[EMERGING THERAPIES FOR LI-FRAUMENI SYNDROME AND RHABDOMYOSARCOMA: CAN CHAT GPT ASSESS THE THERAPIES IT SUGGESTS USING THE BRADFORD-HILL CRITERIA?]

Abstract

Introduction: Li-Fraumeni syndrome is a rare genetic disorder caused by a mutation in the TP53 gene. The syndrome is characterized by a variety of malignancies that manifest in early childhood. In this article, we present a young patient diagnosed with Li-Fraumeni syndrome following recurrent rhabdomyosarcoma. A personalized literature search was performed to identify novel therapeutics that may be appropriate for this patient. In the absence of randomized, double-blind studies, the relevance of each potential treatment was evaluated using the Bradford-Hill criteria. These criteria underlie the medical logic used for assessing causation. This case was then used to explore the ability of the large language model ChatGPT to conduct a similar review and assess its treatment suggestions using the Bradford-Hill criteria. In-depth exploration suggested that ChatGPT treatment suggestions based solely on the presentation of the case and a general question are insufficiently informative. Targeted prompts (instructions for structuring the response, formulated as questions) lead to the generation of more accurate and clinically useful model responses but these, too, should be approached with caution.