Objective:The EEG (electroencephalogram) plays a crucial role in prognosticating outcomes for comatose patients after cardiac arrest (CA), but the precise probability of poor outcome (PPO) associated with specific EEG patterns and how it changes over time is not well understood. We aimed to quantify the PPO for individual EEG patterns, assess its precision, and evaluate time-dependent changes after CA. Methods:We retrospectively analyzed continuous EEGs from comatose adults treated at three Boston hospitals for CA (2010-2023). We evaluated the prognostic significance of 22 ACNS-standard patterns, and one additional pattern, "flat EEG (<2μV)", using PPO (i.e., positive predictive value for poor outcome) with 95 % Wilson confidence intervals as the primary metric; sensitivity, and false-positive rate (FPR) were secondary. Temporal variabilities were assessed in four epochs (0-24, 24-48, 48-72, >72 h) using linear regression. Poor and favorable outcomes were defined as death and survival to discharge, respectively. Results:Among 1,000 patients, the mean age was 58 years for survivors and 60 years for non-survivors; 59% of patients were male. Flat EEG demonstrated PPO of 1.00 across epochs, but CIs ranged from 0.89-1.00 at 24 hours and widened to 0.54-1.00 beyond 72 hours. Status epilepticus (SE) showed PPO of 0.80 (0.44-0.97) at 24 hours, 1.00 (0.74-1.00) at 48 hours, and 0.67 (0.09-0.99) after 72 hours. Suppression (< 10 μV), the most common historically termed highly "malignant" pattern (prevalence 50-65%), had a PPO of 0.76-0.83 and sensitivity of up to 0.80, with 17-24% of patients experiencing a favorable outcome. Time-dependent analysis further demonstrated that PPO for suppression increased over time, while sensitivity declined for burst-suppression and flat EEG. Conclusion:The prognostic uncertainty of EEG patterns post-CA is highly variable, and evolves within the first 72 hours. Flat EEG, and SE are highly specific prognosticators of poor outcome, but, importantly, they are rare and not infallible. Suppression identifies the most patients with poor outcomes but has moderately high FPRs and variable prognostic certainty over time. Future large-scale prospective studies that incorporate long-term outcomes are needed to validate these findings.

Super-refractory status epilepticus (SRSE) is a neurological emergency defined as status epilepticus persisting for more than 24 h despite conventional management with anesthetics or recurrence with withdrawal of anesthesia. It is associated with high morbidity and mortality, and contemporary literature on treatment is limited. We present a systematic review of surgical interventions and outcomes for SRSE. We performed a multidatabase literature search according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, using an International Prospective Register of Systematic Reviews (PROSPERO)-registered protocol, to identify published literature through February 2024. Four reviewers independently screened citations, abstracts, and manuscripts of SRSE with surgical or neuromodulatory interventions in pediatric and adult populations, with a senior reviewer resolving discrepancies. The primary outcomes were the resolution of SRSE and Engel I classification at the last reported follow-up. We screened 1436 citations, reviewed 66 manuscripts, and identified 114 patients who underwent acute neurosurgical intervention for SRSE. Of the 114 cases, 111 had resolution of SRSE after the intervention, and 57 of 114 patients were reported to be Engel class I (free of disabling seizures) at the last follow-up. Among the remaining half (57), 8 patients did not survive, 3 had recurrence of status epilepticus, and 46 continued to have medically refractory seizures, with 29 experiencing fewer seizures than before. Our subgroup analyses also highlighted that surgical interventions were effective even in patients with underlying autoimmune and genetic etiologies. An additional striking finding was that the patients who had surgical interventions delayed beyond 3 weeks from the SRSE onset were less likely to achieve resolution of SRSE and had poorer longitudinal seizure freedom (40.7

