Come visit us at Booth 3700 to see the latest initiatives in data science and informatics from the ACR. Below are just a few of the major initiatives we have been working on.
November 27 - December 1, 2022 -- McCormick Place - Chicago, Illinois
Distributed Research Registries
- With ACR Connect, sites participating in research registries can now choose if their data moves centrally or if it stays local, allowing more sites to participate in research registries as they match their individual needs. All the while, to a researcher, this expanded network allows queries and analysis to be performed across the network.
Local Data Prep
- Using ACR Connect, sites are able to prepare, annotate, validate, and harmonize their data locally before using it to participate in a research registry, AI project, or quality registry. ACR Connect allows sites more individual control and ownership of their data in projects without having to send data offsite or using expensive data harmonizing solutions.
Evaluating Operational AI Models
- AI-LAB has been expanded to support AI models beyond research AI, allowing radiologists and sites to test operational and commercial AI models against their data in a safe and secure way. With this enhanced functionality, members can better understand how solutions will work at their organizations.
- DART, in combination with ACR Connect, allows researchers to produce standardized annotations across the distributed network. This enables researchers to generate high quality, normalized annotations in datasets without the overhead of bringing all dispersed data into a centrally managed repository.
- ACR is collaborating with CAP (College of American Pathologists) to produce detailed concordance scenarios between radiology and pathology. Building upon the IHE profile for rad-path concordance, ACR and CAP are producing data element mappings between pathology and radiology for health care scenarios, such as PI-RADS and BI-RADS. These mappings allow sites and vendors to produce highly effective concordance workflows.