Seminars in Nursing, Vol 21, No 4 (November), 2005: pp 229-235 229

OBJECTIVE:

To summarize the scientific prin- ciples underlying preven- tion. PRINCIPLES OF

DATA SOURCES: Articles, text books, personal com- CANCER munications, and experience. CONCLUSION: PREVENTION The scientific basis of cancer pre- vention is complex and involves experimental and epidemiologic approaches and clinical trials. FRANK L. MEYSKENS,JR AND PATRICIA TULLY IMPLICATIONS FOR NURSING PRACTICE: ANCER prevention has classically encompassed As more information becomes three large areas of clinical practice: prevention, available regarding proven and screening, and early detection. Several excellent potential cancer-prevention strate- reviews and book chapters have been published gies, oncology nurses are regu- involving these topics.1-3 The importance of bio- larly called upon to guide patients Clogical and molecular markers as potential surrogates have been and others in making choices re- emphasized in the past several years,4,5 and more recently precan- garding preventative options. It is cers or intraepithelial neoplasia (IEN) have become a target.6,7 The important for oncology nurses to integration of the well-established approaches of prevention, stay abreast of this growing body screening, and early detection with biological measures of disease of knowledge. risk, progression, and prognosis has become the hallmark of mod- ern (see Fig 1). In this article we will review the principles of cancer prevention and will emphasize the scientific underpinnings of this emerging discipline. The principles of cancer prevention have evolved from three separate scientific disciplines: , epidemiology, and clinical trials. Many other areas From the Department of Medicine and of science and medicine, including and the behavioral Biological Chemistry, Chao Family Com- sciences, are contributing to this evolving and complex field. prehensive Cancer Center, University of California, Irvine, CA. Frank L. Meyskens, Jr, MD: Professor, Department of Medicine and Biological SCIENTIFIC PRINCIPLES Chemistry, Chao Family Comprehensive Cancer Center, University of California, Irvine, CA. Patricia Tully, RN: Research Carcinogenesis Nurse, Department of Medicine and Bio- logical Chemistry, Chao Family Compre- arcinogenesis is the study of factors that contribute to the hensive Cancer Center, University of Cali- fornia, Irvine, CA. Cpathogenesis of cancer formation, the processes that regu- Supported in part by CA 62203 from the late normal differentiation and maturation of cells, and the genetic National Institutes of Health, Bethesda, MD. and epigenetic factors that enhance or exhibit tumor formation. Address correspondence to Frank L. Meyskens, Jr, MD: 101 The City Drive, UCI The overall relationship of the process of carcinogenesis to cancer Medical Center, Chao Family Comprehen- prevention, screening, and early detection is depicted schemati- sive Cancer Center, Building 56, Room cally in Fig 1 and should be referred to liberally throughout this 215, Orange, CA 92868; e-mail: fl[email protected] article. This figure represents an integrated systematic biological and clinical approach in which the process of cancer formation is related to underlying biologic changes and the influence of the environment © 2005 Elsevier Inc. All rights reserved. and the preclinical detection of clinical outcomes by screening and 0749-2081/05/2104-$30.00/0 early detection strategies or the interruption or suppression of the doi:10.1016/j.soncn.2005.06.002 malignant process by chemoprevention. 230 MEYSKENS AND TULLY

FIGURE 1. Prevention of cancer: A synthetic biological and clinical approach.

