Monday, December 1 | 2:00 p.m.-2:10 p.m. | M6-SSMK03-4 | Room E450A
When it comes to knee MRI exams, partial AI integration without workflow embedding came up short for reducing reporting time, according to researchers leading this Monday afternoon scientific session.
For this presentation, radiologist Benoît Rizk, MD, will examine the impact of AI-assisted structured reporting on knee MRI scans, as part of a pilot involving 10 radiology practices in Switzerland. The prospective study was conducted from September 2023 to October 2024 and compared three reporting workflows.
Using 1,285 consecutive knee MRI scans, Rizk and colleagues began by measuring baseline standard reporting without AI, then measuring reporting time after partial diagnostic AI integration, and finally, measuring the time to report using fully integrated AI-generated findings and autopopulated structured report templates.
Particularly for nonspecialist readers, fully integrated AI-assisted reporting reduced overall knee MRI reporting time by 13.4%, compared with standard reporting, with general radiologists observing the largest reduction, the group found. Musculoskeletal specialists also benefited, but to a smaller extent, they said.
"Partial AI integration did not yield time savings and, in some cases, slightly increased reporting time, likely due to workflow modifications," the group stated prior to RSNA. "Effective integration into existing reporting systems is crucial to realizing these benefits in routine clinical practice."
Bring your questions and join the session to discuss AI-assisted structured reporting as a workflow enhancement for knee MRI exams, including suggestions for further optimization.
