A deep learning algorithm that can accurately detect breast cancer on screening mammograms using an "end-to-end" training approach that leverages training datasets with either complete clinical annotation or only the cancer status of the whole image, eliminating the reliance on rarely available lesion annotations.