LLMs like ChatGPT-4o and AmbossGPT can accurately classify bone fractures in CT radiology reports, aiding radiologists.
Key Details
- 1Study assessed four LLMs (ChatGPT-4o, AmbossGPT, Claude 3.5 Sonnet, Gemini 2.0 Flash) on 292 artificial CT reports representing 310 fractures.
- 2ChatGPT-4o and AmbossGPT showed highest overall classification accuracy (74.6% and 74.3%).
- 3Bone recognition rates were high for all models (90%-99%), but fracture subtype classification was lower (71%-77%).
- 4Statistically significant accuracy differences were noted between LLMs by fracture type and anatomical location.
- 5Validation with real-world reports (145 fractures) using LLaMA 3.3-70B yielded similar results to artificial datasets (~70% performance).
- 6Authors note need for further validation on large, multi-center real-world datasets.
Why It Matters
Radiology practices rely heavily on textual reporting, and automating fracture classification could streamline radiological workflows, reduce variability, and improve efficiency. While current LLMs show promise, further validation is necessary before widespread adoption.
Source
AuntMinnie
Related News
•Radiology Business
AMA Survey: Physicians Double AI Use, Wary About Patient-Directed Radiology AI
The share of U.S. doctors using AI in practice has doubled, yet many are hesitant about patients using AI for interpreting radiology results.
•AuntMinnie
Deep Learning Model Enhances MRI Detection of Brain Metastases
A deep learning model significantly improves speed and accuracy of brain metastasis detection on MRI scans.
•AuntMinnie
Google AI Studio's Performance in Lung Cancer Detection on CT Evaluated
Google AI Studio demonstrates moderate accuracy in identifying lung malignancy on CT, but requires further refinement before clinical use.