Classically, carcinogenesis has been separated visual or endoscopic means. During progression, into three distinct phases: initiation, promotion, additional genetic damage occurs, re- and progression.8 Initiation is the first step of sult, and the gradually acquires the properties carcinogenesis and results in irreversible damage of a fully transformed including inva- to DNA by a physical, chemical, or viral agent (ie, sive and metastatic properties. a ). The events of initiation occur rap- idly. In some cases damage to DNA may be re- The Role of Genetics and Genes paired before it is “fixed” as a . In such situations, initiation does not happen and the cell An increased understanding of the malignant returns to its baseline state. In general, any inter- process in the last 15 to 20 years has established, vention that slows cellular proliferation should beyond a reasonable doubt, that hereditable and provide more time for DNA repair and lessen DNA acquired genetic changes play a role in all damage and mutation. Alternatively, initiated (Fig 2). The process is complex and the specifics cells are damaged and, if recognized as such, un- vary from cancer to cancer, but the essential dergo (“cell death”) and are removed elements are probably quite similar.8 Vis-à-vis the before they can evolve into a malignancy. malignant process, two major changes occur to Progression occurs over a prolonged period of genes: a loss or gain in function. Genes that sup- time (many years in ) and results in ex- press proliferation or enhance apoptosis are pansion of an initiated clone of cells without fur- known as tumor suppressor genes.9,10 Genes that ther or minimal genetic change. Exogenous or enhance proliferation or inhibit apoptosis are endogenous factors that enhance the or known as .11 One useful way to think of endocrine responsiveness of the cell play a major these two major types of genes is as the brake and role in this stage of cancer formation. The fault gas pedal of the . line between promotion and progression is gener- When a key gene is altered at a constitutive or ally regarded as the appearance of a clinically hereditable level, a malignant outcome is nearly detectable lesion, frequently referred to as a pre- inevitable early in life (eg, loss of the Rb gene and cancer, premalignancy, or in solid tumors, an IEN. development of a ) or is highly The systematic characterization of IEN and the likely in adult life and occurs earlier (eg, loss or biologic factors that determine risk and mutation of the adenomatosis polyposis gene, have been a major issue in the past few years.6 leading to thousands of colorectal polyps early in Colorectal , cervical IEN, oral leukopla- life and colon cancer at a young age). In both these kia, and actinic keratoses of the skin are repre- cases a key gene that represses growth has been sentative examples of IEN that can be detected by lost; this element is called a tumor suppressor PRINCIPLES OF CANCER PREVENTION 231

FIGURE 2. Constitutive and ac- quired genetic changes play a clini- cal role in cancer formation and pre- vention.

gene. Other genes can be altered in a less severe Almost all human cancers seem to evolve to fashion, for example, the BRCA1 and BRCA2 genes their malignant state over a long period of time, at that, if altered, lead to at an earlier least in immunologically competent individuals. age and at a higher frequency. A large number of Influencing the fate of this genetic background are genes have now been identified that contribute to a wide range of potential direct and indirect car- a high risk of human cancer.12 Taken together, cinogens. Among the more potent direct carcino- however, they contribute perhaps at most 15% to gens to which human populations are exposed 20% of the major risk for human cancers. include , chemicals, and radiation. Potent So what might the contribution of other genetic indirect influences include products of inflamma- factors be? Changes in genes can be lumped into tion and , endogenous and exogenous a general category called polymorphisms, which , and dietary constituents. account for variation in genes. The study of single The nature of the defined genetic changes in polymorphisms, or changes in the sin- many of the major malignancies has now been gle base of a gene and their possible role in cancer characterized with a substantial understanding of risk or outcome, has exploded in the last few the changes in colon and head and neck malignan- years. A wide variety of genes may be affected cies.4,7 Molecular classifications have emerged such as those that regulate and carcin- that will profoundly influence the prognostic clas- ogen metabolism, each an important property that sification of tumors; for example, recent work contributes to risk that is dependent on the par- using chromosomal genomic hybridization has led ticular property of the coded for by the to a new understanding of and its pre- gene in question.9,13 This constellation of heredi- cursor that has etiologic and prognostic implica- table/familial genetic changes provides an individ- tions.16 Although treatment decisions have been ualized baseline “set-point,” from which malig- driven less by specific genetic alterations, the re- nancy or other disease processes evolves in an cent recognition that the same specific genetic individual. Only in the most extreme cases is alteration in diseases as histologically and clini- cancer inevitable, and in most cases the interac- cally diverse as chronic myelogenous tion of environmental parameters with the genetic and gastrointestinal malignancies predict respon- material determines phenotypic outcome and the siveness to the inhibitor imatinib clinical appearance of cancer. With increases in mesylate was eye opening.17 Although the genetic our knowledge of the contributions of genetic vari- changes in most solid tumors are considerably ation to cancer risk, the role of the oncology nurse more complex than in this unique case, even early may well evolve into one of that of advisor regard- in the life history of tumors, this type of observa- ing gene-environment interactions, eg, the role of tion suggests that the development of prevention , diet, physical exercise, , chem- strategies based on an understanding of the basis icals (eg, ), and drugs in determining out- of these molecular changes will be worthwhile. A comes in at risk individuals, a recommendation little explored notion is, “why are specific genetic we made nearly 20 years ago.14,15 changes acquired and what is the role of the 232 MEYSKENS AND TULLY

