Multidimensional preoperative cognitive phenotypes predict postoperative cognition in glioblastoma

Multidimensional preoperative cognitive phenotypes predict postoperative cognition in glioblastoma | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Multidimensional preoperative cognitive phenotypes predict postoperative cognition in glioblastoma Yizhou Wan, Ajay Halai, Tom Manly, Haiyan Zheng, Roxanne Mayrand, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9163008/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose Glioblastoma patients suffer from cognitive deficits across domains. We used principal component analysis and gaussian mixture modelling to identify preoperative cognitive phenotypes which are at risk of postoperative cognitive impairment. Methods 44 patients underwent neuropsychological screening before and after surgery (within 1 week and 4–6 weeks post-surgery). Patient scores were standardised to a normative reference population and compared using one-sided z-scores. Reliable change indices were calculated with correction for hospitalization effects using a cohort of biopsy patients. Gaussian mixture models were fit after principal component analysis of standardised preoperative scores. Linear mixed effects models were performed to assess predictors of postoperative principal component scores using demographic, clinical and imaging variables. Results Three principal components accounted for 55% of the variance in cognitive scores: The first component loaded on Language, Executive Function and Perception (25%). The second, on Prospective and Retrospective Memory (15%). The third component on Recognition Memory and Working Memory (15%). The optimal model identified four preoperative cognitive phenotypes: Minimal (n = 23), Mild (n = 12), Moderate (n = 5) and Severe (n = 4) cognition. Post-surgery, preoperative cognition limited recovery and few patients returned to their preoperative baseline. Moderate and Severe phenotypes were significantly associated with worse postoperative Language, Executive Function as well as Prospective and Retrospective Memory. Conclusion We identified four preoperative cognitive phenotypes which stratify patients into those at risk of postoperative cognitive deficit. Better preoperative cognition was associated with improved postoperative cognition. Cognition Glioma Neuropsychological screening Surgery Figures Figure 1 Figure 2 Figure 3 Figure 4 INTRODUCTION Gliomas are the most common primary brain cancer affecting adults. 1 , 2 Of these, glioblastoma (WHO grade 4) with a median prognosis of only 15 months, is the most common. 3 , 4 Given the poor prognosis, there is increasing recognition that maintenance of cognition during treatment is a crucial goal for patients. 5 – 7 Cognition is important for brain tumour patients. 8 – 10 Motor function, processing speed, executive function, and memory have been linked with functional well-being and role function. 9 – 11 Previous research has found heterogenous patterns of cognitive outcomes in brain tumour patients when tested at different timepoints. 12 , 13 Performance variability may be because of patient, tumour and treatment factors and Study methodology. Comparison of results between studies is difficult due to the variation across neuropsychological tests. Cognitive deficits can be considered a latent system where underlying functional disruption is observed in cognitive test scores. 14 , 15 The dimensionality of the latent system is lower that the number of observed cognitive tests because: 1) test scores are intercorrelated and 2) patient have deficits in similar functional regions, constrained by neuroanatomical organisation and lesion location. 16 Tests vary considerably in task difficulty and requirements. 13 , 17 Easier tests avoids floor effects in severely impaired patients while harder tests are more sensitive to milder deficits. 18 In addition, each task may depend on multiple underlying cognitive systems. 15 , 19 It is important to understand cognitive variability prior surgery to stratify patients into risk-profiles which can guide clinical decision making. For example, different surgical resection strategies may result in different postoperative outcomes in terms of survival and cognition. If cognition is primarily determined by patient or tumour effects, understanding perioperative cognition may help identify which patients can benefit from aggressive surgical strategies to delay tumour progression. 20 Emerging evidence suggests that postoperative cognitive rehabilitation for glioma patients is associated with improved visual attention, memory and executive function. 21 In addition, risk stratification may identify patients who benefit from early clinical trial involvement in the context of window-of-opportunity trials prior to surgery. 22 In this study we used Principal Component Analysis (PCA) and Gaussian Mixture Modelling (GMM) to investigate the underlying structure of cognitive deficits in Glioma patients undergoing surgery. Both dimensionality reduction techniques seek to identify the underlying latent profiles and to explain the observed cognitive test scores. This allows us to stratify patients by baseline risk of postoperative cognitive deterioration. We also examined which clinical factors affected preoperative and postoperative cognition, at the early (within 1-week) and late (within 4–6 weeks) postoperative timepoints. METHODS Participants 44 patients (age > 17) were recruited as part of the Surgically Induced Neurological Deficits in Glioblastomas) (SIND) Study (REC: 19/WM/0152). Written informed consent was obtained from all patients. The inclusion criteria were: 1) assessed by Tumour Board to have high-grade-glioma on imaging, WHO Performance Status (PS) 0–2, suitable for surgery, resection patients were those where the treating neurosurgeon feels that > 90% of the enhancing tumour can be resected. Patients with tumours located near/in speech/sensorimotor regions will undergo speech/language mapping +/- motor mapping at the treatment surgeon’s discretion. Neuropsychological Screen (NPS) was performed at 1) t0- baseline (prior to surgery), 2) t1- early postoperative (72-h postoperative), and t2- delayed postoperative (6-weeks post-surgery). Patients who are unsuitable for a contrast-enhanced MRI were excluded. All patients had WHO 2021 histopathology confirmed glioblastoma or gliosarcoma. Image-guided biopsies with minimal brain disruption were performed in the Biopsy control group. Resection patients underwent 5-aminolevulinic acid (5-ALA) guided-surgery with neuronavigation (StealthStation, Medtronic), and other adjuvants (e.g., cortical and subcortical mapping) for maximal safe resection of fluorescent tumour. Neuropsychological screening tool A Neuropsychological Screen (NPS), validated in glioma patients were administered with a touch-screen tablet as part of research studies by two trained Researchers with Neuropsychology and Neurosurgery qualifications (Y.W and R.M). 23 The Oxford Cognitive Screen and Cambridge Attention, Memory, and Perception Screen (OCS-Bridge) consisted of tests of language, orientation, attention, perception, memory, praxis, and numeracy skills. 23 , 24 Parallel test versions were used to minimise practice effects. Total test time was 25–30 minutes (Supplementary Information, Table 1 ). Imaging analysis MRIs were acquired with a 3-Tesla scanner (Siemens Healthcare, USA) at the Wolfson Brain Imaging Centre, University of Cambridge. (Supplementary Information, Methods). Semi-automated methods were used to segment the whole tumour, FLAIR abnormality and resection cavity on preoperative images to calculate lesion volumes from these regions. 25 Statistical analysis Clinical characteristics were summarised using medians (IQR), and percentages. Wilcoxon rank sum, Fisher Exact and Chi-squared tests were used to compare groups. Multiple-comparison correction was performed using false-discovery rate ( q < 0.05). OCS-BRIDGE values were z-scored using the mean and standard deviation (SD) of reference scores from 268 healthy participants aged 16 to 89 (mean age, 51.44, SD 19.86). 23 Performance ≥ 2 SD below the mean was considered impaired. Individual level changes were calculated using a standardized regression-based reliable change index (RCI), based on mean and SD of performance in biopsy group (n = 6). 26 RCI values exceeding ± 1.645 (corresponding with a two-tailed alpha of 0.10%, 90% confidence interval) indicated meaningful change. The percentage of patients with improved, stable, and declined performance were calculated per domain. 27 , 28 Principal component analysis Tests with constant scores, formed linear combinations with other scores, and a Kaiser, Meyer, Olkin Measure of Sampling Adequacy (KMO) ≤ 0.3 were iteratively removed, retaining overall KMO > 0.6. 29 PCA was performed using standardised baseline NPS scores across all patients (both resection and biopsy). Horn's parallel analysis (5000 iterations) was used to determine the optimal number of components. 30 A varimax rotation was performed to improve loading interpretability. The rotation matrix was multiplied by the standardised scores at other timepoints to calculate postoperative rotated component (RC) scores. Gaussian mixture modelling To identify cognitive phenotypes across time, baseline and postoperative RC scores for the combined cohort (resection and biopsy) were entered into Gaussian Mixture Models (GMMs), using finite mixture models via the Expectation-Maximization (EM) algorithm to identify latent patient groups. The optimal number of clusters was selected using by iteratively computing the BIC across models with differing number of components and covariance parameters with bootstrap likelihood ratio tests (5000 samples). 31 The model with the minimum BIC was selected and local fit parameters checked with the mean cluster assignment posterior probability. Cluster phenotype identities across time were aligned by matching the position of each cluster’s centroid along the principal PCA axis of variation. (Supplementary Information Fig. 7). Linear mixed effects models assessed which predictors were associated with postoperative rotated component (RC) scores. The Order-Norm transformation was used for right skewed postoperative PCA scores. 32 The preoperative cognitive phenotype was evaluated against baseline RC scores to determine if it provided predictive value. For each postoperative RC score, four models were fitted: 1) null (timepoint only), 2) time and baseline RC scores, 3) time and preoperative phenotype and 4) time, preoperative phenotype and RC scores. Fixed effects were added to the time and preoperative phenotype model to assess which clinical variables were significant predictors. Forward selection was used to retain variables which were univariate significant ( p < 0.1) and remained multivariate significance ( p < 0.05). Subject ID was the random effect. 