Cancer Prevention, Screening, and Follow-up

From WiserWiki

[edit] Cancer Prevention, Screening, and Follow-up

Linda M. Sutton

As a result of decades of research in the laboratory and in clinical practice, optimism in cancer care is now well founded. Although nearly 1.25 million Americans will be diagnosed with cancer in 1999, the incidence of cancer overall has declined 0.7% from 1990 to 1995 in contrast to increasing trends in earlier years. The downward trend in incidence has also been seen for the individual leading cancer types. There have been similar improvements in cancer-related mortality. Since the 1950s cancer has been the second leading cause of death in the United States, and an estimated 538,000 deaths in 1999 will be attributed to cancer.^[1]Box 122-1 details the five primary sites of invasive malignancy, which account for nearly 60% of the estimated annual cancer incidence and mortality. However, when compared with the early 1960s, the rate of survival at 5 years after diagnosis has increased for a majority of cancer types. The last half of the 1990s has witnessed the fruition of several approaches to the prevention of cancer mortality. Improvements in cancer treatments, improved supportive care of those with cancer, and better detection of early-stage disease have contributed to decreased mortality. Perhaps more exciting are recent results suggesting that, at least for some individuals, the development of cancer may be prevented through interruption of the progression of neoplastic growth toward invasive malignancy.

[edit] PREVENTION

Prevention of cancer-related mortality has been a health care priority for the United States for a majority of the twentieth century. The deceptive simplicity of neoplastic growth clouded initial progress. Early tools in the preventive armamentarium consisted largely of antineoplastic therapies. Unfortunately, this tertiary prevention led to declines in cancer-related morbidity and mortality for only a minority of malignancies. Lack of curative therapies for the majority of cancers and a limited knowledge base precluded effective use of other preventive approaches.

We now understand that cancer is not a single target at which preventive strategies can be aimed. Rather, cancer is the collective term for many distinct neoplasms with different etiologies, risk factors, and behavioral characteristics. Carcinogenesis is a multistep process with long periods of latency for individual steps. This complicates but does not preclude effective preventive programs. There are several levels in the ontogeny of an invasive neoplasm at which specific interventions may prevent subsequent morbidity and mortality. An improved understanding of the process of carcinogenesis has allowed better identification of high-risk groups and has led to enthusiasm for primary and secondary preventive strategies.

Primary preventive strategies seek to interrupt neoplastic development before a malignant growth is established. This can be accomplished through modification of patient characteristics or environmental factors that have been identified through epidemiologic studies as being associated with increased cancer risk. Although risk factors such as age and gender are not amenable to modification, other risk factors such as tobacco use are believed to play a causative role in the transformation of a normal host cell to a neoplastic clone. Elimination of such causative factors can have a major impact on cancer incidence and mortality. Alternatively, neoplastic development could be reversed or inhibited through chemoprevention. A number of chemical agents, both naturally occurring and pharmaceutically engineered, have been shown to inhibit carcinogenesis in experimental systems. Recent clinical trials, such as the Tamoxifen Prevention Trial conducted by the National Surgical Breast and Bowel Project (NSABP), have documented reduction in short-term breast cancer incidence through chemoprevention. Whether the improvement in incidence rates, seen in the time interval over which the chemoprevention trials have been conducted, will translate into improved cancer-related mortality in the long term is unclear.

[edit] High-risk Groups

Successful preventive interventions require that factors associated with an increased risk of developing cancer can be defined and individuals or groups who possess these high-risk characteristics and behaviors can be identified prior to the diagnosis of cancer. Furthermore, the identified risk factors must be amenable to modification such that for individual patients the development of a particular cancer is delayed or prevented. It has been estimated that between 75% and 80% of cancers in the United States could be avoided through alterations in behavior.^[2]Box 122-2 summarizes the categories of factors that are associated with enhanced cancer risk.

[edit] Tobacco.

Tobacco is responsible for more cancer-related deaths than any other single risk factor and accounts for nearly one third of all cancer deaths.Box 122-3 underscores the impact of tobacco use on cancer incidence. Clearly, tobacco use is a high-risk behavior and its elimination would have a profound impact on cancer-related mortality in the United States. For example, the incidence of lung cancer would decline by approximately 90%, with an estimated 25% reduction in overall cancer mortality if smoking alone were completely eradicated. Elimination of other forms of tobacco use would further decrease cancer incidence and mortality. There would be additional health benefits in the form of decreased incidence of cardiovascular and respiratory disease.

