Received: 3 March 2021 | Accepted: 10 August 2021 DOI: 10.1111/ene.15065 ORIGINAL ARTICLE Magnetic resonance imaging findings in focal-­onset status epilepticus Silvana Sarria-­Estrada1,2,3 | Estevo Santamarina3,4 | Manuel Quintana3,4 | 1,2 5 1 4 Deborah Pareto | Maria Sueiras | Cristina Auger | Manuel Toledo | Alex Rovira1,2 Abstract 1 Neuroradiology Section, Radiology Department, Vall d'Hebron Hospital, Barcelona, Spain Background and purpose: Magnetic resonance imaging (MRI) is commonly used in the 2 Vall d'Hebron Research Institute, diagnostic work-­up for status epilepticus (SE). The purpose of this study was to character- Neuroradiology Research, Barcelona, ize MRI features in SE patients and determine their association with clinical and electro- Spain 3 Department of Medicine, Autonomous encephalography (EEG) findings. The mid-­term consequences of baseline MRI features University of Barcelona (UAB), Barcelona, were also analysed. Spain 4 Methods: This is a prospective study including consecutive patients with SE who un- Epilepsy Unit, Neurology Department, Vall d'Hebron Hospital, Barcelona, Spain derwent brain MRI within 240 h after SE onset. The MRI protocol included T1-­weighted 5 EEG Unit, Neurophysiology Department, (T1WI), T2-­weighted (T2W), fluid-­attenuated inversion recovery (FLAIR) and diffusion-­ Vall d'Hebron Hospital, Barcelona, Spain weighted imaging (DWI) sequences. Follow-­up MRI was performed after SE resolution in Correspondence some patients. Silvana Sarria-­Estrada, Neuroradiology Results: Sixty patients (56.7% men, mean age 58.3 years) were included. SE-­related MRI Section, Radiology Department, Passeig Vall d'Hebron 119–­121, 08035 Barcelona, abnormalities were seen in 31 (51.7%), manifesting as hyperintensities on T2W/FLAIR Spain. imaging (58.1%) and DWI (74.2%) sequences. Hippocampal and pulvinar involvement was Email: silvana.sarria.idi@gencat.cat seen in 58.0% and 25.8% of patients, respectively. MRI abnormalities were associated with a longer SE duration (p = 0.013) and the presence of lateralized periodic discharges (LPDs) on EEG (p < 0.001). Amongst the 33 follow-­up MRIs, nine (27.3%) showed mesial temporal sclerosis (MTS), which was associated with severe clinical status (p = 0.031), hippocampal oedema (p = 0.001) and LPDs (p = 0.001) at baseline. A poorer clinical out- come was associated with baseline T2W/FLAIR imaging hyperintensities (p = 0.003). Conclusion: MRI showed abnormalities in more than half of SE patients. A longer SE dura- tion and LPDs on EEG were associated with SE-­related MRI abnormalities and the devel- opment of MTS. KEYWORDS diffusion-­weighted imaging, magnetic resonance imaging, seizures, status epilepticus; epilepsy This is an open access article under the terms of the Creative Commons Attribution-­NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2021 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology Eur J Neurol. 2022;29:3–11.  wileyonlinelibrary.com/journal/ene | 3 4 | SARRIA-­E STRADA et al. I NTRO D U C TI O N follow-­up MRI. Patients who underwent brain surgery with paren- chymal resection were also excluded for the same reason. Status epilepticus (SE) is defined as single or multiple seizures that Other patients excluded from this study were those who could not persist with incomplete return to the baseline neurological status for undergo an MRI examination during or immediately after the SE episode a certain length of time. The time points for defining SE are based because of MRI contraindications (pacemakers and MR-­incompatible on a duration of 5 min for tonic–­clonic convulsive SE and 10 min prosthetic heart valves or metallic implants), critical illness (severe re- for focal SE with impaired consciousness. The critical duration of spiratory, cardiovascular or neurological derangement), the inability to 30–­60 min is associated with neuronal damage, which can lead to remain still, or MRI unavailability. Patients with low diagnostic quality further consequences, such as brain atrophy or mesial temporal scle- MRIs or incomplete MRI protocols were also excluded (Figure 1). rosis (MTS) [1–­3]. Follow-­up MRI after the SE had resolved was performed be- Magnetic resonance imaging (MRI) is often performed in focal-­ tween 3 and 18 months later in patients who agreed to undergo a onset SE to detect underlying structural lesions. T2-­weighted second MRI for the purpose of the study. imaging (T2WI) and diffusion-­weighted imaging (DWI) signal ab- The SE MRI study protocol (baseline and follow-­up) included the normalities can be seen in most cases with a cortico-­subcortical following sequences