Introduction:More than 50 million people worldwide suffer from epilepsy. Approximately 30% of epileptic patients suffer from medically refractory epilepsy (MRE), which means that over 15 million people must seek extensive treatment. One such treatment involves surgical removal of the epileptogenic zone (EZ) of the brain. However, because there is no clinically validated biomarker of the EZ, surgical success rates vary between 30%-70%. The current standard for EZ localization often requires invasive monitoring of patients for several weeks in the hospital during which intracranial EEG (iEEG) data is captured. This process is time-consuming as the clinical team must wait for seizures and visually interpret the iEEG during these events. Hence, an iEEG biomarker that does not rely on seizure observations is desirable to improve EZ localization and surgical success rates. Recently, the source-sink index (SSI) was proposed as an interictal (between seizure) biomarker of the EZ, which captures regional interactions in the brain and in particular identifies the EZ as regions being inhibited ("sinks") by neighbors ("sources") when patients are not seizing. The SSI only requires 5-min snapshots of interictal iEEG recordings. However, one limitation of the SSI is that it is computed heuristically from the parameters of dynamical network models (DNMs). Methods:In this work, we propose a formal method for detecting sink regions from DNMs, which has a strong foundation in linear systems theory. In particular, the steady-state solution of the DNM highlights the sinks and is characterized by the leading eigenvector of the state-transition matrix of the DNM. To test this, we build patient-specific DNMs from interictal iEEG data collected from 65 patients treated across 6 centers. From each DNM, we compute the average leading eigenvectors and evaluate their potential as a biomarker to accurately predict EZ and surgical success. Results:Our findings show the ability of the leading eigenvector to accurately predict EZ (average accuracy 66.81% ± 0.19%) and surgical success (average accuracy 71.9% ± 0.22%) with data from 65 patients across 6 centers from 5 min of data, which we show is comparable with the current method of localizing the EZ over several weeks. Discussion:This eigenvector biomarker has the potential to assist clinicians in localizing the EZ quickly and thus increase surgical success in patients with MRE, resulting in an improvement in patient care and quality of life.

OBJECTIVE:To determine if implementing the IFCN criteria to define interictal epileptiform discharges (IEDs) improves expert inter-rater reliability (IRR) and diagnostic performance. METHODS:Nine EEG experts rated the same 200 candidate IEDs (100 expert-consensus, 100 epilepsy monitoring unit [EMU]-validated) as epileptiform or not, in random order, in two rounds separated by at least 30 days. During the second round, raters additionally selected the applicable IFCN criteria for each candidate IED. RESULTS:Overall, there were no major differences in performance (AUC; 0.90 vs. 0.91) or IRR (AC1; 0.48 vs. 0.47) between both Parts; nor was there a major difference in calibration within the expert-consensus dataset (median absolute calibration index; 35.5 vs. 30.0). Similarly, there were no major differences in performance or IRR within either dataset. IRR was substantial within the EMU-validated dataset and only fair within the expert-consensus dataset. IRR was fair for criteria 2, 3, 5 and 6, and moderate for criteria 1 and 4. CONCLUSIONS:Our findings suggest that the IFCN criteria to define IEDs may not significantly improve IRR, performance, or overall calibration among experts. SIGNIFICANCE:Increasing expert IRR for each criterion may enhance the utility of the IFCN criteria in clinical practice.

OBJECTIVE:Whereas a scalp electroencephalogram (EEG) is important for diagnosing epilepsy, a single routine EEG is limited in its diagnostic value. Only a small percentage of routine EEGs show interictal epileptiform discharges (IEDs) and overall misdiagnosis rates of epilepsy are 20% to 30%. We aim to demonstrate how network properties in EEG recordings can be used to improve the speed and accuracy differentiating epilepsy from mimics, such as functional seizures - even in the absence of IEDs. METHODS:In this multicenter study, we analyzed routine scalp EEGs from 218 patients with suspected epilepsy and normal initial EEGs. The patients' diagnoses were later confirmed based on an epilepsy monitoring unit (EMU) admission. About 46% ultimately being diagnosed with epilepsy and 54% with non-epileptic conditions. A logistic regression model was trained using spectral and network-derived EEG features to differentiate between epilepsy and non-epilepsy. Of the 218 patients, 90% were used for training and 10% were held out for testing. Within the training set, 10-fold cross validation was performed. The resulting tool was named "EpiScalp." RESULTS:EpiScalp achieved an area under the curve (AUC) of 0.940, an accuracy of 0.904, a sensitivity of 0.835, and a specificity of 0.963 in classifying patients as having epilepsy or not. INTERPRETATION:EpiScalp provides an accurate diagnostic aid from a single initial EEG recording, even in more challenging epilepsy cases with normal initial EEGs. This may represent a paradigm shift in epilepsy diagnosis by deriving an objective measure of epilepsy likelihood from previously uninformative EEGs. ANN NEUROL 2025;97:907-918.

