Intraoperative Neurophysiological Monitoring in Brain Tumor Surgery: Prospective Cohort Studies

BACKGROUND Brain tumor surgery requires balancing two competing priorities: maximizing the extent of tumor resection (EOR) while preserving neurological function to reach maximal safe resection. This balance is particularly challenging in eloquent brain areas, where even minor injuries may result in permanent loss of function such as muscle paralysis, language impairment, or visual loss. Intraoperative neurophysiological monitoring (IOM) combines functional mapping to identify eloquent structures with continuous monitoring of their integrity during tumor resection. IOM with motor evoked potentials (MEP) is widely used during surgery in general anesthesia or asleep-awake-asleep setting to localize and monitor the precentral gyrus (motor cortex) and corticospinal tract (CST). Continuous stimulation of the precentral gyrus monitors CST integrity throughout resection, while subcortical motor mapping within the resection cavity estimates the distance to the CST and guides the extent of safe resection. Significant warning criteria include a \>50% reduction in MEP amplitude or a 20% increase in stimulation threshold, while lower subcortical stimulation thresholds indicate increasing proximity to the CST (approximately 1 mA=1 mm). A subcortical resection limit of 3 mA is usual although lower thresholds can be safe. Combined with diffusion tensor imaging (DTI), MEP monitoring has retrospectively been associated with greater extent of resection and improved long-term survival. Although widely adopted, high-quality evidence demonstrating patient-related benefit remains lacking. IOM with CCEP is promising for language-eloquent brain tumor surgery and has been studied in general anesthesia and asleep-awake-asleep settings. Direct cortical stimulation of one language area (Broca's or Wernicke's area) with recording from the other enables monitoring of subcortical language tract function, including the arcuate fasciculus (AF) and superior longitudinal fasciculus (SLF). During surgery in asleep-awake-asleep setting, CCEP monitoring complements cortical and subcortical language mapping which temporarily disrupts the patients' speech ability, looking for speech arrest and phonological paraphasia which is linked to AF function disruption. CCEP waveforms comprise positive (P1, P2) and negative (N1, N2) deflections. A previous study by Seidel et al. (co-applicant) demonstrated an association between N2 latency changes during awake surgery and postoperative language impairment. However, it remains unclear whether amplitudes, latencies, or both provide the most clinically relevant information, and validated warning thresholds have yet to be established. Consequently, IOM with CCEP lacks standardization and further clinical validation. IOM with VEP can be used to monitor visual pathway function during visual-eloquent surgery. The visual pathways are normally divided into the anterior (structures from the retina to the lateral geniculate body) and posterior visual pathways (from the lateral geniculate body to the primary visual cortex). Flash stimulation delivered through light-emitting diode eye pads evoke cortical responses recorded using scalp electroencephalography (EEG), and when applicable, strip electrodes placed over the visual cortex to improve signal quality. Simultaneous electroretinography confirms adequate retinal light stimulation. Previous studies have demonstrated promising results for preserving anterior visual pathway function during prechiasmal and sellar/suprasellar tumor surgery, including endoscopic transsphenoidal surgery. However, evidence for monitoring posterior visual pathway function remains limited. Furthermore, considerable intra- and interindividual variability and signal instability reduces signal reliability and limits clinical applicability to patients with sufficient baseline visual function. VEP waveforms comprise an initial deflection called N1 and a prominent deflection called P100 (also known as P2). Focus has been on the voltage difference between N1 and P2 (P100) on one hand, and P100 amplitude on the other hand. Although a 50% reduction in the N1-P2 peak-to-peak amplitude has been proposed as a warning criterion, clinically relevant VEP signal changes and validated warning thresholds remain uncertain. Consequently, IOM with VEP lacks standardization and further clinical validation and is therefore not widely adopted. PURPOSE The overall aim is to strengthen the evidence base supporting the standardization and clinical application of IOM in eloquent brain tumor surgery. The project comprises four unprecedented studies (Parts 1-4) with the following objectives: 1. To evaluate current clinical practice and evidence gaps in cortico-cortical evoked potential (CCEP) monitoring. 2. To evaluate the clinical significance of CCEP monitoring by assessing the association between intraoperative signal changes and postoperative language outcomes, and to explore clinically relevant warning thresholds. 