| Electroencephalographic source imaging (ESI) is a kind of technique which uses epileptic discharges collected from scalp electroencephalography (EEG) to perform intracranial epileptogenic zone localization. There are some studies on it around the world, but very few in our country. Recently,256-channel dense array electroencephalographic source imaging (256-ch dESI), which bears high spatial resolution, is emerging in the field of pre-surgical work-up for intractable epilepsy. Although there isn’t any relevant study in our country, there’re already some ones around the world showing that it is promising. Comparing to current pre-surgical work-up tools for intractable epilepsy such as structural magnetic resonance imaging (MRI), positron emission tomography (PET), conventional electroencephalography (cEEG), semiology and magnetoencephalography/magnetic source imaging (MEG/MSI),256-ch dESI shows many advantages.256-ch dESI has high spatial and temporal resolution, in addition, it is non-invasive, atoxic, suitable for long-term monitoring and sensitive to detect deep sources. In this study, we discussed the pre-surgical work-up value for intractable epilepsy of this tool and the effect of different source patterns on the surgical outcome. Also, we performed 256-ch dESI based on individual head model (256-ch dESI IHM) and its coregisteration study with intracranial electroencephalography (icEEG) to discuss its clinical application and value. We achieved several international and domestic innovations in this study.Part I Pre-surgical work-up value of 256-channel dense array electroencephalographic source imaging in intractable epilepsyThis part discussed pre-surgical work-up value of 256-ch dESI based on 53 cases of intractable epilepsy. All cases underwent comprehensive pre-surgical evaluation for epileptogenic zone including 256-ch dESI, and then accepted one-stage surgeries. Defining the resective zone of patients whose surgical outcome is excellent as the scope of epileptogenic zone, we took both sub-lobar and lobar standards to evaluate the localization value of multiple tools. Standard 1:localization results within the resective zone are deemed as correct, and this standard is applied for comparison among 256-ch dESI, PET and MRI; Standard 2:localization results within the same lobe as the resective zone are deemed as correct, this standard is applied for comparison among all the tools. Moreover, we selected 14 cases of mesial TLE (mTLE) according to MRI, semiology and cEEG. For these cases we defined the mesial temporal structures as the epileptogenic zone to evaluate the accuracy of 256-ch dESI and PET. We used statistic methods to analyze the relationship between source patterns and surgical outcome.The results showed that no matter which standards we took to evaluate the diagnositic value of multiple tools,256-ch dESI presented with the best sensitivity and specificity. Among the 14 mTLE cases,256-ch dESI localized completely onto the mesial temporal structures and basal temporal lobe in 78.6% cases, while PET only 36.4%, which indicated that 256-ch dESI can provide more localized results than PET (p<0.05, Fisher’s exact test). We divided the 53 cases into "single source" group and "multiple sources" group:univariate analysis showed "single source" group had better probability of postoperative seizure freedom than "multiple sources" group (Kaplan-Meier survival, p<0.05, Log Rank). We also divided the two groups into "source within resection" and "source outside resection", the results showed that the former had better probability of postoperative seizure freedom than the latter (Kaplan-Meier survival, p<0.05, Log Rank). Using Cox regression as multivariate analysis, we found out that resection of "sources" is correlated to good prognosis.256-ch dESI has high spatial and temporal resolution, and has advantages over other conventional tools. For the increase of electrodes covering face, cheek and neck, it is more sensitive to detect epileptic discharges from basal and mesial temporal lobe. Studies on the patterns of source results showed that cases with "single source" usually have better prognosis, while cases with "multiple sources" may have poor prognosis. For the latter, icEEG may be considerd in the pre-surgical work-up. Cases whose sources are within resective zones may have better prognosis than whose sources are not, and this may indicate that irritative zones generating IEDs may to some extent represent epileptogenic zone. Ictal EEG is not suitable for 256-ch dESI, but it should be routinely acquired by cEEG.Part â…¡ Surgical treatment of PET positive, MRI negative temporal lobe epilepsy and pre-surgical work-up value of 256-channel dense array electroencephalographic source imaging in this clinical entityFirstly we performed a retrospective study to discuss clinical features, surgical treatment and outcome of PET positive, MRI negative temporal lobe epilepsy (PET+MRI-TLE). From the viewpoint of semiology, demography, surgical treatment and prognosis evaluation, we compared 19 PET+MRI-TLE to 41 TLE with hippocampal sclerosis and 18 lesional TLE, and then statistically analyzed the difference among these three cohorts. It appeared that there was no significant difference between the surgical outcome of PET+MRI-TLE (Engle class I:68.4%; Engle class â… +â…¡:84.2%, p>0.05) and TLE with hippocampal sclerosis (Engle class â… : 68.3%; Engle class â… +â…¡:80.5%, p>0.05). The analysis also showed that to some extent PET+MRI-TLE might be distinct from TLE with hippocampal sclerosis as a clinical entity, i.e. the former is not a subtype of the latter. History of febrile convulsion and