The Diagnostic Imagination in Radiology: Part 2.

Developing algorithms for the improve- ment of diagnostic care leverages tech- nologies and techniques developed across industries that are exponentially being improved, developed, and tested. Machine learning means extracting patterns not only from patient level obser- vations or a radiologist's primary diag- nosis, but from secondary diagnoses, incidental findings, claims data and similarities with other patients for predictive benefit. The business model for radiology will be based on deeply knowing and leveraging existing data and generating data on patients that can be reused and made easily accessible for-future algo- rithms and changes in healthcare policy and reimbursement.