free (cancer or premalignancy) individuals sometimes are the least excited or motivated to participate in preventive interventions because the condition has not yet occurred (although in some ways this is an artificial boundary because current research indicates that premalignant lesions are common in people with no symptoms or diagnosis of disease). Most familiar to the public in the area of primary prevention are the recommendations to limit alcohol intake, food consumption, and stress conditions. These recommendations also assert positive effects from eating the proper foods and nutrients and obtaining adequate physical exercise. Although promulgated for a considerable length of time, these recommendations are still minimally effective in changing behavior, perhaps in part for lack of mechanistic understanding. New promise for primary cancer prevention, however, is developing out of behavioral linked with molecular science, for example, in the area of obesity and overweight, energy balance, and caloric restriction. Recent science has discovered the molecular basis, particularly actions of AMP-activated protein kinase and Akt pathways, of energy-balance effects, and caloric restriction is the most consistently positive approach for preventing cancer in animals. The overall study of controlling obesity and overweight to reduce cancer risk built on earlier work by Tannenbaum (1–3) and others in caloric restriction and the effects of various nutrients on animal model carcinogenesis.

The absence of invasive disease also can impede enthusiasm for secondary prevention, which can involve the impediment of invasive procedures (e.g., biopsy) to monitor and measure chemopreventive effects (e.g., in women with breast premalignancy) or the impediment of adverse drug effects that raise the omission bias, or the tendency to do nothing, although this avoids a lesser harm (e.g., a small increased risk of endometrial cancer with tamoxifen), rather than do something, although doing something would bring a greater benefit (e.g., a 50% reduced risk of breast cancer with tamoxifen).

Notwithstanding the aforementioned problems with participation in primary prevention, participation in primary prevention trials of natural agents such as vitamins, minerals, and antioxidants has been strong [e.g., 35,533 men randomized in the Selenium and Vitamin E (prostate) Cancer Prevention Trial]. This area of prevention, however, has encountered disappointing clinical results thus far. A recent blow was the very large, negative-neutral Selenium and Vitamin E Cancer Prevention Trial (4), which also is turning out to be a phoenix for natural agent prevention. The clinical ashes of the Selenium and Vitamin E Cancer Prevention Trial have stimulated tremendous interest in the science behind natural agents, such as pharmacogenomics, for improving their clinical batting average. For example, pharmacogenomic studies have shown that polymorphisms of manganese superoxide dismutase (SOD2) and SEP15 (5) may identify men more likely to benefit from selenium for prostate cancer prevention, thus improving the overall prospects of natural agents for producing future clinical impact.

A critical concept underlying cancer prevention, as well as cardiovascular disease prevention, is that risk and pre–definitive end point conditions are diseases. Cardiology considers high cholesterol and atherogenesis to be diseases, and treating them with statins and aspirin is akin to primary and secondary cancer prevention. As it happens, there also is substantial interest in aspirin and statins [not to mention selective estrogen receptor modulators (SERMs) and vaccines] for cancer prevention, which highlights the cross-discipline aspects of many aging-related diseases including cardiovascular and neurodegenerative diseases, osteoporosis, macular degeneration, and cancer.

New molecular information suggests that powerful prevention efforts may come at the level of secondary prevention in the setting of premalignant lesions. Accumulating evidence suggests that premalignant lesions may be more common than previously appreciated. Studies by several investigators have shown a surprising prevalence of undetected premalignant lesions in “disease-free” individuals. Breast tissue provides a dramatic example of this phenomenon. For example, in one study of double mastectomy specimens from medicolegal autopsies, in which the cause of death was unrelated to breast cancer, nearly one-third (32%) harbored hyperplastic lesions, over one-quarter (27%) contained atypical ductal hyperplasia, almost one-fifth (18%) showed ductal carcinoma in situ, and 2% had invasive breast cancer (6). Furthermore, almost half of the women with ductal carcinoma in situ had bilateral (41%) and/or multifocal (45%) disease. Another study of breast samples from random autopsies confirmed this high prevalence of undetected premalignant breast lesions (7). Similar data have been collected from examinations of prostate, bladder, pancreatic, lung, and colon tissue. These data indicate that the initiation of premalignant lesions, identified by morphologic alterations within the tissue, is by no means a rare event. This abundance of undetected or subclinical premalignant lesions compels us to ask why some premalignant lesions progress to cancer and others do not, a crucial question for cancer preventive interventions.

