CA: A Cancer Journal for Clinicians, Vol 27, 66-87, Copyright
© 1977 by American Cancer Society
Screening for Breast Cancer in Younger Women: Life Expectancy Gains and Losses
An Analysis According to Risk Indicator Groups
Herbert Seidman M.B.A.1
1 Chief, Statistical Analyses, Department of Epidemiology and Statistics, Department of Research, American Cancer Society, New York, New York.
Using data from the ACS Prospective Study and excluding women with a history of breast cancer who are know to be at very high risk of a second breast cancer, breast cancer incidence rates are estimated for age-groups 35-39 through 55-59 years for various risk categories defined by the women's own statements. The women with none of the risk indicators, the "none" group, comprised about 54 percent of the total women and showed breast cancer rates, which though the lowest of any risk indicator group, were still quite appreciable, being 80 percent of the rates for total women of their age group.
Women with menarche prior to age 12 or who had not given birth to a live child before age 30 showed an excess of 15 percent to 50 percent over the breast cancer incidence rates of the "none" group. These risk indicators are termed "minor"(relatively). The "minors only" group comprised about 38 percent of total women and had breast cancer rates much the same as those of the total women of their age.
Current breast lump or discharge, or a history of breast surgery or of breast cancer in mother or sister are termed "major" risk indicators. One such indicator was noted for about eight percent of total women and generally denoted breast cancer incidence rates somewhat over 100 percent in excess of the rates of the "none" group. Two or more "major" risk indicators, observed in less than one percent of the women, meant an excess of 200 percent to over 400 percent compared with the incidence rates of the "none" group.
Using these data in conjunction with information from the initial screening series of the BCDDP women ages 35-59, and similarly excluding women with a history of breast cancer at initial screening, it is found that the BCDDP women are overweighted with high-risk women, having a breast cancer risk about one-quarter higher than the general population.
The rate of breast cancer per 1,000 women detected in the initial screening series ranged from 1.0 for women 35-39 years to 6.5 for those 55-59 years. The detection rates at the second annual screenings were about 25-75 percent of those at initial screenings.
It is estimated that the mean "lead time" was about 1.4 years for the breast cancers detected at initial screening and over two years for those detected at the second annual screening.
Using survival probabilities for the general population from U.S. life tables, and data on the survival of various categories of breast cancer patients, estimates are made of the life expectancies following cancer detection for patients in various categories. As compared with normal expectancy, the usual clinical invasive breast cancer patient is subject to a loss of 21 years of life at ages 35-39 years and about half that at ages 55-59 years.
Combining the estimates of life expectancy with the number of breast cancers detected in the various categories, data are developed showing the gain in life expectancy through screening. This is computed as the difference between years of life expectancy for the cancers detected by screening, and the years of life expectancy for the usual breast cancer case finding plus mean lead time. The mean gain per breast cancer detected at initial screening for those under 50 is about five to seven years, and for those in their 50's, about two to four years. These gains represent a reduction of roughly one-third in the loss of life expectancy of the usual clinical breast cancer patient.
Using the BEIR report estimates of six radiation-induced breast cancers per million women per year per rad of dosage to the breast after a 10-year latent period, calculations are made concerning the number of breast cancers presumably to be induced, and the loss in life expectancy thereby. The median cancer case would occur about 26 years after screening of the 35-39 year-old women, and 19 years after screening of those 55-59 years.
The estimated gains in life expectancy per woman screened through screening detection of breast cancers are seen to be modest. For the initial or second annual screening series, the highest gain among the groups studied is .08 years per screenee (78 years per 1,000 screenees). For a given screening, the upper limit of the presumable loss due to radiation-induced breast cancers is perhaps .006 years per screenee at ages 35-39 (six years per 1,000 screenees) and .001 years per screenee at ages 55-59 (one year per 1,000 screenees).
Except in 35-39 year-old women, the estimated detection benefit is many times the presumed radiation loss at initial screening. The expectation that the greater frequency of breast cancer in higher-risk women would yield an increased gain through screening is validated by these data.
The "major" risk indicators delineate a small proportion of women who are at substantially higher risk of developing breast cancer and who show more gain from screening than the large majority of women of their age. However, even without such risk factors, women have an appreciable risk of developing breast cancer. This risk is of sufficient magnitude to alert all physicians to the possible presence of breast cancer at the time of clinical examination.
