Prognostic Role of Functional Neuroimaging after Multilobar Resection in Patients with Localization-Related Epilepsy

To investigate the usage of functional neuroimaging as a prognostic tool for seizure recurrence and long-term outcomes in patients with multilobar resection, we recruited 90 patients who received multilobar resections between 1995 and 2013 with at least 1-year follow-up (mean 8.0 years). All patients were monitored using intracranial electroencephalography (EEG) after pre-surgical evaluation. Clinical data (demographics, electrophysiology, and neuroimaging) were reviewed retrospectively. Surgical outcomes were evaluated at 1, 2, 5 years after surgery, and at the end of the study. After 1 year, 56 patients (62.2%) became Engel class I and at the last follow-up, 47 patients (52.2%) remained seizure-free. Furthermore, non-localized 18F-fluorodeoxyglucose positron emission tomography (PET), identifying hypometabolic areas not concordant with ictal onset zones, significantly correlated with seizure recurrence after 1 year. Non-lesional magnetic resonance imaging (MRI) and left-sided resection correlated with poor outcomes. In the last follow-up, non-localized PET and left-sided resection significantly correlated with seizure recurrence. Both localized PET and ictal-interictal SPECT subtraction co-registered to MR (SISCOM) predicted good surgical outcomes in the last follow-up (69.2%, Engel I). This study suggests that PET and SISCOM may predict postoperative outcomes for patients after multilobar epilepsy and shows comparable long-term surgical outcomes after multilobar resection.


Introduction
Medically intractable localization-related epilepsy (LRE) patients have potential to eliminate seizures by resection surgery for epilepsy. Multilobar resection is used for patients with multiple epileptogenic foci. However, surgical outcomes for multilobar resection are worse than single lobe resections. [1][2][3][4] Therefore, multilobar resection is regarded to be a poor prognostic factor of surgical outcomes in patients with focal cortical dysplasia [1] and extra-temporal lobe epilepsy. [5,6] Previous studies have reported that 41% of patients undergoing multilobar resection were seizure-free after 10 years. [7] A recent study also showed even better results where 60.6% of patients with two or more lobe resections (54/89) had good outcomes (Engel I-III). [3] However, the duration of follow-up was short (2 years) and the range of 'good outcomes' was broader than other reports.

Methods Participants
The study was conducted at Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea. We retrospectively evaluated 90 patients with multilobar epilepsy who underwent resection surgery between May 1995 and October 2013. Participants had iEEG monitoring to localize the epileptogenic foci and to determine the resection margin. Epilepsy was intractable before surgery despite proper and sufficient antiepileptic drug (AED) treatment. Clinical characteristics registered for each patient included age of seizure onset, age at surgery, duration of epilepsy, history of febrile seizures, monthly seizure frequency, and number of AEDs at surgery.

Ethics Statement
All patients provided written informed consent for their participation in the study. Written informed consent was obtained from the next of kin, caretakers, or guardians on the behalf of the minors/children participants involved in this study. The study was approved by Institutional Review Board of Samsung Medical Center.

Presurgical evaluation
Intractable epilepsy patients received a comprehensive pre-surgical evaluation consisting of complete neurologic examination, scalp video-electroencephalography (EEG) monitoring and brain magnetic resonance imaging (MRI) during the first admission period. Ictal and interictal SPECT studies were performed to lateralize or localize epileptic foci. In case of multilobar epilepsy, the patient underwent PET and neuropsychological tests during the second admission. All data from these admissions were reviewed and discussed in an epilepsy management conference and the surgical strategy, including intracranial EEG (iEEG) monitoring, was established to remove the epileptic foci. In this study, multilobar epilepsy was defined when multilobar surgical resection was performed after multiple ictal onset zones were confirmed from more than two lobes during the iEEG monitoring.

Analyses of clinical seizures during scalp EEG monitoring
We reviewed each patient's seizures carefully. The presence of aura was determined by patient memory or the patient pressing a button before seizures.

