Prognostic Value of Complete Resection of the High-Frequency Oscillation Area in Intracranial EEG

Background and Objectives High-frequency oscillations (HFOs; ripples 80–250 Hz; fast ripples [FRs] 250–500 Hz) recorded with intracranial electrodes generated excitement and debate about their potential to localize epileptogenic foci. We performed a systematic review and meta-analysis on the prognostic value of complete resection of the HFOs-area (crHFOs-area) for epilepsy surgical outcome in intracranial EEG (iEEG) accessing multiple subgroups. Methods We searched PubMed, Embase, and Web of Science for original research from inception to October 27, 2022. We defined favorable surgical outcome (FSO) as Engel class I, International League Against Epilepsy class 1, or seizure-free status. The prognostic value of crHFOs-area for FSO was assessed by (1) the pooled FSO proportion after crHFOs-area; (2) FSO for crHFOs-area vs without crHFOs-area; and (3) the predictive performance. We defined high combined prognostic value as FSO proportion >80% + FSO crHFOs-area >without crHFOs-area + area under the curve (AUC) >0.75 and examined this for the clinical subgroups (study design, age, diagnostic type, HFOs-identification method, HFOs-rate thresholding, and iEEG state). Temporal lobe epilepsy (TLE) was compared with extra-TLE through dichotomous variable analysis. Individual patient analysis was performed for sex, affected hemisphere, MRI findings, surgery location, and pathology. Results Of 1,387 studies screened, 31 studies (703 patients) met our eligibility criteria. Twenty-seven studies (602 patients) analyzed FRs and 20 studies (424 patients) ripples. Pooled FSO proportion after crHFOs-area was 81% (95% CI 76%–86%) for FRs and 82% (73%–89%) for ripples. Patients with crHFOs-area achieved more often FSO than those without crHFOs-area (FRs odds ratio [OR] 6.38, 4.03–10.09, p < 0.001; ripples 4.04, 2.32–7.04, p < 0.001). The pooled AUCs were 0.81 (0.77–0.84) for FRs and 0.76 (0.72–0.79) for ripples. Combined prognostic value was high in 10 subgroups: retrospective, children, long-term iEEG, threshold (FRs and ripples) and automated detection and interictal (FRs). FSO after complete resection of FRs-area (crFRs-area) was achieved less often in people with TLE than extra-TLE (OR 0.37, 0.15–0.89, p = 0.006). Individual patient analyses showed that crFRs-area was seen more in patients with FSO with than without MRI lesions (p = 0.02 after multiple correction). Discussion Complete resection of the brain area with HFOs is associated with good postsurgical outcome. Its prognostic value holds, especially for FRs, for various subgroups. The use of HFOs for extra-TLE patients requires further evidence.


Introduction
Intracranial EEG (iEEG)-guided epilepsy surgery can cure people from drug-resistant epilepsy. 1,2The premise of surgery is accurate localization and complete removal of epileptogenic brain tissue.The epileptogenic zone (EZ) refers to a theoretical brain region that is necessary and sufficient to initiate epileptic seizures.No diagnostic technique so far can localize the EZ on its own, and the EZ is estimated based on various predictive measures. 3For capturing electrographic biomarkers, intraoperative electrocorticography (ioECoG) and diagnostic intracranial monitoring, including long-term electrocorticography (ECoG) and stereo-EEG, each has its advantages and shortcomings.5][6][7] Both types of brain oscillations have limitations: ripples can be physiologic (predominantly in eloquent regions) and pathologic, 3 while FRs are scarce in occurrence. 8][10][11] Patients with recurrent seizures after epilepsy surgery showed widespread HFOs in preresection ECoG recordings and residual FRs in postresection ECoG recordings. 13,14Some iEEG studies do not show advantages for HFOs. 15,16A Cochrane review in 2017 including 2 prospective studies (11 participants), with low evidence quality, concluded that prognostic value of HFOs was unreliable. 17So far HFOs are rarely used for clinical decision-making, although some studies have prospectively analyzed HFOs in iEEG data and evaluated whether crHFOsarea was associated with postoperative seizure freedom. 8,15,18,19he noninferiority randomized controlled "HFO trial," compared surgical guidance with HFOs and IEDs in the ioECoG to seizure outcome.This trial challenges using HFOs as biomarkers for epileptogenic tissue, particularly in temporal lobe epilepsy (TLE). 4meta-analysis identified 13 observational cohort studies and verified FRs as electrographic biomarkers of the EZ with moderate diagnostic accuracy. 20Next to answering the overall question whether HFOs are good predictors for surgical outcome, we have to consider that diverse clinical variables may reflect different results between various studies.These variables, which were not investigated in previous metaanalyses, include the study design, age group, diagnostic type, HFO identification method, whether an HFO rate threshold was used, iEEG state, sex, affected hemisphere, MRI findings, surgery location, and underlying pathology.We performed a 3-step systematic review and meta-analysis of this literature to (1) summarize the clinical characteristics and recording techniques, (2) assess the overall prognostic value of crHFOsarea for surgical outcome, and (3) determine the influence of clinical variables on this prognostic value.

