Background: We aimed to investigate the efficacy of computer-aided detection (CAD) for MRI in the assessment\nof tumor extent, lymph node status, and multifocality in invasive breast cancers in comparison with other breast\nimaging modalities.\nMethods: Two radiologists measured the maximum tumor size, as well as, analyzed lymph node status and\nmultifocality in 86 patients with invasive breast cancers using mammography, ultrasound, CT, MRI with and\nwithout CAD, and 18-fludeoxyglucose positron emission tomography (FDG-PET). The assessed data were compared\nwith pathology.\nResults: For tumor extent, there were no significant differences between pathological size and measured size using\nmammography, ultrasound, CT, or MRI with and without CAD (P > 0.05). For evaluation of lymph node status,\nultrasound had the best kappa coefficients (0.522) for agreement between imaging and pathology, and diagnostic\nperformance with 92.1% specificity and 90.0% positive predictive value. For multifocality, MRI with CAD had the\nhighest area under the receiver operating characteristic curve (AUC = 0.888).\nConclusions: CAD for MRI is feasible to assess tumor extent and multifocality in invasive breast cancer patients.\nHowever, CAD is not effective in evaluation of nodal status.
Loading....