At least part of the answer to the foregoing question probably lies in the interactions between tumor epithelial cells and their stromal neighbors and the extracellular microenvironment. It is well known that stromal components undergo dramatic changes as a premalignant lesion or cancer progresses (8, 9). When and how these changes occur are under intense study. Stromal changes extend beyond the increase in endothelial cells for angiogenic support of tumor cells and beyond the influx of immune cells into the immediate tumor environment. Previous studies have shown that fibroblasts in the vicinity of a cancer, precancer, or inflammatory lesions express altered protein profiles such as increased growth factor and matrix metalloproteinase production (10). Previous studies in both an in vivo recombinant model and an in vitro coculture system have further shown that these fibroblasts can act in a functional manner to facilitate the transition to tumorigenicity in an otherwise nontumorigenic cell (11). These effects are not detected when normal fibroblasts and epithelial cells from the same human tissue specimen are grown under the same experimental conditions as the cancer-associated cells. Given these powerful effects of stromal components, the opportunity to modulate the signals and influence disease progression in these tissues is a new area for cancer prevention. These leading-edge stromal studies add to earlier and recent studies of the microenvironment and targeting angiogenesis and inflammation (12–14).

Recent work from several different directions is indicating a causal relationship between physiologic stress and alteration of molecular markers associated with cancer phenotypes. A differentiated cell's typical response to stress (physical, chemical, or physiologic) is to activate an arrest, senescent, or death response. Only cells that have sustained alterations can bypass this stress response and continue proliferating to form a premalignant or malignant lesion. Recent work has shown that the risk of future tumor formation is low in association with tissue biopsies exhibiting the activation of a stress response and high in association with tissue biopsies exhibiting a bypass of the stress arrest. These stress-related markers are proving to be useful in stratification of risk and will aid in targeting prevention efforts.

Adult somatic stem cells are exempt from this activation of stress-induced arrest processes. Programmed to respond to stress in a different manner than described above, these cells are recruited to sites of stress, stimulated to proliferate and reprogram cell fate, and positioned to reconstitute the injured or stressed tissue site. It has been hypothesized that these properties of adult somatic stem cells may qualify them as candidates for premalignant progression, for generating a group of cells with a minimal number of mutations but that manifest several premalignant properties. Because much of the control of stem cell phenotypes depends on epigenetic reprogramming, epigenetics has become a very promising area of basic and preclinical molecular-targeted prevention science. An important preclinical study of epigenetics showed that reversing methylation can prevent intestinal neoplasia in mice (15). Current epigenetic science is focused on risk markers and developing more selective and less toxic targeting agents. Research in this area also may shed light on the target cells for malignant transformation and novel opportunities for prevention (16).

Other areas of “stress” that are potentially applicable to advancing the science of cancer prevention involve telomere biology. To put this in context first, telomerase hyperactivity has been associated with malignant cells themselves in a wide range of human cancers. But conversely, telomerase in normal cells seems to be protective against cancer development through its ability to sustain telomere maintenance and, hence, genomic stability (17). As early as 2001, studies in rare inherited genetic diseases resulting in telomerase insufficiencies revealed that compromised telomere maintenance over a person's lifetime causes increased cancer risks in humans (18–20). In broader populations without genetic telomerase diseases, several studies show that telomere shortness is associated with major cancer risk factors including smoking, inflammation, and obesity (21). For example, studies have linked shortness of telomeres (and by implication, compromised telomere maintenance), which was measured in white blood cells, to the risk of Barrett's esophagus progressing to esophageal carcinoma (22) and to the risk of gastric cancer (23). Such findings in humans mirror similar conclusions drawn from various mouse-model studies (17).