Search Strategy
The systematic review and meta-analysis were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guideline (eAppendix 1).PubMed, Embase, and Web of Science were searched systematically (up to October 27, 2022) using the following terms as keywords: "high-frequency oscillation," "epilepsy," "surgery," "resect," and multiple variants of these terms (details in eMethods).A full-text review was performed when the screened abstract was considered relevant by Z. Wang and J. Guo.A consensus meeting was held in case of disagreement.We removed duplicates and searched reference lists of included articles to ensure more comprehensive literature retrieval.
Exclusion criteria were as follows: (1) articles reporting on high gamma frequencies and very HFOs (>600 Hz); (2)  unpublished data, nonhuman studies, conference abstracts, editorials, reviews, and studies in languages other than English; (3) studies focusing on nonresective surgery such as vagus nerve stimulation; and (4) HFOs detected by noninvasive techniques, such as scalp EEG and magnetoencephalography.

Data Extraction
Publications that fulfilled eligibility criteria would undergo data extraction by Z. Wang and J. Guo independently, including the following: (1) Publication details (publication year, full title, first author, country of origin, enrollment period, study design).(2) Clinical characteristics (age at surgery, sex, MRI findings, affected hemisphere, diagnostic type, HFO identification method, whether an HFO rate threshold was used, iEEG state (ictal or interictal), surgery location, underlying pathology, duration of postsurgical follow-up).(3) HFO type (FRs and ripples).We did not differentiate between physiologic and epileptic HFOs.(4) Classification of completeness of HFOs-area resection and/or calculation of the HFOs-area removal proportion per patient.Patients without HFOs detected were labeled as crHFOs-area because there are no HFOs remaining.We preferred to prioritize the analysis of HFOs detected post resection in patients with ioECoG subgroups if available; otherwise, we focused on HFOs recorded before resection.(5) Postsurgical seizure outcomes (Engel, ILAE, and seizure-free status).

Surgical Outcome Measure
Favorable surgical outcome (FSO) was defined as Engel class I a-d, ILAE class 1, or seizure-free status at ≥6 months after epilepsy surgery.Nonfavorable surgical outcome (non-FSO) was defined as Engel class II-IV, ILAE class 2-6, or seizure recurrence.
We divided the study patients of each study into 4 groups: (1)  patients with crHFOs-area and FSO; (2) patients with crHFOs-area and non-FSO; (3) patients without crHFOsarea and FSO; (4) patients without crHFOs-area and non-FSO.This was conducted separately for FRs and ripples.

Quality Assessment
For the quality assessment of nonrandomized studies, we used the Newcastle-Ottawa Scale (NOS). 21For RCTs, we used the Cochrane risk of bias tool to categorize the bias as high or low.To assess the risk of bias, Z. Wang and J. Guo conducted an independent assessment, and disagreements were settled by S. Hoogteijling.

