neuroGPT-X: Towards an Accountable Expert Opinion Tool for Vestibular Schwannoma

We hypothesize that a well-trained, context-enriched GPT will perform at the level of or better than an expert surgeon in generating comprehensive answers to questions surrounding commonly posed in day-to-day practice regarding vestibular schwannoma. In this study, we make three key contributions to assessing the feasibility of LLMs as a clinical decision-making adjunct. 1. We develop a framework to context-enrich GPT with context relevant to vestibular schwannoma. 2. We compare the performance of ChatGPT (Jan. 30, 2023 model) and a context-enriched GPT model against leading neurosurgical experts worldwide, evaluating the ability of large language models (LLMs) to assist in clinical decision-making. 3. We introduce a proof-of-concept clinical decision-making tool, neuroGPT-X, which incorporates working memory, sources with each answer, and a web-based chat platform to address challenges in using LLMs in a clinical setting, including interpretability, reliability, accountability, and safety.