environment exterior to the cell in producing making definitive conclusions is risky. While the these mutations?” This area of research, gene- results of a case-control study usually do not pro- environment interactions, is burgeoning and has vide definitive evidence of causation, they may become a focus of many recent investigations.18 suggest that a more rigorous design is worth ex- The potential in the future of this approach for ploring. In some cases in which other types of developing useful agents should be significant. epidemiologic studies are not possible, a larger study (many cases) may clarify the risk. In a cohort study the approach is somewhat Epidemiology different. In general, the characteristics of a target The second major discipline that underpins pre- group is defined at the beginning of the study and vention is epidemiology, which classically has after enough primary events (eg, a disease) have been the study of diseases in populations. With accumulated a comparison of the presence of risk the advent of powerful molecular techniques that factors can be made between those participants in can assess the role of individual genetic variation the cohort who have experienced an event and at risk, there has been an increased interest in those who have not, and a relative risk calculated. assessing individual risk for diseases, including Blood and samples are frequently collected cancers19; although the idea is not a new one.14,15 at the beginning of the study and at times there- The basic principle of epidemiology is that evi- after and the contribution of a variety of biochem- dence for causality can be identified by studying ical and molecular measurements to risk can also one group with a condition or disease and by be assessed. Cohort studies require following a comparing the test group to an appropriate con- large group of individuals over time, frequently trol group, factor(s) can be assessed for their pres- a long time, and requires extensive planning and a ence in the two groups. The relative involvement long-term commitment. A great deal about cancer in the two groups leads to determination of a and other diseases has been learned from cohort relative risk profile for a particular item(s) in the studies, such as the Physicians Health Study, the test group. Nursing Health Studies, and others. There are many types of epidemiologic studies Analysis of secondary events in completed ran- used, and they can be divided into two major domized trials has provided a valuable opportu- types; observational and experimental. In order of nity to identify risk factors because the random- increasing power to accurately predict the obser- ized design provides a critical control population vational studies include case-control, cohort, and that has been matched for key parameters. For secondary analysis of a randomized clinical trial. example, results from a randomized trial of sele- The reader is referred to Jepson et al20 for a more nium supplementation and identified detailed discussion. Many case-control studies are that selenium appeared to protect against prostate preformed because they are relatively easy to do, and other cancers.21,22 Largely based on this ob- can be done rapidly (the cases are available), and servation, a large number of randomized clinical are not expensive. In these types of studies, cases trials are being undertaken using selenium as the of a particular disease or entity are identified and treatment drug. All of these observational studies matched with controls (usually 2 to 3 times the have strengths and weaknesses, but they have number of cases). The presence of a large number been important in identifying possible risk (eg, of particular factors (eg, age, gender, pill con- , many cancers) and protective factors sumption) is enumerated and the frequency de- (eg, oral contraceptive pills, ovarian cancer). With termined in both cases and controls and the ratios sufficiently consistent observational results, pub- of the factors in cases and in controls determined. lic health recommendations are frequently made A ratio of 1.0 implies no effect. Less than 1.0 and modifications in clinical practice undertaken suggests a protective effort of the factor in the even without the results of randomized trials be- cases while a ratio of greater than 1.0 suggests an ing available. However, in those situations in adverse effect for the condition being measured. which a randomized trial can be launched, it is The significance of the association is then tested. frequently needed to sort out conflicting results The power of case-control studies is relatively from observational studies. Two important exam- weak so, unless a marked association is seen (eg, ples include the value of mammography to screen odds ratio Ͻ0.5 or Ͼ2.0) and/or multiple case for breast cancer and the effect of hormone re- control studies of the same entity are consistent, placement therapy on breast cancer inci- PRINCIPLES OF CANCER PREVENTION 233

