Lead Time Gained by Diagnostic Screening for Breast Cancer23

Preliminary findings of a clinical trial of screening for breast cancer have been used to estimate descriptive parameters of preclinical disease. The average duration of preclinical disease, measured from first recognition on mammography or physical examination to diagnosis as clinical cancer in absence of special screening, is estimated to be 20 months. Since cases detected at screening are not all detected at the earliest possible moment, the average lead time gained at a single screening will often be less than the average duration of preclinical disease. If the population were completely homogeneous with respect to the natural course of preclinical disease, the lead time would be exactly half the duration of preclinical disease, or 10 months. In a heterogeneous population it would exceed 10, increasing in proportion to the variance of the natural duration of preclinical disease. Incidence of clinical breast cancer in the unscreened control group of this study is 1.6 per 1,000 per year as compared with corresponding incidences of 1.2–1.4 reported by others. The incidence of preclinical disease estimated from the early results of the present study is 2.2. This comparatively high rate may be due to selection bias or to nonspecificity or insensitivity of the screening procedure. The preliminary data suggest that selection bias is the chief distorting factor.