Interval breast cancers in the screening mammography program of British Columbia: Analysis and classification

Abstract

OBJECTIVE. Interval cancers in an annual mammography screening program are defined as cancers detected within 12 months after a mammographic screening in which findings are considered normal. Our objective was to analyze interval cancers for histologic type, tumor size, staging, and histologic grade. A classification for interval cancers is presented. Detection of interval cancers is an integral part of quality control and is required for the establishment of sensitivity rates for screening mammography, a necessary part of implementing the recommendations for breast cancer screening. MATERIALS AND METHODS. A total of 47,583 mammographic screening examinations of 38,219 women were available for review for the period of July 1988 to March 1991. Linkage with the provincial population-based cancer registry provided identification of all interval cancers of the breast. Histologic analysis of breast cancers was provided by an internationally recognized breast cancer pathologist who interpreted the specimens twice. RESULTS. A total of 207 breast cancers were detected at mammographic screening (true- positive cancers) during the first 33 months of the program, and 37 interval cancers were identified during the 33 months of screening plus a 12-month interval. Blinded review by three radiologists experienced in screening mammography resulted in classification of 21 of 37 cases as true interval cancers and 16 cases as misses. No interval cancers were missed because of poor mammographic technique. The sensitivity of screening mammography was 85%, 63% for patients less than 50 years old and 89% for patients more than 50 years old. The specificity was 93% for all patients (both younger and older than 50 years). CONCLUSION. The availability of a population-based cancer registry in the province of British Columbia makes an accurate determination of interval cancers and sensitivity rates possible. These data are essential for quality analysis, program planning, and education of radiologists in screening mammography. There were more true interval cancers than misses, and histologic analysis demonstrated that the majority of interval cancers were in stages I and II.