Mammographic patterns and breast cancer risk: methodologic standards and contradictory results

The claim that classification of the pattern of the breast parenchyma upon mammography can distinguish groups at different risks for breast cancer has been examined by 17 other studies published in the English language literature with contradictory results; this controversy was assessed by us in examination of these studies for their adherence to commonly accepted methodologic standards for the investigation of causal relationships. The nine standards for the examination of the studies included a description of the way the study population had been assembled and followed up and descriptions of the methods of mammographic pattern classification and breast cancer risk analysis. A strong association was found between the standards adopted by a study and the obtained results. Among nine cohort and case-control studies that found a statistically significant association between mammographic pattern and breast cancer risk, all met at least four standards, whereas only two of six "negative" cohort or case-control studies met as many as four standards. Among prevalence surveys, the association between methods and results was less striking, but several negative prevalence surveys were associated with "positive" cohort studies employing the same group of patients. These results indicate that methodologic differences among studies contribute substantially to the controversy surrounding this subject. Studies that follow the usual scientific methods employed in the epidemiologic investigation of risk generally have confirmed an association between mammographic pattern and breast cancer risk.