[Computer-assisted diagnosis in mammography: the R2 ImageChecker System in detection of spiculated lesions]

Because of the rapid development of computer systems for digitalization and image analysis, they play an increasing important role in computer-assisted diagnosis (CAD). Especially in the field of mammography, the early signs of malignancy are relatively uniform and therefore more easily detected by a computer algorithm. In this study, we tested one of the few commercially available systems for the detection of both microcalcifications and suspicious, spiculated solid lesions on 40 cases of proven breast carcinomas. These mammograms were analyzed by three independent observers with and without knowledge of the computer results, respectively. Depending on the time of their radiologic experience, the sensitivity of the observers alone was 92.4%, 86.1% and 82%. With knowledge of the computer interpretation, sensitivity of all three observers rose significantly to 100%, 92.7%, and 95%, respectively. However, due to a high number of false positive results of the computer algorithm (0.4 markers per image), the positive predictive value of the interpretations worsened from 100%, 92.7%, and 95.5% to 86.4%, 97.3%, and 91.1%, respectively. It can be expected that future developments will soon overcome this problem and CAD will become an effective tool in screening mammography.