Marking what many cancer researchers are calling an important trend, the actual number of cancer deaths in the United States fell by more than 3,000 from 2003 to 2004, the most recent year for which mortality data are available. The steady decline in the rate of cancer deaths also continued. 

The downturn from 556,902 deaths in 2003 to 553,888 in 2004 marks the biggest single-year decrease ever and the second consecutive year in which actual deaths - not just the death rate - have dipped.

"This decrease in actual number of cancer deaths, coupled with our SEER data showing a decline in the rate of cancer deaths since 1991, is extremely encouraging and highlights real progress," said NCI Director Dr. John E. Niederhuber.

The new mortality numbers were released as part of an annual report released by the American Cancer Society (ACS), Cancer Statistics 2007, which reports on the most recent mortality data, as well as providing cancer incidence and mortality estimates for the current year. Last year's publication showed a decrease of 369 cancer deaths from 2002 to 2003, the first-ever reported decline in cancer deaths since such statistics have been reported.

The new report has sparked intense optimism.

"This second consecutive drop in the number of actual cancer deaths, much steeper than the first, shows last year's drop was no fluke," said ACS Chief Executive Officer Dr. John R. Seffrin.

More Data, Better Estimates An important change to this year's ACS report, explained Dr. Linda Pickle, from NCI's Division of Cancer Control and Population Sciences (DCCPS), is that the estimates of cancer incidence for 2007 were based on data from far more cancer registries, using an improved prediction method. Previous reports, she explained, included estimates that relied strictly on the original nine registries in NCI's Surveillance Epidemiology and End Results (SEER) program. That meant it only covered about 10 percent of the U.S. population. For the 2007 report, the estimates are based on data from SEER, CDC, and the North American Association of Central Cancer Registries. "Combined, those registries cover roughly 40 states and 86 percent of the U.S. population," she said. The new method also accounts for geographic variation of factors such as smoking patterns or income in producing incidence estimates.

Even though the mortality rate has been decreasing for some time, added Dr. Gabriel N. Hortobagyi, president of the American Society of Clinical Oncology, the fact that the number of cancer deaths has decreased during a time of continued population growth "is very encouraging."

The drop in deaths was largely attributed to the reductions in smoking prevalence over the past several decades, improvements in cancer screening rates, and the use of increasingly effective treatment regimens. Screening's impact was most evident for colorectal cancer, which had the greatest reductions in cancer deaths among both men and women from 2003 to 2004.

According to Dr. Brenda Edwards, associate director of the NCI Surveillance Research Program, collaborative research is ongoing by the NCI-sponsored Cancer Intervention and Surveillance Modeling Network to better understand the impact of risk factor reduction and prevention, screening, and treatment on trends in death and incidence rates for cancers of the breast, colon-rectum, lung, and prostate.

The report, led by Dr. Ahmedin Jemal, from the ACS Department of Epidemiology and Surveillance Research, notes that cancer still accounted for about 23 percent of all U.S. deaths in 2004, with only heart disease responsible for more. Also, with the exception of a few cancer sites, Dr. Jemal and his ACS colleagues wrote, "[Cancer] incidence and death rates are consistently higher in African Americans than in Whites."

Although improvements in treatment for some cancers have lagged behind those seen in screening, Dr. Hortobagyi said he was optimistic that, with the pattern that's begun to develop with the newer targeted agents, that would start to change.

The pattern, he explained, begins with a "biological signal" in early-stage trials that a drug could be effective. It then moves into larger trials where it demonstrates some prolongation in disease progression and survival in previously treated metastatic disease. Finally, it transitions to testing in an early-stage disease with a curative intent.

"We've seen that with breast cancer and colorectal cancer, for example, and we're starting to see it in other tumors, as well," he said.

By Carmen Phillips