33 , 34 Sensitivity analysis GMM is sensitive to sample size. 35 To assess stability in cluster assignments, bootstrap resampling was performed (n = 1000) with refitted GMMs at each timepoint and centroids re-calculated. The bootstrap distributions of the centroid positions were compared in RC1-RC2 space with the original centroid estimates. RESULTS Patient characteristics 44 patients were recruited. (5/44 [11.4%]) patients dropped out at t1. (13/44 [29.5%]) dropped out at t2 due to logistical difficulty, travelling for test administration, rather than disease progression (Fig. 1 ). The median age of diagnosis was 63.1 years (IQR: 53.4–68.0). Most patients were male (68% [30/44]) with equal percentage of tumours in each hemisphere. Frontal (45% [20/44]) and temporal (32% [14/44]) tumours were the most common locations (Table 1 ). Table 1 Clinical characteristics of patients who completed NPS at each timepoint. Characteristic T0 N = 44 1 T1 N = 39 1 T2 N = 31 1 p-value 2 q-value 3 Age 63.1 (53.4, 68.0) 62.9 (53.2, 68.2) 60.1 (53.2, 67.1) 0.8 > 0.9 Sex > 0.9 > 0.9 0 (0%) 0 (0%) 0 (0%) Female 14 (32%) 12 (31%) 10 (32%) Male 30 (68%) 27 (69%) 21 (68%) Hand dominance > 0.9 > 0.9 Left 6 (14%) 6 (15%) 4 (13%) Right 38 (86%) 33 (85%) 27 (87%) WHO performance > 0.9 > 0.9 0 21 (48%) 20 (51%) 14 (45%) 1 21 (48%) 17 (44%) 15 (48%) 2 2 (4.5%) 2 (5.1%) 2 (6.5%) Education (years) 16.0 (13.0, 18.0) 16.0 (13.0, 19.0) 16.0 (14.0, 19.0) 0.9 > 0.9 Tumour side > 0.9 > 0.9 Left 22 (50%) 19 (49%) 14 (45%) Right 22 (50%) 20 (51%) 17 (55%) Lobe > 0.9 > 0.9 Frontal 20 (45%) 17 (44%) 14 (45%) Occipital 2 (4.5%) 2 (5.1%) 2 (6.5%) Parietal 8 (18%) 6 (15%) 4 (13%) Temporal 14 (32%) 14 (36%) 11 (35%) Lesion volume (cc) 34.8 (16.4, 47.8) 30.8 (15.2, 47.2) 26.3 (15.2, 47.2) 0.6 > 0.9 Awake 6 (14%) 5 (13%) 5 (16%) 0.9 > 0.9 Neurophysiology 21 (48%) 19 (49%) 16 (52%) > 0.9 > 0.9 Complication 6 (14%) 3 (7.7%) 3 (9.7%) 0.7 > 0.9 Preop AED (mg) 0 (0, 0) 0 (0, 0) 0 (0, 1,000) > 0.9 > 0.9 Postop AED (mg) 0 (0, 0) 0 (0, 0) 0 (0, 1,000) > 0.9 > 0.9 Steroid (mg) 5 (4, 8) 6 (4, 8) 6 (4, 8) > 0.9 > 0.9 IDH > 0.9 > 0.9 Mutant 1 (2.3%) 1 (2.6%) 1 (3.2%) Wildtype 43 (98%) 38 (97%) 30 (97%) MGMT > 0.9 > 0.9 Methylated 23 (52%) 22 (56%) 16 (52%) Unmethylated 21 (48%) 17 (44%) 15 (48%) Pathology > 0.9 > 0.9 Glioblastoma 41 (93%) 36 (92%) 29 (94%) Gliosarcoma 3 (6.8%) 3 (7.7%) 2 (6.5%) Radiotherapy 32 (73%) 30 (77%) 27 (87%) 0.3 > 0.9 Chemotherapy 28 (64%) 26 (67%) 24 (77%) 0.4 > 0.9 Surgery > 0.9 > 0.9 Biopsy 6 (14%) 6 (15%) 5 (16%) Resection 38 (86%) 33 (85%) 26 (84%) GAD anxiety 4.0 (2.0, 7.0) 4.0 (1.0, 6.0) 4.0 (0.0, 7.0) 0.6 > 0.9 PHQ depression 4.0 (2.0, 8.0) 3.0 (1.0, 6.0) 4.0 (3.0, 8.0) 0.6 > 0.9 Preop FLAIR volume (cc) 16 (9, 41) 16 (9, 39) 16 (9, 54) > 0.9 > 0.9 1 Median (IQR) or Frequency (%) 2 Kruskal-Wallis rank sum test; Fisher's exact test; Pearson's Chi-squared test 3 False discovery rate correction for multiple testing Median time between t0 test and surgery was similar between Biopsy (6.5 days [IQR: 9.3–4.5]) and Resection patients (5 days [4.0–9.0]) ( p = 0.6). Median time between surgery and t1 test was 1.0 day [IQR: 1.0–2.5] for biopsy patients and 3 days [IQR: 2–3] for resection patients ( p = 0.1). At t2, there was no difference between biopsy (41.8 days [IQR: 34.4–45.8]) and resection patients (40.5 days [IQR: 36.4–42]) ( p = 0.8). There were no significant differences in baseline characteristics by surgery (Supplementary Information Table 2), or between patients completing NPS at each timepoint versus those who failed to complete NPS (Supplementary Information Table 3) ( q > 0.05). Preoperative and postoperative group cognition At baseline, 26–50% of resection patients and 17–83% of biopsy patients had cognitive impairments. The results were similar at t1 and t2 (Fig. 2 ). There is no significant difference between the percentages of patients with impairment at any timepoint, by type of surgery (Resection versus Biopsy) (Supplementary Information Table 4) or tumour hemisphere (Supplementary Information Table 5). Group and individual changes in cognition over time The percentage of impaired patients varied across domains. (Supplementary Information Fig. 1 ). Significantly more resection patients were impaired at t1 versus t0 in Perception (71.1% versus 31.8%) and Calculation (63.2 versus 31.6%) domains ( q < 0.05) (Supplementary Information Fig. 3 ). ( q < 0.05). Between t0 and t1, Resection patients (n = 38) declined between 3% to 53% (i.e. RCI < -1.645), on different tests, most commonly for Calculation (53%). Declines between t1 and t2, ranged between 3% to 42%. Most frequently for Verbal working memory (8–21%) and Executive function (11–42%). Performance was stable across most tests between t0 and t1 (26–82%) and between t1 and t2 (3–61%). (Supplementary Information Fig. 3 ). Early improvement was significantly more likely in Attention (78.9% versus 44.7%) ( q = 0.006), Executive function (63.2% versus 36.8%) ( q = 0.004), Perception (44.7% versus 15.8%) ( q < 0.02) and Recognition memory (71.1% versus 13.2%) ( q < 0.001). Delayed improvement was more likely for Language (5.3% versus 26.3%) ( q = 0.04), and Calculation domains (5.3% versus 52.6%) ( q = 0.001). (Supplementary Information Table 4). Cognitive phenotypes PCA identified three components explaining 55% of the variance (KMO = 0.72, Bartlett Test, p < 0.001) The Scree plot is shown in Supplementary Information Fig. 4 . The first component loaded on Language, Executive Function and Perception (25%). The second, on Prospective and Retrospective Memory (15%). The third component on Recognition Memory and Working Memory (15%) (Supplementary Information Fig. 5). The optimal gaussian mixture model for t0 was ellipsoidal, equal volume and orientation, (EVE) with 4 clusters, yielding the lowest BIC, (log-likelihood − 198.9, BIC − 500.0, mean posterior probability 0.98). 6 clusters were found at t1 and 2 at t2 (Supplementary Information Figs. 6 and 7). Sensitivity analysis shows that the bootstrap centroid points group around the cluster solutions across timepoints. This supports the clusters being representations of patient cognition within latent space (Supplementary Information Fig. 8). Within latent cognitive space, patients separated along the dominant axis of deficit RC1. (Supplementary Information Fig. 7). Minimal deficit is characterised by the highest RC1 scores. This phenotype was only seen at t0 (n = 23). Mild deficit patients had mildly negative RC1 scores. At t1, this phenotype was represented by two clusters (t1 clusters 1 and 2, n = 16) that differed on RC2 but occupied similar positions on RC1. Both converged to a single cluster at t2 (n = 19), representing postoperative recovery. Moderate deficit patients had low RC1 scores, with negative RC2 and mildly positive RC3. In the severe deficit group, patients had the lowest RC1 scores, low RC2. At t1, this phenotype was represented by two clusters (t1 clusters 4 and 5, n = 9). This phenotype was not present at t2. The Transient deficit (t1 cluster 6, n = 3), was assigned to a cluster which was only seen at t1. These patients had mildly negative RC1 but very low RC3 representing an early post-surgical cognitive pattern affecting Recognition Memory and Working Memory. (Supplementary Information Fig. 9). Supplementary Information Table 6 shows the final phenotypes and RC scores. The primary determinant of postoperative cognitive phenotype is preoperative cognition. In general, cognition declined in resection patients in the early postoperative period and did not return to their preoperative baseline by the start of adjuvant chemoradiotherapy (Fig. 3 ). Predictors of postoperative cognition Models incorporating preoperative cognitive phenotype predicted postoperative Language, Executive Function as well as Perception, Recognition Memory and Working Memory. Cognitive phenotype provided similar predictive performance to baseline scores for Prospective and Retrospective Memory (Supplementary Information Table 7). Moderate deficit (-1.2 [CI: -1.9, -0.49], p < 0.001) and Severe deficit (-1.9 [CI: -2.8, -1.1], p < 0.001) phenotypes predicted worse postoperative RC1 scores, with significant recovery from t1 to t2 (0.67 [CI: 0.35, 0.99], p < 0.001). For RC2, Moderate deficit (-1.5 [CI: -2.3, -0.64], p < 0.001) and Severe deficit (-1.3 [CI: -2.3, -0.32], p < 0.001) phenotypes predicted worse scores. Male sex associated with higher scores (0.54 [CI: 0.00, 1.1], p = 0.041). For RC3, baseline depression (PHQ) predicted worse scores (-0.05 [CI: -0.09, -0.01], p = 0.008), temporal lobe tumours predicted better scores compared to frontal (0.95 [CI: 0.31, 1.6], p = 0.024), and neurophysiological monitoring predicted worse scores (-0.64 [CI: -1.2, − 0.08], p = 0.019). The preoperative cluster was not a significant predictor of RC3 ( p = 0.2) (Fig. 4 ). (Supplementary Information Table 9). Univariable screening of predictors is shown in Supplementary Information Table 8. DISCUSSION Cognition in glioma patients has been extensively investigated. 12 , 27 , 36 – 41 But few studies have investigated how surgery affects cognition without the confound of chemoradiotherapy. 37,40,41 Patients performed ≥ 2 SD below the mean compared to normative controls across several domains. Cognitive deficits are common in glioma patients, occurring in up to 55–82% of presurgical and 49–84% of postsurgical patients. 27 , 36 , 41 , 42 The percentage of patients with Perception and Calculation deficits significantly increased early after surgery with a trend towards recovery by six-weeks. Despite improvement in individual test performance in the early postoperative period, most patients remained impaired across several domains including Perception and Calculation. RCI improvement in domains such as Perception may be driven by practice effects in unimpaired patients. A previous meta-analysis found improved cognition in mixed-grade gliomas after surgery. 17 In glioblastoma, studies reported stability, improvement and declines post-surgery. 13 Excluding Biopsy patients may have overestimated the beneficial effects of surgery. 17,36,40,41 We included Biopsy patients to reduce confounding from anaesthesia and hospitalisation. Individual cognitive changes are heterogenous. Early post-surgery (1–3 weeks), Attention, Verbal and Visual working memory were most likely to decline. 40 , 41 3–6 months from surgery, Executive function, Verbal memory and Praxis declined in mixed-grade glioma patients, but were stable in glioblastoma patients. 27 , 39 The differences between studies could be because presurgical deficit is more severe in high-grade glioma, compared to low-grade glioma patients, thus limiting recovery. Alternatively, patients with low or high baseline performance may have the greatest room for improvement and decline respectively. 36 , 39 Three PCA components accounted for half test variance: 1) Language, Executive Function and Perception. 2) Prospective and Retrospective Memory and 3) Recognition Memory and Working Memory. Test scores may reflect latent cognitive processes and lesion anatomy. 15 , 43 We identified four preoperative cognitive-phenotypes, Minimal, Mild, Moderate and Severe cognition, Moderate and Severe patients were more likely to decline postoperatively. Reyes et al. used longitudinal pre and post-radiotherapy cognition scores in brain tumour patients to identify three cognitive phenotypes; global impairment, isolated Verbal memory deficits and Minimal impairment. 12 Compared to their Study, our approach has the advantage of not requiring the phenotype to fixed across time. We show that resection is associated with a disruption to the phenotype in the early postoperative period with a minority of patients (15.8% [6/38]) improving to a better phenotype by 6-weeks post-surgery. No patient recovered to a preoperative Minimal deficit. Models incorporating preoperative phenotype predict postoperative cognition equivalent or better than baseline RC scores. This may be because the phenotype captures a multivariate pattern of cognitive deficit. Temporal lobe gliomas in both hemispheres are associated with deficits in Executive function, Verbal memory and Attention. 44 , 45 Our results highlight that preoperative cognition and tumour location predict postsurgical cognition. 27 , 46 Temporal lobe tumours were associated with improved postoperative Recognition Memory and Working Memory. This may be due to the frontal lobes playing the primary role in cue-based recognition tasks. 47 Future work should investigate whether these phenotypes are correlated with neuroanatomical locations and tumour invasiveness. This suggests cognitive phenotypes may represent latent patterns of brain injury caused by different lesion locations affecting common structural and functional brain networks. 43 , 48 , 49 Limitations The sample size is modest, especially the number of biopsy patients. This reduces the precision of estimates of perioperative effects when calculating reliable change. 30% of patients dropped out of NPS at t2 mainly due to chemoradiotherapy treatment at different centers but this may still have led to overestimation of cognition in the cohort. However, completers and non-completers did not differ significantly on baseline characteristics. In addition, model-based clustering is sensitive to small sample sizes but our bootstrap analysis supports the overall spatial structure of the cognitive phenotypes. Previous work compared the OCS-Bridge NPS tool with traditional pen-and-paper neuropsychological assessments in low-grade glioma patients. They have found that traditional assessments may be more sensitive to deficits in domains such as Attention and Memory. 