Despite recent medical therapeutic advances, approximately one third of patients do not attain seizure freedom with medications. This drug-resistant epilepsy population suffers from heightened morbidity and mortality. In appropriate patients, resective epilepsy surgery is far superior to continued medical therapy. Despite this efficacy, there remain drawbacks to traditional epilepsy surgery, such as the morbidity of open neurosurgical procedures as well as neuropsychological adverse effects. SEEG-guided Radiofrequency Thermocoagulation (SgRFTC) is a minimally invasive, electrophysiology-guided intervention with both diagnostic and therapeutic implications for drug-resistant epilepsy that offers a convenient adjunct or alternative to ablative and resective approaches. We review the international experience with this procedure, including methodologies, diagnostic benefit, therapeutic benefit, and safety considerations. We propose a framework in which SgRFTC may be incorporated into intracranial EEG evaluations alongside passive recording. Lastly, we discuss the potential role of SgRFTC in both delineating and reorganizing epilepsy networks.

Introduction: For patients with drug-resistant epilepsy, successful localization and surgical treatment of the epileptogenic zone (EZ) can bring seizure freedom. However, surgical success rates vary widely because there are currently no clinically validated biomarkers of the EZ. Highly epileptogenic regions often display increased levels of cortical excitability, which can be probed using single-pulse electrical stimulation (SPES), where brief pulses of electrical current are delivered to brain tissue. It has been shown that high-amplitude responses to SPES can localize EZ regions, indicating a decreased threshold of excitability. However, performing extensive SPES in the epilepsy monitoring unit (EMU) is time-consuming. Thus, we built patient-specific in silico dynamical network models from interictal intracranial EEG (iEEG) to test whether virtual stimulation could reveal information about the underlying network to identify highly excitable brain regions similar to physical stimulation of the brain.Methods: We performed virtual stimulation in 69 patients that were evaluated at five centers and assessed for clinical outcome 1 year post surgery. We further investigated differences in observed SPES iEEG responses of 14 patients stratified by surgical outcome.Results: Clinically-labeled EZ cortical regions exhibited higher excitability from virtual stimulation than non-EZ regions with most significant differences in successful patients and little difference in failure patients. These trends were also observed in responses to extensive SPES performed in the EMU. Finally, when excitability was used to predict whether a channel is in the EZ or not, the classifier achieved an accuracy of 91%.Discussion: This study demonstrates how excitability determined via virtual stimulation can capture valuable information about the EZ from interictal intracranial EEG.

Importance:Guidelines recommend seizure prophylaxis for early posttraumatic seizures (PTS) after severe traumatic brain injury (TBI). Use of antiseizure medications for early seizure prophylaxis after mild or moderate TBI remains controversial.Objective:To determine the association between seizure prophylaxis and risk reduction for early PTS in mild and moderate TBI.Data Sources:PubMed, Google Scholar, and Web of Science (January 1, 1991, to April 18, 2023) were systematically searched.Study Selection:Observational studies of adult patients presenting to trauma centers in high-income countries with mild (Glasgow Coma Scale [GCS], 13-15) and moderate (GCS, 9-12) TBI comparing rates of early PTS among patients with seizure prophylaxis with those without seizure prophylaxis.Data Extraction and Synthesis:The Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) reporting guidelines were used. Two authors independently reviewed all titles and abstracts, and 3 authors reviewed final studies for inclusion. A meta-analysis was performed using a random-effects model with absolute risk reduction.Main Outcome Measures:The main outcome was absolute risk reduction of early PTS, defined as seizures within 7 days of initial injury, in patients with mild or moderate TBI receiving seizure prophylaxis in the first week after injury. A secondary analysis was performed in patients with only mild TBI.Results:A total of 64 full articles were reviewed after screening; 8 studies (including 5637 patients) were included for the mild and moderate TBI analysis, and 5 studies (including 3803 patients) were included for the mild TBI analysis. The absolute risk reduction of seizure prophylaxis for early PTS in mild to moderate TBI (GCS, 9-15) was 0.6% (95% CI, 0.1%-1.2%; P = .02). The absolute risk reduction for mild TBI alone was similar 0.6% (95% CI, 0.01%-1.2%; P = .04). The number needed to treat to prevent 1 seizure was 167 patients.Conclusion and Relevance:Seizure prophylaxis after mild and moderate TBI was associated with a small but statistically significant reduced risk of early posttraumatic seizures after mild and moderate TBI. The small absolute risk reduction and low prevalence of early seizures should be weighed against potential acute risks of antiseizure medications as well as the risk of inappropriate continuation beyond 7 days.