dence.23,24 In the former case, the value of screen- trial may be a pivotal study and could lead to the ing has been proven while in the latter situation approval of a drug for a new indication. the therapy, which was presumed to be effective, Secondly, the issue of biomarkers is a difficult was not and may well be harmful. and complex one.26-29 This term is thrown around fairly casually. A biomarker can be any biochem- ical, molecular, histologic, or imaging alteration Clinical Trials that may or may not predict or be on the progres- The clinical trial represents the third major sion pathway to cancer. In contrast, a surrogate discipline that is essential for moving the preven- endpoint biomarker is on the pathway to cancer tion agenda forward. Although the nosology for and, most critically, accurately predicts that a prevention studies is similar to that for treatment drug will modulate the true endpoint (in most trials,1,25 there are substantial differences, the cases cancer). A validated US Food and Drug Ad- most important overall being that the risk/benefit ministration is one that has undergone indepen- ratio in prevention trials must be low if a tech- dent confirmation.28 There are an infinite number nique or drug is to be eventually adapted by well of potential biomarkers but to date no validated or nearly well individuals. There are also several surrogate endpoint biomarker for cancer exists. different sources of intervention agents that add The issues and limitations of developing biomar- complexity to clinical prevention trials: these in- kers has been debated and discussed at length, clude nutriceuticals, dietary compounds used in a although far too many investigators are unaware pharmacologic manner, drugs approved for an- or choose to ignore the problems and launch into other indication and being tested for a cancer conducting trials that are seriously flawed or, even prevention purpose, and compounds specifically worse, uninformative. developed as drugs for cancer chemoprevention. Thirdly, the risk/benefit ratio needs to be care- In general, clinical chemoprevention trials have fully assessed. In a predisposing situation where adhered to US Food and Drug Administration the risk for cancer is low, then the side effects guidelines, but nutriceuticals (and to an extent must be low. In a high-risk situation more side dietary compounds) have not undergone rigorous effects may be acceptable. Both 13–cis retinoic clinical testing before widespread adoption. This acid and have not been adopted in cavalier approach has presented a real problem practice despite randomized trials that showed a because the claims made for many of these com- marked decrease in second malignancies in those pounds has exceeded the evidence, in some cases with a prior aerodigestive malignancy30 or a de- rather markedly so. crease in first breast cancer in women at high risk The nuances of clinical chemoprevention trials based on an epidemiologic profile of high risk,31 are numerous and evolving.26 Four major points respectively. Largely because of real risks that will be made here. First, a critical first step is to were measured in well-conducted trials and/or show, whenever possible, that a candidate drug perceived by the target population as too risky. affects a relevant biomarker in the human of Fourth, the criteria for US Food and Drug Ad- interest at a dose that is achievable and likely to ministration approval for chemoprevention agents have minimal to no side effects at the chosen dose. is not firmly established. Although there are sev- This objective is no easy task and is the major goal eral approved drugs for prevention, these confir- of phase II trials. Most commonly, a relatively mations have to date occurred almost inciden- short-term trial (2 to 4 weeks for presurgical mod- tally. There has also been convened a large task els; 4 to 12 weeks for IEN interventions) is initially force to explore this area, but opinions remain performed in which the effects of several doses of fluid.6 a drug (and placebo whenever possible) on a rel- evant biomarker are assessed. The lowest dose at which a consistent alteration in the marker is TYPES OF PREVENTION produced is advanced to the longer phase IIb trial, in which both the effect on the biomarker and a here have been various classifications pro- clinical endpoint is determined. Phase IIb trials Tposed to categorize the activities that rep- last from 6 months to about 3 years and in addi- resent prevention used. One that has been tion carefully monitor for side effects in a rigorous widely adapted is the idea of primary, second- fashion. In some cases a randomized phase IIb ary, and tertiary prevention1 which is modeled 234 MEYSKENS AND TULLY