Interesting findings indicate that a different type of stress, chronic psychological stress, takes its toll on telomere maintenance in humans (24, 25). This common type of stress has been extensively linked to increased risks of cardiovascular diseases and diabetes and other metabolic syndromes, and there are intriguing potential links between this nongenetic life factor and cancer that merit systematic exploration. Behavioral stresses, such as produced by social isolation, have been elucidated as factors in cancer development and burden in recent animal studies (26, 27). A feeling of loneliness and hypervigilance may be important contributors to racial health disparities, specifically for breast cancer (28). Therefore, the intersection of telomere biology with the biology and behavioral science of stress provides a good example of how integrative, interdisciplinary science can contribute to cancer prevention approaches.

A model for how telomere maintenance could be usefully applied clinically in cancer prevention may be drawn from cardiovascular disease prevention. A West of Scotland study (29) of statins for preventing cardiovascular disease in people with one or more cardiovascular risk factors (but not symptoms of disease) provides an example (and biomarker) with potential relevance to cancer prevention. This study showed that individuals with shorter telomere length in white blood cells at the study outset were selectively protected from future heart disease by statin treatment. Because the statin-associated benefit occurred only in individuals with shorter telomeres, this biomarker not only preidentified the group with sensitivity to clinically beneficial statin effects but also showed potential for use in sparing nonsensitive individuals from the inherent risks associated with taking statins. Scientific advances in telomere biology similarly promise to have implications for cancer prevention, including behavioral/pharmacologic interventions, the development of molecular risk models, and molecular-targeted chemoprevention approaches (e.g., targeting regulation of telomerase).

Although sophisticated imaging techniques are increasing the ability to detect premalignant lesions (30), the majority of these preinvasive lesions are not associated with future tumor formation (31). Over the last four decades, investigators have tried extensively to identify clinicopathologic variables and molecular markers that may be important in predicting which premalignancy patients will progress to invasive cancer. Identification of such variables would help to prevent overtreatment (thus reducing morbidity) and undertreatment (thus reducing mortality). Several recognized lesion variables are routinely assessed in the clinic (e.g., size, nuclear grade, and surgical margins) and some have been found to have predictive value for subsequent in situ but not invasive disease. However, none have proven strong enough to fully support choices for intervention (hazard ratios of from approximately 2–5) and none have predicted a future invasive event. Although focused on associating risk with the differential expression of molecular markers, more recent studies also have found no molecular features that can distinguish between premalignant lesions that do and do not precede a tumor event. Because of the lack of distinguishing markers, current clinical practice offers all individuals diagnosed with premalignant lesions the same treatment options, from complete organ (breast or prostate) removal to “watchful waiting.” Because most premalignant lesions are not associated with subsequent invasive tumors, it is likely that many individuals diagnosed with in situ disease will be overtreated. Conversely, because even with this therapy some initially observed in situ lesions are followed by subsequent invasive carcinomas, some individuals with in situ disease may be undertreated. New approaches to identify true malignant precursors with high sensitivity and specificity are desperately needed and are being developed (31, 32).

Other current science in this area is modeling germline and somatic markers of risk and predictive markers (of agents' beneficial and toxic effects) toward personalizing cancer prevention (5, 33–38). These efforts are leading toward comprehensive, complex modeling in clinical cohorts that will depend on evolving statistical methodology, highlighting the crucial role of biostatistical researchers in developing new methods needed for studies of cancer prevention, detection, and treatment. There are many areas where biostatisticians are responding to new types of biological data and public-health concerns in developing the tools needed for study design, analysis, and interpretation, including, for example, the efficient design and analysis of studies to detect genetic associations with disease, efficient methods to detect subgroups that are particularly susceptible to an environmental exposure (or “gene-environment interaction“), methods to estimate and evaluate absolute disease risk, methods to integrate risk estimation with public-health decisions, and methods to use longitudinal biomarker data to understand evolving patterns of risk. Based on this science, clinical trials of various agents in highest risk people could be developed, early cancer detection could be enhanced, and prevention could be personalized to individuals sensitive to a particular agent and at the highest risk of cancer.

