The Clinical Value of ChatGPT for Epilepsy Presurgical Decision Making: Systematic Evaluation on Seizure Semiology Interpretation.

Background: For patients with drug-resistant focal epilepsy (DRE), surgical resection of the epileptogenic zone (EZ) is an effective treatment to control seizures. Accurate localization of the EZ is crucial and is typically achieved through comprehensive presurgical approaches such as seizure semiology interpretation, electroencephalography (EEG), magnetic resonance imaging (MRI), and intracranial EEG (iEEG). However, interpreting seizure semiology poses challenges because it relies heavily on expert knowledge and is often based on inconsistent and incoherent descriptions, leading to variability and potential limitations in presurgical evaluation. To overcome these challenges, advanced technologies like large language models (LLMs)-with ChatGPT being a notable example-offer valuable tools for analyzing complex textual information, making them well-suited to interpret detailed seizure semiology descriptions and assist in accurately localizing the EZ.

Objective: This study evaluates the clinical value of ChatGPT in interpreting seizure semiology to localize EZs in presurgical assessments for patients with focal epilepsy and compares its performance with epileptologists.

Methods: Two data cohorts were compiled: a publicly sourced cohort consisting of 852 semiology-EZ pairs from 193 peer-reviewed journal publications and a private cohort of 184 semiology-EZ pairs collected from Far Eastern Memorial Hospital (FEMH) in Taiwan. ChatGPT was evaluated to predict the most likely EZ locations using two prompt methods: zero-shot prompting (ZSP) and few-shot prompting (FSP). To compare ChatGPT's performance, eight epileptologists were recruited to participate in an online survey to interpret 100 randomly selected semiology records. The responses from ChatGPT and the epileptologists were compared using three metrics: regional sensitivity (RSens), weighted sensitivity (WSens), and net positive inference rate (NPIR).

Results: In the publicly sourced cohort, ChatGPT demonstrated high RSens reliability, achieving 80-90% for the frontal and temporal lobes, 20-40% for the parietal lobe, occipital lobe, and insular cortex, and only 3% for the cingulate cortex. The WSens, which accounts for biased data distribution, consistently exceeded 67%, while the mean NPIR remained around 0. These evaluation results based on the private FEMH cohort are consistent with those from the publicly sourced cohort. A group -test with 1000 bootstrap samples revealed that ChatGPT-4 significantly outperformed epileptologists in RSens for commonly represented EZs, such as the frontal and temporal lobes (p < 0.001). Additionally, ChatGPT-4 demonstrated superior overall performance in WSens (p < 0.001). However, no significant differences were observed between ChatGPT and the epileptologists in NPIR, highlighting comparable performance in this metric.

Conclusions: ChatGPT demonstrated clinical value as a tool to assist the decision-making in the epilepsy preoperative workup. With ongoing advancements in LLMs, it is anticipated that the reliability and accuracy of LLMs will continue to improve in the future.