Numerous clinical trials have demonstrated that physicians are a powerful force in shaping the behavior of individual patients. Physician-based smoking cessation trials have repeatedly demonstrated that any intervention can effectively reduce tobacco use by smoking patients. A systematic review of 188 randomized controlled trials by Law and Tang demonstrated a modest but highly cost-effective benefit of a single episode of advice and encouragement on smoking cessation from a physician.^[3] Brief advice by a physician to quit smoking can lead to sustained abstinence (at 1 year) for an estimated 2% of patients. Although this degree of efficacy may not seem worth the bother to individual physicians, the time involved is usually less than 5 minutes and the public health impact is significant. Additional supportive measures such as follow-up letters or visits increase efficacy to a variable extent. However, when these simple interventions are offered to populations with enhanced risk, such as pregnant women or patients with ischemic heart disease, there is a dramatic impact on efficacy. While the efficacy of smoking cessation advice increases an impressive fourfold for pregnant women, following a myocardial infarction (MI) efficacy can be as high as 35%. It is unclear whether efficacy is maintained when advice and encouragement are provided by health care providers other than the physician. Nurses in health promotion clinics have variable results, but advice, counseling, and follow-up provided by dental hygienists to those using chewing tobacco were significantly more effective than routine care.^[4] Behavior modification therapy, such as relaxation or visualization techniques or avoidance of stimuli that trigger the urge to smoke, was no more effective than simple advice. Nicotine replacement can be an important adjunct for nicotine-dependent smokers who seek help in stopping. Nicotine replacement products are available as nasal spray, transdermal patch, and gum. These pharmacologic aids may enhance sustained abstinence for 13% of smokers. Finally, a recently reported randomized placebo controlled trial of sustained-release bupropion demonstrated improved 1-year abstinence rates over control, documented with carbon monoxide measurements.^[5] The effect was dose responsive and statistically significant at the higher doses used (300 mg and 150 mg daily). At 1-year following intervention, approximately one fifth of the bupropion-treated groups maintained abstinence, whereas only 12.4% in the placebo group were able to remain smoke-free. Side effects of the bupropion were minimal. One interesting side effect was reduced weight gain, particularly at higher doses, with those in the 300-mg group gaining an average of 1.5 kg.

Recommendations for tobacco cessation are listed in Box 122-4.

[edit] Diet.

There is a substantial body of evidence supporting the role of diet in both the development and inhibition of neoplastic growth. The evidence linking factors in the human diet to neoplastic growth is largely derived from epidemiologic studies conducted either at the population level or through specific cohort or case-control studies. The interpretation of such epidemiologic data is complicated by numerous factors. Retrospective epidemiologic studies attempting to link nutrient intake to risk of subsequent cancer suffer from recall bias, inaccurate knowledge of the composition of prepared foods, and confounding variables such as physical activity. In addition, the culprit in carcinogenesis may well be the intake of foodstuffs many years prior to a particular study. Finally, the role of genetic polymorphisms that may moderate the effects of particular risk factors for individuals is only beginning to be understood. Inherited changes in enzymes known to detoxify carcinogenic metabolites may enhance cancer risk when individuals are exposed to a carcinogen in the presence or absence of another factor. For example, smokers for whom the carcinogen detoxifying enzyme GSTM1 is genetically deleted appear to have an increased risk of developing cancer only when low plasma concentrations of micronutrients such as α-tocopherol and specific carotenoids are also present.

The human diet is composed of macronutrients and micronutrients, as defined in Box 122-5. Both deficiencies and excesses of these elements within the diet appear to contribute to the development of specific malignancies. Epidemiologic studies have identified populations with high calorie and fat intake and low consumption of fiber, fruit, and vegetables, as well as certain micronutrients, in which there is an increased risk of developing a number of malignancies. How this risk should be interpreted for individual patients is unclear. Whether dietary modification for individual patients will alter neoplastic development is controversial. Kritchevsky, in a review of cancer and diet, emphasizes that a high-fiber diet may not be equivalent to a low-fiber diet supplement with fiber.^[6] Vitamin and mineral supplementation may be complicated by unexpected toxicities and may not decrease cancer risk as well as adoption of a diet containing foods rich in these same vitamins and minerals. The increased incidence of lung cancer and lung and cardiovascular deaths among the 18,314 smokers and asbestos-exposed workers in the CARET study of β-carotene and vitamin A supplementation highlights the need for caution regarding dietary supplementation.^[7]

A number of dietary intervention trials designed to address the impact of dietary modifications on neoplastic development and growth are currently underway. The number of these trials is limited by the large number of patients required to achieve statistically reliable results and the long duration between neoplastic incitement and clinically detectable cancer. Several studies have used intermediate markers of neoplastic transformation, such as colonic polyp formation or indicators of proliferative activity within target tissues, as a means of restricting study length. Although colonic polyps are established neoplastic precursors, the import of increased proliferative activity for malignant transformation in other tissues is not clearly established. There is currently a large research effort aimed at the identification and validation of intermediate markers of cancerous growth. An alternative approach is to modify the diet of patients diagnosed with cancer who are at high risk of recurrence or a second primary cancer. The ongoing Women's Intervention Nutrition Study (WINS), which addresses the impact of low-fat diet as an adjuvant to hormonal or chemotherapy for women with early-stage breast cancer, is one example.

[edit] Fat.

The role of fat and fiber intake in carcinogenesis has been widely studied. Clinical studies have demonstrated an association between excessive consumption of fat and calories and cancers of the breast, colon, prostate, ovary, endometrium, kidney, and pancreas. International epidemiologic studies have identified strong associations between per capita fat intake and age-adjusted rates for specific cancers. The prevalence of breast, colon, and prostate cancers is much higher in the Western world than in countries such as Japan. The Western world derives approximately 35% to 40% of its calories from fat, whereas the amount of fat consumed in the Japanese diet is more limited. Interestingly, the incidence of breast cancer has risen in Japan as fat consumption has risen from about 10% of the caloric intake in the 1950s to the current 25%.

