Highlights of the evidence of benefit for women aged 40-49 years from the 14-year follow-up of the Breast Cancer Detection Demonstration Project.

Randomized breast cancer screening trials are helpful in establishing evidence of benefit when they yield statistically significant results. The results of individual randomized screening trials can vary greatly depending on the quality of the images, the frequency of screening, compliance, contamination, sample size, and the length and adequacy of follow-up. For women 40-49 years of age at entry the first randomized breast cancer screening trial, the Health Insurance Program of Greater New York (HIP), showed a statistically significant decrease of 24% in breast cancer mortality at 18 years of follow-up, virtually the same as the 23% decrease seen in women 50-59 years of age. At 10-12 years of follow-up, five of seven randomized trials have shown nonstatistical decreases in breast cancer mortality. No trials with fewer than 8 years of follow-up have shown a decrease in mortality for women 40-49 years of age; whereas seven trials have shown a decrease for women older than 50 years of age. The largest study on the screening of women aged 40-49 (93,471) was the Breast Cancer Detection Demonstration Program (BCDDP). This demonstration program was not a randomized trial but has yielded indirect supportive evidence for the screening of women aged 40-49, by comparing age group results. Three highlights of these studies are presented: (1) Mammography in the BCDDP (1970s) was improved greatly compared to that in the HIP study (1960s). Mammography detected 40% of the cancers in women 40-49 years of age in the HIP compared to 90% in the BCDDP. In women 50-59 years of age, mammography detected 60% of cancers in the HIP but 92% in the BCDDP. (2) Using annual two-view mammography plus clinical breast examination after the first screen of women 45 years of age and older, subsequent detection rates were virtually the same for all age groups. The number of women screened annually yields slightly more than two breast cancers per 1000 women, regardless of age group. (3) Similar types, sizes, stage distributions, and survival and case fatality rates were demonstrated in women aged 40-49, 50-59 and 60-69 years.