20 However, computerized tools are easier to use, especially for longitudinal testing postoperatively when patients may be fatigued from treatment effects. In addition, OCS-Bridge captures reaction times, increasing its sensitivity to non-verbal perceptual deficits. 23 Nonetheless, NPS offer breadth in place of depth which may limit the variability captured in by PCA. Future work Our findings should be validated using an external cohort. Future studies should assess whether preoperative phenotypes predict other outcomes such as survival and HRQOL. The mechanisms underlying different cognitive phenotypes should be investigated by correlating neuroimaging markers of structural and functional brain injury and tumour invasion with cognitive phenotypes. CONCLUSION We show that it is possible to identify high grade glioma patients who are at risk of postoperative cognitive deterioration using preoperative cognitive tests. This is important for counselling patients about expected outcomes from surgery. It may also help identify patients for early cognitive rehabilitation Declarations Funding YW is supported by Cancer Research UK Clinical Research Training Fellowship and by the CRUK Cambridge Centre. This work was supported by the NIHR HealthTech Research Centre in Brain Injury and the NIHR Cambridge Biomedical Research Centre (NIHR203312) and the Assessing impact of surgically-induced deficits on patient functioning and quality of life (SIND study) (19/WM/0152). This publication presents independent research funded by the National Institute for Health and Care Research (NIHR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Author Contribution Y.W led the manuscript development, including data collection, analysis and drafting of the manuscript. A.H, T.M, H.Z, M.LR contributed to data analysis and interpretation, including appraisal of analysis methods. R.S contributed to data analysis and evaluation of the manuscript. A.J, R.M, R.C.M, T.S. contributed to data collection and interpretation of the results. S.J.P provided supervision throughout the research process. All authors contributed to data collection, interpretation and critical appraisal of the manuscript. Data Availability The data that support the findings of this study are not openly available due to patient confidentiality and informed consent. Data available from the corresponding author upon reasonable request with a Data Sharing Agreement. Data is in controlled access data storage at University of Cambridge. References Ostrom QT, Gittleman H, Truitt G, Boscia A, Kruchko C, Barnholtz-Sloan JS (2018) CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2011–2015. Neurooncology 20(suppl4):iv1–iv86. 10.1093/neuonc/noy131 Mansouri A, Lai C, Scales D, Pirouzmand F (2022) A phase II pilot randomized controlled trial to assess the feasibility of the supra-marginal surgical resection of malignant glioma (G-SUMIT: Glioma supra marginal incision trial) study protocol. Pilot Feasibility Stud 8(1):138. 10.1186/s40814-022-01104-1 Burnet NG, Jefferies SJ, Benson RJ, Hunt DP, Treasure FP (2005) Years of life lost (YLL) from cancer is an important measure of population burden — and should be considered when allocating research funds. Br J Cancer 92(2):2. 10.1038/sj.bjc.6602321 Louis DN, Perry A, Wesseling P et al (2021) The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neurooncology 23(8):1231–1251. 10.1093/neuonc/noab106 Bosma I, Vos MJ, Heimans JJ et al (2007) The course of neurocognitive functioning in high-grade glioma patients1. Neurooncology 9(1):53–62. 10.1215/15228517-2006-012 Bosma I, Reijneveld JC, Douw L et al (2009) Health-related quality of life of long-term high-grade glioma survivors. Neurooncology 11(1):51–58. 10.1215/15228517-2008-049 Coomans MB, Dirven L, Aaronson N et al (2022) Factors associated with health-related quality of life (HRQoL) deterioration in glioma patients during the progression-free survival period. Neurooncology 24(12):2159–2169. 10.1093/neuonc/noac097 Henriksson R, Asklund T, Poulsen HS (2011) Impact of therapy on quality of life, neurocognitive function and their correlates in glioblastoma multiforme: a review. J Neurooncol 104(3):639–646. 10.1007/s11060-011-0565-x Nakajima R, Kinoshita M, Okita H, Nakada M (2024) Postsurgical motor function and processing speed as predictors of quality of life in patients with chronic-phase glioblastoma. Acta Neurochir 166(1):357. 10.1007/s00701-024-06245-1 Noll KR, Bradshaw ME, Weinberg JS, Wefel JS (2017) Relationships between neurocognitive functioning, mood, and quality of life in patients with temporal lobe glioma. Psychooncology 26(5):617–624. 10.1002/pon.4046 Giovagnoli AR (2005) Facets and determinants of quality of life in patients with recurrent high grade glioma. J Neurol Neurosurg Psychiatry 76(4):562–568. 10.1136/jnnp.2004.036186 Reyes A, Stasenko A, Hopper A et al (2025) Cognitive phenotypes: Unraveling the heterogeneity in cognitive dysfunction among patients with primary brain tumors receiving radiotherapy. Neurooncology 27(2):466–478. 10.1093/neuonc/noae183 Sinha R, Stephenson JM, Price SJ A systematic review of cognitive function in patients with glioblastoma undergoing surgery. Neuro-Oncology Practice . Published online 2019. 10.1093/nop/npz018 Hope TMH, Halai A, Crinion J, Castelli P, Price CJ, Bowman H (2024) Principal component analysis-based latent-space dimensionality under-estimation, with uncorrelated latent variables. Brain 147(2):e14–e16. 10.1093/brain/awad355 Halai AD, Woollams AM, Lambon Ralph MA (2017) Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic post-stroke aphasia: Revealing the unique neural correlates of speech fluency, phonology and semantics. Cortex 86:275–289. 10.1016/j.cortex.2016.04.016 Zhao Y, Halai AD, Lambon Ralph MA (2020) Evaluating the granularity and statistical structure of lesions and behaviour in post-stroke aphasia. Brain Commun 2(2). 10.1093/braincomms/fcaa062 Ng JCH, See AAQ, Ang TY, Tan LYR, Ang BT, King NKK (2019) Effects of surgery on neurocognitive function in patients with glioma: a meta-analysis of immediate post-operative and long-term follow-up neurocognitive outcomes. J Neurooncol 141(1):167–182. 10.1007/s11060-018-03023-9 Butler RA, Lambon Ralph MA, Woollams AM (2014) Capturing multidimensionality in stroke aphasia: mapping principal behavioural components to neural structures. Brain 137(12):3248–3266. 10.1093/brain/awu286 Woollams AM, Halai A, Lambon Ralph MA (2018) Mapping the intersection of language and reading: the neural bases of the primary systems hypothesis. Brain Struct Function 223(8):3769–3786. 10.1007/s00429-018-1716-z Rauch S, Wan Y, Halai A et al (2026) Multidimensional cognitive reserve and cognitive outcomes in glioblastoma: a pre- and postoperative analysis. J Neurooncol 177(1):11. 10.1007/s11060-026-05475-w Zucchella C, Capone A, Codella V et al (2013) Cognitive rehabilitation for early post-surgery inpatients affected by primary brain tumor: a randomized, controlled trial. J Neurooncol 114(1):93–100. 10.1007/s11060-013-1153-z Vogelbaum MA, Krivosheya D, Borghei-Razavi H et al (2020) Phase 0 and window of opportunity clinical trial design in neuro-oncology: a RANO review. Neuro Oncol 22(11):1568–1579. 10.1093/neuonc/noaa149 Romero-Garcia R, Owen M, McDonald A et al (2022) Assessment of neuropsychological function in brain tumor treatment: a comparison of traditional neuropsychological assessment with app-based cognitive screening. Acta Neurochir (Wien) 164(8):2021–2034. 10.1007/s00701-022-05162-5 Demeyere N, Riddoch MJ, Slavkova ED, Bickerton WL, Humphreys GW (2015) The Oxford Cognitive Screen (OCS): validation of a stroke-specific short cognitive screening tool. Psychol Assess 27(3):883–894. 10.1037/pas0000082 Wan Y, Rahmat R, Price SJ (2020) Deep learning for glioblastoma segmentation using preoperative magnetic resonance imaging identifies volumetric features associated with survival. Acta Neurochir 162(12):3067–3080. 10.1007/s00701-020-04483-7 Maassen GH, Bossema E, Brand N (2009) Reliable change and practice effects: outcomes of various indices compared. J Clin Exp Neuropsychol 31(3):339–352. 10.1080/13803390802169059 Rijnen SJM, Butterbrod E, Rutten GJM, Sitskoorn MM, Gehring K (2020) Presurgical Identification of Patients With Glioblastoma at Risk for Cognitive Impairment at 3-Month Follow-up. Neurosurgery 87(6):1119–1129. 10.1093/neuros/nyaa190 Rijnen SJM, Meskal I, Bakker M et al (2019) Cognitive outcomes in meningioma patients undergoing surgery: individual changes over time and predictors of late cognitive functioning. Neurooncology 21(7):911–922. 10.1093/neuonc/noz039 Storopoli J, FactorAssumptions Set of Assumptions for Factor and Principal Component Analysis. Published online March 6, 2020:2.0.1. 10.32614/CRAN.package.FactorAssumptions Dinno A paran: Horn’s Test of Principal Components/Factors. Published online August 12, 2007:1.5.5. 10.32614/CRAN.package.paran Arno Fritsch (2012) mcclust: Process an MCMC Sample of Clusterings. Published online July 23. 10.32614/cran.package.mcclust Peterson RA, Cavanaugh JE Ordered quantile normalization: a semiparametric transformation built for the cross-validation era. J Appl Stat 47(13–15):2312–2327. 10.1080/02664763.2019.1630372 Bartoń K, MuMIn (2025) Multi-Model Inference. Published online April 1, 2025. Accessed July 9. https://cran.r-project.org/web/packages/MuMIn/index.html Bates D, Mächler M, Bolker B, Walker S (2015) Fitting Linear Mixed-Effects Models Using lme4. J Stat Softw 67:1–48. 10.18637/jss.v067.i01 Psutka JV, Psutka J (2019) Sample size for maximum-likelihood estimates of Gaussian model depending on dimensionality of pattern space. Pattern Recogn 91:25–33. 10.1016/j.patcog.2019.01.046 Van Loenen IS, Rijnen SJM, Bruijn J, Rutten GJM, Gehring K, Sitskoorn MM (2018) Group Changes in Cognitive Performance After Surgery Mask Changes in Individual Patients with Glioblastoma. World Neurosurg 117:e172–e179. 10.1016/j.wneu.2018.05.232 Sinha R, Masina R, Morales C et al (2023) A Prospective Study of Longitudinal Risks of Cognitive Deficit for People Undergoing Glioblastoma Surgery Using a Tablet Computer Cognition Testing Battery: Towards Personalized Understanding of Risks to Cognitive Function. JPM 13(2):278. 10.3390/jpm13020278 Noll KR, Bradshaw M, Sheppard D, Wefel JS (2024) Perioperative Neurocognitive Function in Glioma Surgery. Curr Oncol Rep 26(5):466–476. 10.1007/s11912-024-01522-9 Santini B, Talacchi A, Squintani G, Casagrande F, Capasso R, Miceli G (2012) Cognitive outcome after awake surgery for tumors in language areas. J Neurooncol 108(2):319–326. 10.1007/s11060-012-0817-4 Talacchi A, Santini B, Savazzi S, Gerosa M (2011) Cognitive effects of tumour and surgical treatment in glioma patients. J Neurooncol 103(3):541–549. 10.1007/s11060-010-0417-0 Habets EJJ, Kloet A, Walchenbach R, Vecht CJ, Klein M, Taphoorn MJB (2014) Tumour and surgery effects on cognitive functioning in high-grade glioma patients. Acta Neurochir 156(8):1451–1459. 10.1007/s00701-014-2115-8 van Kessel E, Baumfalk AE, van Zandvoort MJE, Robe PA, Snijders TJ (2017) Tumor-related neurocognitive dysfunction in patients with diffuse glioma: a systematic review of neurocognitive functioning prior to anti-tumor treatment. J Neurooncol 134(1):9–18. 10.1007/s11060-017-2503-z Facchini S, Favaretto C, Castellaro M et al (2023) A common low dimensional structure of cognitive impairment in stroke and brain tumors. NeuroImage: Clin 40:103518. 10.1016/j.nicl.2023.103518 Noll KR, Ziu M, Weinberg JS, Wefel JS (2016) Neurocognitive functioning in patients with glioma of the left and right temporal lobes. J Neurooncol 128(2):323–331. 10.1007/s11060-016-2114-0 Noll KR, Weinberg JS, Ziu M, Benveniste RJ, Suki D, Wefel JS (2015) Neurocognitive Changes Associated With Surgical Resection of Left and Right Temporal Lobe Glioma. Neurosurgery 77(5):777–785. 10.1227/NEU.0000000000000987 Zangrossi A, Silvestri E, Bisio M et al (2022) Presurgical predictors of early cognitive outcome after brain tumor resection in glioma patients. NeuroImage: Clin 36:103219. 10.1016/j.nicl.2022.103219 McFarland CP, Glisky EL (2009) Frontal lobe involvement in a task of time-based prospective memory. Neuropsychologia 47(7):1660–1669. 