performed with 3 T or 1.5 T magnets: (1) sagittal distribution ipsilateral to the epileptogenic hemisphere; such abnor- 3D T1-­weighted (T1W) magnetization prepared rapid gradient-­echo malities are potentially reversible and related to the duration of the (MPRAGE) with multiplanar reconstructions along the main hippo- SE [4–­7]. The abnormalities are presumed to be a consequence of campus axes; (2) transverse and coronal 2D T2W fast spin-­echo; (3) the excessive metabolic demand during prolonged epileptic activ- transverse and coronal 2D T2W fluid-­attenuated inversion recovery ity, as seen in functional neuroimaging with nuclear medicine tracers (FLAIR); and (4) transverse single-­shot spin-­echo echo-­planar DW involving positron emission tomography or single-­photon emission with diffusion sensitivity values of 0, 500 and 1000 s/mm2 along computed tomography [8]. all three orthogonal axes with the corresponding apparent diffusion Although several publications describe SE-­related MRI abnor- coefficient (ADC) maps. Gadolinium enhanced sequences were used malities, there are few available data on the correlation between as part of the diagnostic work-­up in selected cases and were not the electroencephalography (EEG) patterns in SE and acute lesions included in the study assessment. Details of the acquisition parame- seen on MRI. The main finding of these studies is that DWI signal ters are included in Table S1. abnormalities are more frequent in patients with lateralized periodic A staff neuroradiologist (SS, 15 years of experience) blinded to discharges (LPDs) on EEG [9,10]. Nonetheless, prospective MRI stud- any clinical information other than the SE diagnosis visually assessed ies in a large series of patients with different ictal EEG patterns are all the studies immediately after the imaging acquisition of each case. lacking. Structural and signal changes on T1W imaging (T1WI), T2WI, FLAIR Some current topics of debate include the timing and duration of and DWI were identified. High-­signal areas on DWI, indicating re- signal abnormalities on T2WI and DWI in SE, as well as their prog- stricted diffusivity, had to be confirmed on the corresponding ADC nostic value at long-­term follow-­up. maps as low-­signal areas. SE-­related abnormalities were defined on The purpose of this study was to characterize MRI features in T2W/FLAIR images with or without restricted diffusion as follows: (1) SE patients and determine their association with clinical and EEG increased cortical signal intensity with or without thickening and blur- findings. The radiological consequences of SE within the first few ring of the cortical grey and/or white matter borders (subcortical), (2) months after its resolution were also analysed. increased signal intensity in the pulvinar nucleus, (3) increased signal intensity and enlarged hippocampus and (4) increased signal intensity in the claustrum/peri-­insular regions. Contralateral cerebellar T2W/ M E TH O D S FLAIR imaging hypersignals and DWI abnormalities were considered to be crossed cerebellar diaschisis caused by a prolonged seizure. SE-­ This is a prospective, single-­centre and longitudinal study approved related MRI findings in aetiologies that may present local hypersignal by the local ethics committee (PR (AG) 140/2018). From July 2017 in T2WI or DWI (tumours, encephalitis, infections or strokes) were to May 2019, patients older than 16 years with a diagnosis of focal-­ included in the analysis only if they showed T2WI or DWI abnormali- onset SE according to clinical and/or EEG criteria (Salzburg criteria) ties independent from the primary lesion. admitted to our epilepsy unit were included [11]. All included pa- In addition, the observer visually assessed the presence of focal tients underwent a standardized brain MRI examination within the areas of restricted diffusivity associated with signal abnormalities first 240 h after the diagnosis of SE. detected on T2W/FLAIR. The observer specifically sought SE-­ All patients with hypoxic-­ischaemic encephalopathy were ex- related abnormalities, differentiating them from non-­SE-­related cluded to avoid misdiagnosis of SE-­related DWI findings. Patients structural lesions. Signal abnormalities on T2W/FLAIR and DW im- were also excluded who presented with acute/subacute or chronic ages and hippocampal volume changes were visually assessed on extensive brain infarction (>two-­thirds of a vascular territory) or the follow-­up scans, with particular attention given to the identifi- lobar haematoma or chronic traumatic injuries that produced a vol- cation of MTS (decreased hippocampal volume associated