Ha R, et al. - In the present study, researchers tested the premise that convolutional neural network (CNN) can predict Oncotype Dx recurrence score (RS) using an MRI dataset. Participants in the study were 134 patients with estrogen receptor positive/human epidermal growth factor receptor-2 negative (ER+/HER2–) invasive ductal carcinoma who had both breast MRI and Oncotype Dx RS evaluation. Findings from the present study suggested that it was feasible for current deep CNN architecture to be trained to forecast Oncotype DX RS. CNN had an overall accuracy of 84% in two-class prediction with specificity at 81%, and sensitivity at 87%.