10.1016/j.neuropsychologia.2009.02.023 Jütten K, Mainz V, Delev D et al (2020) Asymmetric tumor-related alterations of network‐specific intrinsic functional connectivity in glioma patients. Hum Brain Mapp 41(16):4549–4561. 10.1002/hbm.25140 Arbula S, Ambrosini E, Della Puppa A et al (2020) Focal left prefrontal lesions and cognitive impairment: A multivariate lesion-symptom mapping approach. Neuropsychologia 136:107253. 10.1016/j.neuropsychologia.2019.107253 Additional Declarations No competing interests reported. Supplementary Files SupplementaryInformation.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9163008","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":608862257,"identity":"c3fd5078-6702-4c72-b881-2d66226e8087","order_by":0,"name":"Yizhou Wan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIiWNgGAWjYHCC5Mc/KuCcBKK0PDNmOEOaFsYH0oxtpGgxuJGcYFw4z05et4H54QfGtjRitKQlPJ65Ldlw2wE2YwnGthwitNzOSTDg3XaAcdsBBjMGxrYKYrTkf5DgnXPAftsB9m/EaklIkOZtOJC47QAPyBYiHCZ5/0Ga4YxjycnbDvMUSyScI8L7fGcOJD/4UGNnu+14+8YPH8qSCWtROABjMTMQGZHyDcSoGgWjYBSMgpENAK5zPcEVLTxiAAAAAElFTkSuQmCC","orcid":"","institution":"University of Cambridge","correspondingAuthor":true,"prefix":"","firstName":"Yizhou","middleName":"","lastName":"Wan","suffix":""},{"id":608862259,"identity":"c5b7dd57-b41d-4b6a-876d-d315b8b0b298","order_by":1,"name":"Ajay Halai","email":"","orcid":"","institution":"MRC Cognition and Brain Sciences Unit","correspondingAuthor":false,"prefix":"","firstName":"Ajay","middleName":"","lastName":"Halai","suffix":""},{"id":608862264,"identity":"25f9d533-b851-4753-9333-202d5086e1d2","order_by":2,"name":"Tom Manly","email":"","orcid":"","institution":"MRC Cognition and Brain Sciences Unit","correspondingAuthor":false,"prefix":"","firstName":"Tom","middleName":"","lastName":"Manly","suffix":""},{"id":608862271,"identity":"d4278000-26a0-4e27-ba57-dbece3764eb6","order_by":3,"name":"Haiyan Zheng","email":"","orcid":"","institution":"University of Bath","correspondingAuthor":false,"prefix":"","firstName":"Haiyan","middleName":"","lastName":"Zheng","suffix":""},{"id":608862275,"identity":"3a61fd1c-23a5-4907-ad22-679980112ee7","order_by":4,"name":"Roxanne Mayrand","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Roxanne","middleName":"","lastName":"Mayrand","suffix":""},{"id":608862277,"identity":"8639e25f-baa4-4156-8620-ed269f4a8040","order_by":5,"name":"Rohitashwa Sinha","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Rohitashwa","middleName":"","lastName":"Sinha","suffix":""},{"id":608862281,"identity":"87ce4a16-5a9a-40c6-973a-25cae202aa39","order_by":6,"name":"Alexis Joannides","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Alexis","middleName":"","lastName":"Joannides","suffix":""},{"id":608862282,"identity":"63bdbddd-67b7-47ee-b047-54206e026bac","order_by":7,"name":"Richard Mair","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Richard","middleName":"","lastName":"Mair","suffix":""},{"id":608862283,"identity":"c91ab25f-3dcb-483f-b30e-392147e9b415","order_by":8,"name":"Robert Morris","email":"","orcid":"","institution":"Addenbrooke's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Robert","middleName":"","lastName":"Morris","suffix":""},{"id":608862284,"identity":"d68773f9-a22f-4ae5-9de4-671fd946a80f","order_by":9,"name":"Thomas Santarius","email":"","orcid":"","institution":"Addenbrooke's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Santarius","suffix":""},{"id":608862286,"identity":"5c7277cf-301e-4867-a960-364591430b67","order_by":10,"name":"Matthew Lambon Ralph","email":"","orcid":"","institution":"MRC Cognition and Brain Sciences Unit","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"Lambon","lastName":"Ralph","suffix":""},{"id":608862287,"identity":"91239fe2-0398-4541-9b19-a8ab4f8e3d87","order_by":11,"name":"Stephen John Price","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"John","lastName":"Price","suffix":""}],"badges":[],"createdAt":"2026-03-18 21:23:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9163008/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9163008/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105038185,"identity":"1bd40dcd-578b-4ddd-92f2-401ace4a5b65","added_by":"auto","created_at":"2026-03-20 07:42:40","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":72850,"visible":true,"origin":"","legend":"\u003cp\u003eNumber of patients completing Neuropsychological Screen (NPS) at each timepoint.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9163008/v1/d207b0c93cc798b02c4d3130.jpeg"},{"id":105038032,"identity":"7bbac882-a3ea-4df4-a717-c6e79f81f3e9","added_by":"auto","created_at":"2026-03-20 07:41:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2171717,"visible":true,"origin":"","legend":"\u003cp\u003ePercentages of impairments over all cognitive domains at each timepoint (%).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9163008/v1/f5c63bb66f85d4ca36ee68d8.png"},{"id":105038007,"identity":"11c8d8b4-e905-4f03-9a64-cd8bdfed6db1","added_by":"auto","created_at":"2026-03-20 07:41:30","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":477915,"visible":true,"origin":"","legend":"\u003cp\u003eShankey plot showing change in cognitive phenotype over time for A) resection (n = 38) and B) Biopsy patients (n = 6).\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9163008/v1/372b3445eb7d2ea7972a4788.jpeg"},{"id":105038008,"identity":"006bcf74-f1a5-49c0-b56f-9d1b0ae24e92","added_by":"auto","created_at":"2026-03-20 07:41:32","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":343328,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot showing predictors of postoperative latent cognitive scores.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9163008/v1/5bd40b2d9845d793a265fb06.jpeg"},{"id":106612437,"identity":"2fedde4c-3c5c-45dc-a74b-8587d210f33f","added_by":"auto","created_at":"2026-04-10 12:28:47","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3838004,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9163008/v1/43bb343f-b476-4737-b2f7-73a377598764.pdf"},{"id":105038005,"identity":"c793397b-7a54-401e-bd6e-c11b2c6e20db","added_by":"auto","created_at":"2026-03-20 07:41:30","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":37061573,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-9163008/v1/101e93bb6afa35c324adaff5.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Multidimensional preoperative cognitive phenotypes predict postoperative cognition in glioblastoma","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eGliomas are the most common primary brain cancer affecting adults.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e Of these, glioblastoma (WHO grade 4) with a median prognosis of only 15 months, is the most common.\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e Given the poor prognosis, there is increasing recognition that maintenance of cognition during treatment is a crucial goal for patients.\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eCognition is important for brain tumour patients.\u003csup\u003e\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e Motor function, processing speed, executive function, and memory have been linked with functional well-being and role function.\u003csup\u003e\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e Previous research has found heterogenous patterns of cognitive outcomes in brain tumour patients when tested at different timepoints.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e Performance variability may be because of patient, tumour and treatment factors and Study methodology. Comparison of results between studies is difficult due to the variation across neuropsychological tests. Cognitive deficits can be considered a latent system where underlying functional disruption is observed in cognitive test scores.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e The dimensionality of the latent system is lower that the number of observed cognitive tests because: 1) test scores are intercorrelated and 2) patient have deficits in similar functional regions, constrained by neuroanatomical organisation and lesion location.\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eTests vary considerably in task difficulty and requirements.\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e Easier tests avoids floor effects in severely impaired patients while harder tests are more sensitive to milder deficits.\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e In addition, each task may depend on multiple underlying cognitive systems.\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eIt is important to understand cognitive variability prior surgery to stratify patients into risk-profiles which can guide clinical decision making. For example, different surgical resection strategies may result in different postoperative outcomes in terms of survival and cognition. If cognition is primarily determined by patient or tumour effects, understanding perioperative cognition may help identify which patients can benefit from aggressive surgical strategies to delay tumour progression.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e Emerging evidence suggests that postoperative cognitive rehabilitation for glioma patients is associated with improved visual attention, memory and executive function.\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e In addition, risk stratification may identify patients who benefit from early clinical trial involvement in the context of window-of-opportunity trials prior to surgery.\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eIn this study we used Principal Component Analysis (PCA) and Gaussian Mixture Modelling (GMM) to investigate the underlying structure of cognitive deficits in Glioma patients undergoing surgery. Both dimensionality reduction techniques seek to identify the underlying latent profiles and to explain the observed cognitive test scores. This allows us to stratify patients by baseline risk of postoperative cognitive deterioration. We also examined which clinical factors affected preoperative and postoperative cognition, at the early (within 1-week) and late (within 4\u0026ndash;6 weeks) postoperative timepoints.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003e44 patients (age\u0026thinsp;\u0026gt;\u0026thinsp;17) were recruited as part of the Surgically Induced Neurological Deficits in Glioblastomas) (SIND) Study (REC: 19/WM/0152). Written informed consent was obtained from all patients. The inclusion criteria were: 1) assessed by Tumour Board to have high-grade-glioma on imaging, WHO Performance Status (PS) 0\u0026ndash;2, suitable for surgery, resection patients were those where the treating neurosurgeon feels that \u0026gt;\u0026thinsp;90% of the enhancing tumour can be resected.\u003c/p\u003e \u003cp\u003ePatients with tumours located near/in speech/sensorimotor regions will undergo speech/language mapping +/- motor mapping at the treatment surgeon\u0026rsquo;s discretion. Neuropsychological Screen (NPS) was performed at 1) t0- baseline (prior to surgery), 2) t1- early postoperative (72-h postoperative), and t2- delayed postoperative (6-weeks post-surgery). Patients who are unsuitable for a contrast-enhanced MRI were excluded. All patients had WHO 2021 histopathology confirmed glioblastoma or gliosarcoma.\u003c/p\u003e \u003cp\u003eImage-guided biopsies with minimal brain disruption were performed in the Biopsy control group. Resection patients underwent 5-aminolevulinic acid (5-ALA) guided-surgery with neuronavigation (StealthStation, Medtronic), and other adjuvants (e.g., cortical and subcortical mapping) for maximal safe resection of fluorescent tumour.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eNeuropsychological screening tool\u003c/h3\u003e\n\u003cp\u003eA Neuropsychological Screen (NPS), validated in glioma patients were administered with a touch-screen tablet as part of research studies by two trained Researchers with Neuropsychology and Neurosurgery qualifications (Y.W and R.M).\u003csup\u003e23\u003c/sup\u003e The Oxford Cognitive Screen and Cambridge Attention, Memory, and Perception Screen (OCS-Bridge) consisted of tests of language, orientation, attention, perception, memory, praxis, and numeracy skills.