3. To evaluate the clinical effectiveness of IOM with motor evoked potentials (MEP) compared with surgery without IOM. 4. To evaluate the clinical significance of VEP monitoring by assessing the association between intraoperative signal changes and postoperative visual outcomes, and to explore clinically relevant warning thresholds. HYPOTHESIS 1. CCEP monitoring lacks standardization and clinical validation, and its use varies considerably across clinical practice. 2. Intraoperative CCEP signal changes are associated with postoperative language impairment and can be used to establish clinically relevant warning thresholds. 3. IOM with MEP improves motor function preservation and/or increases the extent of safe tumor resection compared with surgery without IOM. 4. Intraoperative VEP signal changes are associated with postoperative visual impairment and can be used to establish clinically relevant warning thresholds. METHODS Patients scheduled for brain tumor surgery at Odense University Hospital will be screened for eligibility according to the inclusion criteria for each clinical study (Parts 2-4) and recruited during routine outpatient assessment. Part 1: Intraoperative neurophysiological monitoring with cortico-cortical evoked potentials (CCEP) during language-eloquent intracranial surgery: A scoping review Aim: To systematically review the literature on intraoperative CCEP monitoring in brain surgery to identify current clinical practices, methodological variability, and remaining knowledge gaps that may guide future research and clinical implementation. Methods: A scoping review will be conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines. PubMed will be systematically searched using terms including "cortico-cortical evoked potential", "CCEP", and "intraoperative". Eligible studies will be reviewed to evaluate: 1) Clinical applications of intraoperative CCEP monitoring; 2) Proposed warning thresholds; 3) Use during surgery in general anesthesia; 4) Integration with preoperative imaging; 5) Pre- and postoperative language assessment; 6) Patient outcome assessment methods. Part 2: Intraoperative neurophysiological monitoring with cortico-cortical evoked potentials (CCEP) during surgical resections of language-eloquent brain tumors: A prospective observational cohort study Aim: To prospectively evaluate the association between intraoperative CCEP signal changes and postoperative language impairment in patients undergoing language-eloquent brain tumor surgery in either general anesthesia or asleep-awake-asleep setting, and to explore clinically relevant warning thresholds that support safe tumor resection while preserving language function. Methods: Preoperative contrast-enhanced MRI combined with DTI will be used to determine tumor location relative to language-eloquent areas. Baseline language function will be assessed using validated language tests (e.g. the DuLIP language test battery and the Western Aphasia Battery). Patients scheduled for surgery in asleep-awake-asleep setting will additionally undergo neuropsychological assessment and training for intraoperative language testing. During surgery in general anesthesia, Broca's and Wernicke's areas will be identified intraoperative using preoperative imaging and anatomical landmarks. Strip electrodes will be placed over the presumed Broca's and Wernicke's areas, with stimulation of Broca's area and recording from Wernicke's area (BtW-asleep). CCEP recordings will be obtained using two adjacent electrodes, two repeated trials of 30-60 averaged responses, and continuous bipolar and referential monitoring. A valid CCEP will be defined by the presence of distinct P1 and N1 waveforms. Amplitudes and peak latencies of P1, N1, P2 and N2 will be monitored throughout surgery. During surgery in asleep-awake-asleep setting, cortical mapping with bipolar Penfield language mapping will additionally be performed in the awake phase to identify a safe cortical entry site. Stimulation will be increased from 2 mA in 1 mA increments until speech arrest is elicited without afterdischarges on electrocorticography, after which the same stimulation intensity (typically 3-6 mA) will be used for further cortical language mapping. CCEP monitoring (BtW-awake) will be initiated immediately after cortical Penfield language mapping, preferably before cortical incision, and continued throughout the subsequent asleep phase as described above. Subcortical language mapping will be performed to guide tumor resection in the awake phase using monopolar probe stimulation within the resection cavity to localize language eloquent structures and estimate their distance from the resection margin. The stimulation intensity identified during cortical language mapping will be used, and speech arrest or phonological paraphasia will guide cessation of tumor resection at the corresponding site. Statistical analyses for this observational study will compare the difference in mean value of CCEP amplitude, latency and stimulation threshold changes