This study investigates the feasibility of using ChatGPT and GPT-4 to translate radiology reports into plain language for patients and healthcare providers. Radiology reports from 62 low-dose chest CT lung cancer screening scans and 76 brain MRI metastases screening scans were used. ChatGPT successfully translated reports with an average score of 4.27 (on a 5-point scale), with minimal missing or misinformation. GPT-4 significantly improved translation quality. The study demonstrates the feasibility of using large language models in clinical education, but also highlights limitations such as randomness in responses and occasional oversimplification, which can be mitigated with detailed prompts.

Qing Lyu, Josh Tan, Michael E Zapadka, Janardhana Ponnatapura, Christopher T Whitlow, Chuang Niu, Ge Wang, Kyle J Myers