\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e Parallel test versions were used to minimise practice effects. Total test time was 25\u0026ndash;30 minutes (Supplementary Information, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eImaging analysis\u003c/h3\u003e\n\u003cp\u003eMRIs were acquired with a 3-Tesla scanner (Siemens Healthcare, USA) at the Wolfson Brain Imaging Centre, University of Cambridge. (Supplementary Information, Methods).\u003c/p\u003e \u003cp\u003eSemi-automated methods were used to segment the whole tumour, FLAIR abnormality and resection cavity on preoperative images to calculate lesion volumes from these regions.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eClinical characteristics were summarised using medians (IQR), and percentages. Wilcoxon rank sum, Fisher Exact and Chi-squared tests were used to compare groups. Multiple-comparison correction was performed using false-discovery rate (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eOCS-BRIDGE values were z-scored using the mean and standard deviation (SD) of reference scores from 268 healthy participants aged 16 to 89 (mean age, 51.44, SD 19.86).\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e Performance\u0026thinsp;\u0026ge;\u0026thinsp;2 SD below the mean was considered impaired. Individual level changes were calculated using a standardized regression-based reliable change index (RCI), based on mean and SD of performance in biopsy group (n\u0026thinsp;=\u0026thinsp;6).\u003csup\u003e26\u003c/sup\u003e RCI values exceeding\u0026thinsp;\u0026plusmn;\u0026thinsp;1.645 (corresponding with a two-tailed alpha of 0.10%, 90% confidence interval) indicated meaningful change. The percentage of patients with improved, stable, and declined performance were calculated per domain.\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePrincipal component analysis\u003c/h3\u003e\n\u003cp\u003eTests with constant scores, formed linear combinations with other scores, and a Kaiser, Meyer, Olkin Measure of Sampling Adequacy (KMO) \u0026le; 0.3 were iteratively removed, retaining overall KMO\u0026thinsp;\u0026gt;\u0026thinsp;0.6.\u003csup\u003e29\u003c/sup\u003e PCA was performed using standardised baseline NPS scores across all patients (both resection and biopsy). Horn's parallel analysis (5000 iterations) was used to determine the optimal number of components.\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e A varimax rotation was performed to improve loading interpretability. The rotation matrix was multiplied by the standardised scores at other timepoints to calculate postoperative rotated component (RC) scores.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eGaussian mixture modelling\u003c/h2\u003e \u003cp\u003eTo identify cognitive phenotypes across time, baseline and postoperative RC scores for the combined cohort (resection and biopsy) were entered into Gaussian Mixture Models (GMMs), using finite mixture models via the Expectation-Maximization (EM) algorithm to identify latent patient groups. The optimal number of clusters was selected using by iteratively computing the BIC across models with differing number of components and covariance parameters with bootstrap likelihood ratio tests (5000 samples). \u003csup\u003e31\u003c/sup\u003e The model with the minimum BIC was selected and local fit parameters checked with the mean cluster assignment posterior probability.\u003c/p\u003e \u003cp\u003eCluster phenotype identities across time were aligned by matching the position of each cluster\u0026rsquo;s centroid along the principal PCA axis of variation. (Supplementary Information Fig.\u0026nbsp;7).\u003c/p\u003e \u003cp\u003eLinear mixed effects models assessed which predictors were associated with postoperative rotated component (RC) scores. The Order-Norm transformation was used for right skewed postoperative PCA scores.\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThe preoperative cognitive phenotype was evaluated against baseline RC scores to determine if it provided predictive value. For each postoperative RC score, four models were fitted: 1) null (timepoint only), 2) time and baseline RC scores, 3) time and preoperative phenotype and 4) time, preoperative phenotype and RC scores.\u003c/p\u003e \u003cp\u003eFixed effects were added to the time and preoperative phenotype model to assess which clinical variables were significant predictors. Forward selection was used to retain variables which were univariate significant (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.1) and remained multivariate significance (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Subject ID was the random effect.\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSensitivity analysis\u003c/h3\u003e\n\u003cp\u003eGMM is sensitive to sample size.\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e To assess stability in cluster assignments, bootstrap resampling was performed (n\u0026thinsp;=\u0026thinsp;1000) with refitted GMMs at each timepoint and centroids re-calculated. The bootstrap distributions of the centroid positions were compared in RC1-RC2 space with the original centroid estimates.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003ePatient characteristics\u003c/h2\u003e \u003cp\u003e44 patients were recruited. (5/44 [11.4%]) patients dropped out at t1. (13/44 [29.5%]) dropped out at t2 due to logistical difficulty, travelling for test administration, rather than disease progression (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe median age of diagnosis was 63.1 years (IQR: 53.4\u0026ndash;68.0). Most patients were male (68% [30/44]) with equal percentage of tumours in each hemisphere. Frontal (45% [20/44]) and temporal (32% [14/44]) tumours were the most common locations (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClinical characteristics of patients who completed NPS at each timepoint.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eT0 \u003c/p\u003e \u003cp\u003eN\u0026thinsp;=\u0026thinsp;44\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eT1 \u003c/p\u003e \u003cp\u003eN\u0026thinsp;=\u0026thinsp;39\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eT2 \u003c/p\u003e \u003cp\u003eN\u0026thinsp;=\u0026thinsp;31\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep-value\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eq-value\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e63.1 (53.4, 68.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e62.9 (53.2, 68.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60.1 (53.2, 67.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (32%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12 (31%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10 (32%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 (68%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27 (69%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e21 (68%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHand dominance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeft\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (14%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 (15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (13%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38 (86%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33 (85%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27 (87%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWHO performance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20 (51%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14 (45%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17 (44%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (4.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (5.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (6.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.0 (13.0, 18.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.0 (13.0, 19.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.0 (14.0, 19.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTumour side\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeft\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22 (50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (49%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14 (45%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22 (50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20 (51%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17 (55%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLobe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFrontal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20 (45%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17 (44%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14 (45%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOccipital\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (4.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (5.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (6.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eParietal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (18%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 (15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (13%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTemporal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (32%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14 (36%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11 (35%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesion volume (cc)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e34.8 (16.4, 47.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30.8 (15.2, 47.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26.3 (15.2, 47.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAwake\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (14%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (13%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5 (16%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeurophysiology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (49%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16 (52%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eComplication\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (14%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (7.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (9.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePreop AED (mg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0, 0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0, 0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0, 1,000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePostop AED (mg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0, 0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0, 0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0, 1,000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSteroid (mg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (4, 8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 (4, 8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6 (4, 8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIDH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMutant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (2.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (2.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (3.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWildtype\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e43 (98%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38 (97%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30 (97%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMGMT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMethylated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23 (52%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22 (56%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16 (52%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnmethylated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17 (44%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlioblastoma\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e41 (93%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36 (92%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29 (94%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGliosarcoma\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (6.