occurrence of secondary GTCS may possibly differentiate them from each other. In general, this study demonstrated that PET+MRI-TLE can be surgically treated by one-stage surgeries.Although PET is a sensitive tool for pre-surgical evaluation of TLE, it has poor specificity and usually shows broad localization results.256-ch dESI is newly emerged with high spatial and temporal resolution, and we designed relative studies on its pre-surgical evaluation value in PET+MRI-TLE. By inclusion and exclusion criterion, we selected 12 cases of PET+MRI-TLE patients. We defined resective zone in patients who underwent good prognosis as scope of epileptogenic zone. Localization results within this ranges are considered correct (standard 1), or if the results are within the same lobe of the resective zone, it is considerd correct (standard 2). Using standard 1, accuracy of 256-ch dESI was 77.8%, while two of the cases were localized to posterior temporal lobe or tempo-occipital zone; PET’s accuracy was also 77.8%, two of the cases were localized very broadly. When taking standard 2, as for the lobar level, accuracy of 256-ch dESI was raised to 88.9%, the only case which was still "incorrect" was localized to the tempo-occipital zone; PET was correct in all cases, semiology was 33.3% and cEEG was 55.6%. Cases with source of anterior temporal lobe showed better prognosis than the non-anterior temporal lobe ones. (100% VS 40%, p<0.05, Fisher’s exact test)We are the first to discuss the pre-surgical work-up value of 256-ch dESI in PET+MRI-TLE. The results showed that 256-ch dESI is a promising method for MRI negative TLE. If sources are localized to the anterior temporal lobe, then for PET+MRI-TLE, the one-stage resective surgeries may yield good surgical outcome; but if the sources are localized to the non-anterior temporal region, one-stage surgeries should be cautious. These results are instructive for selection of surgical candidates. Although ictal EEG is not suitable for source imaging, yet ictal EEG data should be collected by cEEG routinely, especially for MRI negative epilepsy. Surgical plan for MRI negative epilepsy should be made by a comprehensive evaluation of multiple pre-surgical work-up tools, rather than by a single tool. If necessary, icEEG should be performed to localize the epileptogenic focus.Part III 256-channel dense array electroencephalographic source imaging based on individual head model and its co-registeration study with intracranial electroencephalography in pre-surgical work-up of intractable epilepsyGenerally, forward head models applied in dESI are usually spherical atlas models. Although some centers are using models with individual MRI incorporation (SMAC), yet the models are still based on a spherical geometry. In this study we used high resolution individual head models in 256-ch dESI (256-ch dESI IHM). Based on clinical epilepsy cases, we discussed the clinical application of 256-ch dESI IHM and performed a multi-modal imaging coregistration study, which’s first in our country and in a leading postion internationally.After acquisition of 3D-SPGR sequence of structural MRI (from top of head to chin) and 256-channel dEEG data, we used computational methods to build individual head models and generate source imaging results. The case underwent intracranial electrodes implantation had a thin-sliced CT scan in the next day after surgery. Coregistration of intracranial electrodes and the individual head model were then performed. We applied two different standards to evaluate the tools. Standard 1: resective zone of patients whose surgical outcome was good; Standard 2: epileptogenic zone confirmed by icEEG. When taking standard 1 evaluation, accuracy of 256-ch dESI IHM was 100%,256-ch dESI AHM was 85.7%, PET was 42.9% and MRI was 71.4%. Only 256-ch dESI IHM correctly localized all the cases. Four "incorrect" PET results were due to broad localization; Two "incorrect" MRIs was due to no meaningful structural lesions. One "incorrect" 256-ch dESI AHM was due to results of neighboring the resective zone. When taking standard 2, only 256-ch dESI IHM is concordant with icEEG, while 256-ch dESI AHM, PET and MRI showed disconcordance.Although 256-ch dESI AHM was less accurate, yet its convenience can make it incorporate into the pre-surgical evaluation workflow quickly and supplement conventional tools. With reconstruction of individual head geometry, distribution of oriented dipoles according to real cortical orientations, design of reasonable tissue conductivity and acquisition of dense array EEG data, it is feasible to perform the highly accurate 256-ch dESI IHM technique. In addition, this technique permits one-view analysis of multiple tools including icEEG, which is beneficial for precisely making surgical plan. We applied this technique in pre-surgical work-up of intractable epilepsy and performed its co-registration study with icEEG, testifying its accuracy and clinical value. Again, we testified with icEEG evidence that correctly recoganized and repeated IEDs representing the irritative zone are related closely to epileptogenic zone. However, similar to other regions like symptomatogenic zone, ictal onset zone, epileptogenic lesion and functional deficit zone, irritative zone can not completely represent epileptogenic zone. Therefore, surgical plan should be made by comprehensive evaluation of multiple tools.256-ch dESI AHM/IHM cannot replace other tools (especially for now,256-ch dESI is mainly based on interictal EEG, ictal EEG should be acquired by cEEG or even icEEG), but one shouldn’t ignore its advantages and instruction significance in epileptogenic zone localization. |