The increased risk of these cancers may not be uniform for all types and amounts of fat consumed. The components of a high-fat diet that is causally related to increased cancer risk clearly remain to be identified. It is known that the incidence of colorectal, breast, prostate, pancreatic, and possibly lung cancers increases directly in proportion to the amount of meat, particularly red and processed meat, consumed. Esophageal cancer is increased by ingestion of barbecued meat. Whereas diets rich in omega-3 fatty acids have been shown to retard tumor development in animal models, those relatively rich in omega-6 fatty acids enhance tumor development.

The mechanism by which dietary fat effects carcinogenesis is unclear. The heterocyclic amines present in cooked meat may be partially responsible for the carcinogenic effect. Omega-3 fatty acids, abundant in fish products, interfere with the inflammatory response through competitive inhibition of the cyclooxygenase pathway in prostaglandin formation. Overexpression of the gene for inducible cyclooxygenase was identified as an early critical step in colon carcinogenesis.^[8] In addition, dietary fat may enhance the development of colon cancer by increasing the concentration of bile salts within the colon. Bile acids alter the metabolic activity of microbial flora within the colon such that the colonic mucosa is increasingly exposed to weakly carcinogenic bile acid metabolites. Mechanisms of mutagenesis in other malignancies are less well defined.

Whether dietary alteration has the potential to abrogate malignant growth remains controversial. The prospective Nurses' Health Study failed to document a decline in breast cancer risk for 88,795 women when the lowest quintile was compared with highest quintile for intake of total, polyunsaturated, or monounsaturated fat. More optimistically, the Lyon Diet Heart Study of 605 patients with coronary heart disease, who were randomized to a Mediterranean-type diet or a control diet (approximating a Step 1 American Heart Association prudent diet), demonstrated a statistically significant 65% reduction in cancer incidence for those on the Mediterranean-type diet during 4 years of follow-up.^[9] The diet consisted of higher levels of bread, cereals, fresh fruit and vegetables, legumes, and fish. The diet prescribed fewer delicatessen foods, less meat, and no butter and cream, which were to be replaced with an experimental canola oil–based margarine rich in oleic and α-linolenic acids. Compliance with this α-linolenic acid–rich diet was excellent.

[edit] Fiber.

Numerous case-control and retrospective studies have suggested a link between a low-fiber diet and subsequent development of colorectal cancers. It is hypothesized that a high-fiber diet may lower the risk of colon cancer by increasing fecal bulk and diluting the concentration of carcinogens within the colon, thereby limiting the exposure of colonic mucosa to potential mutagens. However, prospective trials have failed to demonstrate a protective effect of dietary fiber against the development of colorectal cancer or adenomas. The largest prospective trial to address this, the Nurses' Health Study, showed no association between dietary fiber intake and colorectal risk in 88,757 women, in the context of a Western diet.

[edit] Fruits and Vegetables.

Fruits and vegetables are almost invariably protective for the major cancers. Conjecture regarding the micronutrients responsible for this beneficial effect has been extensive. The antioxidant vitamins (vitamin A and related compounds such as β-carotene, as well as vitamins C and E) are prominent components of many fruits and vegetables. The antioxidant vitamins function as scavengers for DNA-damaging, mutagenic oxygen free radicals. Epidemiologic studies suggest that diets deficient in vitamins A, C, and D have been associated with increased cancer risk, whereas those diets composed of foods rich in vitamin A, C, and E are associated with decreased cancer incidence.

Vitamin A is essential for the normal growth and development of epithelial tissue. A deficiency of this important vitamin can increase the susceptibility of normal tissues to mutagens. Dietary deficiency of vitamin A has been implicated in the development of cancers of the lung, breast, oropharynx, stomach, bladder, prostate, and colon. Squamous tissues deficient in vitamin A exhibit metaplastic differentiation that can be reversed by administration of vitamin A and related compounds. There is evidence to suggest that diets rich in vitamin A and related compounds not only diminish the risk but are also protective against the development of certain cancers. While vitamin C appears to inhibit the formation of carcinogenic nitrosamines that have been associated with the development of gastric cancer, vitamin E inhibits mutagenesis and cell transformation mainly through its antioxidant function. Nonetheless, the role of vitamins C and E in neoplastic development remains particularly unclear. Although vitamins C and E function as antioxidants, there is little evidence to support any direct role for these vitamins in the inhibition or reversal of neoplastic growth and development.

Among the many minerals required for normal tissue development, calcium and selenium have received the most attention with regard to carcinogenesis. There are laboratory as well as preliminary clinical data to support a role for calcium deficiency in the development of colon cancer. Results of the numerous studies focusing on the role of dietary selenium are contradictory and inconclusive, but consensus on its beneficial effects is building. Data on other micronutrients, such as molybdenum and zinc, are scant, and further investigations are necessary.