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (7.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (6.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiotherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32 (73%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30 (77%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27 (87%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChemotherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28 (64%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26 (67%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e24 (77%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSurgery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (14%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 (15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5 (16%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResection\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38 (86%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33 (85%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26 (84%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGAD anxiety\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.0 (2.0, 7.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.0 (1.0, 6.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.0 (0.0, 7.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePHQ depression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.0 (2.0, 8.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.0 (1.0, 6.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.0 (3.0, 8.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePreop FLAIR volume (cc)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16 (9, 41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16 (9, 39)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16 (9, 54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003e\u003csup\u003e1\u003c/sup\u003eMedian (IQR) or Frequency (%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003e\u003csup\u003e2\u003c/sup\u003eKruskal-Wallis rank sum test; Fisher's exact test; Pearson's Chi-squared test\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003csup\u003e3\u003c/sup\u003eFalse discovery rate correction for multiple testing\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eMedian time between t0 test and surgery was similar between Biopsy (6.5 days [IQR: 9.3\u0026ndash;4.5]) and Resection patients (5 days [4.0\u0026ndash;9.0]) (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.6). Median time between surgery and t1 test was 1.0 day [IQR: 1.0\u0026ndash;2.5] for biopsy patients and 3 days [IQR: 2\u0026ndash;3] for resection patients (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.1). At t2, there was no difference between biopsy (41.8 days [IQR: 34.4\u0026ndash;45.8]) and resection patients (40.5 days [IQR: 36.4\u0026ndash;42]) (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.8). There were no significant differences in baseline characteristics by surgery (Supplementary Information Table\u0026nbsp;2), or between patients completing NPS at each timepoint versus those who failed to complete NPS (Supplementary Information Table\u0026nbsp;3) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003ePreoperative and postoperative group cognition\u003c/h2\u003e \u003cp\u003eAt baseline, 26\u0026ndash;50% of resection patients and 17\u0026ndash;83% of biopsy patients had cognitive impairments. The results were similar at t1 and t2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThere is no significant difference between the percentages of patients with impairment at any timepoint, by type of surgery (Resection versus Biopsy) (Supplementary Information Table\u0026nbsp;4) or tumour hemisphere (Supplementary Information Table\u0026nbsp;5).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eGroup and individual changes in cognition over time\u003c/h2\u003e \u003cp\u003eThe percentage of impaired patients varied across domains. (Supplementary Information Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Significantly more resection patients were impaired at t1 versus t0 in Perception (71.1% versus 31.8%) and Calculation (63.2 versus 31.6%) domains (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Supplementary Information Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eBetween t0 and t1, Resection patients (n\u0026thinsp;=\u0026thinsp;38) declined between 3% to 53% (i.e. RCI \u0026lt; -1.645), on different tests, most commonly for Calculation (53%). Declines between t1 and t2, ranged between 3% to 42%. Most frequently for Verbal working memory (8\u0026ndash;21%) and Executive function (11\u0026ndash;42%). Performance was stable across most tests between t0 and t1 (26\u0026ndash;82%) and between t1 and t2 (3\u0026ndash;61%). (Supplementary Information Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eEarly improvement was significantly more likely in Attention (78.9% versus 44.7%) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.006), Executive function (63.2% versus 36.8%) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004), Perception (44.7% versus 15.8%) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.02) and Recognition memory (71.1% versus 13.2%) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Delayed improvement was more likely for Language (5.3% versus 26.3%) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.04), and Calculation domains (5.3% versus 52.6%) (\u003cem\u003eq\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001). (Supplementary Information Table\u0026nbsp;4).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eCognitive phenotypes\u003c/h2\u003e \u003cp\u003ePCA identified three components explaining 55% of the variance (KMO\u0026thinsp;=\u0026thinsp;0.72, Bartlett Test, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) The Scree plot is shown in Supplementary Information Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eThe first component loaded on Language, Executive Function and Perception (25%). The second, on Prospective and Retrospective Memory (15%). The third component on Recognition Memory and Working Memory (15%) (Supplementary Information Fig.\u0026nbsp;5).\u003c/p\u003e \u003cp\u003eThe optimal gaussian mixture model for t0 was ellipsoidal, equal volume and orientation, (EVE) with 4 clusters, yielding the lowest BIC, (log-likelihood\u0026thinsp;\u0026minus;\u0026thinsp;198.9, BIC\u0026thinsp;\u0026minus;\u0026thinsp;500.0, mean posterior probability 0.98). 6 clusters were found at t1 and 2 at t2 (Supplementary Information Figs.\u0026nbsp;6 and 7). Sensitivity analysis shows that the bootstrap centroid points group around the cluster solutions across timepoints. This supports the clusters being representations of patient cognition within latent space (Supplementary Information Fig.\u0026nbsp;8).\u003c/p\u003e \u003cp\u003eWithin latent cognitive space, patients separated along the dominant axis of deficit RC1. (Supplementary Information Fig.\u0026nbsp;7). Minimal deficit is characterised by the highest RC1 scores. This phenotype was only seen at t0 (n\u0026thinsp;=\u0026thinsp;23). Mild deficit patients had mildly negative RC1 scores. At t1, this phenotype was represented by two clusters (t1 clusters 1 and 2, n\u0026thinsp;=\u0026thinsp;16) that differed on RC2 but occupied similar positions on RC1. Both converged to a single cluster at t2 (n\u0026thinsp;=\u0026thinsp;19), representing postoperative recovery.\u003c/p\u003e \u003cp\u003eModerate deficit patients had low RC1 scores, with negative RC2 and mildly positive RC3. In the severe deficit group, patients had the lowest RC1 scores, low RC2. At t1, this phenotype was represented by two clusters (t1 clusters 4 and 5, n\u0026thinsp;=\u0026thinsp;9). This phenotype was not present at t2. The Transient deficit (t1 cluster 6, n\u0026thinsp;=\u0026thinsp;3), was assigned to a cluster which was only seen at t1. These patients had mildly negative RC1 but very low RC3 representing an early post-surgical cognitive pattern affecting Recognition Memory and Working Memory. (Supplementary Information Fig.\u0026nbsp;9). Supplementary Information Table\u0026nbsp;6 shows the final phenotypes and RC scores.\u003c/p\u003e \u003cp\u003eThe primary determinant of postoperative cognitive phenotype is preoperative cognition. In general, cognition declined in resection patients in the early postoperative period and did not return to their preoperative baseline by the start of adjuvant chemoradiotherapy (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003ePredictors of postoperative cognition\u003c/h2\u003e \u003cp\u003eModels incorporating preoperative cognitive phenotype predicted postoperative Language, Executive Function as well as Perception, Recognition Memory and Working Memory. Cognitive phenotype provided similar predictive performance to baseline scores for Prospective and Retrospective Memory (Supplementary Information Table\u0026nbsp;7).\u003c/p\u003e \u003cp\u003eModerate deficit (-1.2 [CI: -1.9, -0.49], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and Severe deficit (-1.9 [CI: -2.8, -1.1], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) phenotypes predicted worse postoperative RC1 scores, with significant recovery from t1 to t2 (0.67 [CI: 0.35, 0.99], p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). For RC2, Moderate deficit (-1.5 [CI: -2.3, -0.64], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and Severe deficit (-1.3 [CI: -2.3, -0.32], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) phenotypes predicted worse scores. Male sex associated with higher scores (0.54 [CI: 0.00, 1.1], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.041). For RC3, baseline depression (PHQ) predicted worse scores (-0.05 [CI: -0.09, -0.01], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008), temporal lobe tumours predicted better scores compared to frontal (0.95 [CI: 0.31, 1.6], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.024), and neurophysiological monitoring predicted worse scores (-0.64 [CI: -1.2, \u0026minus;\u0026thinsp;0.08], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.019). The preoperative cluster was not a significant predictor of RC3 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.2) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). (Supplementary Information Table\u0026nbsp;9). Univariable screening of predictors is shown in Supplementary Information Table\u0026nbsp;8.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eCognition in glioma patients has been extensively investigated.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan additionalcitationids=\"CR37 CR38 CR39 CR40\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e But few studies have investigated how surgery affects cognition without the confound of chemoradiotherapy. \u003csup\u003e37,40,41\u003c/sup\u003e\u003c/p\u003e \u003cp\u003ePatients performed\u0026thinsp;\u0026ge;\u0026thinsp;2 SD below the mean compared to normative controls across several domains. Cognitive deficits are common in glioma patients, occurring in up to 55\u0026ndash;82% of presurgical and 49\u0026ndash;84% of postsurgical patients.\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e The percentage of patients with Perception and Calculation deficits significantly increased early after surgery with a trend towards recovery by six-weeks. Despite improvement in individual test performance in the early postoperative period, most patients remained impaired across several domains including Perception and Calculation. RCI improvement in domains such as Perception may be driven by practice effects in unimpaired patients.\u003c/p\u003e \u003cp\u003eA previous meta-analysis found improved cognition in mixed-grade gliomas after surgery.\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e In glioblastoma, studies reported stability, improvement and declines post-surgery.