with a T2 ume loss greater than one-­third of a lobe to avoid misinterpreting hypersignal). MRI FINDINGS IN STATUS EPILEPTICUS | 5 F I G U R E 1 Patient recruitment flowchart. EEG, electroencephalography; MRI, magnetic resonance imaging; SE, status epilepticus The type of SE was diagnosed and classified according to the to chronic brain lesions. Progressive SE occurs in the context of a International League Against Epilepsy (ILAE-­2015), where SE is di- progressive neurological disease, such as CNS neoplasm, mitochon- vided into SE with prominent motor symptoms (generalized convul- drial encephalopathies and CNS storage disorders [1]. sive SE, myoclonic SE and focal motor SE) or SE without prominent The severity of the SE was determined by two scores: the mod- motor symptoms (non-­convulsive status epilepticus [NCSE] without ified Status Epilepticus Severity Score (mSTESS) calculated for each coma, NCSE with coma and focal SE) [1]. patient at admission [14] and the EMSE determined at 24 h [15]. The SE video-­EEG protocol included 21 scalp electrodes using Refractory SE was established when seizures persisted despite the the 10–­20 international system, electrocardiogram, electromyo- use of at least two drugs (benzodiazepine and one antiepileptic drug) gram and respiratory band with a duration range between 45 min at appropriate doses. and 4 h. Serial video-­EEGs with time gaps from 6 to 24 h were The duration of the SE was determined clinically or by EEG. SE performed as part of the routine diagnostic work-­up. The duration onset was the time at which symptoms began or the last time the and frequency of the consecutive video-­EEG recordings depended patient was seen to be asymptomatic. The conclusion or end-­point upon the clinical evolution up to the cessation of seizures. An ex- was the time EEG monitoring recorded seizure suppression. In co- pert neurophysiologist (MS) read the EEGs according to the 2012 matose NCSE patients with an unknown starting time, SE onset was consensus of the American Clinical Neurophysiology Society [12] considered the last time they were seen to be asymptomatic. The and Salzburg's Consensus EEG Criteria [11]. The EEGs were classi- time interval from SE onset to the MRI study was also determined. fied according to the Epidemiology-­based Mortality Score in Status Poor outcome was defined as the development of a new disabil- Epilepticus (EMSE), which in brief includes the following patterns: ity or an increase in the modified Rankin Scale score. Good outcome spontaneous burst-­suppression, after status ictal discharge (ASID), was considered when the functional status returned to baseline. generalized periodic discharges (GPDs), LPD or none of the previ- ous patterns [13]. Demographics, clinical data and baseline functional status Statistical analysis using the modified Rankin Scale were prospectively recorded on admission. Descriptive and frequency statistical analyses were performed, and The aetiology of the SE was classified as acute symptomatic, re- comparisons were made using IBM SPSS Statistics 22.0 software. mote symptomatic, progressive and unknown (cause not identified Categorical variables are reported as frequencies (percentages), after complete study). In acute symptomatic SE, the precipitating ae- continuous variables as the mean ± standard deviation (SD) and tiology occurs within a few weeks prior to SE onset and can include, non-­normally distributed quantitative variables as the median and for example, stroke, traumatic brain injury or central nervous system interquartile range (IQR). The normality assumption for quantitative (CNS) infection. In remote symptomatic SE, the aetiology is related variables was checked with the use of quantile–­quantile (Q-­Q) plots. 6 | SARRIA-­E STRADA et al. TA B L E 1 Demographic and clinical variables of the patients Statistical significance in the comparisons of MRI abnormalities in pa- tients with a poor prognosis and development of MTS was assessed by Men 34 (56.7%) Pearson's chi-­squared or Fisher's exact test for categorical variables, Mean age, years, mean ± SD 58.3 ± 20.6 Student's t test for continuous variables and the Mann–­Whitney U test Aetiology for non-­normally distributed quantitative variables. A receiver operat- Acute symptomatic 26 (43.3%) ing characteristic curve was plotted to calculate the cut-­off point for Remote symptomatic 14 (23.3%) SE duration with the best sensitivity and specificity in predicting MRI Progressive symptomatic 13 (21.7%) abnormalities; the optimal cut-­off point was obtained using the maxi- Unknown 7 (11.7%) mum value of Youden's index (sensitivity + specificity − 1). Variables Aetiology in the univariable analysis associated with a p value <0.1 were entered Cerebral neoplasm 16 (26.7%) into multiple logistic regression models to identify factors indepen- Toxic/drug overdose 9 (15%) dently associated with SE-­related MRI abnormalities and MTS on fol- low-­up MRI. Because of the small follow-­up sample size, the results of Chronic cerebrovascular disease or chronic 8 (13.3%) traumatic brain injury this last regression model should be considered exploratory. p values Unknown cause 7 (11.7%) <0.05 were considered statistically significant. Non-­adherence 6 (10%) Acute stroke 6 (10%) R E S U LT S Acute infection 3 (5%) Inflammatory/autoimmune 3 (5%) Demographics and clinical characteristics Metabolic 2 (3.3%) Pre-­existing seizure 22 (36.7%) In all, 252 patients were admitted during the study period, of whom mSTESS score, median (IQR) 3 (1–­4) 60 met the inclusion criteria of the study (Figure 1). They were 56.7% EMSE score, median (IQR) 62 (39–­81) (n = 34) men, and the mean age of the enrolled cohort was 58.3 EEG (±20.6) years. The patients' demographic and clinical characteristics Ictal focal activity 48 (80%) are summarized in Table 1. The median mSTESS score was 3 (IQR Findings on EEG (EMSE EEG): 1–­4), and the median EMSE was 62 (IQR 39–­81). Video-­EEG findings showed ictal patterns in 48 (80%) patients as focal seizures (Table 1), LPDs 18 (30%) 18 (30%) LPDs plus and eight (13.3%) rhythmic delta activity plus. ASID 20 (33.3%) According to the EMSE EEG score 18 (30%) LPDs, 20 (33.3%) ASIDs GPDs 1 (1.7%) and one (1.7%) GPDs were observed (Table 1). SE duration, h, median (IQR) 26.4 (9.1–­ The median SE duration was 26.4 h (IQR 9.1–­92.9 h), and 36 92.9) (60%) episodes were refractory/super-­refractory SE with a median Refractory/super-­refractory SE duration, h, median 46.2 (15.2–­ duration of 46.2 h (IQR 15.2–­108.3 h). The median time interval from (IQR) 108.3) SE onset to the MRI study was 123 h (IQR 76.8–­183.8 h) (Table 1). ILAE classification Prominent motor symptoms Generalized seizures 16 (26.7%) Radiological findings Focal motor seizures 15 (25%) No prominent motor symptoms In the overall MRI analysis, 49 patients (81.7%) presented with abnor- Without coma 26 (43.3%) mal findings, but these were considered to be SE related in 31 (63.3%). With coma 3 (5%) SE-­related features included DWI abnormalities in 23 patients (74.2%) Time interval to MRI examination, h, median (IQR) 123 (76.8–­ and hyperintensities on T2W/FLAIR imaging suggestive of oedema in 183.8) 18 (58.1%) (Figure 2). The most frequent finding was an abnormality Outcome at discharge in the ipsilateral hippocampus (Figure 3). A unilateral claustrum/peri-­ Unfavourable 23 (38.3%) insular T2W/FLAIR hyperintensity was seen in a single patient with SE Returned to baseline status 37 (61.7%) secondary to herpes simplex virus 1 encephalitis. Crossed cerebellar diaschisis was seen in five patients (15%) (Table 2). Abbreviations: ASID, after status ictal discharge; EEG, electroencephalography; EMSE, Epidemiology-­based Mortality Score Analysis of the factors associated with the presence of SE-­related in Status Epilepticus; GPDs, generalized periodic discharges; ILAE, hyperintensities on T2W/FLAIR or DW images revealed a relation- International League Against Epilepsy; IQR, interquartile range; LPDs, ship with higher EMSE scores (median 74, IQR 61–­88 vs. 42, IQR lateralized periodic discharges; mSTESS, modified Status Epilepticus 21–­65; p < 0.001) and specific aetiologies, such as chronic cerebro- Severity Score; MRI, magnetic resonance imaging; SE, status vascular disease or traumatic brain injury (19.4% vs. 6.9%; p = 0.037). epilepticus. MRI FINDINGS IN STATUS EPILEPTICUS | 7 SE-­related MRI signal abnormalities were also associated with the 86.2%) was >60 h duration (p = 0.002). No significant association was median SE duration (69.6 h, IQR 13–­109 vs. 15 h, IQR 3.1–­43.2; found with the time interval between symptom onset and MRI exam- p = 0.013), and the best cut-­off point (sensitivity 51.6%, specificity ination (p = 0.985) (Table 3). There was a strong association between SE-­related abnormalities and the presence of LPDs (48.4% vs. 10.3%) or ASIDs (41.9% vs. 24.1%) (p < 0.001) on EEG (Table 3). All MRI findings in our series were unilateral and ipsilateral to the epileptic source. The single patient with GPDs on EEG had a normal MRI. According to the seizure onset localization on EEG, lesions on DWI (73.9% vs. 37.8%, p = 0.007) and in the hippocampus (78.6% vs. 43.5%, p = 0.021) were more probably seen in temporal lobe origin SE. After performing