Several recent studies of dietary supplementation highlight the importance of clinical trials in defining benefit.^[10] Both the CARET (β-carotene and retinol) and the α-tocopheral, β-carotene trials suggest that pharmacologic doses of β-carotene increase the risk of lung cancer among smokers or those with asbestos exposure. β-Carotene, vitamin C, and vitamin E were also unable to prevent the development of colorectal adenomas in a placebo-controlled trial of 864 patients. A large, four-arm clinical trial of multiple dietary supplements in 30,000 subjects in Linxian, China, demonstrated no significant effect on cancer incidence. However, those subjects who received treatment with a combination of selenium, β-carotene, and α-tocopherol enjoyed statistically significantly lower total and gastric cancer-specific mortality rates. A notable study of selenium supplementation in 1312 patients was clearly unable to protect against the recurrent development of basal and squamous cell carcinomas in the skin. Nonetheless, secondary end point analyses reveal a statistically compelling reduction in prostate, lung, and colon cancers, as well as significant reductions in total cancer mortality. On the other hand, a nonsignificant increase was noted in leukemias/lymphomas as well as breast and bladder cancers.

[edit] Alcohol.

The chronic consumption of alcohol is strongly associated with cancers of the oropharynx, larynx, and esophagus. The risk of these cancers is greatly magnified by concomitant use of both alcohol and tobacco. Nonetheless, there is a dose-response relationship between alcohol consumption and the development of oral and esophageal cancers that is independent of concomitant tobacco use. In addition, there is some evidence—although not conclusive—that cancers of the liver, stomach, pancreas, colon, and breast are also associated with increased alcohol consumption. The risk of moderate or occasional alcoholic intake for cancer development is not well established. The mechanism through which alcohol influences carcinogenesis is currently under investigation. Animal studies have failed to identify pure ethanol as a carcinogen. Alcohol may exert its influence through alterations in cellular metabolism and permeability that permit carcinogenic disruption of normal cellular behavior by other chemical substances. Alternatively, alcohol consumption may contribute to malnutrition and increase consumption of other substances associated with enhanced cancer risk.

[edit] Dietary Recommendations.

Many authorities believe that the complexity of the human diet with too many undefined interactions with other factors such as physical activity precludes firm dietary recommendations for individual patients. Other authorities such as the American Cancer Society (ACS), the National Research Council, and the National Cancer Institute (NCI) recommend general dietary guidelines for dietary components with sufficient data. The NCI recommends a diet that includes five or more servings of fruits and vegetables per day. Specific dietary instructions must be individualized, with consideration to comorbid illnesses and conditions. Notably most authorities that offer dietary guidelines include limitation of alcohol intake.

[edit] Radiation Exposure.

Radiation comes in a variety of forms and energies. Nearly all tissues within the body are susceptible to its carcinogenic effects. It is an omnipresent risk factor within the environment. Background radiation provides the largest source of exposure for the general population. Light emitted from the sun contains a broad spectrum of radiation energies. The ultraviolet portion of these energies is a major risk factor for the development of basal and squamous cell carcinomas of the skin, as well as melanoma. Ultraviolet radiation is thought to be directly carcinogenic, and risk of subsequent malignancy is directly proportional to received dose. The risk of developing a skin cancer, particularly melanoma, is highest in populations who have the highest exposure by virtue of geographic latitude and skin type. The worldwide incidence of melanoma increases with increasing proximity to the equator. Skin cancers occur most frequently on those areas of the body with the greatest exposure to sunlight. In addition, there are genetically determined differences in susceptibility in ultraviolet radiation. Patients with conditions such as xeroderma pigmentosum are exquisitely sensitive to sunlight-induced skin damage, and are at very high risk of developing skin cancer.

Ionizing radiation is also thought to be directly mutagenic. Much of the data regarding the role of ionizing radiation in neoplastic development is derived from studies of individuals exposed to the radiation fallout of the atomic bomb and those receiving medical x-rays for either diagnostic or therapeutic purposes. These populations generally have exposure in excess of 50 Gy and clearly have an increased risk of leukemias and lung and breast cancers. There is also an increased incidence of thyroid and bone cancers within the field of radiation in some studies. Age at the time of exposure appears to be an important ameliorating factor, since the latency period for radiation-induced cancers is often years to decades.

Although exposure to ultraviolet radiation and the consequent risk of skin cancers can be minimized through the use of protective clothing and sunscreens, the largest source of radiation exposure is from the general background. This exposure is not only difficult to quantify and almost impossible to limit, it is likely to increase with progressive deterioration of the atmospheric ozone layer. However, exposure to medical and dental radiation, the second largest source of radiation exposure, can be limited. Physicians should demonstrate restraint in the use of diagnostic and therapeutic radiation, particularly in patients who are young, who are pregnant, or who have benign medical conditions. The cumulative exposure to medical radiation can be monitored by carefully recording within a patient's medical record all diagnostic radiographic studies as well as the total dose and target of any therapeutic radiation administered. The careful recording of radiation exposure as routinely practiced by radiation oncologists might serve as a paradigm for primary care physicians as a mechanism for monitoring individual patient exposures.

[edit] Chemical Exposure.

The number of chemicals with carcinogenic potential seems to be endless. The daily exposure of individual patients to carcinogenic chemicals is difficult to accurately assess given the number and variety of chemicals present in the air, water, and food. Exposures occur within the home as well as through recreational and occupational activities. Certain chemicals have well-documented associations with human neoplasms (Table 122-1). In addition, some medical therapies are also closely linked with increased risk of subsequent cancer. Alkylating agents, such as chlorambucil, cyclophosphamide, and melphalan, were originally developed for use in antineoplastic regimens. These agents are associated with increased risk of subsequent malignancies when used to treat patients with collagen vascular diseases such as rheumatoid arthritis as well as those with cancers.

