Breast cancer screening is associated with a high recall rate (whereby further images or biopsy after an abnormal mammogram are required). In the Province of Quebec, 22.8% of women having a 1st screening mammogram and 9.7% of women with subsequent mammograms will receive a referral for further work-up and these numbers are increasing over time. In nearly 75% of the instances, this further work-up is negative (no cancer is detected), which represents close to 10% of the total Quebec screening program budget. The main objective of our study is to use Imagia’s artificial intelligence (AI) EVIDENS platform to classify mammogram abnormalities more accurately into benign or malignant lesions without the need for a biopsy (or close follow-up) when benign lesions are identified. We will also explore whether AI is able to more accurately predict the likelihood of response to preoperative chemotherapy in aggressive cancers and to predict the risk of recurrence in smaller cancer. To achieve our goal, we will use the breast biopsy database available at the Centre des maladies du sein du CHU de Québec – Université Laval, that collects over 3,000 cases per year, in addition to normal and abnormal mammograms (over 200,000 mammograms). Since the number of images is critical to the success of AI, the robustness of the database positions us wellfor this project. The potential socio-economical impact of this project is significant. Results of this project will be of great value for all screening programs. For the industrial partner, Imagia, the potential for commercialization of the software is important. For the healthcare system, this project has the potential to positively impact resource utilization; if AI can differentiate benign from malignant mammographic lesions, this will save health care resources by avoiding unnecessary workup for benign lesions. Most importantly for women, it will decrease the costs associated from personal and professional absences for work-up and minimize anxiety associated with false-positive mammograms.