ImportanceGuidelines recommend seizure prophylaxis for early posttraumatic seizures (PTS) after severe traumatic brain injury (TBI). Use of antiseizure medications for early seizure prophylaxis after mild or moderate TBI remains controversial.ObjectiveTo determine the association between seizure prophylaxis and risk reduction for early PTS in mild and moderate TBI.Data SourcesPubMed, Google Scholar, and Web of Science (January 1, 1991, to April 18, 2023) were systematically searched.Study SelectionObservational studies of adult patients presenting to trauma centers in high-income countries with mild (Glasgow Coma Scale [GCS], 13-15) and moderate (GCS, 9-12) TBI comparing rates of early PTS among patients with seizure prophylaxis with those without seizure prophylaxis.Data Extraction and SynthesisThe Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) reporting guidelines were used. Two authors independently reviewed all titles and abstracts, and 3 authors reviewed final studies for inclusion. A meta-analysis was performed using a random-effects model with absolute risk reduction.Main Outcome MeasuresThe main outcome was absolute risk reduction of early PTS, defined as seizures within 7 days of initial injury, in patients with mild or moderate TBI receiving seizure prophylaxis in the first week after injury. A secondary analysis was performed in patients with only mild TBI.ResultsA total of 64 full articles were reviewed after screening; 8 studies (including 5637 patients) were included for the mild and moderate TBI analysis, and 5 studies (including 3803 patients) were included for the mild TBI analysis. The absolute risk reduction of seizure prophylaxis for early PTS in mild to moderate TBI (GCS, 9-15) was 0.6% (95% CI, 0.1%-1.2%; P = .02). The absolute risk reduction for mild TBI alone was similar 0.6% (95% CI, 0.01%-1.2%; P = .04). The number needed to treat to prevent 1 seizure was 167 patients.Conclusion and RelevanceSeizure prophylaxis after mild and moderate TBI was associated with a small but statistically significant reduced risk of early posttraumatic seizures after mild and moderate TBI. The small absolute risk reduction and low prevalence of early seizures should be weighed against potential acute risks of antiseizure medications as well as the risk of inappropriate continuation beyond 7 days.

To determine if implementing the operational definition of interictal epileptiform discharges (IEDs) proposed by the International Federation of Clinical Neurophysiology (IFCN) improves expert diagnostic performance and interrater reliability (IRR).

OBJECTIVE:In mesial temporal lobe epilepsy (MTLE), the ideal surgical approach to achieve seizure freedom and minimize morbidity is an unsolved question. Selective approaches to mesial temporal structures often result in suboptimal seizure outcomes. The authors report the results of a pilot study intended to evaluate the clinical feasibility of using an endoscopic anterior transmaxillary (eATM) approach for minimally invasive management of MTLEs.METHODS:The study is a prospectively collected case series of four consecutive patients who underwent the eATM approach for the treatment of MTLE and were followed for a minimum of 12 months. All participants underwent an epilepsy workup and surgical care at a tertiary referral comprehensive epilepsy center and had medically refractory epilepsy. The noninvasive evaluations and intracranial recordings of these patients confirmed the presence of anatomically restricted epileptogenic zones located in the mesial temporal structures. Data on seizure freedom at 1 year, neuropsychological outcomes, diffusion tractography, and adverse events were collected and analyzed.RESULTS:By applying the eATM technique and approaching the far anterior temporal lobe regions, mesial-basal resections of the temporal polar areas and mesial temporal structures were successfully achieved in all patients (2 with left-sided approaches, 2 with right-sided approaches). No neurological complications or neuropsychological declines were observed. All 4 patients achieved Engel class Ia outcome up to the end of the follow-up period (19, 15, 14, and 12 months). One patient developed hypoesthesia in the left V2 distribution but there were no other adverse events. The low degree of white matter injury from the eATM approach was analyzed using high-definition fiber tractography in 1 patient as a putative mechanism for preserving neuropsychological function.CONCLUSIONS:The described series demonstrates the feasibility and potential safety profile of a novel approach for medically refractory MTLE. The study affirms the feasibility of performing efficacious mesial temporal lobe resections through an eATM approach.