from baseline between patients with and without postoperative language impairment/worsening of language impairment at 1- and 3-month follow-up. Between-group comparisons will be performed using two-sample t-test (two-sided) or Mann-Whitney U test, depending on data distribution. Adjusted logistic regression models will be fitted including EOR and histopathologic diagnosis, and Receiver Operating Characteristic and Area Under the Curve (ROC/AUC) and diagnostic accuracy measures will be computed. Clinically relevant warning thresholds will be explored as the greatest amplitude reduction, latency prolongation, or stimulation threshold increase not associated with significant postoperative language impairment. Sample Size calculations are based on Welch's two-sample t-test corresponding to the comparison of mean N2 latency during resection between patients with and without aphasia. Assumed effect sizes are based on observations from Seidel et al., which recorded N2 latency to be 64.0 ± 12.6ms (mean ±SD) in patients with postoperative aphasia and 48.8 ± 14.2ms in patients without aphasia. The study included 19 patients out of which 7 (i.e. 36.8%) had postoperative aphasia. The observational nature of the grouping variable causes random group sizes and corresponding additional variation is considered for sample size calculations using the approach of Ambrosius et al. The latter considers both the uncertainty in the population group ratio (i.e. the 36.8%) as well as the future realized group allocation. For the parameters θ1and θ2 of the Beta prior we have chosen the values 7.5 and 12.5 which corresponds to the posterior parameters obtained after updating a Jeffrey's prior with data from the study (i.e. 7 vs. 12 patients with/without aphasia). The resulting required sample size corresponding to a significance level of 0.05 and an expected power of 0.8 is N=33 patients. Other clinically relevant secondary outcomes are not used to determine sample size and will be considered explorative or descriptive. Part 3: Intraoperative neurophysiological monitoring (IOM) with motor evoked potentials (MEP) versus no IOM during surgical resections motor-eloquent brain tumors: A randomized controlled trial Aim: To prospectively evaluate the effect of IOM with MEP on postoperative motor function and extent of tumor resection (EOR) compared with surgery without IOM in patients undergoing motor-eloquent brain tumor surgery in general anesthesia setting. Methods: Patients undergoing resection of motor-eloquent brain tumors will be randomized in a 1:1 ratio assigned to surgery with or without IOM (IOM or no IOM). Allocation will be concealed using a computer-generated block randomization sequence with variable block sizes. Patients at highest risk of function loss (direct tumor involvement of motor-eloquent structures) will be excluded and treated using IOM. Preoperative contrast-enhanced MRI combined with DTI and functional MRI will be used to determine tumor location relative to motor-eloquent structures1. Motor function will be assessed using the Medical Research Council (MRC) scale, National Institutes of Health Stroke Scale (NIHSS) motor subscore, Six-Spot-Step Test (SSST), 9-Hole Peg Test (NHPT), hand dynamometry, and 10-Meter Walk Test (10MWT). Overall functional status will be assessed using the Karnofsky Performance Scale (KPS). During surgery without IOM, the precentral gyrus, CST, and a safe incision zone will be identified intraoperative using preoperative imaging studies, anatomical landmarks, and navigated intraoperative ultrasonography in all patients. During surgery with IOM, cortical mapping will additionally be performed using monopolar probe stimulation. Durin surgery with IOM, MEP monitoring of CST integrity will be with direct cortical stimulation and transcranial stimulation with scalp electrodes. A \>50% reduction in MEP amplitude or a 20% increase in stimulus threshold will be considered significant warning criteria and guide cessation of tumor resection at the specific site. Subcortical motor mapping of the CST will be performed using monopolar probe stimulation within the resection cavity, initially at 10-15 mA and gradually reduced as resection approaches the CST. An MEP threshold of 3 mA will guide cessation of tumor resection at the specific site, subject to the surgeon's assessment (critical threshold: 1 mA). Statistical analyses will compare the primary outcome (MRC sum score) between treatment groups (IOM vs. no IOM) using independent two-sample t-tests or Mann-Whitney U tests, as appropriate. Treatment effects will be reported as mean differences with 95% confidence intervals and standardized effect sizes. Adjusted analyses will be performed using linear regression (ANCOVA) including baseline MRC score, age, tumor volume, distance to the corticospinal tract, and histopathological diagnosis as prespecified covariates. Key secondary outcomes are not individually powered and will be analysed using the same two-stage approach of unadjusted group comparisons followed by adjusted