multiple logistic regression analysis, the presence of LPDs (odds ratio [OR] 7.485, 95% confidence interval [CI] 1.075–­ 52.120; p = 0.042) and prolonged SE duration (>60 h) (OR 4.467, 95% CI 1.005–­19.843; p = 0.049) were found to predict SE-­related abnor- malities on DWI and T2W/FLAIR imaging. All patients with LPDs and SE duration >60 h showed MRI abnormalities (Figure S1). The 14 patients (23.3%) who underwent MRI during the active SE seizure were more likely to have T2W/FLAIR imaging hyperin- tensities (44.4% vs. 14.3%, p = 0.019), but there were no significant differences regarding DWI abnormalities. In this subset of patients, the median SE duration was 165.6 h (IQR 72–­319 h), and 57.1% had a temporal lobe origin. At hospital discharge, 23 (38.3%) patients had a poor outcome, and 37 (61.7%) returned to their baseline clinical status. Patients with MRI abnormalities tended to have poorer outcomes (65.2% vs. 43.2%, p = 0.098), particularly those with SE-­related T2W/FLAIR imaging hyperintensities (52.2% vs. 16.2%, p = 0.003). F I G U R E 2 A 46-­year-­old man with new-­onset status epilepticus. (a), (c) Axial T2WI and (b), (d) DWI show an irregularly shaped node Magnetic resonance imaging findings at follow-­up (10-­mm diameter) with peripheral haemosiderin staining in the left parietal lobe, consistent with a cavernous malformation (a, white Follow-­up MRI was performed in 33 of the 60 patients (55%) at a arrow). The hypersignal and cortical thickening beyond the oedema mean interval of 9.7 months (3.2–­17.4 months) after SE cessation. surrounding the cavernoma (a, b, black arrows) are probably T2W/FLAIR and DW MRI abnormalities recovered in 15 of the 33 due to cytotoxic oedema. (c), (d) Follow-­up MRI at 6 months shows resection of the cavernoma (c, arrow) and resolution of patients (45%), whereas in the 10 patients who had normal baseline abnormalities associated with status epilepticus (d, arrow) findings MRI remained normal at follow-­up. TA B L E 2 Radiological findings MRI examinations N (%) MRI abnormal 49 (81.7%) Cortical lesions 24 (49%) MTS at baseline 2 (4%) SE-­related 31 (63.3%) DWI abnormalities (restricted diffusivity) 23 (74.2%) ↑ T2W/FLAIR imaging (with/without cortical thickening) 18 (58.1%) Hippocampal ↑ T2W/FLAIR imaging + enlargement 13 (41.9%) Pulvinar nucleus ↑ T2W/FLAIR imaging (+ DWI) 8 (25.8%) Crossed cerebellar diaschisis (↑ T2W/FLAIR) 5 (16%) Claustrum/peri-­insular ↑ T2W/FLAIR 1 (3.2%) Abbreviations: ↑, increased signal; DWI, diffusion-­weighted imaging; FLAIR, T2W fluid-­attenuated inversion recovery; MRI, magnetic resonance imaging; MTS, mesial temporal sclerosis; SE, status epilepticus; T2W, T2-­weighted. 8 | SARRIA-­E STRADA et al. (a) (b) (c) (d) (e) (f) (g) (h) F I G U R E 3 Two illustrative cases with temporal onset SE. (a)–­(d) Case 1: a 76-­year-­old woman with an SE secondary to a subacute left posterior cerebral artery infarction (a, b, arrowheads). In baseline MRI, FLAIR hyperintensity (a, arrow) and DWI restriction (b, arrow) were observed in the left hippocampus. Moreover, a coronal FLAIR image showed hippocampal enlargement and hyperintensity (c, arrow). The 6-­month follow-­up MRI showed left mesial temporal sclerosis (d, arrow). (e)–­(h) Case 2: a 65-­year-­old man with nonconvulsive SE of unknown aetiology. In baseline MRI, right hippocampus hyperintensity and enlargement on axial T2WI (e, arrow) and coronal FLAIR (g, arrow), as well as restriction on DWI (f, arrow) were observed. The 9-­month follow-­up MRI showed right mesial temporal sclerosis (h, arrow) MTS is was seen in two patients on baseline MRI, one of which and certain EEG patterns. Unlike other studies, a mid-­term follow-­up was related to the epileptogenic focus (Table 2). On follow-­up MRI, MRI in some patients showed that acute SE-­related abnormalities in eight patients (24.2%) with a mean age of 58 (±5.5) years had de- the temporal lobe implied a risk for further MTS development. veloped unilateral MTS, whereas one patient with unilateral MTS Status-­epilepticus-­related MRI abnormalities were seen at a at baseline showed progressive hippocampal atrophy, considered a higher rate than in previous studies [5,10,16,17]. As expected, ab- consequence of the SE. One patient had a history of SE 1 year before normal findings were commonly observed in patients with an acute and another one had active epilepsy. symptomatic aetiology, malignant tumours and chronic cerebral le- The development of MTS was associated with a higher percent- sions [18–­20]. The two main factors associated with the develop- age of LPDs on EEG (77.8% vs. 12.5%, p < 0.001) and a lower