Scalp video EEG monitoring
The 10/10 system for scalp electrodes was used. AEDs were usually reduced or stopped to facilitate the recording of seizures.
-Interictal EEG classification: Interictal epileptiform discharges (IED) were defined as unilobar when 75% or more preponderance in one lobe and as multilobar when over 2 lobes with less than 75% preponderance in any single lobe.
-Ictal EEG classification: Unilobar was defined when the location of scalp ictal EEG onset (sEEG onset) was confined to one lobe, the amplitude ratio of one lobe versus the other lobes was greater than 2:1 in bipolar montages, and greater than 2:1 for the two sides in referential montages. Multilobar was defined when the sEEG onset initiated from 2 lobes over both hemispheres independently or synchronously.
iEEG monitoring iEEG was performed using a combination of grids/strips or depth electrodes. Anatomical targeting of electrodes was established in each patient according to available non-invasive information and hypotheses regarding localization of the epileptogenic zone. Concomitant implantation of depth electrodes was performed in nine patients (temporo-occipital [TO], n = 3; fronto-temporal [FT], n = 4; fronto-temporo-parietal [FTP], n = 1; temporo-parietooccipital [TPO] lobe epilepsies, n = 1). Depth electrodes were inserted uni-or bilaterally into the mesial temporal region (amygdala and/or hippocampus) and the electrode locations were confirmed by intraoperative photographs and post-implantation surface-rendered electrode images. Ictal onset zone (IOZ) was defined when any paroxysmal, sustained ictal EEG pattern during the iEEG monitoring was distinct from background activity and accompanied by clinical seizures. [16] The detailed classification among multilobar epilepsies was determined through the iEEG monitoring of the IOZ site during at least three seizures, regardless of brain lesions.
Neuroimaging studies Brain MRI. MRI was performed using a GE Signa 1.5-Tesla scanner (GE Medical Systems, Inc., Milwaukee, WI, USA) or a 3.0-Tesla scanner (Philips, Best, the Netherlands). All patients underwent Spoiled Gradient Echo, T2-weighted and Fluid Attenuated Inversion Recovery imaging protocols. Also, MRI results were classified as lesional or non-lesional according to the presence of visible lesions with potential epileptogenicity.
FDG-PET studies. PET images were obtained (GE Advance PET scanner, GE Medical Systems, Inc.) after patients had fasted for four or more hours followed by intravenous injection of 7-10 mCi (259-370 MBq) of FDG. EEG during the uptake period demonstrated no EEG seizure activity in any patient. Hypometabolism was determined semi-quantitatively by visual assessment using calibrated color scales. A graduated color scale in 2% increments was used for display and analysis. When the metabolism of the lobe showed a 20% or more reduction compared with the other areas of metabolism, it was regarded as abnormal hypometabolism. [17] Interictal and ictal SPECT studies. Brain SPECT scans were performed 30-60 min after injection of 25 mCi 99mTc-ethyl cysteinate dimer (ECD) using a 3-headed Triad XLT system (Trionix Research Laboratory, Inc., Twinsburg, OH). Interictal SPECT studies were performed when the patients had no documented seizure activity for 24 hours. For ictal studies, patients received radiotracer injections during seizures. The mean radiotracer injection time during the ictal SPECT was 30.0 ± 14.8 seconds (range, 14-105) and the mean seizure duration during injection was 85.8 ± 36.5 seconds (range, . Ictal-Interictal SPECT Subtraction Co-registered to MR images (SISCOM) analysis was performed on an offline workstation with ANA-LYZE 7.5 software (Biomedical Imaging Resource, Mayo Foundation, Rochester, MN) and as previously described. [18] If hyperperfused areas include the IOZ, it is defined as localized. Non-localized indicated hyperperfused spots that were not found within the IOZ. Localized PET was defined when the IOZ was included in the hypometabolic areas, whereas, non-localized PET was when the IOZ was not involved in the hypometabolic areas.
Two neurologists (Cho EB and Joo EY) reviewed the neuroimaging results independently without patient information. In case of discrepancies, the final decision was made by formal reports from the Departments of Radiology or Nuclear Medicine.