on, but different from, the public health desig- Tertiary prevention represents an attempt to de- nations. Primary prevention addresses the initia- tect cancers early. Early detection in which can- tion phase of cancer and has the overall goal to cers are found secondary to follow-up of symp- decrease the appearance of cancer. This approach toms is classified as tertiary prevention and has includes such strategies as “sun sense,”32 anti-smok- become increasingly incorporated into the main- ing education, smoking cessation, dietary and phys- stream practice of clinical oncology. Chemopre- ical activity changes, and condom usage, among vention of a secondary malignancy in an individ- others. Primary prevention also includes counseling ual who has already had a malignancy also of genetically-at-high-risk individuals and interven- represents tertiary prevention. The techniques tion with dietary or pharmacologic compounds in used for screening also used early detection and individuals at high risk for cancer. This approach include radiologic imaging techniques. also includes screening in which no symptomatol- ogy is evident, but in which precancers or cancers CONCLUSION are found. Some well-accepted screening modalities include mammography, colonoscopy, and Papanico- he disease is not cancer, but carcinogene- laou cytology.33 Tsis—the process by which tumors and ma- Secondary prevention addresses the promo- lignancies develop. Modern cancer prevention tional phase of carcinogenesis. Once the initial integrates our understanding of the biological genotypic change has occurred, there is a likeli- and molecular features associated with cancer hood that the malignant process will eventually formation. Application of the principles of evolve to a cancer. Secondary prevention includes screening and early detection has contributed to management of precancers (IEN) either by abla- the decline of morbidity and mortality from tive (, laser) or non-ablative (eg, drugs) several cancers and we can anticipate success in means. The goal is to abrogate or reverse the others. The field of chemoprevention is strongly precancer or IEN. This area has been a major supported by findings from epidemiology and focus of the pharmaceutical approach to preven- preclinical experimental models, and “proof of tion, also called chemoprevention. Nearly 15 principle” has been established in several can- years ago one of us (F.M.) wrote that the field of cers including breast, colon, and prostate. De- chemoprevention of cancer was “coming of age,”34 veloping risk profiles and less toxic agents will the process has indeed been a slow one. The allow the eventual adoption of chemoprevention systematic development of strategies by organ site as a standard practice similar to the use of is an important one; a particularly organized effort antihypertensives for high blood pressure and is that of the Arizona group and skin cancer.35 statins for elevated lipids.

REFERENCES

1. Meyskens FL Jr. Strategies for prevention of cancer in the development of cancer prevention drugs. Clin Cancer Res humans. Oncology 1992;6(suppl):16-24. 2004;10:3908-3918. 2. Greenwald P. Cancer prevention clinical trials. J Clin 8. Bertram JS, Kolonel LN, Meyskens FL Jr. Rationale and Oncol 2002;20(suppl 18):14S-22S. strategies for chemoprevention of cancer in humans. Cancer 3. Meyskens FL Jr. Cancer prevention. In: Abeloff M, Armit- Res 1987;47:3012-3031. age J, Lichter A, Niederhuber J, eds: Clinical Oncology. Ed 3. 9. Ewart-Toland A, Balmain A. The genetics of cancer sus- New York: Churchill Livingstone: 2004;425-472. ceptibility: From mouse to man. Toxicol Pathol 2004;32(suppl 4. Sidransky D. Emerging molecular markers of cancer. Nat 1):26-30. Rev Cancer 2002;2:210-219. 10. Fearn ER, Vogelstein B. A genetic model for colorectal 5. Meyskens FL Jr. Biomarkers intermediate endpoints and tumorigenesis. Cell 1990;61:759-767. cancer prevention. J Natl Cancer Inst Monogr 1992;13: 11. Labazi M, Phillips AC. Oncogenes as regulators of apo- 177-182. ptosis. Essays Biochem 2003;39:89-104. 6. O’Shaughnessy JA, Kelloff FJ, Godon GB, et al. Treatment 12. Greco KE, Mahon S. Common hereditary cancer syn- and prevention of intraepithelial neoplasia: An important tar- dromes. Semin Oncol Nurs 2004;20:164-177. get for accelerated new agent development. Clin Cancer Res 13. Wu X, Zhao H, Suk R, et al. Genetic susceptibility to 2002;8:314-346. tobacco-related cancer. 2004;23:6500-6523. 7. Kelloff GJ, Schilsky RL, Alberts DS, et al. Colorectal 14. Lippman SM, Bassford TL, Meyskens FL Jr. Quantitative adenomas: A prototype for the use of surrogate endpoints in assessment of cancer risk. Texas Med 1990;8:48-53. PRINCIPLES OF CANCER PREVENTION 235