Objective: To investigate the outcome of seizure freedom among patients with DRE who underwent LITT for temporal lobe epilepsy (TLE) and extratemporal lobe epilepsy (ETLE) and identify factors contributing to seizure recurrence after LITT. Background: Surgical resections offer the highest chance for seizure freedom for patients with drug-resistant epilepsy (DRE); however, over the last few years, there has been increased interest in exploring more minimally invasive surgical techniques, such as laser interstitial thermal therapy (LITT). Design/Methods: This retrospective chart review included patients who underwent LITT from 2015–2021. We collected data on demographics, clinical history, presurgical testing, postoperative complications, and seizure freedom outcomes at various time points. We conducted t-tests and Fisher's exact tests to compare the percentages of seizure freedom for each factor of interest and a linear regression to determine whether these factors are strong predictors of seizure recurrence. Results: 19 out of 29 patients (65.5%) were seizure free one year after LITT and 13/29 (44%) were seizure free at the last visit (mean FU duration = 49.5 months). Seizure freedom was 50% (12/24) among the TLE group and 20% (1/5) among the ETLE group at the last visit (p=0.343). Five patients (17%) had acute post-operative complications, which were reversible. Two patients (7%) reported memory and comprehension impairments. Gender, family history of epilepsy, handedness, etiology, and epilepsy duration were not significantly associated with seizure recurrence. Six patients underwent a re-operation, with five anterior temporal lobectomies and one additional ablation. Three of the ATL patients were seizure free after the operation, but the additional ablation did not result in seizure freedom. Conclusions: LITT can potentially offer seizure freedom to carefully selected patients with DRE. Our study showed the TLE group had higher proportions of patients with seizure freedom in comparison to the ETLE group, but it was not statistically significant given the small sample size. Disclosure: Ms. In has nothing to disclose. Dr. El Refaey has nothing to disclose. The institution of Dr. Castellano has received research support from NIH. Mark Richardson has received personal compensation in the range of $500-$4,999 for serving as a Consultant for SetPoint Medical. Mark Richardson has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Neurocrine. Mark Richardson has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Biogen. Mark Richardson has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for AskBio. Mark Richardson has received personal compensation in the range of $5,000-$9,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Voyager Therapeutics. Dr. Bagic has nothing to disclose. Dr. Barot has nothing to disclose.

Objective: We aimed to validate a recently published technique that used neural resonance to localize seizure onset zones (SOZ) with a retrospective dataset collected at an external institution. Further, we critically examined the need for a larger, tailored prospective validation study. Background: Accurate localization of the SOZ could improve surgical outcomes in medically refractory epilepsy patients. A recent study successfully used a novel metric of neural resonance to retrospectively identify SOZ regions and predict surgical success. The technique leverages distinct electrophysiological features of evoked responses during single pulse electrode stimulation (SPES) and incorporates them into a dynamical model. We aimed to validate these findings with a retrospective cohort of six patients to assess whether neural resonance predicted surgical outcomes. Design/Methods: We collected intracranial electroencephalographic (iEEG) data from six patients that underwent intracranial monitoring, single-pulse electrical stimulation protocol, and resection surgery between June 2020 and May 2021. Four out of six patients were stimulated in both SOZ regions and non-SOZ regions and thus were tested using the logistic regression model trained on the original dataset at the primary study institution. Results: Three of the four patients had successful surgical outcomes. Of the four patients tested with the logistic regression model, only one out of four was predicted correctly; the model predicted success for one true success and the failure patient. Conclusions: We hypothesize that the non-reproducibility of the original findings are resultant of differences in a research-driven SPES protocol and a clinically driven stimulation protocol. The external validation dataset stimulated a fewer number of sites, used a different clinical EEG and stimulation system, and varied stimulation parameters in a different subspace. However, this work highlights the need for a large, prospective validation dataset that is multi-center and encompasses a wide range of parameters; the gathering of this dataset is currently underway. Disclosure: Mr. Barnagian has nothing to disclose. Dr. Smith has nothing to disclose. Dr. Kang has nothing to disclose. Mark Hays has nothing to disclose. Dr. Gonzalez has nothing to disclose. Dr. Sarma has stock in Neurologic Solutions, Inc.. Dr. Barot has nothing to disclose.