Table 122-1 Chemically Induced Tumors

Hormonal therapy with estrogens or androgens has also led to increases in the incidence of particular malignancies. The use of androgens in male athletes has been associated with increased risk of liver cancers. In women the use of exogenous estrogens has been associated with an increased risk of endometrial and breast cancers. Use of diethylstilbestrol (DES) in pregnant women is associated with the development of vaginal clear cell adenocarcinoma in female offspring. Tamoxifen, a selective estrogen receptor modifier used in the treatment—and, most recently, prevention—of breast cancer, stimulates endometrial proliferation that contributes to an increased risk of endometrial cancer.

The use of immunosuppressive therapy following organ transplantation is also associated with a dramatic increase in the incidence of lymphoma.

The potential neoplastic risk of any proposed medical therapy must be clearly understood and weighted carefully against the expected benefit. In a fashion similar to documentation for radiation exposures, patients who receive treatment with medications, such as alkylating agents and hormones, that carry an increase risk of subsequent malignancy should have the dose and duration of treatment carefully monitored in the medical record. Prominent labeling of individual patient charts can facilitate a monitoring program. The risks and benefits of these potentially carcinogenic therapies should be reviewed periodically in any patient who receives treatment for prolonged periods.

Exposure to occupational and industrial carcinogens can be minimized by the use of equipment such as gloves, protective clothing, and masks in appropriate situations. Physicians should encourage the use of such protective gear. However, often the hazards of chemical substances are unknown or patients are unaware of the specific exposures they have experienced. Recent legislation requiring employers to inform workers of potentially hazardous substances in the workplace may improve patient's cognizance of exposures. This in turn might lead to a better understanding of the role some chemical substances play in carcinogenesis in humans. Discussion of the occupational environment of individual patients and documentation of reported exposures within the medical record may serve several purposes. The patient may not be aware of the risk inherent in agents that are currently known to have carcinogenic potential and therefore may not adequately employ protective equipment. Secondly, documentation of exposures may allow associations to be identified at a future date.

Infectious agents play an important role in neoplastic development worldwide. There are well-documented associations between a number of agents and specific cancers (Table 122-2). The contribution of infectious agents to cancer morbidity and mortality within the United States is more limited. Nonetheless, the prevalence of certain infectious agents, particularly herpes simplex II (HSV II), hepatitis B virus, and the human immunodeficiency virus (HIV), mandates heightened public awareness of the role of these agents in neoplastic development. In addition, the recent association of Helicobacter pylori infection and subsequent gastric cancer may allow significant reduction in gastric cancer incidence through aggressive antibiotic therapy. H. pylori gastric infection has been strongly linked to mucosa-associated lymphoid tissue (MALT) neoplasms. Eradication of H. pylori leads to histologic regression of early MALT lymphoma. Attempts to eradicate H. pylori are underway in China, where gastric cancer is the second most common malignancy.

Table 122-2 Cancers Associated With Infectious Agents

Alterations in sexual behavior might well lead to diminished risk for those cancers that are associated with sexually transmitted infectious agents such as HSV II and HIV. Patients who engage in unprotected sexual activities with multiple partners are at particularly high risk of becoming infected. Physicians should ask about sexual practices and advise patients of the associated risks of infection and subsequent malignancy. This is particularly important in those patients who have no current evidence of infection with these agents. Patients who do manifest evidence of infection should receive regular follow-up, with attention to evaluating the presence of neoplastic processes.

Patients with exposure to blood-borne agents should be encouraged to take appropriate precautions. Those individuals with occupational exposures to blood and body fluids should be advised to routinely employ protective gear. Intravenous drug users should be discouraged from sharing hypodermic needles. Finally, the hepatitis B vaccine should be offered to all individuals at high risk of exposure to this infectious agent.

[edit] SCREENING

Early detection of cancers through screening provides secondary prevention of cancer-related mortality. It is estimated that successful implementation of screening programs could reduce the mortality of specific cancers by 3% to 60%.^[11] Screening for breast and cervical cancers alone could reduce overall cancer mortality by 3%, for an equivalent of nearly 15,000 lives saved per year. The potential public health impact of early diagnosis and detection of malignancy is clear. Yet, not all malignancies are amenable to screening programs. Although lung cancer is the leading cause of cancer-related mortality in the United States, no screening program has proven to be effective for reduction of lung cancer mortality. Indeed, for many cancers, the utility of secondary prevention is hampered by the lack of adequately sensitive and specific screening tests as well as the low prevalence of cancers at individual primary sites. Even for those malignancies that yield to screening programs, the divergent guidelines and recommendations of professional bodies such as the ACS and the NCI can be confusing. Finally, noncompliance by both patients and physicians further compromises the efficacy of available screening tests.

