The 254 color-coded maps in the atlas will make it easy for researchers and state health departments to identify places where high or low rates occur, and to uncover patterns of cancer that would escape notice if larger areas, such as states, were mapped. The atlas will not tell researchers why death rates are higher in certain localities than in others, but it will provide important clues for further in-depth studies into the causes and control of cancer.

For the first time, maps will be presented for both white and black populations, since earlier mortality statistics lacked data that would permit a separation of blacks from the nonwhite category. In addition, the patterns for liver cancer and for biliary tract cancer will be shown for the first time. Previous disease classification schemes did not permit separate analyses of these cancers.

"It's not easy to sort out why cancer mortality rates vary among geographic areas and why they might be changing," said Joseph F. Fraumeni Jr., M.D., director of NCI's Division of Cancer Epidemiology and Genetics, and the senior author of the current and earlier atlases. "It's natural to want to know why cancer rates are elevated in a particular area, but it's important not to jump to conclusions before careful studies can be conducted in these areas."

Many of the patterns displayed in the current atlas are very similar to previous ones. High breast cancer rates, for instance, have been seen for four decades in urban centers in the Northeast. Studies have shown that regional variations in breast cancer rates are partially, but not entirely, due to established risk factors, including late age at first birth, early menarche, and late menopause, and to certain other factors, including education and mammography history.

Likewise, colon cancer mortality rates have been elevated in the Northeast for at least four decades. Dietary and nutritional factors are thought to be involved, but the specific causative elements are not clear.

"There are some geographic changes over time, however, that are particularly provocative," said Susan S. Devesa, Ph.D., lead author from NCI's Division of Cancer Epidemiology and Genetics. "The greatest changes are seen with lung cancer. The national annual lung cancer rate among white men rose from 39 per 100,000 during 1950-69 to 69 per 100,000 during 1970-94. In earlier atlases, elevated rates were seen in men in the Northeast and Southern Coastal areas of the United States. The highest rates now occur among white men in broad stretches across the South, among white women in the far West, and among the black population in northern urban areas," she explained. "These changes generally coincide with regional and time trends in cigarette smoking."

The patterns of prostate cancer deaths, which in the past showed little distinct geographic variation, have also changed. High rates among white men are much more prominent now in the north central areas. In contrast, among black men, rates are excessive in the southeastern United States, particularly in rural areas.

"We don't know right now why the patterns for prostate cancer are changing," said Robert N. Hoover, M.D., also at the Division of Cancer Epidemiology and Genetics and an author of both current and past atlases. "But the atlas provides clues that will stimulate efforts to find out what characteristics of these populations and their exposures are responsible for the higher rates."

Earlier atlases have been successful in generating leads that prompted further studies, particularly in high-risk areas of the country. The results of these studies are reported in the new atlas. Some of the findings include:

• The high rates of lung cancer among men in Southern Coastal areas were related to asbestos exposure resulting from work in shipyards, particularly during World War II.

• Elevated death rates for mouth and throat cancers among women living in the rural South were associated with use of smokeless tobacco.

• High death rates of esophageal cancer in Washington, D.C. and the coastal areas of South Carolina were linked to alcohol consumption and tobacco use, along with deficiencies in fruit and vegetable consumption.

• High lung cancer death rates were seen not only among smelter workers but also among people who live close to arsenic-emitting smelters.

• High colon cancer death rates in eastern Nebraska occurred mainly among persons of Czechoslovakian background, in whom dietary factors appeared to contribute to the risk.

"We've discovered that the reasons for high rates are quite varied," said Hoover. "Sometimes they're due to occupational exposures. For example, in addition to lung cancer in shipyard workers, we've seen elevated rates of nasal cancer among furniture and textile workers in particular areas of the Southeast. On the other hand, increased rates may be due to general environmental exposures, such as arsenic and lung cancer, or lifestyle differences, such as the regional patterns for breast cancer and, most likely, colon cancer."

One study has already been initiated in response to the new atlas. Bladder cancer among men has tended to cluster in the urban Northeast since the 1950s, particularly in areas with chemical industries. Previous studies in high-risk areas have also shown elevated risks among truck drivers and other workers exposed to motor exhausts. The main risk factor is cigarette smoking, which accounts for one-half of bladder cancer. But the new atlas has shown that Maine, Vermont, New Hampshire, and upstate New York have elevated rates in both sexes that have become more pronounced over time.

"We have been working with the states and with some of the academic departments in these states to develop a pilot study to pursue some of the current hypotheses for why the rates for bladder cancer have become more pronounced," said Hoover.

There are, however, some limitations to the atlas data. The atlas is less useful in generating research leads for cancer sites in which death rates do not vary much across the country, such as cancers of the pancreas and brain. In addition, for cancer sites with higher survival rates, such as breast and cervix, it is difficult to tell whether the geographic variation reflects environmental influences, factors related to medical care and health care delivery systems, reporting practices, migration patterns, or combinations of these variables.

The Atlas of Cancer Mortality in the United States, 1950-94 is a continuation of the cancer mapping project in NCI's epidemiology program. The first atlas with color-coded mortality maps at the county level was published in 1975 and covered the years 1950-69. The current atlas adds 25 years of data to the original atlas and compares the patterns for 1950-69 with those for 1970-94. In addition, for the first time, an interactive version of the data will be available on the Internet.

Web accessibility makes several new features possible. Not only can the maps, text, tables, and figures from the hard copy be downloaded from the Web site, but national and state mortality rates are also available, as are the tabulated data used to generate the maps. Another feature of the atlas Web site is that the user can create customized maps. For example, the user can compare rates in different time periods, look at rates for any cancer in any county, zoom and pan different areas of the country, and make color selections. Visit the atlas Web site at http://www.nci.nih.gov/atlas .

To order a single printed copy of the atlas, call NCI's Cancer Information Service at 1-800-4-CANCER (1-800-422-6237). It may also be ordered on NCI's on-line Publications Locator Service at http://publications.nci.nih.gov. The number for callers with TTY equipment is 1-800-332-8615.

For more information about cancer visit NCI's Web site for patients, public, and the mass media at http://www.nci.nih.gov.

Related Press Releases:
Questions and Answers for Atlas of Cancer Mortality in the United States, 1950-94

Research Contributions from Earlier Atlases