We provide health analytics to enhance radiologist productivity, reduce healthcare costs, and improve quality of care. Radiologists currently have difficulties getting value from patient data laying around in the hospital. Some of it is structured in databases but a lot of it is coming in from unstructured sources: physician notes, image annotations, and complex patient images. Hospitals want to find ways to deliver this patient data in a more actionable and accountable form throughout the radiology workflow. So we have developed a new way to improve accuracy and reduce cost around radiological image interpretation using machine learning and computer vision technologies. We provide a patient-centric view of the data that is more accountable and more personalized to the radiologist. Our personalized medical knowledge graph uses both patient data in the hospital as well as the user's semantic understanding of patient conditions to provide the most relevant information to the radiologist.