[edit] Screening Program Development

The impetus for the development of cancer screening programs and early detection efforts is the direct correlation between the extent of disease (or stage) at diagnosis and cancer-related mortality. Patients diagnosed with early stages of cancer have improved survival and decreased mortality over those patients diagnosed with more advanced stages of disease. Thus the fundamental goal of screening is to reduce cancer-related mortality through early detection and diagnosis. However, for a screening program to successfully reduce mortality, several prerequisites must be fulfilled (Box 122-6). Mammography and cervical cytologic sampling are examples of moderately priced screening tests with high patient acceptability that can identify asymptomatic individuals within a population who have a higher probability of having a particular cancer than the population as a whole. Unfortunately, there are many common malignancies such as ovarian cancer for which screening tests are overly sensitive and inadequately specific, with a positive predictive value too low for use in the general public. Other cancers, such as leukemia with no identifiable localized phase, elude early detection.

The cancer screening protocols advocated by professional agencies are based on demonstration of improved outcomes over what would be expected without screening. It is important to note that the definition of a better outcome underlies the divergence in cancer screening recommendations and guidelines by these well-respected agencies. A reduction in mortality in a screened population over that of an unscreened population, as demonstrated in a randomized clinical trial, provides the strongest evidence of improved outcome and establishes the efficacy of a screening technique. Currently, evidence of this strength only exists for breast and colon cancer. Other measures of improved outcome, such as a shift to earlier stage at diagnosis, an increase in survival, and an increase in detection rates, serve only as flawed substitutes for the gold standard of mortality reduction. These alternative measures suffer from inherent biases (Table 122-3), minimized by randomized clinical trials using mortality reduction as a primary end point. However, randomized clinical screening trials are not feasible for every malignancy and screening procedure. The number of patients required to perform randomized studies that will generate statistically and clinically significant results is prohibitive except for the three or four most common malignancies. For example, a screening study designed to detect a 25% reduction in mortality from prostate cancer would require >100,000 participants. A screening study of the same statistical design in testicular cancer would require over 6.5 million participants.^[12] Furthermore, a screening study designed to detect mortality differences often requires many years of follow-up with the inherent complexity and expense of long-term follow-up.

Table 122-3 Evidence of Benefit for Cancer Screening Strategies

Consequently, cancer screening recommendations based only on those screening protocols that manifest mortality reduction in randomized trials would abrogate all efforts at early detection through public education and all physical examinations on asymptomatic individuals. Self-examinations of the skin, testes, and breasts also would be eliminated by this strict approach. However, many authorities agree that this strict approach is not warranted for all screening techniques, in view of the fact that randomized clinical trials are not feasible for every malignancy. Although invasive or expensive screening strategies should be supported by the strongest possible evidence, the evidence required to document the efficacy of public education or self-examinations need not be as stringent.

[edit] Screening Methods

There is a broad array of methods that can be employed to enhance early detection of cancers. The spectrum includes public education to heighten awareness of malignancies and increase familiarity with early warning signs (Box 122-7), as well as risk assessment, health surveys, instruction in self-examination, physician examination, and mass screening. The vast majority of these methods are directly within the purview of the primary care physician. However, the extent of screening activity varies widely among physicians' practices. Those with a systematic approach to the preventive services are often the most successful at achieving and maintaining high levels of screening.Box 122-8 outlines the components of a successful practice-based implementation of cancer screening.

[edit] Early Detection Guidelines

The ACS was among the first professional medical organizations to advocate screening protocols. Numerous other groups such as the American Medical Association, the American College of Obstetricians and Gynecologists, and the American College of Radiology have promulgated recommendations or guidelines for screening of one or more malignancies. In 1987, the NCI convened representatives from many of these and similar organizations with the goal of establishing working guidelines for early cancer detection based on the best available statistical and clinical evidence. The NCI sought to establish working guidelines rather than firm recommendations in recognition of the fact that much of the available data was imperfect, with guidelines subject to change as new evidence develops. In addition, physicians might well adjust screening programs appropriately for individual patient circumstances that would include family and personal history as well as concomitant medical illnesses.

The NCI effort led to working guidelines for the early detection of cancers of the skin, oropharynx, breast, colon, rectum, cervix, prostate, and testes. Data for early detection of other malignancies were insufficient to support guidelines. These guidelines, compared with those of the ACS, are presented in Table 122-4. The individual organ sites are discussed below.

Table 122-4 Cancer Screening Recommendations/Guidelines of the American Cancer Society and the National Cancer Institute

[edit] Breast Cancer.

The landmark study opened in 1963 by the Health Insurance Plan of New York (HIP) demonstrated a statistically significant mortality reduction following annual screening mammography and clinical breast examination for women ages 40 to 64. In the years since the inception of the HIP trial, numerous controlled trials have confirmed the benefit of screening mammography for women older than 50 years of age.^[2] Until recently, only the HIP study has provided direct evidence of benefit (mortality reduction) for screening mammography in women 40 to 49 years of age. This paucity of evidence supporting a direct benefit for younger women persisted despite population surveillance data (SEER) demonstrating a decline in breast cancer mortality in the face of rising incidence for women under age 50. Consequently there has been persistent controversy over screening mammography guidelines. The controversy was fueled in late 1993 when the NCI dropped its guideline for screening mammography in women ages 40 to 49 after several published trials failed to demonstrate benefit for women of this age group. However, evidence of benefit from screening mammography for women in their forties has emerged with longer follow-up of previously completed trials. The current NCI guidelines reflect the updated trial results.

