ChatGPT in the diagnosis and management of complex polyneuropathies: comparative analysis with neurologists using real-world cases - PubMed
3 hours ago
- #polyneuropathy diagnosis
- #ChatGPT-4o
- #comparative analysis
- The study evaluated ChatGPT-4o's diagnostic performance on 100 real-world polyneuropathy cases compared to specialists and non-specialist neurologists.
- ChatGPT-4o achieved 65.5% leading diagnosis accuracy, similar to non-specialists (63.0%) but lower than specialists (74.0%).
- For differential diagnoses, ChatGPT-4o outperformed non-specialists (82.0% vs. 77.5%) and recommended more appropriate confirmatory tests (68.0% vs. 53.0%).
- Non-specialists revised their assessments in 21.8% of cases after viewing ChatGPT-4o's outputs, leading to improved diagnostic accuracy.
- The findings suggest ChatGPT-4o has potential as a diagnostic aid, especially in non-specialist or resource-limited clinical settings.