linear regression models and interpreted as supportive evidence. Longitudinal primary and key secondary outcomes will additionally be analysed using linear mixed-effects models including fixed effects for treatment, time, and treatment-by-time interaction, and a random intercept for participant. Ordinal exploratory secondary outcomes (Engel seizure classification, PGIC, and Surgeon-rated usefulness) will be analysed using ordinal logistic regression or rank-based tests. Binary exploratory outcomes (completeness of fluorescein-guided tumor resection and tumor progression) will be analysed using chi-square or Fisher´s exact tests, with generalized linear models used to estimate relative risks with 95% confidence intervals. Time to tumor progression will additionally be explored using Kaplan-Meier curves and Cox proportional hazards regression. Safety endpoints will be summarized descriptively by treatment group. All tests will be two-sided with a significance level of α=0.05 for the primary outcome. Secondary and exploratory outcomes will be interpreted primarily using effect estimates and 95% confidence intervals. Sample size calculations indicate that n=66 participants (33 per group) are required to detect a clinically meaningful difference in postoperative motor function assuming a standard deviation of 1.0 on the MRC score and a conservative literature-supported standardized mean difference of Cohen's d=0.704 for continuous motor outcome, with a two-sided significance level of α=0.05 and 80% power. As this kind of randomized controlled trial is unprecedented, these assumptions are intentionally conservative. Secondary outcomes and safety endpoints are not included in the sample size calculation and will be analysed as supportive, exploratory, or descriptive outcomes. Part 4: Intraoperative signal changes of visual evoked potentials (VEP) during transsphenoidal and transcranial tumor resections: Clinical correlation and warning thresholds Aim: To prospectively evaluate the association between intraoperative VEP signal changes and postoperative visual impairment in patients undergoing visual-eloquent transsphenoidal pituitary tumor or transcranial brain tumor surgery in general anesthesia setting, and to explore clinically relevant warning thresholds that support safe tumor resection while preserving visual function. Methods: Preoperative contrast-enhanced MRI combined with DTI will be used to determine tumor location relative to the visual pathways. Visual function will be assessed by best-corrected visual acuity and perimetry. Intraoperative binocular flash stimulation will be delivered using light-emitting diode eye pads placed over the closed eyelids and secured with a protective seal. Flash visual evoked potentials will be recorded using low amplitude scalp EEG over the occipital lobes and simultaneous electroretinography to confirm adequate retinal light stimulation. During transcranial surgery, a strip electrode will additionally be placed over the occipital cortex when permitted by the craniotomy, locating the visual cortex with the use of preoperative imaging and adjusting electrode placement to optimize VEP recordings. Stimulation intensity will be 7,000-8,000 lx with alternating light-on and light-off stimulation. VEP monitoring will be performed by continuous flash stimulation with monitoring of N1 and P2 (P100) amplitudes and latencies using low amplitude scalp EEG, and strip electrodes when applicable. Statistical analyses for this observational study will compare the difference in mean value of VEP amplitude, N1-P2 peak-to-peak amplitude, and VEP latency changes from baseline between patients with and without postoperative visual impairment/worsening of visual impairment at 1-2 weeks and 3 months. Between-group comparisons will be performed using two-sample t-test (two-sided) or Mann-Whitney U test, depending on data distribution. Adjusted logistic regression models will be fitted including EOR, and ROC/AUC and diagnostic accuracy measures will be computed. Clinically relevant warning thresholds will be explored as the greatest VEP or N1-P2 peak-to-peak amplitude reduction or VEP latency prolongation not associated with significant postoperative visual impairment. Sample Size calculations indicate that n=64 participants are required to detect the expected difference in the incidence of VEP signal changes associated with postoperative visual impairment compared with those not associated with visual impairment (9.4% vs. 56.3%, two-sample proportion test, two-sided α=0.05, power 0.8). EXPECTED OUTCOME AND IMPACT Brain tumors are associated with substantial morbidity, permanent loss of function, and mortality. Even modest improvements in functional preservation and the extent of safe tumor resection may significantly improve survival, independence, quality of life, and return to daily activities. The project is expected to generate unprecedented clinical evidence on IOM in eloquent brain tumor surgery through the first randomized controlled trial evaluating the