\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e Excluding Biopsy patients may have overestimated the beneficial effects of surgery. \u003csup\u003e17,36,40,41\u003c/sup\u003e We included Biopsy patients to reduce confounding from anaesthesia and hospitalisation. Individual cognitive changes are heterogenous. Early post-surgery (1\u0026ndash;3 weeks), Attention, Verbal and Visual working memory were most likely to decline.\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e 3\u0026ndash;6 months from surgery, Executive function, Verbal memory and Praxis declined in mixed-grade glioma patients, but were stable in glioblastoma patients.\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e The differences between studies could be because presurgical deficit is more severe in high-grade glioma, compared to low-grade glioma patients, thus limiting recovery. Alternatively, patients with low or high baseline performance may have the greatest room for improvement and decline respectively.\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThree PCA components accounted for half test variance: 1) Language, Executive Function and Perception. 2) Prospective and Retrospective Memory and 3) Recognition Memory and Working Memory. Test scores may reflect latent cognitive processes and lesion anatomy.\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eWe identified four preoperative cognitive-phenotypes, Minimal, Mild, Moderate and Severe cognition, Moderate and Severe patients were more likely to decline postoperatively. Reyes et al. used longitudinal pre and post-radiotherapy cognition scores in brain tumour patients to identify three cognitive phenotypes; global impairment, isolated Verbal memory deficits and Minimal impairment.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e Compared to their Study, our approach has the advantage of not requiring the phenotype to fixed across time. We show that resection is associated with a disruption to the phenotype in the early postoperative period with a minority of patients (15.8% [6/38]) improving to a better phenotype by 6-weeks post-surgery. No patient recovered to a preoperative Minimal deficit. Models incorporating preoperative phenotype predict postoperative cognition equivalent or better than baseline RC scores. This may be because the phenotype captures a multivariate pattern of cognitive deficit.\u003c/p\u003e \u003cp\u003eTemporal lobe gliomas in both hemispheres are associated with deficits in Executive function, Verbal memory and Attention.\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e Our results highlight that preoperative cognition and tumour location predict postsurgical cognition.\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e Temporal lobe tumours were associated with improved postoperative Recognition Memory and Working Memory. This may be due to the frontal lobes playing the primary role in cue-based recognition tasks.\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e Future work should investigate whether these phenotypes are correlated with neuroanatomical locations and tumour invasiveness. This suggests cognitive phenotypes may represent latent patterns of brain injury caused by different lesion locations affecting common structural and functional brain networks.\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e,\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThe sample size is modest, especially the number of biopsy patients. This reduces the precision of estimates of perioperative effects when calculating reliable change. 30% of patients dropped out of NPS at t2 mainly due to chemoradiotherapy treatment at different centers but this may still have led to overestimation of cognition in the cohort. However, completers and non-completers did not differ significantly on baseline characteristics. In addition, model-based clustering is sensitive to small sample sizes but our bootstrap analysis supports the overall spatial structure of the cognitive phenotypes.\u003c/p\u003e \u003cp\u003ePrevious work compared the OCS-Bridge NPS tool with traditional pen-and-paper neuropsychological assessments in low-grade glioma patients. They have found that traditional assessments may be more sensitive to deficits in domains such as Attention and Memory.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e However, computerized tools are easier to use, especially for longitudinal testing postoperatively when patients may be fatigued from treatment effects. In addition, OCS-Bridge captures reaction times, increasing its sensitivity to non-verbal perceptual deficits.\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e Nonetheless, NPS offer breadth in place of depth which may limit the variability captured in by PCA.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eFuture work\u003c/h2\u003e \u003cp\u003eOur findings should be validated using an external cohort. Future studies should assess whether preoperative phenotypes predict other outcomes such as survival and HRQOL. The mechanisms underlying different cognitive phenotypes should be investigated by correlating neuroimaging markers of structural and functional brain injury and tumour invasion with cognitive phenotypes.\u003c/p\u003e \u003c/div\u003e"},{"header":"CONCLUSION","content":"\u003cp\u003eWe show that it is possible to identify high grade glioma patients who are at risk of postoperative cognitive deterioration using preoperative cognitive tests. This is important for counselling patients about expected outcomes from surgery. It may also help identify patients for early cognitive rehabilitation\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eYW is supported by Cancer Research UK Clinical Research Training Fellowship and by the CRUK Cambridge Centre. This work was supported by the NIHR HealthTech Research Centre in Brain Injury and the NIHR Cambridge Biomedical Research Centre (NIHR203312) and the Assessing impact of surgically-induced deficits on patient functioning and quality of life (SIND study) (19/WM/0152).\u003c/p\u003e \u003cp\u003e This publication presents independent research funded by the National Institute for Health and Care Research (NIHR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.\u003c/p\u003e \u003cp\u003eThe authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eY.W led the manuscript development, including data collection, analysis and drafting of the manuscript. A.H, T.M, H.Z, M.LR contributed to data analysis and interpretation, including appraisal of analysis methods. R.S contributed to data analysis and evaluation of the manuscript. A.J, R.M, R.C.M, T.S. contributed to data collection and interpretation of the results. S.J.P provided supervision throughout the research process. All authors contributed to data collection, interpretation and critical appraisal of the manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are not openly available due to patient confidentiality and informed consent. Data available from the corresponding author upon reasonable request with a Data Sharing Agreement. Data is in controlled access data storage at University of Cambridge.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eOstrom QT, Gittleman H, Truitt G, Boscia A, Kruchko C, Barnholtz-Sloan JS (2018) CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2011\u0026ndash;2015. Neurooncology 20(suppl4):iv1\u0026ndash;iv86. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuonc/noy131\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noy131\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMansouri A, Lai C, Scales D, Pirouzmand F (2022) A phase II pilot randomized controlled trial to assess the feasibility of the supra-marginal surgical resection of malignant glioma (G-SUMIT: Glioma supra marginal incision trial) study protocol. Pilot Feasibility Stud 8(1):138. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s40814-022-01104-1\u003c/span\u003e\u003cspan address=\"10.1186/s40814-022-01104-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurnet NG, Jefferies SJ, Benson RJ, Hunt DP, Treasure FP (2005) Years of life lost (YLL) from cancer is an important measure of population burden \u0026mdash; and should be considered when allocating research funds. Br J Cancer 92(2):2. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/sj.bjc.6602321\u003c/span\u003e\u003cspan address=\"10.1038/sj.bjc.6602321\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLouis DN, Perry A, Wesseling P et al (2021) The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neurooncology 23(8):1231\u0026ndash;1251. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuonc/noab106\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noab106\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBosma I, Vos MJ, Heimans JJ et al (2007) The course of neurocognitive functioning in high-grade glioma patients1. Neurooncology 9(1):53\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1215/15228517-2006-012\u003c/span\u003e\u003cspan address=\"10.1215/15228517-2006-012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBosma I, Reijneveld JC, Douw L et al (2009) Health-related quality of life of long-term high-grade glioma survivors. Neurooncology 11(1):51\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1215/15228517-2008-049\u003c/span\u003e\u003cspan address=\"10.1215/15228517-2008-049\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCoomans MB, Dirven L, Aaronson N et al (2022) Factors associated with health-related quality of life (HRQoL) deterioration in glioma patients during the progression-free survival period. Neurooncology 24(12):2159\u0026ndash;2169. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuonc/noac097\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noac097\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHenriksson R, Asklund T, Poulsen HS (2011) Impact of therapy on quality of life, neurocognitive function and their correlates in glioblastoma multiforme: a review. J Neurooncol 104(3):639\u0026ndash;646. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-011-0565-x\u003c/span\u003e\u003cspan address=\"10.1007/s11060-011-0565-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNakajima R, Kinoshita M, Okita H, Nakada M (2024) Postsurgical motor function and processing speed as predictors of quality of life in patients with chronic-phase glioblastoma. Acta Neurochir 166(1):357. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00701-024-06245-1\u003c/span\u003e\u003cspan address=\"10.1007/s00701-024-06245-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoll KR, Bradshaw ME, Weinberg JS, Wefel JS (2017) Relationships between neurocognitive functioning, mood, and quality of life in patients with temporal lobe glioma. Psychooncology 26(5):617\u0026ndash;624. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/pon.4046\u003c/span\u003e\u003cspan address=\"10.1002/pon.4046\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiovagnoli AR (2005) Facets and determinants of quality of life in patients with recurrent high grade glioma. J Neurol Neurosurg Psychiatry 76(4):562\u0026ndash;568. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/jnnp.2004.036186\u003c/span\u003e\u003cspan address=\"10.1136/jnnp.2004.036186\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReyes A, Stasenko A, Hopper A et al (2025) Cognitive phenotypes: Unraveling the heterogeneity in cognitive dysfunction among patients with primary brain tumors receiving radiotherapy. Neurooncology 27(2):466\u0026ndash;478. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuonc/noae183\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noae183\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinha R, Stephenson JM, Price SJ A systematic review of cognitive function in patients with glioblastoma undergoing surgery. \u003cem\u003eNeuro-Oncology Practice\u003c/em\u003e. Published online 2019. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nop/npz018\u003c/span\u003e\u003cspan address=\"10.1093/nop/npz018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHope TMH, Halai A, Crinion J, Castelli P, Price CJ, Bowman H (2024) Principal component analysis-based latent-space dimensionality under-estimation, with uncorrelated latent variables. Brain 147(2):e14\u0026ndash;e16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/brain/awad355\u003c/span\u003e\u003cspan address=\"10.1093/brain/awad355\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHalai AD, Woollams AM, Lambon Ralph MA (2017) Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic post-stroke aphasia: Revealing the unique neural correlates of speech fluency, phonology and semantics. Cortex 86:275\u0026ndash;289. