Search papers, labs, and topics across Lattice.
EpiScreen, a novel approach for early epilepsy detection, leverages fine-tuned large language models on routinely collected clinical notes from electronic health records. The model achieves high performance, with AUC scores of 0.875 on MIMIC-IV and 0.980 on a private dataset. Furthermore, EpiScreen-assisted neurologists outperformed unaided experts by up to 10.9%, demonstrating the potential for improved diagnostic accuracy and efficiency.
LLMs can sift through routine clinical notes to detect epilepsy with high accuracy, even boosting expert neurologists' diagnostic performance by over 10%.
Epilepsy and psychogenic non-epileptic seizures often present with similar seizure-like manifestations but require fundamentally different management strategies. Misdiagnosis is common and can lead to prolonged diagnostic delays, unnecessary treatments, and substantial patient morbidity. Although prolonged video-electroencephalography is the diagnostic gold standard, its high cost and limited accessibility hinder timely diagnosis. Here, we developed a low-cost, effective approach, EpiScreen, for early epilepsy detection by utilizing routinely collected clinical notes from electronic health records. Through fine-tuning large language models on labeled notes, EpiScreen achieved an AUC of up to 0.875 on the MIMIC-IV dataset and 0.980 on a private cohort of the University of Minnesota. In a clinician-AI collaboration setting, EpiScreen-assisted neurologists outperformed unaided experts by up to 10.9%. Overall, this study demonstrates that EpiScreen supports early epilepsy detection, facilitating timely and cost-effective screening that may reduce diagnostic delays and avoid unnecessary interventions, particularly in resource-limited regions.