Data Analysis
We synthesized study-level data and conducted whole-group analysis separately for complete resection of FRs-area (crFRsarea) and ripples-area (crRs-area) to evaluate surgical outcomes.The following data analyses were performed: (1)  we calculated the overall pooled proportion of patients attaining FSO after crHFOs-area, using a random-effects model-addressing the potential interstudy heterogeneitywhich integrated the Freeman-Tukey double-arcsine transformation to stabilize proportions and the Der Simonian-Laird technique to estimate between-study variance.This approach yielded a pooled proportion along with Clopper-Pearson 95% CIs.(2) We conducted a 2 × 2 contingency analysis using the Cochran-Mantel-Haenszel procedure with a random-effects model to compare the associations between FSO with and without crHFOs-area.Outcomes were presented as odds ratios (ORs) with 95% CIs based on the binomial distribution.(3) We used bivariate hierarchical modeling to evaluate the predictive performance of crHFOs-area.We plotted the summary receiver operating characteristic (SROC) curve and calculated the pooled sensitivity, specificity, and area under the curve (AUC).The AUC value was defined between 0.5 and 0.75 as a weak predictive performance and between 0.76 and 0.92 as a good predictive performance. 22,23xt, we extracted study-level data from the included studies and assessed the prognostic value for the following 12 subgroups categorized as follows: study design (prospective and retrospective), age group (children [≤20 years] and adults [>20 years]), diagnostic type (ioECoG and long-term iEEG [stereo-EEG or long-term ECoG]), HFO identification method (visual and automated), whether an HFO rate threshold was used (threshold and nonthreshold), iEEG state (interictal and ictal recording).The aforementioned subanalyses were performed separately for FRs and ripples.We classified the combined prognostic value for each subgroup as high if the following criteria were met: (1) pooled FSO proportion after crHFOs-area of ≥80%; (2) a significant association (lower bound of the 95% CI > 1) between crHFOs-area and FSO compared with without crHFOs-area; and (3) good predictive performance (AUC value >0.75) of the crHFOsarea.Meeting none of the 3 criteria was classified as a low combined prognostic value.
We compared surgery location (TLE vs extra-TLE) stratified by diagnostic type (ioECoG and long-term iEEG) at study level through dichotomous variable meta-analysis.Clinical variables of individual patients attaining FSO included sex (female, male), affected hemisphere (left, right), surgery location (TLE, extra-TLE), MRI findings (lesional, nonlesional), and underlying pathology.These variables were compared between crHFOs-area and without crHFOs-area patients for FRs and ripples separately (in total 10 groups), with categorical variables evaluated using Pearson χ 2 tests and continuity correction.We applied the Benjamini-Hochberg procedure to calculate false discovery rates (FDRs) to correct for the effect of multiple comparisons among the 10 groups.
Statistics were conducted using R studio version 4.1.0(meta, metafor, and meta4diag package) and Stata version 16.0.We calculated ORs and 95% CIs of all dichotomous variables for study-level data and considered p < 0.05 as statistically significant.Interstudy heterogeneity was assessed using Cochran Q and I 2 statistics, in which significant heterogeneity was defined at p < 0.10 and I 2 >50%.Individual patient analyses were assessed with a significance threshold of p < 0.05 after FDR correction.

Publication Bias and Sensitivity Analysis
The study effect and the extent of potential publication bias were assessed by visual inspection of funnel plots for the pooled proportion and dichotomous variables meta-analysis in the whole group.The Egger test was conducted as a quantitative assessment of the funnel plot symmetry.Sensitivity analysis was conducted through leave-one-out crossvalidation to evaluate the stability of dichotomous variable meta-analysis results.

Standard Protocol Approvals, Registrations, and Patient Consents
The study was registered in PROSPERO (CRD42022353768).Because this was a systematic review and meta-analysis of publications, ethical standards committee approval and patient informed consent were not applicable.

