Harrisons Manual of Medicine, 18th Ed.

CHAPTER 216. Prevention and Early Detection of Cancer

One of the most important functions of medical care is to prevent disease or discover it early enough that treatment might be more effective. All risk factors for cancer have not yet been defined. However, a substantial number of factors that elevate risk are within a person’s control. Some of these factors are listed in Table 216-1. Every physician visit is an opportunity to teach and reinforce the elements of a healthy lifestyle. Cancer screening in the asymptomatic population at average risk is a complex issue. To be of value, screening must detect disease at a stage that is more readily curable than disease that is treated after symptoms appear. For cervix cancer and colon cancer, screening has been shown to save lives. For other tumors, benefit is less clear. Screening can cause harm; complications may ensue from the screening test or the tests done to validate a positive screening test or from treatments for the underlying disease. Furthermore, quality of life can be adversely affected by false-positive tests. Evaluation of screening tools can be biased and needs to rely on prospective randomized studies. Lead-time bias occurs when the natural history of disease is unaffected by the diagnosis, but the pt is diagnosed earlier in the course of disease than normal; thus, the pt spends more of his/her life span knowing the diagnosis. Length bias occurs when slow-growing cancers that might never have come to medical attention are detected during screening. Overdiagnosis is a form of length bias in which a cancer is detected when it is not growing and is not an influence on length of survival. Selection bias is the term for the fact that people who volunteer for screening trials may be different from the general population. Volunteers might have family history concerns that actually elevate their risk, or they may be generally more health-conscious, which can affect outcome.



The various groups that evaluate and recommend screening practice guidelines have used varying criteria to make their recommendations (Table 216-2). The absence of data on survival for a number of diseases has led to a lack of consensus. In particular, four areas are worth noting.






1. Prostate cancer: Prostate-specific antigen (PSA) levels are elevated in prostate cancer, but a substantial number of the cancers detected appear to be non-life-threatening. PSA screening has not been shown to improve survival. Efforts are underway to develop better tests (predominantly using bound vs. free and rate of increase of PSA) to distinguish lethal and nonlethal cancers.

2. Breast cancer: The data on annual mammography support its use in women age >50 years. However, the benefit for women age 40–49 years is quite small. One study shows some advantage for women who are screened starting at age 40 that appears 15 years later; however, it is unclear if this benefit would not have also been derived by starting screening at age 50 years. Women age 40–49 years have a much lower incidence of breast cancer and a higher false-positive rate on mammography. Nearly half of women screened during their forties will have a false-positive test. Refined methods of screening are in development.

3. Colon cancer: Annual fecal occult blood testing after age 50 years is felt to be useful. Colonoscopy is the gold standard in colorectal cancer detection, but it is expensive and has not been shown to be cost-effective in asymptomatic people.

4. Lung cancer: Chest radiographs and sputum cytology in smokers appear to identify more early-stage tumors, but paradoxically, the screened pts do not have improved survival. Low-dose spiral CT scanning performed annually for 3 years reduces lung cancer death in older smokers by 20% compared with annual chest x-ray. However, 96% of the positive tests are false-positives and overall survival is improved by only 6.7%.



Risk factors include age, early menarche, nulliparity or late first pregnancy, high body-mass index, radiation exposure before age 30 years, hormone-replacement therapy (HRT), alcohol consumption, family history, presence of mutations in BRCA1 or BRCA2, and prior history of breast neoplasia. Risk assessment models have been developed to predict an individual’s likelihood of developing breast cancer (see www.cancer.gov/cancertopics/pdq/treatment/breast/healthprofessional#Section_627).


MRI scanning is a more effective screening tool than mammography in women with a familial breast cancer risk.


Women whose risk exceeds 1.66% in the next 5 years have been shown to have a 50% reduction in breast cancer from taking tamoxifen or raloxifene. Aromatase inhibitors have generally been superior to tamoxifen in the adjuvant treatment of hormone-sensitive breast cancer, and one of them (exemestane) reduces the risk of breast cancer by 65% in postmenopausal women at increased risk. Women with strong family histories should undergo testing for mutations in BRCA1 and BRCA2. Mutations in these genes carry a lifetime probability of >80% for developing breast cancer. Bilateral prophylactic mastectomy prevents at least 90% of these cancers but is a more radical prevention than the usual treatment for the disease. In addition, bilateral salpingo-oophorectomy reduces ovarian and fallopian tube cancer risk by about 96% in women with BRCA1 or BRCA2 mutations.