Women at increased risk for breast cancer through an inherited predisposition as a result of one or more germline mutations in BRCA1 or BRCA2 deserve special consideration with regard to cancer screening. The Cancer Genetics Studies Consortium points out that the benefit of cancer surveillance and other measures to reduce inherent risk in individuals who possess cancer-predisposing mutations is presumptive at best.^[13] Nonetheless, monthly self-examination beginning in early adulthood and annual or semiannual physician examinations in conjunction with annual mammography beginning at age 25 to 35 are suggested recommendations for this population whose lifetime risk of breast cancer may exceed 85%.

[edit] Colorectal Cancer.

Colorectal cancer affecting both men and women is the second leading cause of cancer-related deaths in the United States. Both incidence and mortality are decreasing. Although the recommendations of the ACS are in concert with the guidelines of the NCI, several agencies, such as the U.S. Preventive Services Task Force, have found insufficient evidence to support firm recommendations. Virtually all screening studies using any of these modalities, either singly or in combination, have demonstrated a shift to earlier stage at diagnosis as well as an increase in survival. Yet, despite a wealth of indirect evidence supporting screening for colon cancer, direct evidence of benefit in the form of a mortality reduction was lacking until the Minnesota Colon Cancer Control Study demonstrated a 33% decrease in mortality following annual fecal occult blood testing (FOBT) in a randomized study in over 46,000 volunteers.^[14] The paucity of direct evidence has prompted alternative approaches to the formulation of screening recommendations for colon cancer. Based on a mathematical model of 11 possible colon cancer screening strategies that incorporates the natural history of the disease, individual screening test performance, and costs associated with the individual strategies, Eddy suggests that annual FOBT and flexible sigmoidoscopy every 5 years might reduce mortality by 40% in individuals of average risk.^[15] Notably, the estimated reduction in mortality is even greater when colonoscopy or barium enemas are substituted for flexible sigmoidoscopy in the screening strategy but are more costly and less well tolerated by patients. Other investigators confirm that the benefits of colonoscopy in detecting treatable neoplasms are compromised by the costs of the procedure. A cost-effectiveness model for colon cancer screening created by Lieberman illustrates the compromises required during selection of appropriate screening strategies for individual patients. In his model, FOBT alone was clearly the most cost-effective, but with fewer cancer deaths prevented than other strategies. Flexible sigmoidoscopy in conjunction with FOBT increases the rate of cancer prevention, but one-time colonoscopy (e.g., age 60) had the greatest impact on colorectal cancer mortality.^[16]

Individuals with cancer-predisposing germline mutations, such as with hereditary nonpolyposis colorectal cancer, should undergo colonoscopy every 1 to 3 years beginning at age 25 years. In addition, endometrial cancer screening is also recommended. Although individuals who possess BRCA1 mutations are possibly at increased risk of colorectal cancer, no special screening strategies are recommended at this time.

[edit] Prostate Cancer.

Screening and early detection of prostate cancer remain very controversial. To date, no randomized controlled studies have demonstrated prostate cancer mortality reduction for any test or procedure, including digital rectal examination (DRE). Consequently, many organizations decline to recommend any specific prostate screening strategies. The NCI makes no recommendations for screening based on the profound lack of evidence to establish prostate cancer–specific mortality reduction with DRE, transrectal ultrasound (TRUS), or serum markers (i.e., prostate- specific antigen [PSA]). However, several, although not all, studies have suggested improved survival and a shift to earlier stage at diagnosis for men diagnosed with prostate cancer following annual DRE. Nonetheless, some experts feel that the DRE alone is too insensitive to significantly alter prostate cancer mortality. The ACS has included PSA screening in its recommendations for early cancer detection. Results from the National Prostate Cancer Detection Project, conducted by the ACS on over 2400 men undergoing DRE, PSA, and TRUS screening, have suggested that DRE used in combination with PSA screening may be more cost-effective than DRE alone. The cost of DRE as the only screening technique rose significantly when DRE was not performed by a highly skilled examiner such as a urologist.^[17] Interestingly, the ACS recommends that prostate cancer screening only be performed on men who have at least a 10-year life-expectancy. The ACS further recommends that younger African-American men and those at increased risk with two or more affected first-degree relatives also undergo regular screening.

[edit] Cancer of the Cervix.

The guidelines and recommendations for early detection of cervical cancer of the various professional organizations are generally concordant. The NCI notes that strong evidence from case-controlled and cohort studies suggests a decline in mortality following regular screening with Pap tests. However, the ACS suggests that the recommendation for pelvic examination be similar to the recommendation for Pap testing (i.e., a pelvic examination should be performed with Pap testing every 1 to 3 years for women between the ages of 18 and 40). In addition, the ACS recommends that yearly pelvic examinations be performed on women over the age of 40. The U.S. Preventive Services Task Force has suggested that Pap testing may be discontinued at 65 if previous examinations have been consistently normal.

Although direct evidence of mortality reduction is lacking, several large studies in Sweden, Finland, the United States, and Canada have demonstrated reductions in cervical cancer incidence and mortality rates following initiation of screening programs.^{[18][19][20][21]} In Canada, the reduction in mortality correlated with the intensity of screening. Cost-sensitive scrutiny of the available data has led several independent investigators to conclude that healthy women, with prior normal examinations, can be screened every 3 years, with the optimal screening interval potentially even less frequent. The wealth of indirect evidence of benefit and the widespread acceptance of regular Pap testing in conjunction with pelvic examination in screening for cervical cancer make a randomized controlled trial of the efficacy of this screening strategy improbable.