Data Availability
This article and its supplementary files encompass all the original and analyzed data during the course of this study.

Clinical Characteristics
Among the included 31 studies, 6 studies 27,[29][30][31][32][33] did not list individual patient characteristics.We requested data from corresponding authors or extracted relative data from figures and tables.Table 1 summarizes clinical characteristics and recording techniques of all included studies.
Eight studies (FRs in 140 patients; ripples in 77 patients) described crHFOs-area and surgical seizure outcome in children (median 10.5 years [range 0.5-20]) and 9 studies (FRs in 129 patients; ripples in 108 patients) in adults (median 30 years [range 21-79]).Thirteen patients from 2 studies had follow-up durations between 6 and 12 months, 28,32 with a mean duration of 8 months.Twenty-seven studies followed up patients for ≥12 months after surgery, and 10 of these followed up patients for ≥24 months.Two studies stated "adequate time" for follow-up, 9,27 rather than providing precise durations.The median follow-up time was 25 months (range 6-214 months) in the FR analysis group and 18 months (8-76 months) in the ripple analysis group.Seventeen (55%) of these studies reported postsurgical seizure outcomes using Engel classification; 8 (26%) studies used the seizure-free or not; and 6 (19%) studies used the ILAE classification of postsurgical seizure outcomes.On average, 64% of all included patients achieved postoperative favorable outcomes, with rates ranging from 5% to 84% per study.Notably, 1 study reported 1 seizure-free patient of 20 patients. 34

Comparison of crHFOs-Area With Without crHFOs-Area and FSOs
The likelihood of FSO was higher for crHFOs-area compared with without crHFOs-area (FRs OR 6.38, 95% CI 4.03-10.09,p < 0.001, Figure 2B; ripples OR 4.04, 95% CI 2.32-7.04,p < 0.001, Figure 3B).Pooling all results in the meta-analysis for FRs and ripples yielded a positive difference without any overlap with 1 on x-coordinate.The random-effects model yielded low heterogeneity for FRs and ripples with I 2 = 0%, p = 0.67 and I 2 = 0.4%, p = 0.55.

Subgroup Meta-Analyses Fast Ripples
The pooled FSO proportion in patients with crFRs-area was <80% in 5 predefined subgroups: prospective studies, adults, ioECoG, visual detection, and nonthreshold.The proportion was ≥80% in the other 7 subgroups (retrospective, children, long-term iEEG, automatic detection, threshold, ictal, and interictal).CrFRs-area showed no association with FSO for the ictal recording subgroup and significant   4).For details of the subgroup analysis for FRs, see eTable 1.  a Study #5: the author used a concept of "majority removal" of HFOs for this multiple-center study.We extracted data from PEP-UCLA only because they provided the data of CrHFOs-area.
b Study #13 and study #16: the authors used the definition of "HFOs" as activity >80 Hz, and ripples and FRs were not separately distinguished.

Ripples
The pooled FSO proportion in patients with crRs-area is <80% in 3 predefined subgroups: prospective, adults, and nonthreshold; the proportion is ≥80% in the other 9 subgroups.CrRs-area was not associated with FSO for 3 subgroups (prospective, IoECoG, and ictal recording) and significantly associated in the other 9 subgroups.The SROC curves showed weak predictive performance of crRs-area for FSO in 7 subgroups: prospective, adults, IoECoG, visual, automatic, nonthreshold, and interictal recording and good predictive performance in the other 5 subgroups (Figure 4).For details of the subgroup analysis for ripples, see eTable 2.   Meta-Analyses for Surgical Location in Patients With crHFOs-Area We summarized surgical location in patients with crFRs-area from 11 studies, including 64 patients with TLE and 124 patients with extra-TLE.FSO was achieved less often in TLE compared with extra-TLE for the whole group (OR 0.37; 95% CI 0.15-0.89,p = 0.006) and the ioECoG subgroup (OR 0.20; 95% CI 0.05-0.85,p = 0.01, Figure 5A).There was low interstudy heterogeneity for the whole group (I 2 = 27%, p = 0.18) and significant heterogeneity for the ioECoG subgroup (I 2 = 56%, p = 0.06).In the subgroup of long-term iEEG, 6 studies including 71 patients compared TLE with extra-TLE.There were no between-group differences (TLE vs extra-TLE; OR 0.66; 95% CI 0.21-2.06;p = 0.35).Low interstudy heterogeneity was found for the subgroup of long-term iEEG (I 2 = 0%, p = 0.64).