OBJECTIVE:Posttraumatic epilepsy (PTE) develops in as many as one third of severe traumatic brain injury (TBI) patients, often years after injury. Analysis of early electroencephalographic (EEG) features, by both standardized visual interpretation (viEEG) and quantitative EEG (qEEG) analysis, may aid early identification of patients at high risk for PTE. METHODS:We performed a case-control study using a prospective database of severe TBI patients treated at a single center from 2011 to 2018. We identified patients who survived 2 years postinjury and matched patients with PTE to those without using age and admission Glasgow Coma Scale score. A neuropsychologist recorded outcomes at 1 year using the Expanded Glasgow Outcomes Scale (GOSE). All patients underwent continuous EEG for 3-5 days. A board-certified epileptologist, blinded to outcomes, described viEEG features using standardized descriptions. We extracted 14 qEEG features from an early 5-min epoch, described them using qualitative statistics, then developed two multivariable models to predict long-term risk of PTE (random forest and logistic regression). RESULTS:We identified 27 patients with and 35 without PTE. GOSE scores were similar at 1 year (p = .93). The median time to onset of PTE was 7.2 months posttrauma (interquartile range = 2.2-22.2 months). None of the viEEG features was different between the groups. On qEEG, the PTE cohort had higher spectral power in the delta frequencies, more power variance in the delta and theta frequencies, and higher peak envelope (all p < .01). Using random forest, combining qEEG and clinical features produced an area under the curve of .76. Using logistic regression, increases in the delta:theta power ratio (odds ratio [OR] = 1.3, p < .01) and peak envelope (OR = 1.1, p < .01) predicted risk for PTE. SIGNIFICANCE:In a cohort of severe TBI patients, acute phase EEG features may predict PTE. Predictive models, as applied to this study, may help identify patients at high risk for PTE, assist early clinical management, and guide patient selection for clinical trials.

INTRODUCTION: Up to one-third of severe traumatic brain injury (TBI) patients develop post-traumatic epilepsy (PTE), often years after their injury. Currently, there is no validated predictive model to identify patients at high risk for PTE. METHODS: We retrospectively analyzed a prospective database of severe TBI patients treated at a single level one trauma center from 2012 through 2018. Outcomes were recorded with the Expanded Glasgow Outcomes Scale (GOSE) at 6- and 12-months post-injury. We identified a cohort of patients who survived to two years and did not have an early post-traumatic seizure, defined as a seizure within 7 days of initial injury. From this cohort, we matched patients with PTE to those without using age and Glasgow coma scale score at admission. All patients underwent continuous EEG for 3-5 days upon admission, and a board-certified epileptologist identified 5-minute artifact-free segment for analysis. We extracted qEEG features, described them with qualitative statistics, and developed a logistic regression model using backwards selection. RESULTS: We identified 27 patients with PTE and 23 without who survived two years post-injury. The median time to onset of PTE was 7.2 months post-trauma and GOSE was similar when stratified by PTE at 6- and 12-months (p > 0.73). On qEEG analysis, patients with PTE had higher spectral power in the delta frequencies (p = 0.003), more variance in the delta and theta frequencies (p < 0.04), a higher mean amplitude (p = 0.01), and peak envelope (p = 0.01). On multivariate modeling, delta power (Odds Ratio [OR] 1.8), theta power (OR 0.2), and rhythmic spectrum (1.7) all increased risk of PTE and had an AUC of 0.85 for predicting PTE. CONCLUSIONS: In a cohort of severe TBI patients, quantitative qEEG features in the acute phase after injury identified future risk of PTE.

