There is no question that breast cancer death rates have fallen in recent decades. Because mammography and adjuvant therapies gained in popularity and use at about the same time, it is not a simple matter to separate out the relative contribution of each to improved survival.

The National Cancer Institute (NCI) supported the research, which was conducted by a total of 43 investigators working in seven groups, including the University of Texas MD Anderson Cancer Center.

Contribution of Each Method Debated

There has been a controversy among experts over what factor or factors to credit for improved breast cancer survival: screening mammography, better treatment or a combination of the two.

Each of the seven teams designed its own statistical model to determine the contribution of each method. These independent models used the same sources of data -- some of which had not been mined before -- but their approaches and assumptions differed.

The teams reached somewhat different conclusions but were closest to each other in estimating how much the adjuvant therapies tamoxifen and chemotherapy reduced mortality in patients (12 percent to 21 percent, with a median of 19 percent).

The range for screening mammography, however, was 7 percent to 23 percent (with a median of 15 percent), reflecting the greater uncertainty associated with estimating the benefit of screening.

Still, according to the models, the combination of screening and adjuvant therapy together reduced the breast cancer death rate by an estimated 25 percent to 38 percent, with a median of 30 percent -- which explains the drop in breast cancer mortality from 1975 to 2000, says the study's lead author, Donald Berry, PhD, chair of the Department of Biostatistics and Applied Mathematics at MD Anderson Cancer Center.

"While we didn’t agree with each other as to the percentages of benefit, all seven groups concluded that the decline in the rate of death from breast cancer is a combination of screening and therapy and not restricted to one or the other," he says.

"Screening would have no benefit if not followed by treatment, including surgery, and treatment has the potential to be more effective if cancer is detected at earlier stages by screening," he notes.

24% Fewer Deaths Than in 1990

The differences in conclusions reflect uncertainties in the interpretation of available information and differing modeling approaches, Berry stresses, rather than contradictions among the models.

"These are seven top modeling groups applying their efforts to the best data that we have available in this country," he says.

He adds that the survival benefits offered by screening and adjuvant therapy both may be lower than researchers had expected.

"Some people think the benefit of screening is huge, and others say that the reduction in death rates is due primarily to adjuvant therapy," Berry says. "No one has known for sure, and although we still don't know for sure, this is the best set of analyses that is possible given the available information."

There is no question that breast cancer death rates have fallen in recent decades, both in the United States and in other countries. In 1975, the rate of death from breast cancer among women 30 to 79 years of age, adjusted for age to the 2000 population, was 48.3 deaths per 100,000 women. By 1990, the rate had increased slightly to 49.7 per 100,000, but then fell to 38.0 per 100,000 by 2000, a decrease of 24 percent from 1990.

The two major changes in breast oncology over that time have been the advent of screening mammography (which 70 percent of women over age 40 received in 2000) and the use of adjuvant therapies -- chemotherapy and hormone therapy used in conjunction with primary treatment by surgery and/or radiation.

But because each of these factors gained in popularity and use at about the same time, it is not a simple matter to separate out the relative contribution of each to improved survival, Berry says.

Mammography vs. Chemotherapy and Tamoxifen

To find out, the CISNET groups used data that reflected "real life" use of screening and adjuvant therapy, including some population databases that had not been tapped before for this purpose.

Their analysis relied on the incidence of breast cancer as reported by the Surveillance, Epidemiology, and End Results (SEER) program and the rate of death from breast cancer as reported by the National Center for Health Statistics (NCHS), as well as additional databases concerning uses of screening and treatment and their efficacy in the population.

The models reached somewhat different estimates. Dana-Farber’s model found that screening mammography accounted for 65 percent of the reduced breast cancer death rate (with 35 percent due to use of chemotherapy and tamoxifen) while the MD Anderson model reached the opposite conclusion -- that 65 percent of the mortality benefit is due to adjuvant chemotherapy and 35 percent is due to screening.

Researchers in the study are comfortable with the level of disagreement concerning their point estimates and with the level of uncertainty reflected in their overall conclusions, says Berry, whose earlier work on mammography has focused on its downside. Women should be informed of the risks of screening as well as the benefits, he maintains.

Still, "the evidence makes clear that the benefit of screening on breast cancer mortality is very likely," Berry acknowledges, "while the benefit of providing adjuvant therapy is certain."

Models May Help Determine Best Strategies

The models now in place can be used to answer a number of questions related to screening and treatment, according to Berry.

"They provide a way for assessing the benefits of screening and treatment strategies different from the ones that were actually in place over the time period of the study," he notes.

"For example, our models can address what the effect on mortality would be if all women in particular age groups were to be screened at regular intervals," he says. "And it allows for addressing the impact on overall mortality if effective therapies were made available to particular types of patients with breast cancer."

The models may "help determine what strategies for delivering medical care are best for patients," suggests Berry, "and that is necessary, since our efforts, taken as a whole, haven't come close to eliminating breast cancer mortality."

The study was undertaken by the Cancer Intervention and Surveillance Modeling Network (CISNET), an NCI-sponsored consortium of investigators whose purpose is to measure the effect of cancer-control interventions on the incidence of and risk of death from cancer in the general population.

The seven breast cancer teams involved in this study are from Dana-Farber Cancer Institute in Boston; Erasmus University Medical Center in Rotterdam, the Netherlands; Georgetown University; MD Anderson Cancer Center; Stanford University; the University of Rochester; and the University of Wisconsin-Madison.