There are imposing barriers to the identification of early causal alterations in tumor cells. Because of these difficulties, many laboratories use tumor cells in culture that are derived from highly progressed metastatic lesions. Although these cells are useful for studying later steps in the transition to malignancy, they are inadequate for studying the early steps of carcinogenesis or primary or secondary prevention efforts. These tumor cells have already accumulated numerous gross, chromosomal abnormalities, as well as many molecular alterations, thus making it difficult to reconstruct the early events that lead to malignancy.

An alternative approach to identifying early causal events is to grow normal human epithelial cells in culture and then define conditions that would convert them to tumorigenicity. Hence, the development of in vitro model systems became a focused effort in the field. These models need to preserve the functional biology of transformation and allow the examination of cells throughout the process of initiation and progression. Efforts from several laboratories over the past three decades have provided culture systems that allow for the isolation and propagation of normal human epithelial cells, for example from human breast tissue (39, 40), and allow investigators to examine the processes required for transformation. Some studies have treated breast cells with chemical carcinogens and obtained mutant mammary epithelial cells that are tumorigenic (41). Other laboratories have introduced viral carcinogens such as the SV40 large T antigen, in conjunction with other molecular alterations (e.g., activated ras and overexpression of telomerase), to obtain malignant cells (42). An alternative approach has been to identify a variant subpopulation of human epithelial cells that possess properties of premalignant lesions without any exposure of the cells to viral, chemical, or physical carcinogens (43). These variant cells exist in vivo and express markers often seen in tumors (44, 45).

A limitation of this approach is the scarcity of disease-free tissue. Recognizing this barrier, the U.S. National Cancer Institute is currently launching a program to procure human tissue at each stage of the disease continuum, starting with disease-free tissue and encompassing malignant tissues. This initiative should support novel, previously impossible prevention studies. Although surrogate studies in murine cells and tissues have been educational about general principles of malignant transformation, critical details of cellular signaling are not conserved between mouse and human tissues. Therefore, human tissue is needed.

A rich stream of scientific discovery has informed past clinical advances and impact of cancer prevention. Presented in the following paragraphs, a few of these many exciting, leading scientific channels are loosely grouped as molecular targeting, vaccines, and epidemiology and behavioral science.

Molecular targeting involving sex hormones has produced major clinical advances. The SERMs tamoxifen and raloxifene reduced preinvasive and/or invasive ER-positive breast cancer risk by ∼50% in women with a higher-than-average risk (46, 47). This overall reduction comprised a far higher reduction (69%) in ER-positive disease, consistent with tamoxifen effects in premalignancy (atypical hyperplasia, ductal carcinoma in situ), adjuvant therapy (contralateral second primary cancers and recurrence), and advanced cancer therapy (48). Long-term (post–primary report) follow-up within the large primary prevention trials of these SERMs highlights the important role that such extended analyses can have in strengthening and/or modifying the interpretation of primary findings. The major science underlying these clinical impacts included work of Lathrop and Loeb in the early 1900s (49) showing that castration (oophorectomy) in female mice reduced mammary tumor incidence, follow-on clinical studies of oophorectomy in genetically high-risk women (50–52), and the discovery of the ER in 1966 (53) and tamoxifen prevention of mammary tumors in rats in 1974 (54). There also have been clinical advances with large prevention trials of the 5-α-reductase inhibitors finasteride and dutasteride (55–57) that sprang from work of Huggins (58) in the 1940s showing that castration benefitted patients with metastatic prostate cancer and subsequent study of androgen metabolism and signaling. The current science in hormonal prevention is working to develop newer and better SERMs and other approaches for interfering with estrogen metabolism, such as with aromatase inhibitors for reducing ER-positive breast cancer risk. New approaches for the much-needed prevention of ER-negative breast cancer include targeting HER1/2 signaling.