Background and ObjectiveNearly one-third of patients with severe traumatic brain injury (TBI) develop posttraumatic epilepsy (PTE). The relationship between PTE and long-term outcomes is unknown. We tested whether, after controlling for injury severity and age, PTE is associated with worse functional outcomes after severe TBI.MethodsWe performed a retrospective analysis of a prospective database of patients with severe TBI treated from 2002 through 2018 at a single level 1 trauma center. Glasgow Outcome Scale (GOS) was collected at 3, 6, 12, and 24 months postinjury. We used repeated-measures logistic regression predicting GOS, dichotomized as favorable (GOS 4-5) and unfavorable (GOS 1-3), and a separate logistic model predicting mortality at 2 years. We used predictors as defined by the International Mission for Prognosis and Analysis of Clinical Trials in TBI (IMPACT) base model (i.e., age, pupil reactivity, and GCS motor score), PTE status, and time.ResultsOf 392 patients who survived to discharge, 98 (25%) developed PTE. The proportion of patients with favorable outcomes at 3 months did not differ between those with and without PTE (23% [95% Confidence Interval [CI]: 15%-34%] vs 32% [95% CI: 27%-39%]; p = 0.11) but was significantly lower at 6 (33% [95% CI: 23%-44%] vs 46%; [95% CI: 39%-52%] p = 0.03), 12 (41% [95% CI: 30%-52%] vs 54% [95% CI: 47%-61%]; p = 0.03), and 24 months (40% [95% CI: 47%-61%] vs 55% [95% CI: 47%-63%]; p = 0.04). This was driven by higher rates of GOS 2 (vegetative) and 3 (severe disability) outcomes in the PTE group. By 2 years, the incidence of GOS 2 or 3 was double in the PTE group (46% [95% CI: 34%-59%]) compared with that in the non-PTE group (21% [95% CI: 16%-28%]; p < 0.001), while mortality was similar (14% [95% CI: 7%-25%] vs 23% [95% CI: 17%-30%]; p = 0.28). In multivariate analysis, patients with PTE had lower odds of favorable outcome (odds radio [OR] 0.1; 95% CI: 0.1-0.4; p < 0.001), but not mortality (OR 0.9; 95% CI: 0.1-1.9; p = 0.46).DiscussionPosttraumatic epilepsy is associated with impaired recovery from severe TBI and poor functional outcomes. Early screening and treatment of PTE may improve patient outcomes.

Introduction: Guidelines recommend use of computerized tomography (CT) and electroencephalography (EEG) in post-arrest prognostication. Strong associations between CT and EEG might obviate the need to acquire both modalities. We quantified these associations via deep learning. Methods: We performed a single-center, retrospective study including comatose patients hospitalized after cardiac arrest. We extracted brain CT DICOMs, resized and registered each to a standard anatomical atlas, performed skull stripping and windowed images to optimize contrast of the gray-white junction. We classified initial EEG as generalized suppression, other highly pathological findings or benign activity. We extracted clinical information available on presentation from our prospective registry. We trained three machine learning (ML) models to predict EEG from clinical covariates. We used three state-of-the-art approaches to build multi-headed deep learning models using similar model architectures. Finally, we combined the best performing clinical and imaging models. We evaluated discrimination in test sets. Results: We included 500 patients, of whom 218 (44%) had benign EEG findings, 135 (27%) showed generalized suppression and 147 (29%) had other highly pathological findings that were most commonly (93%) burst suppression with identical bursts. Clinical ML models had moderate discrimination (test set AUCs 0.73-0.80). Image-based deep learning performed worse (test set AUCs 0.51-0.69), particularly discriminating benign from highly pathological findings. Adding image-based deep learning to clinical models improved prediction of generalized suppression due to accurate detection of severe cerebral edema. Discussion: CT and EEG provide complementary information about post-arrest brain injury. Our results do not support selective acquisition of only one of these modalities, except in the most severely injured patients.