Surgery and outcomes
Complete resection was defined as when resection margins include the IOZ with or without frequent interictal spikes in adjacent brain regions and early ictal propagation on the iEEG monitoring. All patients had a minimum follow-up period of 1 year, up to 19 years. When the IOZ was diffuse (non-localized), or included eloquent areas within the resection margins, incomplete resection was performed. Patients were classified as "seizure-free" if they achieved Engel Class I by the last year of follow-up and "completely seizure-free" no seizures occurred after surgery. Early seizures were defined when seizures occurred within 6 months after surgery. The patients were instructed to visit the clinic 1 month after surgery, then every 3 months. If patients became seizure-free, they visited the clinic every 6 months. Postoperative seizure frequency and any possible provocative factors were documented. Surgical outcomes were evaluated at 6 months, 1, 2, and 5 years and at the end of the study period. Postoperative seizure outcomes were determined by outpatient clinic or telephone interviews using Engel's classification.

Statistical analyses
For comparison between seizure-free (Engel I) patients and patients with recurrent seizures (Engel II-IV), Chi-square or Fisher's exact test were applied for categorical variables. A Student's t test or Mann-Whitney U test was performed for continuous variables. Logistic regression analyses were used to verify independent risk factors for seizure recurrence. Variables with p values 0.05 in the simple logistic regression were tested for multiple logistic regression analysis. Statistical significance was accepted at p < 0.05. The time to first seizure recurrence was plotted using a Kaplan-Meier survival curve to estimate the proportion of individuals remaining seizure-free at various time points, according to several prognostic factors. A log rank test and a comparison of 95% confidence intervals were used to establish differences between good and bad prognostic factors of seizure recurrence after surgery.

Postoperative progress and outcomes
The median postoperative follow-up time was 7.3 years (interquartile range, 2.9-12.6) and approximately 77% of patients completed a 5-year follow-up. In the first year after surgery, 56  Table 3 depicts seizure outcomes in each group. A Kaplan-Meier plot of time to the first seizure after surgery revealed 72% (n = 65, 95% CI 63-82) of patients were seizure-free at 6 months, 60% (n = 54, 95% CI 50-70) at 1 year, and 40% (n = 35, 95% CI 28-52) at 5 years (Fig 1). Early seizures occurred within 6 months for 25 patients (50%). Another 25 patients (50%) showed seizure recurrence 6 months after surgery. However, the final outcomes of patients with early seizures were not different from patients with late seizure recurrence. In the last follow-up, the number of patients with early seizures was four for Engel I, 10 for Engel II, one for Engel III, and 10 for Engel IV. Similarly, the number of patients with late seizures was three for Engel I, nine for Engel II, four for Engel III, and nine for Engel IV.
Analysis based on resection type revealed that TO and PT resections had better outcomes. At 5 years, 55% of TO and 56% of PT resections were seizure-free, while 34% of FT and 14% of FP resections were seizure-free. In log rank tests, FP resections (n = 6, 85.7%) had more recurrent seizures than TO resections (n = 12, 46.2%) (p = 0.027).

Risk factors for seizure recurrence
Significant prognostic factors for seizure recurrence at 1-year post-operation were age at surgery, absence of aura, left-sided resection, non-lesional MRI, and non-localized PET (Table 4). In the last year of follow-up, left-sided resection, non-localized PET, and incomplete resection were associated with recurrent seizures (Table 5, Fig 2A-2C).
Postoperatively, the presence of interictal epileptiform discharges on scalp EEG were found to be a potent predictor of poor seizure outcomes in the last year of follow-up (OR = 3.62; 95% CI = 1.46-8.97; p = 0.005) (Fig 2D).
There was a significant correlation between incomplete resection and IED on postoperative scalp EEG (p = 0.001). Among 35 patients showing IED on postoperative EEG, 12 were seizure-free (Engel I) and 23 were not (Engel II-IV) in the last year of follow-up.