Clinical trials have established the preventive activity of cyclooxygenase-2 (COX-2)–specific inhibitors and other NSAIDs in familial adenomatous polyposis (59) and sporadic colorectal adenomas (60, 61) and show that the NSAID aspirin may reduce colorectal cancer risk (62). These clinical advances built on studies showing that colorectal screening and polypectomy can reduce colorectal cancer by 90% (63), epidemiology showing consistent NSAID associations with reduced colorectal adenoma and cancer risk and mortality, and work of the mid-1990s showing COX-2 upregulation in human colorectal adenomas (64) and that COX-2 targeting was effective in APC-knockout mice (65). Current scientific research involving the signaling pathways targeted by these agents includes work on germline and somatic molecular predictors of NSAID activity (66–68), work on COX-2 regulation and downstream and related targets (69, 70), and work to elucidate the molecular risk factors involved in the highly publicized adverse cardiovascular effects of COX-2 inhibitors (71), which have essentially stopped interest in these agents for clinical cancer prevention. The problem of unexpected serious adverse effects also has limited public acceptance of the other highly active targeted agents tamoxifen for breast cancer prevention (largely because of an increased risk of endometrial cancer; ref. 72) and finasteride for prostate cancer prevention (largely because of an apparently increased risk of high-grade prostate cancer; ref. 56).

A major clinical trial of combined difluoromethylornithine and the NSAID sulindac (73) prevented colorectal adenomas/polyps and is one approach that began to address the issue of adverse events associated with molecular-targeted cancer prevention. This trial built on and validated earlier concepts of combination chemoprevention (74) and built on the science behind COX-2/NSAIDs cited above. The promise of this approach is synergistic or additive preventive effects when the agents are combined, which allows the lowest active dose of each agent and thus lowers the potential for adverse effects. The toxicity of molecular-targeted prevention is being addressed by other approaches as well, including more selective agents (e.g., raloxifene, which has a lower endometrial neoplasia risk than does tamoxifen; ref. 47), local delivery (e.g., intravesical rapamycin; ref. 75), and intermittent, short-term regimens (e.g., with tumor antigen vaccines or other approaches in colorectal cancer prevention; refs. 76, 77).

One of the most successful approaches for preventing disease, vaccines prevent ∼6 million annual deaths worldwide. The science of vaccines is one of the most promising areas of ongoing cancer prevention research. Clinical studies established the efficacy of hepatitis B virus (HBV) vaccination of children in preventing hepatocellular cancer in Taiwan (78, 79). The science driving this clinical advance included Blumberg's discovery of HBV in 1967 (80) and its link with hepatocellular carcinoma (81). clinical trials established that human papillomavirus (HPV) vaccines reduced the risks of cervical, vulvar, and vaginal neoplasia (82, 83), and HPV screening has led to major advances in early detection and surgical prevention of cervical cancer (84). These vaccines generate mainly antibody (type II) responses and were effective primarily before and not after HPV exposure. These clinical impacts were driven by zur Hausen's work (85) showing the link between HPV and cervical cancer in 1974 and the discovery of the first specific human HPV (HPV-16) in cervical cancer patients (86). Recent studies have documented a causal link between HPV-16 and oropharyngeal carcinogenesis (87) and suggested that a well-established racial disparity in the outcome of oropharyngeal cancer treatment is related to a lower prevalence in blacks of HPV-positive oropharyngeal cancer, which responds better to therapy (88). The importance of this work is highlighted by an alarming recent increase in oropharyngeal cancer incidence in the United States and other countries, which has important implications for HPV vaccines in this setting.

