Nathaniel I. Berlin, MD, is currently Professor Emeritus at the Sylvester Comprehensive Cancer Center, University of Miami. He is the former director of the NCIs Division of Cancer Biology and Diagnosis as well as the former chairman of the NCIs Breast Cancer Task Force. He is also the former director and Professor Emeritus of the Robert H. Lurie Comprehensive Cancer Center, Northwestern University, Chicago.
Dr. Berlin was the senior physician at the NCI responsible for managing the Breast Cancer Demonstration Detection Task Force from its inception until he left the Institute in 1975.
ONI: The controversy over breast cancer screening in women aged 40 to 49 continues today after more than 2 decades of debate. How did it all start?
DR. BERLIN: The breast cancer screening controversy first began in 1975 as a result of criticisms surrounding the Breast Cancer Detection Demonstration Projects (BCDDP), a multi-institutional effort started in 1968 by the National Cancer Institute and the American Cancer Society to promote the use of mammography for breast cancer screening.
Criticisms included concerns that x-radiation to the breast would produce a significant number of breast cancers and that, based on analysis of existing data, there was no screening benefit to women under age 50.
Concern over x-radiation was dispelled in later years. Conversely, concern over analyses showing no screening benefit to women under 50 sparked a controversy that has yet to be resolved after almost a quarter century.
It should be noted that in 1975 the BCDDP was relying solely on data from a study being conducted by the Health Insurance Plan (HIP) of New York, an early health maintenance organization. While the HIP study was an important benchmark, the lack of data from comparable trials at the time was a large limiting factor for the BCDDP.
ONI: So the early criticisms of the BCDDP were in large part a criticism of data from the HIP study. How did the HIP study originate and what were its goals?
DR. BERLIN: The HIP study, which was the first randomized controlled clinical trial of cancer screening, began in 1963 with the support of the National Cancer Institute. Three questions were asked:
Would female members of HIP respond to a screening offer of a mammography and physical examination of the breast?
- Would women initially responding to the screening
offer comply with long-term screening efforts by returning annually
for 4 years?
- Would mammography and physical examination of the breast be associated with reduced breast cancer mortality?
At the time of the study, 62,000 women who were enrolled in HIP were randomly assigned to either the study or control group. The 31,000 women in the study group received a letter offering them the opportunity to receive breast cancer screening through mammography and physical examination of the breast. The offer was for a total of four screens that would be given annually. The 31,000 women in the control group received the usual and customary care of women enrolled in HIP and were not notified of the ongoing study.
ONI: Initially, what were the major findings of the HIP study?
DR. BERLIN: There was a positive response to the HIP study: 67% (20,700) of the women contacted came in for the first examination; 58% (17,980) returned for the second examination; 50% (15,530) returned for the third examination; and 39% (12,276) completed the series of four examinations.
Early data from the HIP study, reported in 1971, showed a 30% reduction in breast cancer mortality in the screened group. However, this analysis included the entire age spectrum, and a later subset analysis of women aged 40 to 49 at the time of entry into the study found that there was an insufficient number of women in this age group to draw any meaningful conclusions.
The inconclusive findings in the 40 to 49 age group were questioned in 1982, and the possibility of a screening benefit was raised. Repeated analyses of the HIP data in the 1980s and 1990s yielded a general acknowledgment in 1997 that there was a 25% statistically significant reduction in mortality for women in the 40 to 49 age group. This was the same reduction seen in women aged 50 to 59, although the reduction appears 3 to 5 years later in the younger age group.
ONI: There is a great deal of controversy surrounding the HIP study. What is this attributed to?
DR. BERLIN: The controversy stems from a deep-seated polarization dividing the medical and health policy communities. The health policy community argues that the reduction in mortality for younger women is either nonexistent or so small as to be insufficient for screening recommendations. These critics question the validity of the statistical methods used to show a screening benefit.
The medical community, usually represented by diagnostic radiologists, contends that a screening benefit is present in the study data and that they can detect breast cancer in women under age 50, often at a favorable stage.
ONI: The HIP study spurred the development of several large randomized screening studies. What are some of the other major breast cancer screening studies, and how have their results compared to those of the HIP trial?
DR. BERLIN: Eight different randomized screening studies followed the HIP trial: five in Sweden, one in Canada, and two in Great Britain. With the notable exception of the Canadian study, there is general agreement with the data from the HIP study.
ONI: In 1993, the NCI responded to continued controversy over the HIP study and data from newer screening studies by convening an International Workshop on Screening for Breast Cancer. What were the major conclusions of the NCI workshop?
DR. BERLIN: The NCI workshop found that the data from all eight international trials showed a reduction in mortality in the 50 to 69 age group of about one third. This benefit, however, was not seen in women under 50, which was largely a result of the studies having data for only 5 to 7 years of follow-up.
The HIP study was looked at separately, as it was the only study with data for 10 to 18 years of follow-up. In this time period, the HIP study showed a 25% reduction in mortality in women 40 to 49.
Given that other screening trials did not conclusively show a benefit in women under 50, and that interpretation of the HIP data was still being disputed, the NCI workshop concluded that women aged 40 to 49 showed no screening benefit 5 to 7 years after entry, and at 10 to 12 years after entry, the screening benefit is uncertain.
ONI: How did the NCI report impact the debate over whether to screen women in the 40 to 49 age bracket for breast cancer?
DR. BERLIN: The 1993 NCI report was followed by a recommendation from the Board of Scientific Counselors of the NCIs Division of Cancer Prevention and Control that the NCI no longer recommend screening in women under 50. This recommendation was then rejected by the National Cancer Advisory Board, whose advice was rejected, in turn, by the director of the National Cancer Institute. This was followed by congressional hearings that were critical of the NCI.
In response, the NCI convened the 1997 consensus development panel meeting, which issued two reports, a majority report that did not recommend screening in women under age 50, and a minority report that did recommend screening for this population.
The 1997 consensus meeting was followed by Senatorial hearings, at the conclusion of which the NCI revised its screening position and issued a final recommendation that women under age 50 should be screened for breast cancer.
ONI: What are the implications of the controversy over breast cancer screening for other screening programs and the development of health policy?
DR. BERLIN: The major implications concern which criteria should be used in evaluating a screening trial. Public health policy advocates appear to recognize as valid only randomized trials that have mortality as an endpoint. Yet there are other important endpoints, one of which would be stage at diagnosis, with all that staging implies as a predictor of outcome.
ONI: How would you sum up the breast cancer screening debate?
DR. BERLIN: The underlying issues that have driven the breast cancer screening debate for years are, first of all, does screening result in a reduction in mortality in women who entered the study under age 50, and, if so, how can the statistics showing this be validated?
Second, if there is a benefit that is statistically valid, how large of a benefit is necessary to justify adoption of a screening policy in light of other public needs? For example, there are those who would argue that money being spent on breast cancer screening would be better spent on other public health needs, such as pediatric immunizations for all children, which may show a greater benefit. Given that public health resources are often finite, it is extremely difficult to define the threshold at which one public health need outweighs another.