Patient-specific CT scans are being pulled from the hospital's PACS and bones of interest are being segmented by deep-learning (DL) algorithms. These are virtually loaded by physiological loads and their response is simulated by finite element analysis (FEA). Another DL algorithm analyses the relevant FEA results, presenting to the orthopedic surgeon the risk of fracture and assists in patient-specific fixation and implant selection.We validate our simulations by in-vitro experiments on fresh frozen human bones, and by clinical retrospective and prospective studies. Our algorithms were implemented and are being used at Tel Aviv Sourasky MC to monitor the femoral risk of fracture in patients with tumors in their femurs. We also completed a clinical study that demonstrates the high sensitivity and specificity of our algorithms for predicting osteopenic/osteoporotic risk of femoral fractures in the elderly population.