A Closer Look at Automated LLM Protocoling for Abdominal and Pelvic CT

As an abdominal radiologist, Rajesh Bhayana, M.D., said image protocoling for hundreds of cases may take up to an hour a day and noted general estimates that protocoling comprises about five to six percent of a radiologist’s time at work.

Accordingly, Dr. Bhayana and colleagues recently compared prompting-only and fine-tuned GPT-4o model selection of computed tomography (CT) protocols for abdominal and pelvic CT exams to that of radiologists.

For the retrospective study, the researchers found that the prompting-only GPT-4o model selected optimal CT protocols in 96.2 percent of patients in comparison to 88.3 percent for unassisted radiologists.

In a recent interview with Diagnostic Imaging, Dr. Bhayana acknowledged the “monster prompt” instructions for the prompting-only GPT-4o model that covered seven pages and over 40-plus protocols for abdominal and pelvic CTs. That said, Dr. Bhayana called large language models (LLMs) a “natural fit for automated protocoling” that can be adapted as necessary with institutional protocol changes.

“ … Evidence emerges and as we decide, for instance, to give less oral contrast, LLMs, given their flexibility, give you the opportunity — especially given that (the) prompting only (LLM model) performs so well — to adjust or add new protocols relatively easily when they come about,” noted Dr. Bhayana, an assistant professor of radiology and radiologist technology lead at the Joint Department of Imaging at the University of Toronto.

Noting that he is currently using LLM-automated CT protocoling in practice, Dr. Bhayana emphasized that it can bolster consistency and efficiency in supervised settings.

“ … The LLM will preselect the protocol, urgency, etc., and the radiologist can either agree and verify or adjust. So obviously that makes us potentially more efficient but also helps us align across the department as to how we select protocols in various clinical situations, which has also been helpful so far,” added Dr. Bhayana.

(Editor’s note: For related content, see “Could LLMs Enhance Protocoling Efficiency for Abdominal and Pelvic CTs?,” “Current and Emerging Concepts with LLMs in Radiology: An Interview with Rajesh Bhayana, MD” and “Clinical Applications of LLMs in Radiology: Key Takeaways from RSNA 2025.”)

For more insights from Dr. Bhayana, watch the video below.