ina Kolata asked two experts on opposite sides of the debate on early screening to look at breast cancer data and discuss what they indicate about screening and treatment.

DR. BARNETT KRAMER, director of the Office of Disease Prevention at the National Institutes of Health, looks at national trends in breast cancer and sees hints of overdiagnosis, meaning that mammography is finding many cancers that either would go away if left alone or that would be found and treated when they became larger. He points to two features of the breast cancer data that give him pause.
First, the number of new cases that are being found has increased over the past 20 years, since mammography came into widespread use. In 1983, there were 111.0 cases per 100,000 women, while in 1998 that number had risen to 140.5.

"There are very few things that can so dramatically increase the incidence of new disease," he said. One is the introduction of a powerful new carcinogen, like tobacco.

But, Dr. Kramer said, "we don't know of any carcinogens out there that could explain the increase in breast cancer." That, he said, leaves screening as an explanation. Mammography appears to find many cancers that would not otherwise have been found in a woman's lifetime.
Second, mammography can find very small tumors, and that ability explains why the number of women with in situ cancers, confined to the breast ducts, and small cancers that have not spread beyond the breast has risen. But there has been little decrease in the number of women with more advanced cancers.

In 1983, there were 6.1 women per 100,000 with in situ breast cancers. By 1998, there were 30.7 per 100,000. In 1983, there were 22.01 women per 100,000 with small cancers that were confined to the breast. In 1998, that number was 60.08 per 100,000.

In 1983, there were 26.83 women per 100,000 with large cancers, at least two centimeters in diameter but confined to the breast, which have a worse prognosis than smaller tumors. In 1998, that number was 23.81 per 100,000.

In 1983, there were 42.16 women per 100,000 with cancers that had spread to the underarm lymph nodes or had spread as a continuous tumor to the chest wall or muscle. These usually have a poor prognosis. In 1998, there were 38.24 per 100,000.

The number of women with breast cancers with the worst prognosis, those that spread to other organs, had been fairly constant in the years before mammography was introduced, and that trend did not change after the introduction of mammography.

The problem, Dr. Kramer said, is the numbers of cancers found early and found later, corrected for size and age of the population. If screening worked perfectly, every cancer found early would correspond to one fewer cancer found later. That, he said, did not happen. Mammography, instead has resulted in a huge new population of women with early stage cancer but without a corresponding decline in the numbers of women with advanced cancer.

DR. LARRY NORTON, a breast cancer expert, president of the American Society for Clinical Oncology and the chief of medical oncology at Memorial Sloan-Kettering Cancer Center, looked at the breast cancer data and said that mammography was allowing women to have less disfiguring surgery and less chemotherapy.

"Twenty five years ago, I saw big tumors," he said. "Now the whole nature of the disease has changed. In general, I do not think I have seen a Stage 3 cancer in years, and I used to see them every week," referring to disease so advanced that the entire tumor cannot be removed.

He said women should continue to have mammograms. "I have seen nothing yet that tells us we shouldn't screen while this issue is being thought about," he said.

The incidence rate has increased, but that does not mean that mammography's possible downside outweighs the good it can do. "Every time you are looking at an early intervention process, you are looking at overtreatment," Dr. Norton said.

Screening does find some tiny cancers that may not become deadly. But it also finds tiny tumors that will become untreatable or that will require drastic surgery and large doses of chemotherapy if they are found later.

"In some situations, it may be that early diagnosis is important and in others it is not," Dr. Norton said. But until there is a way of sorting out which tiny tumor is dangerous or until treatment gets so good that it does not matter when a tumor is found, early diagnosis makes sense, he said.

There has been a decline in later stage disease — the cancers that have spread beyond the breast — and although it is not as pronounced as the rise in early stage disease, it does indicate that mammography is finding cancers when they are more easily treated.

In discussing these data, Dr. Norton said he relied on analyses by Dr. Eric J. Feuer, a National Cancer Institute statistician, and others.

Dr. Feuer said that a number of factors could complicate an analysis. Since the mid-1930's, the incidence of breast cancer has been rising by about 1 percent a year, for unknown reasons, but perhaps including factors like delayed childbearing, Dr. Feuer said. In addition, the way doctors examine women to see if their cancers have spread has improved.

For example, the number of women with cancers that had spread to their lymph nodes fell in the late 1980's and early 1990's to 35.63 from 42.16 per 100,000 women, probably reflecting the effects of mammography, he said. Then the rate rose again to 38.24 per 100,000, probably reflecting better detection methods.

"We're trying as best we can to tease it apart," Dr. Feuer said.

Dr. Norton stressed that early detection and better treatments went together. Screening, he said, is needed to find the cancers most likely to be aided by better treatments.

"There's a package of rational behavior, and that's what I'm in favor of," Dr. Norton said. "I think early diagnosis is part of a package that is making advances against cancer."