Radiology Report
Radiology reports, crucial for medical diagnosis and treatment, are increasingly being analyzed and generated using artificial intelligence. Current research focuses on improving the accuracy and efficiency of automated report generation and extraction of structured data from unstructured reports, employing large language models (LLMs) like LLAMA and GPT, often augmented with retrieval-augmented generation (RAG) and other techniques like contrastive learning. This work aims to reduce radiologist workload, improve diagnostic accuracy, and facilitate data analysis for research and clinical decision support, with a strong emphasis on ensuring data privacy and clinical validity. The development of new evaluation metrics specifically tailored for radiology reports is also a key area of ongoing investigation.
Papers
a2z-1 for Multi-Disease Detection in Abdomen-Pelvis CT: External Validation and Performance Analysis Across 21 Conditions
Pranav Rajpurkar, Julian N. Acosta, Siddhant Dogra, Jaehwan Jeong, Deepanshu Jindal, Michael Moritz, Samir Rajpurkar
ReXTrust: A Model for Fine-Grained Hallucination Detection in AI-Generated Radiology Reports
Romain Hardy, Sung Eun Kim, Pranav Rajpurkar
The Impact of AI Assistance on Radiology Reporting: A Pilot Study Using Simulated AI Draft Reports
Julián N. Acosta, Siddhant Dogra, Subathra Adithan, Kay Wu, Michael Moritz, Stephen Kwak, Pranav Rajpurkar
LLM-RG4: Flexible and Factual Radiology Report Generation across Diverse Input Contexts
Zhuhao Wang, Yihua Sun, Zihan Li, Xuan Yang, Fang Chen, Hongen Liao
Overview of the First Shared Task on Clinical Text Generation: RRG24 and "Discharge Me!"
Justin Xu, Zhihong Chen, Andrew Johnston, Louis Blankemeier, Maya Varma, Jason Hom, William J. Collins, Ankit Modi, Robert Lloyd, Benjamin Hopkins, Curtis Langlotz, Jean-Benoit Delbrouck
Enhancing disease detection in radiology reports through fine-tuning lightweight LLM on weak labels
Yishu Wei, Xindi Wang, Hanley Ong, Yiliang Zhou, Adam Flanders, George Shih, Yifan Peng