About 50% of breast cancers detected on mammography are visible retrospectively on prior studies; many of these cancers are obscured by dense breast tissue, are subtle on mammography, or are missed through human error, according to a research team led by Dr. Alyssa Watanabe of the University of Southern California (USC). In addition, a high percentage of mammography results and biopsies are false positives.
As a result, there's a need for methods and techniques that can improve the sensitivity and specificity of mammography interpretation. In their study, the researchers investigated the impact of the cmAssist (CureMetrix) AI-based CAD software on the sensitivity of radiologists in breast cancer screening and detection. Both before and after viewing the AI-based CAD results, seven radiologists were asked to read a set of mammograms that included 90 previously missed breast cancers and had been initially interpreted with conventional CAD software.
All radiologists achieved a significant improvement in their cancer detection rate from using the AI-based CAD application, according to the researchers.
"The results of this study show that AI can be used for earlier detection of breast cancer on mammography and can raise the cancer detection rate of an inexperienced general radiologist to the level of a breast fellowship-trained full-time mammographer," Watanabe told AuntMinnie.com. "This is the first study, to my knowledge, that shows that AI can provide a significant clinical benefit for radiologists."
Stop by this Wednesday morning talk to get all the details.