[edit] Skin Cancer.

Skin cancers, melanoma in particular, are ideal candidates for early detection. Not only can physical examination detect over 90% of melanomas arising in the skin, but the vast majority of melanomas have a prolonged horizontal growth phase prior to the vertical invasion that correlates with prognosis. This prolonged period of horizontal growth provides lead time in which locally confined lesions can be detected. Aggressive public and professional education in Australia has led to an increase in the proportion of patients diagnosed with early-stage melanoma, as well asimprovement in survival. Similar trends have been noted in Scotland and the United States. Although studies designed to document mortality reduction following screening are lacking, the indirect evidence of benefit, in conjunction with the low cost, absence of morbidity, ease of performance, and patient acceptability, provides ample justification for this guideline. Individuals with pigmented nevi or personal or family history of skin cancer or dysplastic nevi should be targeted for special surveillance.

[edit] Oral Cancer.

Routine annual oral examination can lead to a shift in the proportion of patients diagnosed with early-stage disease. Physicians should pay special attention to those at high risk due to socioeconomic status and tobacco and alcohol use. More than 90% of oral cancers occur in patients over the age of 45. The oral cavity is unique in that careful physical examination can detect premalignant lesions such as leukoplakia and erythroplasia that are asymptomatic. The white patches that constitute leukoplakia occur most commonly on the lower lip, floor of the mouth, buccal mucosa, lateral tongue back, and retromolar region. Erythroplasia occurs as velvety patches on the floor of the mouth, lateral tongue, and soft palate. Over 90% of these lesions are found to harbor severe epithelial dysplasia, carcinoma in situ, or frankly invasive carcinoma. Identification of these early lesions is an important means of distinguishing high-risk patients who can be targeted for intensive counseling (tobacco and alcohol cessation) and close follow-up. Unfortunately, there is only scant evidence supporting a role for screening. In the future, patients with premalignant lesions may benefit from chemopreventative strategies.

[edit] Testicular Cancer.

The most common malignancy in men aged 20 to 34, testicular cancer is most often discovered by patients themselves. Localized and regional testicular cancer is highly curable. The accessibility of the testicles allows for earlier detection of testicular masses than might otherwise be possible. The benefit of testicular examination is unlikely to be conclusively demonstrated in clinical trials given the low prevalence of this disease. Nonetheless, the low cost and morbidity of self-examination and physician examination support incorporation of these techniques into an early detection program.

Finally, many sites of potentially malignant growth are within reach of a skilled examiner's eyes and hands. Sarcomas, lymphoid malignancies, and thyroid cancers initially may be palpable or even visible as subtle alterations in body contours. Every opportunity should be exploited to detect and diagnose abnormalities that might otherwise go unnoticed until more advanced disease creates symptoms that demand medical attention.

[edit] FOLLOW-UP

Cancer is increasingly curable. Approximately 40% to 50% of those diagnosed with cancer will survive the disease. Countless others will live for extended periods of time with treatment-responsive disease. These numbers are likely to increase in the future as the anti cancer armamentarium expands with new strategies and methodologies. All of these patients require follow-up to monitor for distant effects of cancer therapy, recurrence of disease, or development of additional primary tumors.

A reasonable follow-up strategy for the most common and/or most treatable malignancies is suggested in Table 122-5. The natural history of the primary cancer determines the frequency and procedures of follow-up. In addition, cancer survivors may be at increased risk for primary cancers in second or third sites and should continue to undergo routine cancer screening. It is important to note that few follow-up strategies have been subjected to the same rigorous analyses that have been required for screening strategies. Nonetheless, early detection of recurrent tumors may allow some patients to be salvaged with additional treatment. For example, a rising carcinoembryonic antigen (CEA) following resection of a colon cancer might detect isolated hepatic or pulmonary metastases. Some patients with hepatic metastases confined to one lobe or region of the liver can be cured by hepatic resection. Whereas the mean survival for patients with metastatic disease is 24 months, the 5-year survival for patients with isolated hepatic or pulmonary metastases is approximately 30%.

Table 122-5 Possible Strategies for Follow-up of Selected Cancers

Some investigators question the benefit of periodic follow-up, suggesting that the improved survival associated with detection of occult recurrences may result from lead time and length bias. The increasing use of tumor markers for periodic follow-up has fueled the controversy over what constitutes appropriate follow-up. Many tumor markers are nonspecific and can be elevated by nonmalignant as well as malignant processes. An elevated CEA following breast cancer can lead to an extensive, costly search for a pathologic etiology. The sensitivity of the CEA assay may exceed the ability to detect recurrent lesions. The resultant fruitless search may leave the alarmed patient and physician unsatisfied and frustrated. Some tumor markers have well-defined roles in cancer follow-up, as outlined in Table 122-6. Although controversy over appropriate follow-up regimens persists, detection of abnormalities in any follow-up study should prompt a complete diagnostic evaluation and referral when appropriate.

Table 122-6 Tumor-associated Antigens With Efficacy in Follow-up

[edit] REFERENCES