P4.022: Novel Smartphone-Based Screening Tool for Absence Seizures in Pediatric Settings: A Comparative Study with video EEG

Background: Accurate differentiation between ADHD (affecting 7 million US children) and Childhood Absence Epilepsy (CAE) is crucial, as overlapping symptoms often lead to diagnostic delays. While AAP guidelines recommend ruling out conditions like CAE during ADHD evaluations and management, subtle absence seizures are frequently missed by history alone. The gold standard for diagnosing CAE is evaluation by a neurologist and confirmatory EEG, but this is often impractical because of expense and long wait times. Hyperventilation (HV) is a reliable method to provoke absence seizures for diagnosis and monitoring, eliciting seizures in 90% of affected patients. However, HV is underutilized by pediatricians due to limited training and perceived risks. To address this, Eysz has developed a smartphone-based HV Recorder, standardizing HV and providing visual seizure evidence.

Methods: 46 participants used the Eysz HV Recorder during concurrent EEG observation. Three neurologists reviewed the EEG recording and marked seizure presence. A second set of three epileptologists reviewed the video from the Eysz HV Recorder (without access to the EEG data) for seizure presence. We evaluated the performance of the Eysz HV app reviewers in identifying seizures, using the EEG reviews as the ground truth. Our study has captured 25 seizures from 12 participants and recorded 37 participants who did not have seizures. Seizure presence was determined by majority agreement of the 3 EEG reviewers.

Results: Sensitivity of the HV recording video reviewers was 96.3% (two reviewers had 100% sensitivity) for detecting patients that had seizures ≥ 7 s. Only 3 participants had seizures that were all < 7 s, and each only had one seizure. Mean sensitivity for these participants was 44.4%. Mean specificity across Eysz HV app reviewers was 90.1%. We also compared how often individual EEG reviewers agreed with each other to how often the HV video reviewers agreed with the EEG reviewers, and found similar rates of agreement: EEG <> EEG agreement was 94.4% (SD: 9.6%), while HV app <> EEG agreement was 83.3% (SD: 7.2%).

Conclusion: These results show that neurologists are highly accurate at identifying seizures ≥ 7 s long in videos recorded during guided HV using the Eysz HV Recorder. Our findings support the use of the Eysz HV Recorder as a screening tool for CAE which could facilitate earlier diagnosis and treatment, which will lead to improvements in quality of life. Additionally, this screening tool can facilitate the ruling out of CAE, which can mimic conditions like ADHD, can co-occur with ADHD, and can also contribute to other common symptoms seen by pediatricians such as impaired academic performance, daydreaming, zoning out, and memory issues.