patient-related benefits of IOM with MEP, and the first prospective studies on CCEP monitoring in awake and general anesthesia setting and VEP monitoring of the entire visual pathway. By addressing major evidence gaps, the project will strengthen the evidence base for IOM, improve interpretation of IOM signals and surgical decision-making, and support rapid implementation into clinical practice locally as well as impacting the use of IOM world wide. Ultimately, the project is expected to enable safer and more effective brain tumor surgery with a lower risk of permanent motor, language, and visual impairment, improved survival and quality of life, and reduced need for rehabilitation and supportive care, thereby improving patient safety and enabling more precise and effective surgical treatment. Furthermore, the project is related to the 3rd United Nations sustainable development goal: "Ensure healthy lives and promote well-being for all at all ages." FEASIBILITY AND COLLABORATIONS A multidisciplinary collaboration is well established between the Departments of Neurosurgery and Clinical Neurophysiology at Odense University Hospital (OUH) and Professor Kathleen Seidel who is an internationally recognized expert in IOM. Our department has extensive experience with IOM, access to a large and relevant patient population, and a strong research infrastructure, facilitating both project completion and early clinical implementation. Parts 2 and 4 are prospective observational studies, enabling early data acquisition and statistical model development. Based on our patient volume, we expect to recruit 25 patients/year for Part 2 (n = 33), 20-25 patients/year for Part 3 (n = 66), and 48 patients/year for Part 4 (n = 64). The sample size for Part 3 is based on conservative literature-supported assumptions, and a larger treatment effect may reduce the required sample size. Overall, the project is considered highly feasible within the planned three-year timeframe. DATA MANAGEMENT, SHARING AND DISSEMINATION A Research Electronic Data Capture database will be used for safe data storage, OPEN Analyze for statistical data storage, and Stata IC for analysis. All systems are hosted by OPEN on secure servers in the Region of Southern Denmark (RSYD). The project will be registered in RSYD´s record of Data Processing Activities and conducted in accordance with the EU General Data Protection Regulation (GDPR), the Danish Data Protection Act, the Danish Health Act, and Good Clinical Practice standard. All project data will be made available through Zenodo open data repository. Results will be published in international peer-reviewed Open Access journals and presented at national and international conferences. Lay summaries will be disseminated through patient organizations, including the Danish Brain Tumor Association, and presented to ensure broad accessibility for both healthcare professionals and the public. ROLES Dr. Grønhøj (Principal Applicant): Overall responsibility for project management, administration, and coordination. Performs surgeries, oversees patient recruitment and follow-up, and supervises the PhD student in funding acquisition, data curation, statistical analyses, patient and public involvement, and manuscript preparation. Björnsson, MD (PhD student): Responsible for protocol development, funding acquisition, project coordination, patient recruitment and follow-up, data curation, statistical analyses, scientific and public dissemination, and manuscript preparation, all under supervision. Dr. Binzer (Co-applicant): Performs intraoperative monitoring and supervises patient recruitment, funding acquisition, data curation, statistical analyses, and manuscript preparation. Prof. Seidel (Co-applicant): Contributes to data interpretation, scientific discussion, and manuscript preparation, and supervises the PhD student during a two-week research stay in Bern. PERMITS The project proposal has been approved as the basis for PhD enrolment at the Graduate School of Health Sciences, SDU, and Björnsson has been accepted as the PhD candidate. Ethical approval has been submitted to the Danish National Committee on Health Research Ethics and is currently under review. ABBREVIATIONS EOR: extent of tumor resection IOM: Intraoperative neurophysiological monitoring MEP: motor evoked potentials CST: corticospinal tract DTI: diffuse tension imaging CCEP: cortico-cortical evoked potentials VEP: visual evoked potentials AF: arcuate fasciculus SLF: superior longitudinal fasciculus EEG: electroencephalography PRISMA-ScR: Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews mA: milliampere MRI: magnetic resonance imaging DuLIP: the Dutch Linguistic Intraoperative Protocol BtW: Broca to Wernicke μV: microvolts ms: milliseconds ROC/AUC: Receiver Operating Characteristic and Area Under the Curve SD: standard deviation MRC: Medical Research Council NIHSS: the National Institutes of Health Stroke Scale ANCOVA: Analysis of Covariance lx: Lux OUH: Odense University Hospital RSYD: Region of Southern Denmark MD: Medical doctor