Abstract
There are important trade-offs in the practice of interpreting mammography. Radiologists do not want to miss identifying breast cancer, yet performing additional imaging to rule out cancer increases false-positive rates. False-positive mammograms generate anxiety, excess costs and, at times, morbidity from subsequent biopsies. The false-positive rate for screening mammography is higher in the United States than in European countries. Reducing the false-positive rates and maintaining high levels of sensitivity, as suggested by Esserman et al. (1) in this issue of the Journal, is appealing. They hypothesize that interpreting a high volume of mammograms, as is the norm for radiologists in the U.K., results in higher sensitivity than interpreting a low volume, as is often the norm for some U.S. radiologists. In their study (1), a standardized test set of 60 screening films from asymptomatic women, 13 of which included nonoccult breast cancers, was interpreted by 60 U.S. radiologists and also by 194 U.K. radiologists. The U.S. radiologists were assigned to one of three categories on the basis of their self-reported monthly volume of mammography interpretations: low (≤100/month), moderate (101–300/month), or high volume (≥301/month). Their results suggest that both sensitivity and specificity are better among higher volume readers, although the differences for specificity are not statistically significant (1).