median ment of MRI abnormalities were the duration of the SE episode and EMSE score (55, IQR 33–­72, vs. 69, IQR 64–­77; p = 0.031). The tem- the presence of LPDs on EEG examination. The relevance of the du- poral lobe was the most likely epileptic source of the SE in patients ration of the ictal activity is in line with previous publications, which with MTS on follow-­up MRI (seven patients, 77.8%). showed that signal abnormalities on T2WI and DWI were observed Following multiple regression analysis, the factors inde- more frequently in patients with prolonged SE [2,4,16,17]. pendently associated with MTS on follow-­up MRI were T2W/FLAIR Our study showed a large heterogeneity of EEG findings. and DWI abnormalities in the hippocampus (OR 28.074, 95% CI However, the main EEG pattern observed was the presence of LPDs 1.990–­396.131; p = 0.014) and LPDs on EEG (OR 14.061, 95% CI in line with other studies [10]. LPDs commonly described in SE are 1.095–­180.486; p = 0.042) during the SE episode (Table S2). attributed to disconnection of the cortex from subcortical struc- tures in the presence of brain lesions such as strokes, encephalitis or malignancies [21,22]. In our study, LPDs were strongly associated DISCUSSION with SE-­related MRI abnormalities, particularly in the hippocampal and pulvinar regions, even in the absence of underlying lesions [21]. This is a prospective study of patients who underwent MRI during As has been observed in previous studies, SE-­related MRI ab- an episode of SE, where it was observed that the presence of SE-­ normalities occur ipsilateral to ictal activity on EEG [2,21,22]. It related MRI abnormalities was associated with some clinical features was found that specific alterations in the hippocampus (41.9%) and MRI FINDINGS IN STATUS EPILEPTICUS | 9 TA B L E 3 Status-­epilepticus-­related SE-­related abnormalities hyperintensities on DWI and T2W/FLAIR imaging n Yes No p Mean age, years, mean ± SD 55.87 ± 12.29 60.97 ± 22.81 0.343 Type of seizures (worst) Focal 14 (45.2%) 11 (37.9%) 0.478 Generalized 14 (45.2%) 17 (58.6%) NCSE with coma 3 (9.7%) 1 (3.4%) mSTESS, median (IQR) 2 (1–­3) 3 (2–­4) 0.142 EMSE, median (IQR) 74 (61–­88) 42 (21–­65) <0.001 ILAE classification Prominent motor symptoms 14 (45.2%) 17 (58.6%) 0.297 No prominent motor symptoms 17 (54.8%) 12 (41.4%) ILAE aetiology Acute 19 (32.3%) 16 (55.2%) 0.363 Remote 9 (29.0%) 5 (17.2%) Progressive 8 (25.8%) 5 (17.2%) Unknown 4 (12.9%) 3 (10.3%) Findings on EEG (EMSE EEG) LPDs 15 (48.4%) 3 (10.3%) <0.001 ASID 13 (41.9%) 7 (24.1%) GPDs 0 1 (3.4%) No LPD/ASID/GPD 3 (9.7%) 18 (62.1%) SE duration, h, median (IQR) 69.6 (13–­109) 15 (3.1–­43.2) 0.013 Time interval to MRI examination, h, 120.7 (73.6–­162.9) 126.5 0.985 median (IQR) (90.6–­212.9) Duration >60 h 25 (80.1%) 11 (37.9%) 0.002 Abbreviations: ASID, after status ictal discharge; DWI, diffusion-­weighted imaging; EEG, electroencephalography; EMSE, Epidemiology-­based Mortality Score in Status Epilepticus; FLAIR, T2W fluid-­attenuated inversion recovery; GPDs, generalized periodic discharges; ILAE, International League Against Epilepsy; IQR, interquartile range; LPDs, lateralized periodic discharges; MRI, magnetic resonance imaging; mSTESS, modified Status Epilepticus Severity Score; NCSE, non-­convulsive SE; SD, standard deviation; SE, status epilepticus; T2W, T2-­weighted. pulvinar nucleus (25.8%) were strongly related to the temporal origin imaging increased the risk of a poor functional outcome at hospital of the SE. However, this type of involvement can be seen regardless discharge and the development of MTS, which supports previous of the origin of SE as a consequence of seizure propagation [4,22–­ findings [2]. Due to the limited number of patients in whom a fol- 24]. The involvement of mesial temporal structures is widely rec- low-­up MRI was obtained, no firm conclusions can be drawn regard- ognized in epilepsy, whilst the thalamic pulvinar nucleus has gained ing the value of the follow-­up findings as a marker of permanent less attention. Although not directly involved in epileptogenesis, this functional and structural damage. structure is especially prone to changes induced by seizure propaga- Due to the high number of refractory SE, the lack of early avail- tion [4,22–­24]. The claustrum/peri-­insular involvement described in ability for MRI and incomplete or suboptimal MRI quality, a dynamic aggressive forms of febrile SE was rare in our series [25]. evaluation of SE-­related abnormalities on MRI was not possible for all A quarter of the patients with follow-­up MRI presented with patients. However, according to a previous retrospective