Variable Analyses With Individual Patient Data
Using individual patient data with FSO, clinical variables between crHFOs-area and without crHFOs-area groups were compared in 10 groups (Table 2).From available data of FR analysis group, 93 patients with TLE surgery and 107 patients with eTLE surgery achieved FSO; In patients with FSO, crFRs-area seemed to appear more often in the extra-TLE surgery group than in the TLE surgery group (p = 0.04).This result was not significant after correction for multiple comparison (p FDR adjusted = 0.19).One hundred fifty-six patients with lesional MRI findings attained FSO, and 12 patients with nonlesional MRI attained FSO.In patients with FSO, CrFRs-area was observed more frequently in lesional MRI group compared with nonlesional group (p = 0.002; p FDR adjusted = 0.02).We did not find such statistical difference in ripple analysis group (surgical location p = 0.14, p FDR adjusted = 0.36; MRI findings p = 0.16, p FDR adjusted = 0.36).
Quality of Assessment eTable 3 summarizes that the NOS scores ranged from 7 to 9 stars, and the mean total score was 8.4 stars, considered as good quality on average.The RCT study showed a low risk of bias in terms of selection, performance, detection, attrition, reporting, and others (eTable 4).

Bias Analysis and Sensitivity Analysis
Funnel plots of the dichotomous variables and pooled proportion meta-analysis in the whole group showed no asymmetry (Egger tests p > 0.05 in all groups); No group had obvious publication bias (eFigure 1 and 2).
Sensitivity analysis of the dichotomous variables yielded similar results in the whole group analysis (eFigure 3) and among 20 predefined subgroups (eFigure 4-9).In 4 additional subgroups (prospective ripples; ioECoG ripples; ictal FRs, and ripples), the results of the sensitivity analysis did not match those of the primary studies (eFigure 4, 6, and 9).

