Screening and Breast Cancer Incidence

Abstract

During the past decade, the reported incidence of breast cancer in the United States, as well as in other westernized countries, has been gradually increasing. In addition, a recent sharp increase in incidence in older women (>60 years) has occurred. However, before one may attribute an increase in reported incidence to a change in risk, other causes of an apparent increase in incidence, such as more accurate registration, earlier diagnosis, and overdiagnosis of breast cancer cases, must be excluded. Relevant to this consideration is the increasing availability of screening mammography. It is a well-known fact that screening asymptomatic women by periodic mammographic examinations detects breast cancer at an earlier stage than when the disease is symptomatic. Screendetected cancers are more frequently non-invasive; and, if invasive, are of smaller size and less commonly metastasized to axillary lymph nodes than those which are diagnosed without screening. Also, there is evidence of a shift towards cancers of a more differentiated and specialized type (e.g., papillary, cribriform) and a redistribution of estrogen receptor concentrations towards higher concentrations (1-3). Examination of the distribution of tumor stage-at-presentation is, therefore, one method of monitoring the efficiency of a screening program. In two recent reports in this Journal (one in today's issue) tumor stage-atpresentation has been used to determine the contribution of mammographic screening to the increased incidence rates (4,5). Both studies confirm that a substantial increase in the detected incidence rate for breast cancer has occurred in older women living on the west coast of the US during the 1980s. However, whereas Glass and Hoover reported that this increase comprised both those with cancer confined to the breast and those with regional spread of disease (e.g., to axillary lymph nodes), White and her colleagues, in their much larger study, found that for women between the ages of 45 and 65 years, the increase was limited to those with localized cancer. They suggest that earlier diagnosis may be responsible for the apparent increase in this age group. In the Glass and Hoover study, only 9% of the 178 invasive cancers diagnosed during 1985 were first detected by screening mammography. Similar information is not available in today's report, but the proportion of normal women living in the locality who used mammography was assessed by a telephone survey. This survey indicated that of those aged 45-74 years, 30% of the respondents had undergone a screening mammogram during 1987 and 15% a first mammogram. The predicted yield of new cases, calculated from the ratio of the incidence rate in screenees compared with that in controls and the proportion of women screened for the first time in a given year, was then compared with the observed increase in incidence rate. This comparison led to the conclusion that all of the increased incidence observed in women aged 45-64 years could be accounted for by the increased use of screening mammography. But the increased use of mammography alone did not explain the increased incidence rates for younger or for older women. Although this method for determining use of mammography was not free from selection bias, these findings support the value of screening mammography in detecting cases at an earlier stage. There is now unequivocal evidence that mammography screening reduces breast cancer mortality (6-5). But screening can be effective in reducing total mortality from breast cancer within a nation only if it is applied to a greater number of women than a self-selected 30% of the population who are likely to be healthaware, educated, motivated, and economically stable. Mammographic screening on a population basis can be achieved only when an organized effort is made to reach out to all women in a target group, and when it is accepted that their need overrides that of preserving traditional, and at times vested, professional interests. The fact that mammography screening can be offered universally has been demonstrated in several European countries, including the United Kingdom, which have freely available health care. Sadly, despite the remarkably far-seeing initiative taken by Shapiro and Strax in launching the first-ever trial of population screening for breast cancer in New York in 1963, a similar service is not presently available in this country (6).