Solution Accelerator
Automating Radiology Workflows With Large Language Models (LLMs)
Pre-built code, sample data and step-by-step instructions ready to go in a Databricks Notebook
Relieve the administrative burden of radiology workflows
Generative AI, particularly large language models (LLMs), is sparking a transformative shift in the healthcare industry. Researchers, clinicians and medical professionals are harnessing the power of LLMs to analyze and distill vast troves of medical data to extract actionable insights that can improve patient outcomes.
Within radiology, one opportunity to deploy LLMs is to automate manual processes, from assigning protocol labels to imaging orders, which can be time-consuming and prone to human error.
With this Solution Accelerator, radiology departments can alleviate manual administrative tasks associated with examination protocols and labeling by:
- Leveraging open source LLMs, like Meditron, trained on PubMed articles
- Fine-tuning with Parameter-Efficient Fine-Tuning (PEFT) for cost-effectiveness
- Automatically assigning radiology protocol labels to imaging, e.g., a CT scan