OIDM-Based Next-gen Reporting Assistance Framework

One of the key ideas of the OIDM framework is to enable the creation of tools that work within the systems that make up the imaging informatics ecosystem — the reporting tool, PACs, viewer, the worklist manager, and data exploration tools. What would this look like in the context of the reporting tool?

Reporting Assistance Framework

A proposed approach would be that developers would you be able to author plugins  in a standard programming language, like JavaScript.

The reporting tool would then set up a container that would "host" the plugins.

Execution Cycle

The reporting container would then run each script. The providing access to a standard data structure—defined  by OIDM!—that would contain all the relevant contextual information—about the findings in the current report, about prior reports, about the exams that were being reported on. Each script in the container could then examine that data context and decide whether it had any useful assistance to offer in the reporting process. The assistance would take the form of standard commands generated by the scripts back to the reporting environment, which could then insert generated text into the existing report, ask the radiologist for additional information, alert the radiologist to a particular problem, or send data to an external system. As the report context changes (more AI data comes in, the radiologist updates the report manually), the reporting tool would re-run each script with the updated data context so it could take appropriate actions.

The result is a suite of "assistant" scripts that are continually reviewing the relevant data associated with the radiologist creating the report and issuing commands back to the reporting system to provide assistance to the radiologist.