AI-Powered Neurogenetics: Supporting Patient's Evaluation with Chatbot

Area of Science:

• Neurogenetics
• Artificial Intelligence in Medicine
• Clinical Decision Support

Background:

• Artificial intelligence (AI) and large language models (LLMs) offer potential benefits for healthcare professionals.
• This study investigates the utility of ChatGPT and Google's Gemini in the initial assessment of patients with suspected neurogenetic disorders.

Purpose of the Study:

• To evaluate the performance of ChatGPT and Gemini in assisting clinicians with neurogenetic disorder assessments.
• To determine if these AI tools can serve as valuable adjuncts in identifying clinical features, suggesting differential diagnoses, and guiding genetic testing strategies.

Main Methods:

• Ninety questions were posed to ChatGPT (Versions 4o, 4, and 3.5) and Gemini.
• Questions covered clinical diagnosis, genetic inheritance, recurrence risks, genetic testing, and patient management.
• The assessment focused on six rare neurogenetic disorders: Hereditary Spastic Paraplegia (types 4 and 7), Huntington Disease, Fragile X-associated Tremor/Ataxia Syndrome, Becker Muscular Dystrophy, and FacioScapuloHumeral Muscular Dystrophy.

Main Results:

• GPT chatbots demonstrated superior performance compared to Gemini.
• All AI chatbots exhibited significant limitations in diagnostic accuracy.
• A concerning frequency of hallucinations was observed across all tested AI models.

Conclusions:

• AI tools can potentially enhance clinician capabilities in evaluating neurogenetic disorders.
• Effective utilization of these AI tools necessitates close collaboration and supervision by neurologists and geneticists.
• Meticulous oversight is crucial to mitigate diagnostic inaccuracies and AI-generated hallucinations.