Machine learning can reduce a radiologists workload by lowering the number of screening mammograms they’re required to read while preserving accuracy, according to results of a feasibility study published in the Journal of the American College of Radiology.