study pub- MTS. This feature was strongly associated with baseline hippocam- lished by our group, the main predictive factor for the presence of SE-­ pal abnormalities on DW and T2W/FLAIR images and with LPDs and related MRI abnormalities was the duration of the SE [2], which could temporal-­lobe-­onset seizures on EEG. Although the duration of the explain the lack of relation between early MRI acquisition after SE onset SE was not found to be a risk factor for developing MTS, more evi- and the presence of SE-­related abnormalities on MRI. Several reports dence is required to confirm this. describe the dynamics of quantitative changes in multi-­parametric Despite the potential reversibility of the MRI abnormalities in SE, MRI. However, all these studies were carried out in rats [6,26], and T2W/FLAIR imaging hyperintensities may indicate a poor prognosis. therefore the factors and timing of MRI abnormalities in humans are Moreover, it was found that the presence of lesions on T2W/FLAIR not yet known. Moreover, gadolinium contrast enhancement was not 10 | SARRIA-­E STRADA et al. included in the protocol, which has been shown to be a promising di- 2. Requena M, Sarria-­Estrada S, Santamarina E, et al. Peri-­ictal mag- agnostic tool in recent studies of new-­onset SE. An initial diffuse lep- netic resonance imaging in status epilepticus: temporal relationship and prognostic value in 60 patients. Seizure. 2019;71:289-­294. tomeningeal enhancement has been documented [27]. 3. Hocker S, Nagarajan E, Rabinstein AA, Hanson D, Britton JW. In summary, our study shows that SE-­related MRI abnormalities Progressive brain atrophy in super-­refractory status epilepticus. are associated with the duration of SE and the presence of LPDs on JAMA Neurol. 2016;73:1201-­1207. EEG examination. Mesial temporal structures and the pulvinar nucleus 4. Mendes A, Sampaio L. Brain magnetic resonance in status epilepti- cus: a focused review. Seizure. 2016;38:63-­67. are the most susceptible to ictal epileptic damage. The presence of 5. Goyal MK, Sinha S, Ravishankar S, Shivshankar JJ. Peri-­ictal signal hyperintensities in the hippocampus on T2W/FLAIR imaging together changes in seven patients with status epilepticus: interesting MRI with LPDs on EEG may be predictive of a poor functional outcome. observations. Neuroradiology. 2009;51:151-­161. 6. Eidt S, Kendall EJ, Obenaus A. Neuronal and glial cell populations in the piriform cortex distinguished by using an approximation of AU T H O R C O N T R I B U T I O N S q-­space imaging after status epilepticus. AJNR Am J Neuroradiol. Silvana Sarria Estrada: Conceptualization (lead); formal analysis (lead); 2004;25:1225-­1233. investigation (lead); methodology (lead); validation (lead); visualiza- 7. Toledo M, Munuera J, Sueiras M, Rovira R, Álvarez-­Sabín J, tion (lead); writing—­original draft (lead); writing—­review and editing Rovira A. MRI findings in aphasic status epilepticus. Epilepsia. 2008;49:1465-­1469. (lead). Estevo Santamarina: Conceptualization (equal); formal analysis 8. von Oertzen TJ. PET and ictal SPECT can be helpful for localizing (supporting); investigation (supporting); methodology (supporting); epileptic foci. Curr Opin Neurol. 2018;31:184-­191. supervision (equal); validation (equal); writing—­original draft (sup- 9. Narayanan J. Can diffusion-­weighted imaging be used as a tool porting); writing—­review and editing (supporting). Manuel Quintana: to predict seizures in patients with PLEDS? Epileptic Disord. 2016;18:440-­4 46. Conceptualization (supporting); formal analysis (equal); investigation 10. Giovannini G, Kuchukhidze G, McCoy MR, Meletti S, Trinka E. (supporting); methodology (supporting); supervision (supporting); Neuroimaging alterations related to status epilepticus in an adult validation (supporting); writing—­original draft (supporting); writing—­ population: definition of MRI findings and clinical-­EEG correlation. review and editing (supporting). Deborah Pareto: Conceptualization Epilepsia. 2018;59(Suppl 2):120-­127. (supporting); formal analysis (supporting); software (supporting). Maria 11. Leitinger M, Trinka E, Gardella E, et al. Diagnostic accuracy of the Salzburg EEG criteria for non-­convulsive status epilepticus: a retro- Sueiras: Conceptualization (supporting); investigation (supporting); spective study. Lancet Neurol. 2016;15:1054-­1062. supervision (supporting); writing—­review and editing (supporting). 12. Hirsch LJ, LaRoche SM, Gaspard N, et al. American Clinical Cristina Auger: Conceptualization (supporting); investigation (sup- Neurophysiology Society's standardized critical care EEG terminol- porting); writing—­review and editing (supporting). Manuel Toledo: ogy: 2012 version. J Clin Neurophysiol. 