Translating Radiology Reports into Plain Language using ChatGPT and GPT-4 with Prompt Learning: Promising Results, Limitations, and Potential

Original Paper: https://arxiv.org/abs/2303.09038

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

Abstract:

The large language model called ChatGPT has drawn extensively attention because of its human-like expression and reasoning abilities. In this study, we investigate the feasibility of using ChatGPT in experiments on using ChatGPT to translate radiology reports into plain language for patients and healthcare providers so that they are educated for improved healthcare. Radiology reports from 62 low-dose chest CT lung cancer screening scans and 76 brain MRI metastases screening scans were collected in the first half of February for this study. According to the evaluation by radiologists, ChatGPT can successfully translate radiology reports into plain language with an average score of 4.27 in the five-point system with 0.08 places of information missing and 0.07 places of misinformation. In terms of the suggestions provided by ChatGPT, they are general relevant such as keeping following-up with doctors and closely monitoring any symptoms, and for about 37% of 138 cases in total ChatGPT offers specific suggestions based on findings in the report. ChatGPT also presents some randomness in its responses with occasionally over-simplified or neglected information, which can be mitigated using a more detailed prompt. Furthermore, ChatGPT results are compared with a newly released large model GPT-4, showing that GPT-4 can significantly improve the quality of translated reports. Our results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.

Summary Notes

Simplifying Radiology Reports with AI: The Role of ChatGPT and GPT-4

In the world of artificial intelligence (AI), tools like ChatGPT and GPT-4 are transforming various industries, healthcare included.

Their latest feat? Turning complex radiology reports into plain language. This endeavor aims to make medical information more accessible to patients and healthcare providers.

The Problem: Dense Medical Language

Radiology reports are essential for patient care but often come in a language filled with medical terms that are hard for non-specialists and patients to understand.

This complexity can confuse patients, causing anxiety and possibly leading to poor healthcare decisions.

AI to the Rescue: Simplifying Medical Text

Thanks to advancements in AI and natural language processing (NLP), ChatGPT and GPT-4 offer a groundbreaking solution.

These AI models can translate the complicated texts of radiology reports into simpler, understandable language, improving communication and patient care.

How It Works: Technical to Simple Language

• Process: Collect and anonymize radiology reports, like chest CT scans and brain MRIs, from clinical databases.
• AI Translation: Use ChatGPT or GPT-4 to convert these reports into plain language.
• Evaluation: Experienced radiologists assess these translations for accuracy, completeness, and clarity.

Results: Promising Yet Imperfect

The AI translations have delivered impressive results:

• Accuracy: High levels of accuracy were observed, reducing misinformation.
• Clarity and Brevity: The reports became shorter and clearer, making medical terms easier to understand.
• Consistency: There was some variability, with occasional oversimplifications or missing details.

Discussion: Benefits and Improvement Areas

Using AI to translate medical documents into patient-friendly language shows great promise. It aids in patient education and allows for personalized recommendations. However, the inconsistency in translations highlights the need for better prompts and possibly more specialized model training.

Conclusion: The Future of AI in Healthcare

The potential of AI models like ChatGPT and GPT-4 to translate radiology reports into plain language is clear.

The focus now shifts to improving the reliability of these translations and integrating AI tools into clinical settings for real-time use. As AI evolves, it promises to make healthcare information more accessible and understandable, thus enhancing patient care.

Final Thoughts

The use of AI in making medical documents easier to understand marks a significant advancement in patient care. Simplifying radiology reports can lead to better-informed patients, fostering meaningful conversations with healthcare providers and empowering individuals in their health decisions.

The ongoing development of AI technology holds the key to a future where healthcare is more transparent, accessible, and centered around the patient's needs.