The specific aim of this study is to develop an add-on module for the ONCAD breast MRI CAD system that will provide radiologists with additional assistance when they have suspicion of the presence of infiltrating lobular carcinoma (ILC). The proposed research will investigate the possibility of developing, training and evaluating an ILC- specific algorithm that uses both kinetics and morphology to increase the sensitivity of ONCAD to ILC. This ILC-specific algorithm will enable the radiologist to further evaluate cases in which ILC is suspected and to improve sensitivity on this subset of breast cancers. Infiltrating lobular carcinoma (ILC), which is found in a clinically significant fraction of breast cancers, is difficult to detect clinically or with mammography and frequently fails to exhibit suspicious kinetics on MRI. In prior studies, ONCAD, the only FDA-approved CAD system for MRI that marks breast lesions that have suspicious morphologies, identified 7 of 8 biopsy-proven cases of ILC. These results provide preliminary evidence that the ONCAD algorithms may provide a basis of researching methods of identifying sufficient percentage of ILCs to be clinically significant. This ILC-specific algorithm will enable the radiologist to further evaluate cases in which ILC is suspected and to improve sensitivity on this subset of breast cancers. ? ? ?