Discussion
Barriers to consider HFOs in iEEG recordings for actual surgical decision-making include ethical and practical issues, accompanied by incongruent studies with relatively small sample sizes.This comprehensive systematic review and meta-analysis of 31 publications from 18 tertiary epilepsy centers, amassing more than 700 patients makes several key contributions to assess the prognostic value of removing the whole brain area with HFOs for FSO in patients with epilepsy surgery.First, the complete resection of FR and ripple areas give invariantly similar pooled FSO proportions (81% for FRs; 82% for ripples).Second, we demonstrated that while complete resection of both FR and ripple areas significantly contributed to FSO, FRs performed as more accurate biomarkers of FSO than ripples.Third, our subanalyses for various clinical factors validated the high combined prognostic value of complete resection of FRs and/or ripples in 10 of 24 predefined subgroups; this value was absent for ripples in prospective study designs and ioECoG studies.Fourth, our study-level results suggest that in patients with complete FRsarea removal, ioECoG-guided tailoring proved superior to achieve FSO in extra-TLE compared with TLE.Fifth, the individual patient analysis showed that complete FRs-area removal occurred most often if MRI lesions were present.Finally, the robustness of our results was supported and explained through publications bias assessment and sensitivity analysis.
Controversy surrounds the question which type of HFOs is better to delineate the margins of the surgical resection.FRs have been suggested to be more pathologic and more strongly linked to epileptogenicity than ripples. 31An earlier metaanalysis conducted by Höller et al. 35 in 2015 showed a higher resection ratio of both types of HFOs in seizure-free patients compared with seizure-recurrent patients.This is in line with our findings.The SROC plots revealed lower but not statistically different predictive performance of ripples.The existence of physiologic ripples is the main flaw for ripples as a prognostic biomarker. 36Hussain et al. 18 showed in 60 patients that residual FRs independently induce a greater than 25-fold increased risk of seizures.FRs and ripples were analyzed separately in this study.The scattered and small sample size limited analysis for complete resection of simultaneous occurring FR and ripple areas at study level.The combination of FRs and ripples may be more enabling than either alone. 19,37e prognostic value of complete resection of HFOs-area differed among the 24 predefined subgroups.Most included studies were observational retrospective studies.Besides the "HFO trial," we included 7 "prospective studies" in which clinicians were blinded to the marked HFO results and postoperative seizure outcomes.We found a higher pooled FSO proportion for FRs and ripples in retrospective than prospective studies.None of these "prospective studies" used HFOs for surgical planning, which means that a causal link between HFO resection and postoperative outcomes could not be explored.We found that complete ripple area removal in prospective and ioECoG subgroups had low combined prognostic value.Conversely, complete HFOs-area (FRs and ripples) removal in retrospective and long-term iEEG subgroups showed high association and predictive performance for FSO, as previously reported. 8,38Future large-size prospective studies could clarify whether the low value observed in ripple analysis is genuine or influenced by the small cohort.In addition, direct comparisons of ioECoG, stereo-EEG, and long-term ECoG are needed.
The "HFO trial" did not show noninferiority of HFOs compared with IEDs for the whole group but showed noninferiority for the subgroup of extra-TLE after confounder correction. 4In this study, we explored the effect surgery location at study level and patient level.The study-level analysis showed that patients with extra-TLE surgery who underwent complete resection of the area showing FRs were more likely to achieve FSO compared with those who underwent TLE surgery, for the whole group and the ioECoG subgroup, but not for the long-term iEEG subgroup.We extracted patients with complete FRs-area removal from the "HFO trial" (29 of Abbreviations: CrHFOs-area = complete resection of HFOs-area; CrFRs-area = complete resection of fast ripple area; CrRs-area = complete resection of ripple area; FCD = focal cortical dysplasia; FDR = false discovery rate; FR = fast ripple; FSO = favorable surgical outcome; HFO = high-frequency oscillation; TLE = temporal lobe epilepsy.a p < 0.05 was considered significant.b Continuity correction was used for χ 2 tests because the minimum expected count is 2.43. 39 patients), 4 which showed an odds ratio opposite to the other 4 ioECoG studies 8 (OR 1.73; Figure 5), yielding significant heterogeneity. 4Our patient-level analysis indicated that complete FRs-area removal in patients with FSO seemed more frequent in the extra-TLE group than in the TLE group; this trend did not survive FDR correction.Multiple comparison correction can be conservative in small sample sizes, making it challenging to detect significant differences.The current findings supporting the use of HFOs for extra-TLE patients seems relatively weak and requires further studies that include direct comparisons and large sample sizes.
We observed a high combined prognostic value of complete HFOs-area removal for FSO in children vs adults.
Commonly, children with complete removal of MRI lesions are more prone to achieve FSO than adults. 39This phenomenon can be partly explained by a typically larger resected volume in children than in adults, increasing the chance on good outcome.HFOs predicted surgical outcomes better in the group using ioECoG recordings in children than in the long-term iEEG recordings in adults. 40The epileptogenic tissue is often close to the cortical surface in the pediatric population, which enables detecting HFOs-area with grids/strips and even scalp EEG. 41Age-related changes seem to influence HFOs, which needs a next step to assess across childhood and adolescence. 42O rate thresholding is vital when presenting HFO results.
Our subanalyses identified a high combined prognostic value of complete resection of FR and ripple areas in HFO rate threshold subgroup, only if an HFO rate threshold was applied.Jiang et al. 43 and Li et al. 44 set up a quantitative threshold (HFO rate >1 event per minute).Thresholding can also be established using statistical guideline, including 95% HFO rate distributions, 19,37 statistical dispersion (e.g., Tukey Fence), 10,15 and standard deviation (e.g., Kittler method). 45,46niform methods are needed to define the threshold for HFOs algorithms to enable direct statistical comparisons across studies.
The EZ remains theoretical due to uncertainties; seizure freedom after surgery indicates the entire EZ is removed, but does not reflect its size.MRI lesions enable an improvement in preoperative localization of the epileptogenic focus. 47The absence of MRI-visible lesions has been identified as a key predictor of surgical failure.These cases require a more comprehensive presurgical examination including intracranial monitoring. 489][50] This result was limited by the small nonlesional MRI sample size.
Our study has limitations.The foremost methodologic weakness is the varying definition of complete resection across studies, which should be considered when interpreting the results because it may signify incomplete HFOs-area removal in certain cases.Second, FRs and ripples were mixed in several included publications.There were 100 patients from 10 publications who underwent complete resection of combined FR and ripple areas.The repeat inclusion might have led to an indistinguishable effect, increasing the weight of these studies into both FR and ripple analyses.Third, iEEG HFO recording and analysis methodology is not standardized and the inter-reviewer reliability is often inadequate.e4 HFO characteristics, including duration, amplitude, and peak frequency vary, which means that patients categorized as having complete HFOs-area removal in 1 center may have residual HFOs signal in another center.
There should be a push for standardization of HFOsrecording and analysis to improve study outcome comparison.Fourth, we must consider physiologic and epileptic HFO overlap in frequency.Physiologic HFOs are not limited to mesial temporal structures and occur in nonepileptic functionally eloquent regions such as the occipital lobes and perirolandic region.e3,e4 Our statistical findings should not be mistaken for the implication that resection of every individual electrode site exhibiting HFO is necessary or sufficient to achieve FSO; An all-purpose interpretation for the prognostic value of HFOs may not exist, and further studies are required to investigate the value of this biomarker in a case-by-case fashion.By searching ClinicalTrials.gov,we discovered 2 ongoing interventional clinical trials: 1 double-blinded RCT assessing whether HFO-generating tailored surgery can lead to better seizure outcomes (NCT03790280; n = 30); another small interventional clinical trial focusing on the functional utility of stereotyped HFOs with the Brain Interchange implantable system (NCT05439655; n = 20).Future multi-institutional collaborations collating individual patient data are required to investigate factors that could not be assessed in our analysis, including specific surgery location and more pathologic categories.