Another approach is endogenous tumor antigen vaccines (which induce a predominant type I cellular response), which is a burgeoning area of scientific research with potential applications to cancer prevention. This approach has been used in advanced cancer, unsuccessfully, but has shown recent promising results in the adjuvant setting or in minimal cancer and so may work in premalignancy/prevention settings. Current prevention science involving antitumor vaccines is studying HER2 (breast) and MUC1 (colon rectum) antigen vaccines in mice (77, 89, 90). An important potential advantage of cancer prevention vaccines is limited toxicity due to short-term dosing, which is being used with the antivirus vaccines and will be used with tumor antigen vaccines if, as planned, they can sustain a tumor antigen–specific immune response after only a few doses. Recent data indicate that a new type of HPV vaccine (eliciting a predominant type I cellular response) is active against HPV-associated vulvar intraepithelial neoplasia (91). Limited, short-term dosing regimens differentiate vaccines from other active chemopreventive agents such as tamoxifen or COX-2 inhibitors, which require long-term, daily dosing to achieve effect.

Defending against infection-related cancer includes approaches besides vaccination. For example, a clinical trial of combined antibiotics eradicated Helicobacter pylori in patients and reduced rates of new gastric cancer (92). This clinical advance was built on the 1984 science of Marshall and Warren showing the link of Helicobacter pylori with duodenal and gastric diseases (93). Although only 20% of current cancers are known to be infection-related, the portfolio of such cancers is growing (94). For example, the incidence of HPV-related oropharyngeal cancer is increasing dramatically (87) and a link between Merkel polyomavirus and Merkel cell carcinoma is emerging (95), highlighting the increasing importance of vaccines and other anti-infection approaches for preventing cancer.

Epidemiology, behavioral science, and public policy have produced a major clinical impact in reducing smoking rates. The first pharmacologic intervention for overcoming smoking dependence, the nicotine patch, was approved by the U.S. Food and Drug Administration in 1984. This clinical impact was founded on a profound body of scientific study of the relationship of tobacco and cancer. Key epidemiologic studies linking smoking to lung cancer were reported in 1950 (96, 97), and Auerbach (98) established the histopathologic link of smoking with bronchial epithelial changes in 1957, leading to important studies of tobacco carcinogens and molecular field effects including new molecular risk models, target discovery, and preventive agent development approaches (99–101). The addictive nature of smoking was established in the 1990s, including the important link of tobacco dependence with germline genetic variation (102). Current work in this area is focused on developing better pharmacologic means, including nicotine vaccines, for overcoming tobacco dependence (103).

The previous paragraphs describe examples of great science leading to great clinical impact. Not all science leads to clinical homeruns, however, and it is crucial that we accept some scientific dead ends in the overall quest to push back the frontiers of cancer prevention science (104). The not inconsiderable discovery of the Americas by 15^th century Europeans, after all, began as a misguided effort to discover an oceanic trade route between Europe and the Indies.

Cancer biology intersects with several areas and disciplines of cancer prevention, pointing to opportunities for and the importance of integrative, interdisciplinary efforts to advance clinical cancer prevention through hard-won science. Current studies are elucidating the extent and effect of premalignancy and the molecular underpinnings of carcinogenesis involving stromal interactions and epigenetic and other alterations. The basic study of carcinogenesis for prevention also is advancing with the development of in vitro model systems of the progression of normal human epithelial cells to tumorigenesis. Molecular risk-stratification and pharmacogenomic approaches promise to identify the populations with the greatest need for, and the greatest potential to benefit from, clinical cancer prevention. Cancer prevention science and practice will gain from better educating and communicating to (a) the general research community about the mechanistic underpinnings of prevention advances such as linking obesity/overweight to cancer development and (b) everyone, especially at-risk populations and their primary care providers, about the overall benefits, sight of which can be lost in the face of adverse aspects, of clinical advances such as colorectal screening (which reduces colorectal neoplasia by up to 90%), and tamoxifen (which reduces breast cancer risk by 50%).

The clinical impact of cancer prevention in several areas, for example, molecular-targeted drugs in preventing various major cancers, vaccines in preventing virus-related cancers, and smoking control measures in reducing lung cancer risk, not only facilitates new science on the aforementioned and many more important fronts but also has highlighted the tremendous stage-setting science that preceded the impact. From epidemiologists to behavioral and basic scientists to clinical trialists to biostatisticians—every prevention discipline contributes key research to this science.

No potential conflicts of interest were disclosed.