2013;30:1-­27. 13. Leitinger M, Höller Y, Kalss G, et al. Epidemiology-­based Mortality Conceptualization (lead); formal analysis (equal); investigation (equal); Score in Status Epilepticus (EMSE). Neurocrit Care. 2015;22:273-­282. methodology (equal); supervision (equal); writing—­original draft (sup- 14. González-­Cuevas M, Santamarina E, Toledo M, et al. A new clinical porting); writing—­review and editing (equal). Alex Rovira: Formal score for the prognosis of status epilepticus in adults. Eur J Neurol. analysis (supporting); investigation (supporting); methodology (sup- 2016;23:1534-­1540. 15. Ciurans J, Grau-­López L, Jiménez M, Fumanal A, Misis M, Becerra porting); supervision (supporting); validation (supporting); writing—­ JL. Refractory status epilepticus: impact of baseline comorbidity original draft (supporting); writing—­review and editing (equal). and usefulness of STESS and EMSE scoring systems in predicting mortality and functional outcome. Seizure. 2018;56:98-­103. DATA AVA I L A B I L I T Y S TAT E M E N T 16. Nakae Y, Kudo Y, Yamamoto R, et al. Relationship between cortex and pulvinar abnormalities on diffusion-­weighted imaging in status Data available on request from the authors. The data that support epilepticus. J Neurol. 2016;263:127-­132. the findings of this study are available from the corresponding au- 17. Raghavendra S, Ashalatha R, Krishnamoorthy T, et al. Reversible thor upon reasonable request. periictal MRI abnormalities: clinical correlates and long-­term out- come in 12 patients. Epilepsy Res. 2007;73:129-­136. 18. Giovannini G, Pasini F, Orlandi N, Mirandola L, Meletti S. Tumor-­ ORCID associated status epilepticus in patients with glioma: clinical char- Silvana Sarria-­Estrada https://orcid.org/0000-0003-0083-362X acteristics and outcomes. Epilepsy Behav. 2019;101:106370. Estevo Santamarina https://orcid.org/0000-0003-1915-0335 19. Szabo K, Poepel A, Pohlmann-­Eden B, et al. Diffusion-­weighted and Manuel Quintana https://orcid.org/0000-0003-0288-9088 perfusion MRI demonstrates parenchymal changes in complex par- Deborah Pareto https://orcid.org/0000-0001-7356-3769 tial status epilepticus. Brain. 2005;128:1369-­1376. 20. Sarria-­Estrada S, Toledo M. (2019). Status epilepticus. In: Barkhof F, Maria Sueiras https://orcid.org/0000-0003-3184-5615 Jager R, Thurnher M, Rovira Cañellas A (eds), Clinical Neuroradiology. Cristina Auger https://orcid.org/0000-0001-9846-4767 Springer. 10.1007/978-­3-­319-­61423​-­6_54-­1. Epub 2019 March 9. Manuel Toledo https://orcid.org/0000-0003-1550-3316 Online ISBN: 978-­3-­319-­61423-­6. Alex Rovira https://orcid.org/0000-0002-2132-6750 21. Rennebaum F, Kassubek J, Pinkhardt E, et al. Status epilepticus: clinical characteristics and EEG patterns associated with and without MRI diffusion restriction in 69 patients. Epilepsy Res. REFERENCES 2016;120:55-­6 4. 1. Trinka E, Cock H, Hesdorffer D, et al. A definition and classification 22. Kim SE, Lee BI, Shin KJ, et al. Characteristics of seizure-­induced of status epilepticus -­report of the ILAE task force on classification signal changes on MRI in patients with first seizures. Seizure. of status epilepticus. Epilepsia. 2015;56:1515-­1523. 2017;48:62-­68. MRI FINDINGS IN STATUS EPILEPTICUS | 11 23. Jabeen SA, Cherukuri P, Mridula R, et al. A prospective study of diffusion weighted magnetic resonance imaging abnormalities in S U P P O R T I N G I N FO R M AT I O N patients with cluster of seizures and status epilepticus. Clin Neurol Additional supporting information may be found online in the Neurosurg. 2017;155:70-­74. Supporting Information section. 24. Chatzikonstantinou A, Gass A, Förster A, Hennerici MG, Szabo K. Features of acute DWI abnormalities related to status epilepticus. Epilepsy Res. 2011;97:45-­51. How to cite this article: Sarria-­Estrada S, Santamarina E, 25. Meletti S, Slonkova J, Mareckova I, et al. Claustrum damage and refractory status epilepticus following febrile illness. Neurology. Quintana M, et al. Magnetic resonance imaging findings in 2015;85:1224-­1232. focal-­onset status epilepticus. Eur J Neurol. 2022;29:3–­11. 26. Wall CJ, Kendall EJ, Obenaus A. Rapid alterations in diffusion-­ https://doi.org/10.1111/ene.15065 weighted images with anatomic correlates in a rodent model of sta- tus epilepticus. AJNR Am J Neuroradiol. 2000;21:1841-­1852. 27. Kim HJ, Lee SA, Kim HW, Kim SJ, Jeon SB, Koo YS. The timelines of MRI findings related to outcomes in adult patients with new-­onset refractory status epilepticus. Epilepsia. 2020;61:1735-­1748.