Figure 2
Figure 2 Whole-Group Meta-Analysis Results of the Prognostic Value of crFRs-Area for FSO

Figure 3
Figure 3 Whole-Group Meta-Analysis Results of the Prognostic Value of crRs-Area for FSO

Figure 4
Figure 4 Subgroup Meta-Analysis Results of the Combined Prognostic Value of crHFOs-Area for FSO in Predefined Subgroups

Figure 5
Figure 5 Meta-Analysis for Surgical Location in Patients With crHFOs-Area

Table 1
Clinical Characteristics and Recording Techniques of the 31 Articles Included in the Meta-Analysis Continuedassociation for the other 11 subgroups.The SROC curves showed weak predictive performance of crFRs-area for FSO in 3 subgroups: ioECoG, adults; no HFO-rate thresholding and good predictive performance in the other 9 subgroups (Figure

Table 1
Clinical Characteristics and Recording Techniques of the 31 Articles Included in the Meta-Analysis (continued)

Table 2
Clinical Characteristics of Individual Patients Attaining Favorable Surgical Outcome, Stratified by Sex, Affected Hemisphere, Surgery Locations, MRI Findings, and Pathologies Funding Z.Wangand J. Guo were supported by the China Scholarship Council, grant number 202106170027 and 202106170024.M. Zijlmans, M. van 't Klooster, and S. Hoogteijling were supported by the ERC grant number 803880.M. van 't Klooster was supported by the Health Holland LSHM19080.