Examining 25 years of cancer records nationwide, researchers concluded that smaller tumor size accounted for 61 percent of the improvement in survival when cancer had not spread beyond the breast, and 28 percent when it had spread just a little.

For women 65 and older with early stage tumors -- the most common scenario -- the shift in size accounted for virtually all of the improvement in survival.

"We don't in any way want to diminish the benefits we've seen from advances in treatment because they've been enormous," said lead researcher Elena Elkin. "But not all of the improvement in survival is due to treatment when important characteristics like size have also changed over time."

Importance of Screening

The study wasn't designed to determine the value of mammograms or treatments. But it implies much about the value of early detection.

"This really helps to show the importance of screening," said Debbie Saslow, who heads breast cancer research at the American Cancer Society. "In addition to finding more small tumors, we're also finding less big tumors."

Saslow had no role in the study, which was being published Monday online by the society's journal Cancer and will be in its Sept. 15 print edition.

It was conducted by doctors at Memorial Sloan-Kettering Cancer Center in New York and used a federal government's database that includes nine cancer registries covering 10 percent of the US population. More than 265,000 breast tumors were analyzed.

Breast cancer is the most common cancer in American women. An estimated 211,240 new cases and 40,400 deaths from it are expected this year.

Survival has increased, but experts have argued over how much of that is because of better drugs or tumors being found at earlier stages. Two-thirds of breast cancers today are diagnosed at the local stage, when they're still confined to the breast; in the 1970s, only half were.

Bigger Difference for Older Women

However, this is the largest study in American women to look at size within those stages.

"Even within the same stage category, the average tumor size is smaller today than it was 25 years ago," Elkin said.

For example, the number of local-stage breast cancers that were smaller than 1 centimeter rose from less than 10 percent from 1975 through 1979 to 25 percent from 1995 through 1999. An inch is about 2.5 centimeters.

Of regional-stage cancers -- those that spread to nearby tissue or lymph nodes but not widely throughout the body -- the portion that were smaller than 2 centimeters rose from one-fifth to one-third.

Next, researchers compared five-year survival rates for these time periods, taking into account the shift in tumor size.

For women with local-stage breast cancers, survival rose from nearly 91 percent to more than 97 percent, but was only 93 percent after adjusting for smaller tumors. Looked at another way, the shift in size accounted for 61 percent of the improvement in survival.

For regional cancers, survival rose from about 68 percent to about 80 percent, but was only 76 percent once tumor size was factored in.

Size made a much bigger difference for older women than younger ones. A whopping 96 percent of the survival improvement for women 65 and older with local-stage cancers was explained by this. Only 38 percent of the improvement in women under 50 was due to the shift in tumor size.

Mammogram Benefit

"It isn't necessarily because treatment works better for certain women, it reflects who's getting more," because younger women are more likely to receive chemotherapy, Elkin said.

It also shows that older women have benefited from mammograms, Saslow said.

Federal surveys show that the percent of women 40 and older who had a mammogram in the previous two years increased from 29 percent in 1987 to 70 percent in 2000. The federal government recommends that women over 40 get mammograms every 1 to 2 years, with or without a breast exam by their doctor.

However, Dr. Barnett Kramer, associate director for disease prevention at the National Institutes of Health, noted that the study did not have information on how many of these tumors were found through mammograms or what treatments various groups of women received, so no direct conclusions about the value of these can be drawn.

"You don't know how the cancers were diagnosed so there's no way to link any type of testing or physical examination ... with the results of this study," he said.

Attributing the benefit of smaller tumors to early detection and mammograms "makes a lot of intuitive sense, but that's of course why we do studies as opposed to depend on leaps of logic," he said.