The significance of concordance in mammographic interpretations.

Twenty-eight women participating in a screening program had a breast cancer diagnosed subsequent to the report of a negative mammogram which was read by only one of a group of radiologists. Fourteen occurred prior to a scheduled routine screening visit (interval cancers) and 14 were detected during such an examination. The negative mammograms from the 28 cancer patients, together with those from 120 women without cancer (controls) were independently reviewed by each member of a panel of three radiologists. Forty-six percent of the cancer cases and 5.8% of the controls were interpreted as positive by two or more of the radiologists. These findings suggest that agreement among several independent reviewers enchances the value and accuracy of mamography by reducing the number of false negative interpretations.