Risk factors include diets high in saturated fats and low in fruits and vegetables, smoking, and alcohol consumption. Stronger but less prevalent risk factors are the presence of inflammatory bowel disease or genetic disorders such as familial polyposis (autosomal dominant germline mutation in APC) and hereditary nonpolyposis colorectal cancer (mutations in DNA mismatch repair genes hMSH2 and hMLH1).


Pts with ulcerative colitis and familial polyposis generally undergo total colectomy. In familial polyposis, nonsteroidal anti-inflammatory drugs (NSAIDs) reduce the number and size of polyps. Celecoxib, sulindac, and even aspirin appear to be effective, and celecoxib is approved by the U.S. Food and Drug Administration for this indication. Calcium supplementation can lead to a decrease in the recurrence of adenomas, but it is not yet clear that the risk of colorectal cancer is decreased and survival increased. The Women’s Health Study noted a significant reduction in the risk of colorectal cancer in women taking HRT, but the increase in thrombotic events and breast cancers counterbalanced this benefit. Studies are underway to assess NSAIDs with and without inhibitors of the epidermal growth factor (EGF) receptor in other risk groups.


Risk factors include smoking, exposure to radiation, asbestos, radon.


Smoking cessation is the only effective prevention (Chap. 217). NSAIDs and EGF receptor inhibitors are being evaluated. Carotenoids, selenium, retinoids, and α-tocopherol do not work.


Risk factors include age, family history, and possibly dietary fat intake. African Americans are at increased risk. The disease is highly prevalent, with autopsy studies finding prostate cancer in 70–80% of men over age 70.


In a group of men age ≥55 years with normal rectal examinations and PSA levels <3 ng/mL, daily finasteride reduced the incidence of prostate cancer by 25%. Finasteride also prevents the progression of benign prostate hyperplasia. However, some men experience decreased libido as a side effect. The Gleason grade of tumors seen in men taking finasteride prevention was somewhat higher than the controls; however, androgen deprivation alters the morphology of the cells and it is not yet clear that the Gleason grade is a reliable indicator of tumor aggressiveness in the setting of androgen deprivation. Dutasteride, another 5α-reductase inhibitor, had similar effects. The FDA has reviewed the data and concluded that the reduction in risk is primarily in the group of pts with low-grade tumors whose risk from prostate cancer is unclear. One additional high-grade tumor emerges for every 3–4 low-grade tumors averted. More follow-up is necessary to see if higher-grade tumors emerging on preventive therapy have the same aggressive behavior as those occurring in the absence of preventive hormone blockade.


Risk factors include early age at first intercourse, multiple sexual partners, smoking, and infection with human papillomavirus (HPV) types 16, 18, 45, and 56.


Regular Pap testing can detect nearly all cases of the premalignant lesion called cervical intraepithelial neoplasia. Untreated, the lesion can progress to carcinoma in situ and invasive cervical cancer. Surgical removal, cryotherapy, or laser therapy is used to treat the disease and is effective in 80%. Risk of recurrence is highest in women over age 30, those with prior HPV infection, and those who have had prior treatment for the same condition. A vaccine (Gardasil) containing antigens of strains 6, 11, 16, and 18 has been shown to be 100% effective in preventing HPV infections from those strains. The vaccine is recommended for all females age 9–16 years and could prevent up to 70% of all cervical cancer. The vaccine is not effective after infection has been established.


Risk factors include smoking, alcohol consumption, and possibly HPV infection.


Oral leukoplakia, white lesions of the oral mucosa, occur in 1–2 persons in 1000, and 2–3% of these pts go on to develop head and neck cancer. Spontaneous regression of oral leukoplakia is seen in 30–40% of pts. Retinoid treatment (13-cis retinoid acid) can increase the regression rate. Vitamin A induces complete remission in ~50% of pts. The use of retinoids in pts who have been diagnosed with head and neck cancer and received definitive local therapy has not produced consistent results. Initial studies claimed that retinoids prevented the development of second primary tumors, a common feature of head and neck cancer. However, large randomized studies did not confirm this benefit. Other studies are underway combining retinoids and NSAIDs with and without EGF receptor inhibitors.


Pts can be taught to look for early warning signals. The American Cancer Society has identified seven major warning signs of cancer:

• A change in bowel or bladder habits

• A sore that does not heal

• Unusual bleeding or discharge

• A lump in the breast or other parts of the body

• Chronic indigestion or difficulty in swallowing

• Obvious changes in a wart or mole

• Persistent coughing or hoarseness


For a more detailed discussion, see Crosswell JM, Brawley OW, Kramer BS: Prevention and Early Detection of Cancer, Chap. 82, p. 655, in HPIM-18.