EMIRE (EEG Monitoring In Rare Epilepsies) – Results of the first Italian experience in detecting electrophysiological patterns through electroencephalographic monitoring with subcutaneous electrodes

Objective: Accurate long-term seizure monitoring remains a major challenge in epilepsy care, particularly in patients with developmental and epileptic encephalopathies (DEEs) and drug-resistant epilepsy (DRE), in whom seizure diaries are often unreliable and conventional electroencephalographic (EEG) methods are limited by short recording duration. Subcutaneous EEG (sqEEG) systems offer the possibility of continuous, objective monitoring over months in real-world settings. Aims: This thesis aimed to evaluate the clinical utility of ultra-long-term sqEEG monitoring through the EMIRE (EEG Monitoring in Rare Epilepsies) study, the first Italian experience using subcutaneous electrodes to identify electrophysiological patterns and quantify seizure burden in patients with DEEs and DRE. Other objectives included assessing the performance of automated seizure detection, comparing it with patient-reported seizure diaries, evaluate the tolerability of the device and describing electroclinical patterns recorded by the device. Methods: EMIRE is a prospective, multicenter, interventional study involving four Italian tertiary epilepsy centers. Participants aged ≥12 years with uncontrolled epilepsy underwent implantation of the UNEEG 24/7 EEG SubQ device. At the time of analysis, 11 patients with DEEs and/or DRE had been enrolled at the Modena center, and these are the patients described in this thesis. Following a validation phase with simultaneous scalp video-EEG and sqEEG recording, patients were monitored continuously at home for a period ranging of a minimum of three months to a maximum of six months. EEG data were analyzed using the EpiSight platform, which combines artificial intelligence–based seizure detection with expert neurophysiological review. Clinical, electroencephalographic, and monitoring data from enrolled patients were examined to assess seizure detection accuracy, feasibility, and clinical applicability. Results: The sqEEG system successfully recorded a wide range of epileptic events and interictal abnormalities, allowing the identification of patient-specific electrographic patterns. Recordings demonstrated a high degree of concordance with scalp EEG and enabled prolonged monitoring in naturalistic conditions. The average daily use of the electrode in total was 76.18% (18.3h), ranging from 37% (8.8h) to 98% (23.5h). Overall sensitivity of the automated EpiSight detection system was 35.9% and positive predictive value was 1.58%. Detection performance varied according to seizure type, reaching sensitivity of 91.3% for generalized tonic-clonic seizures and 44.4% for focal seizures, while lower sensitivity was observed in DEEs (29.2%). Sensitivity of patients’ seizure diaries resulted 28.9%, ranging from 6.45% to 100%. The percentage of false positive for patients’ diaries was 52.49%. Long-term monitoring also revealed circadian seizure patterns and provided clinically relevant information regarding seizure frequency, treatment response, and disease evolution. Conclusions: Ultra-long-term sqEEG monitoring represents a feasible and well-tolerated approach for continuous seizure assessment in patients with complex epilepsies. Compared with traditional self-reporting methods, it provides a more objective characterization of seizure burden and electrophysiological activity, supporting clinical decision-making and therapeutic management. Although further improvements in automated detection algorithms are required, subcutaneous EEG monitoring shows considerable potential to become an important tool for precision epilepsy care and future seizure forecasting strategies.