Discussion
We observed long-term surgical outcomes in patients with multilobar epilepsy (median 7.4 years, 2.9-12.6) and found 40% complete seizure freedom (no postsurgical seizures) at 5 years after surgery. In the last year of follow-up, 52.2% of the 90 patients remained seizure-free (Engel I). This is comparable to a prior study demonstrating 41-44.4% seizure freedom after multilobar resection. [6] The discrepancy of seizure-free rates may result from different subtypes of populations (multilobar vs. extra-temporal, adults vs. pediatric or mixed subjects) and different surgical techniques. Follow-up of up to 19 years in this study provides a more appropriate prognosis of multilobar surgeries than a shorter follow-up. We found that patients with excellent outcomes 1 year after surgery might experience seizures within this follow-up. Most, however, had achieved seizure freedom by the end of follow-up (47/55, 85.4%). Similarly, patients with the worst outcomes (Engel IV) still had persistent seizures in the end (19/20, 95%). Engel II patients showed the most fluctuant course, however, their seizures finally decreased in the last year of follow-up (6.7% ! 21.1%). Over half of patients (50/90, 55.5%) experienced at least one seizure after surgery. Early seizures occurred in 25 patients (25/50, 50%), similar to previous studies. [7,19] However, final outcome of patients with early seizures did not differ from patients with late seizures. Fourteen   Non-localized FDG-PET was a significant prognostic factor for patients with poor outcomes in both the first and the last year of follow-up. To the best of our knowledge, this is the first study presenting a prognostic role for PET findings before multilobar resection. It suggests that localized hypometabolic findings on PET are closely associated with good surgical outcomes. PET reveals interictal brain dysfunction with decreased metabolic need can possibly detects invisible pathologic tissues. [19,20] PET is widely utilized for cryptogenic neocortical epilepsy surgery [12,21,22] and good surgical outcomes were reported after resection of localized hypometabolic areas in several surgical studies. [11,22,23] This means that concordantly localized hypometabolism on PET guides the insertion of subdural grids with confidence and leads to complete resection of epileptic foci, which applies to cases with multilobar epilepsy. Sarkis et al. also reviewed PET for presurgical evaluation. However, they classified findings as normal versus abnormal (ipsilateral vs. bilateral), which showed no differences. [7] Yu et al. reviewed head-PET-computed tomography records of patients with uni-or multilobar epilepsies, but they did not evaluate the role of imaging in prognosis in detail. [6] In this study, we classified PET findings as localized or non-localized according to the IOZ on scalp EEG monitoring. This process is routinely performed during presurgical evaluation in our epilepsy clinic and the results guide surgical planning for patients. We try to cover hypometabolic lesions as much as possible with subdural grids or strips. Mismatches between IOZ and hypometabolic zones (i.e. non-localized PET) may result in mislocation of subdural electrodes or incomplete coverage of epileptic foci, leading to poor surgical outcomes.
Another functional neuroimaging technique, the SPECT (SISCOM) alone, was not a predictive factor during the first or last year of follow-up, contradicting previous reports. [11,24,25] In the whole, we confirmed that patients with both localized PET and SISCOM (n = 26) had good surgical outcomes for multilobar epilepsy in the last year of follow-up (Engel I). Non-lesional MRI was correlated with seizure recurrence at 1-year follow-up, although it was not a potent predictor for poor outcomes in the last year. Rates of good surgical outcomes for non-lesional epilepsy ranged from 26 to 66% according to surgical location, with extra-temporal lobe epilepsy being less effective (particularly for frontal and parietal lobes). [19][20][21]26] Eleven of 24 patients with non-lesional MRI (11/24, 45.8%) showed good seizure outcomes (Engel I) and all showed localized PET and SISCOM. This supports the complementary roles of functional neuroimaging to achieve good surgical outcomes in patients with cryptogenic multilobar LRE.
Seizure recurrence was more frequent in patients with left-sided resection than right-sided resection. The location and laterality of surgery may be associated with surgical outcomes. [2,3,27,28] Surgical outcomes of left-sided frontal lobe epilepsy are much worse than rightsided, [26,27] which is explained by a higher frequency of incomplete resection (left vs. right; 30.8% vs. 0%; p = 0.039). Since most patients are right-handed, left-sided resection may be limited by eloquent areas (language areas in the dominant hemisphere). This is congruent with the finding that incomplete resection of epileptic foci was an independently poor prognostic factor in the last year of follow-up and FP resection had the worst surgical prognosis.
In previous studies, postoperative IEDs on EEG were used as a prognostic factor for seizure recurrence. [7,29,30] Also, we were able to reproduce these results and found a correlation between IED spikes and incomplete resection of IOZ (p = 0.01), suggesting residual epileptogenic tissue. [7] In conclusion, we observed surgical outcomes in patients with multilobar epilepsy during long-term follow-up. Functional neuroimaging, PET and SISCOM may be potent prognostic factors for postoperative outcomes following multilobar resection in patients with LRE.