15. Lippman SM, Bassford TL, Meyskens FL Jr. A quantita- 25. Goodman GE. The clinical evaluation of cancer chemo- tively scored cancer-risk assessment tool: Its development and prevention agents: Defining and contrasting phase I, II, and III use. J Cancer Educ 1992;7:15-36. objectives. Cancer Res 1992;52(suppl):2752s-2757s. 16. Bastian BC. Molecular genetics of melanocytic neopla- 26. Meyskens FL Jr, Szabo E. How should we move the filed sia: practical applications for diagnosis. Pathology 2004;36: of chemopreventive agent development forward in a produc- 458-461. tive manner? Recent Results Cancer Res 2005;166:113-24. 17. Hochhaus A. Imatinib mesylate (Glivec, Gleevec) in the 27. Armstrong WB, Taylor TH, Meyskens FL Jr. Can a treatment of chronic myelogenous leukemia (CML) and gas- marker be a surrogate for development of cancer, and would trointestinal stromal tumors (GIST). Ann Hematol 2004; we know it if it exists? Recent Results Cancer Res 2005;166: 83(suppl 2):S65-S66. 99-112. 18. Schottenfeld D, Beebe-Dimmer JL. Advances in cancer epidemiology: Understanding causal mechanisms and the evi- 28. Ransohoff DF: Rules of evidence for cancer molecular- dence of implementing interventions. Annu Rev Public Health marker discovery and validation. Nat Rev Cancer 2004;4: 2005;26:37-60. 309-314. 19. Culler D, Grimes SJ, Acheson LS, et al. Cancer genetics 29. Schatzkin A, Gail M: The promise and peril of surrogate in primary care. Prim Care 2004;31:649-683. and end points in . Nat Rev Cancer 2002;2:19- 20. Jepsen P, Johnsen SP, Gillman MW, et al. Interpretation 27. of observational studies. Heart 2004;90:956-960. 30. Hong WK, Lippman SM, Itri LM, et al. Prevention of 21. Duffield-Lillico AJ, Slate EH, Reid ME, et al. Nutri- second primary tumors with in squamous-cell carci- tional Prevention of Cancer Study Group. Selenium supple- noma of the head and neck. N Engl J Med 1990;323:795-801. mentation and secondary prevention of nonmelanoma skin 31. Fisher B, Constantino JP, Wickerham DL, et al. Tamox- cancer in a randomized trial. J Natl Cancer Inst 2003;95: ifen for prevention of breast cancer: Report of the National 1477-1481. Surgical Adjuvant Breast and Bowel Project P-1 Study. J Natl 22. Duffield-Lillico AJ, Reid ME, Turnbull BW, et al. Baseline Cancer Inst 1998;90:1371-1388. characteristics and the effect of selenium supplementation on 32. Ramstack JL, White SE, Hazelkorn KS, et al. Sunshine cancer incidence in a randomized clinical trial: A summary and skin cancer: A school-based skin cancer prevention report of the Nutritional Prevention of Cancer Trial. Cancer project. J Cancer Educ 1986;1:001-008. Epidemiol Biomarkers Prev 2002;11:630-639. 23. Smith RA, Duffy SW, Gabe R, et al. The randomized 33. Meyskens FL Jr: Screening for cancer: Valuable or not? trials of breast : What have we learned? Radiol Curr Oncol Report 2004;6:485-490. Clin North Am 2004;42:793-806. 34. Meyskens FL Jr: Coming of age: Chemoprevention of 24. Manson JE, Hsia J, Johnson KC, et al. Women’s Health cancer. N Engl J Med 1990;323:825-827. Initiative Investigators. plus progestin and the risk of 35. Harris RB, Alberts DS: Strategies for skin cancer preven- coronary heart disease. N Engl J Med 2003;349:523-534. tion. Int J Dermatol 2004;43:243-251.