In the proposed study, our objective is to improve accuracy of breast cancer screening by applying textural analysis and machine learning to contrast-enhanced digital mammography (CEDM). Standard digital mammography often results in false positives that require unnecessary additional testing that may be costly and physically uncomfortable to patients. Studies have demonstrated that CEDM, a relatively new modality, achieves increased sensitivity in comparison to standard mammography as it enables a radiologist to combine the visual signatures of malignancies observed on mammography with quantitative signal intensity information that is available in contrast-enhanced studies. Additional quantitative information (e.g., texture analysis) derived from CEDM may offer additional sensitivity gains. Although there are no existing studies that have examined the performance of textural analysis when applied to CEDM, our preliminary findings are promising. In the proposed study, we will retrospectively analyze a population of 120 patients with the goal of demonstrating that textural analysis coupled with machine learning methodologies will further improve the specificity of CEDM. With the framework created with the seed grant, we hope to develop a more comprehensive externally-funded study that uses a larger sample size to develop integrated software and decision-making tools that enable improved breast cancer detection.