Radiologists outperformed a convolutional neural network (CNN) and radiomic analysis (RA) at classifying contrast-enhancing lesions on multiparametric breast MRI, according to a Nov. 13 study published in Radiology. With more training, however, CNNs may soon close that gap.