• 7/10/2004
  • Bethesda, MD
  • Karen Antman, MD
  • National Cancer Institute

Since about 1990, cancer mortality per 100,000 population in the United States has been falling. This trend has been driven by decreasing mortality in the four most common malignancies: prostate, breast, lung, and colon cancers.1 Although many debate the impact of either cancer screening or treatment, the decrease in mortality for breast, colon, and prostate cancers is attributable to better treatment, to screening, or, probably, to both. The fall in lung cancer deaths almost certainly results from the fact that significant numbers of Americans quit smoking 15 to 20 years ago. Most of those who quit were men; consequently, lung cancer mortality in men peaked in about 1990 and is now falling, whereas mortality in women has only leveled off. Women have particular difficulty quitting smoking, perhaps because our society places great value on thinness, and women who quit tend to gain weight. Depression, which also decreases the success rate of quitting, is approximately twice as common in women as in men.

Mortality per 100,000 is a clean statistical end point. Although 5-year survival rates have also improved, from about 40% in the 1950s to 62% today, these rates are subject to early detection bias; they can be inflated by the implementation of screening programs, which increase the detection of small lesions with a better prognosis or of premalignant lesions of uncertain clinical significance.1 Nevertheless, because large numbers of baby boomers are now entering their 50s—an age when the risk of cancer increases substantially—the absolute numbers of Americans with cancer will almost certainly rise over the next decade.

Although mortality from the most common cancers is falling, a number of malignancies are increasing in incidence and mortality, among them pancreatic cancer, esophageal cancer, non-Hodgkin lymphoma, and melanoma. Melanoma may be related to sun exposure, and pancreatic cancer is partly associated with smoking. The rise in distal adenocarcinomas is associated with Barrett esophagus, gastric reflux, and obesity; the risk of more proximal squamous esophageal cancers has remained stable.

In terms of mortality from cancer, not all Americans are created equal. Mortality for breast and colon cancers is significantly higher in the Northeast. Mortality from prostate cancers is higher in the Southeast. Despite Medicare, some patients older than 65 years do not receive adjuvant chemotherapy for colon cancer2,3 or the current standard treatments for ovarian cancer,4 even though randomized trials have documented significant survival benefit.

Furthermore, different racial and ethnic groups have substantially different mortalities from various cancers. In children with acute lymphoblastic leukemia, survival is better in whites than in African Americans, Hispanics, or Native Americans.5 Mortality from prostate cancer is twice as high in African-American men as in white men.1 Breast cancer incidence is lower in African-American women (96 per 100,000) than in white women (113 per 100,000), but breast cancer mortality for African-American women is higher. Although breast cancer mortalities were similar for African Americans and whites in the United States in the 1970s, the rates diverged as better treatments were developed, falling in whites but not in African Americans.6,7

African-American women often have more advanced breast cancer at diagnosis, with lower percentages of well-differentiated and estrogen receptor-positive disease.8 In addition, breast cancer survival is higher in African-American women treated in military facilities than in African-American women as a whole, as reported in the Surveillance, Epidemiology, and End Results (SEER) Program. Those statistics suggest that differences in access to quality care, insurance, and social support account for much of the racial disparity.9,10 Nevertheless, among military beneficiaries, whites still survived significantly longer than African Americans, so some component may be biologic.11

Biologic differences between ethnic groups are inadequately studied, but differences in drug metabolism and transport are well described and could affect sensitivity to carcinogens, as well as responses to chemotherapy.12,13 A smaller percentage of breast tumors from African Americans express estrogen and progesterone receptors; higher receptor levels are associated with a better prognosis.10,14 In at least one study, however, other molecular indices associated with breast cancer prognosis did not differ significantly among whites, African Americans, and Hispanics.15

To have the greatest impact on the burden of cancer, we must not only discover and develop new effective therapies but also ensure equal access to established prevention, diagnostic, and treatment services.