[Skip to Content]
[Skip to Content Landing]
November 1, 1985

Sensitivity and Specificity in Silicone Breast Models-Reply

Author Affiliations

University of North Carolina at Chapel Hill

JAMA. 1985;254(17):2409. doi:10.1001/jama.1985.03360170049014

In Reply.—  We agree with Dr Gump that manufactured breast models do not simulate human breast tissue in all ways. Nevertheless, the models we used in our study, developed by Pennypacker and associates,1,2 did not have a uniform matrix. They were constructed of potting gel and silicon elastomer to simulate the amount of background nodularity found in breast tissue of an average 50-year-old woman. This background material probably accounted for the false-positive detection rate of 0.75 detections per physician in our study. Also, our study did not demonstrate that examination with fingertips (as opposed to pads) was more effective. There were no statistically significant differences between the approaches. We did find that recommended techniques, including using finger pads, were associated with fewer false-positive detections.It is likely, as Dr Gump implies, that the use of silicone breast models may be more useful in helping physicians improve the sensitivity than