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cortex.2016.04.016\u003c/span\u003e\u003cspan address=\"10.1016/j.cortex.2016.04.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao Y, Halai AD, Lambon Ralph MA (2020) Evaluating the granularity and statistical structure of lesions and behaviour in post-stroke aphasia. Brain Commun 2(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/braincomms/fcaa062\u003c/span\u003e\u003cspan address=\"10.1093/braincomms/fcaa062\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNg JCH, See AAQ, Ang TY, Tan LYR, Ang BT, King NKK (2019) Effects of surgery on neurocognitive function in patients with glioma: a meta-analysis of immediate post-operative and long-term follow-up neurocognitive outcomes. J Neurooncol 141(1):167\u0026ndash;182. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-018-03023-9\u003c/span\u003e\u003cspan address=\"10.1007/s11060-018-03023-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eButler RA, Lambon Ralph MA, Woollams AM (2014) Capturing multidimensionality in stroke aphasia: mapping principal behavioural components to neural structures. Brain 137(12):3248\u0026ndash;3266. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/brain/awu286\u003c/span\u003e\u003cspan address=\"10.1093/brain/awu286\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWoollams AM, Halai A, Lambon Ralph MA (2018) Mapping the intersection of language and reading: the neural bases of the primary systems hypothesis. Brain Struct Function 223(8):3769\u0026ndash;3786. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00429-018-1716-z\u003c/span\u003e\u003cspan address=\"10.1007/s00429-018-1716-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRauch S, Wan Y, Halai A et al (2026) Multidimensional cognitive reserve and cognitive outcomes in glioblastoma: a pre- and postoperative analysis. J Neurooncol 177(1):11. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-026-05475-w\u003c/span\u003e\u003cspan address=\"10.1007/s11060-026-05475-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZucchella C, Capone A, Codella V et al (2013) Cognitive rehabilitation for early post-surgery inpatients affected by primary brain tumor: a randomized, controlled trial. J Neurooncol 114(1):93\u0026ndash;100. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-013-1153-z\u003c/span\u003e\u003cspan address=\"10.1007/s11060-013-1153-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVogelbaum MA, Krivosheya D, Borghei-Razavi H et al (2020) Phase 0 and window of opportunity clinical trial design in neuro-oncology: a RANO review. Neuro Oncol 22(11):1568\u0026ndash;1579. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuonc/noaa149\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noaa149\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRomero-Garcia R, Owen M, McDonald A et al (2022) Assessment of neuropsychological function in brain tumor treatment: a comparison of traditional neuropsychological assessment with app-based cognitive screening. Acta Neurochir (Wien) 164(8):2021\u0026ndash;2034. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00701-022-05162-5\u003c/span\u003e\u003cspan address=\"10.1007/s00701-022-05162-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemeyere N, Riddoch MJ, Slavkova ED, Bickerton WL, Humphreys GW (2015) The Oxford Cognitive Screen (OCS): validation of a stroke-specific short cognitive screening tool. Psychol Assess 27(3):883\u0026ndash;894. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/pas0000082\u003c/span\u003e\u003cspan address=\"10.1037/pas0000082\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWan Y, Rahmat R, Price SJ (2020) Deep learning for glioblastoma segmentation using preoperative magnetic resonance imaging identifies volumetric features associated with survival. Acta Neurochir 162(12):3067\u0026ndash;3080. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00701-020-04483-7\u003c/span\u003e\u003cspan address=\"10.1007/s00701-020-04483-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaassen GH, Bossema E, Brand N (2009) Reliable change and practice effects: outcomes of various indices compared. J Clin Exp Neuropsychol 31(3):339\u0026ndash;352. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/13803390802169059\u003c/span\u003e\u003cspan address=\"10.1080/13803390802169059\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRijnen SJM, Butterbrod E, Rutten GJM, Sitskoorn MM, Gehring K (2020) Presurgical Identification of Patients With Glioblastoma at Risk for Cognitive Impairment at 3-Month Follow-up. Neurosurgery 87(6):1119\u0026ndash;1129. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuros/nyaa190\u003c/span\u003e\u003cspan address=\"10.1093/neuros/nyaa190\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRijnen SJM, Meskal I, Bakker M et al (2019) Cognitive outcomes in meningioma patients undergoing surgery: individual changes over time and predictors of late cognitive functioning. Neurooncology 21(7):911\u0026ndash;922. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/neuonc/noz039\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noz039\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStoropoli J, FactorAssumptions Set of Assumptions for Factor and Principal Component Analysis. Published online March 6, 2020:2.0.1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.32614/CRAN.package.FactorAssumptions\u003c/span\u003e\u003cspan address=\"10.32614/CRAN.package.FactorAssumptions\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDinno A paran: Horn\u0026rsquo;s Test of Principal Components/Factors. Published online August 12, 2007:1.5.5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.32614/CRAN.package.paran\u003c/span\u003e\u003cspan address=\"10.32614/CRAN.package.paran\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArno Fritsch (2012) mcclust: Process an MCMC Sample of Clusterings. Published online July 23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.32614/cran.package.mcclust\u003c/span\u003e\u003cspan address=\"10.32614/cran.package.mcclust\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeterson RA, Cavanaugh JE Ordered quantile normalization: a semiparametric transformation built for the cross-validation era. J Appl Stat 47(13\u0026ndash;15):2312\u0026ndash;2327. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/02664763.2019.1630372\u003c/span\u003e\u003cspan address=\"10.1080/02664763.2019.1630372\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBartoń K, MuMIn (2025) Multi-Model Inference. Published online April 1, 2025. Accessed July 9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cran.r-project.org/web/packages/MuMIn/index.html\u003c/span\u003e\u003cspan address=\"https://cran.r-project.org/web/packages/MuMIn/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBates D, M\u0026auml;chler M, Bolker B, Walker S (2015) Fitting Linear Mixed-Effects Models Using lme4. J Stat Softw 67:1\u0026ndash;48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18637/jss.v067.i01\u003c/span\u003e\u003cspan address=\"10.18637/jss.v067.i01\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePsutka JV, Psutka J (2019) Sample size for maximum-likelihood estimates of Gaussian model depending on dimensionality of pattern space. Pattern Recogn 91:25\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.patcog.2019.01.046\u003c/span\u003e\u003cspan address=\"10.1016/j.patcog.2019.01.046\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Loenen IS, Rijnen SJM, Bruijn J, Rutten GJM, Gehring K, Sitskoorn MM (2018) Group Changes in Cognitive Performance After Surgery Mask Changes in Individual Patients with Glioblastoma. World Neurosurg 117:e172\u0026ndash;e179. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.wneu.2018.05.232\u003c/span\u003e\u003cspan address=\"10.1016/j.wneu.2018.05.232\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinha R, Masina R, Morales C et al (2023) A Prospective Study of Longitudinal Risks of Cognitive Deficit for People Undergoing Glioblastoma Surgery Using a Tablet Computer Cognition Testing Battery: Towards Personalized Understanding of Risks to Cognitive Function. JPM 13(2):278. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/jpm13020278\u003c/span\u003e\u003cspan address=\"10.3390/jpm13020278\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoll KR, Bradshaw M, Sheppard D, Wefel JS (2024) Perioperative Neurocognitive Function in Glioma Surgery. Curr Oncol Rep 26(5):466\u0026ndash;476. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11912-024-01522-9\u003c/span\u003e\u003cspan address=\"10.1007/s11912-024-01522-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSantini B, Talacchi A, Squintani G, Casagrande F, Capasso R, Miceli G (2012) Cognitive outcome after awake surgery for tumors in language areas. J Neurooncol 108(2):319\u0026ndash;326. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-012-0817-4\u003c/span\u003e\u003cspan address=\"10.1007/s11060-012-0817-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalacchi A, Santini B, Savazzi S, Gerosa M (2011) Cognitive effects of tumour and surgical treatment in glioma patients. J Neurooncol 103(3):541\u0026ndash;549. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-010-0417-0\u003c/span\u003e\u003cspan address=\"10.1007/s11060-010-0417-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHabets EJJ, Kloet A, Walchenbach R, Vecht CJ, Klein M, Taphoorn MJB (2014) Tumour and surgery effects on cognitive functioning in high-grade glioma patients. Acta Neurochir 156(8):1451\u0026ndash;1459. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00701-014-2115-8\u003c/span\u003e\u003cspan address=\"10.1007/s00701-014-2115-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Kessel E, Baumfalk AE, van Zandvoort MJE, Robe PA, Snijders TJ (2017) Tumor-related neurocognitive dysfunction in patients with diffuse glioma: a systematic review of neurocognitive functioning prior to anti-tumor treatment. J Neurooncol 134(1):9\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-017-2503-z\u003c/span\u003e\u003cspan address=\"10.1007/s11060-017-2503-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFacchini S, Favaretto C, Castellaro M et al (2023) A common low dimensional structure of cognitive impairment in stroke and brain tumors. NeuroImage: Clin 40:103518. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.nicl.2023.103518\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2023.103518\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoll KR, Ziu M, Weinberg JS, Wefel JS (2016) Neurocognitive functioning in patients with glioma of the left and right temporal lobes. J Neurooncol 128(2):323\u0026ndash;331. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11060-016-2114-0\u003c/span\u003e\u003cspan address=\"10.1007/s11060-016-2114-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoll KR, Weinberg JS, Ziu M, Benveniste RJ, Suki D, Wefel JS (2015) Neurocognitive Changes Associated With Surgical Resection of Left and Right Temporal Lobe Glioma. Neurosurgery 77(5):777\u0026ndash;785. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1227/NEU.0000000000000987\u003c/span\u003e\u003cspan address=\"10.1227/NEU.0000000000000987\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZangrossi A, Silvestri E, Bisio M et al (2022) Presurgical predictors of early cognitive outcome after brain tumor resection in glioma patients. NeuroImage: Clin 36:103219. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.nicl.2022.103219\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2022.103219\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcFarland CP, Glisky EL (2009) Frontal lobe involvement in a task of time-based prospective memory. Neuropsychologia 47(7):1660\u0026ndash;1669. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuropsychologia.2009.02.023\u003c/span\u003e\u003cspan address=\"10.1016/j.neuropsychologia.2009.02.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ\u0026uuml;tten K, Mainz V, Delev D et al (2020) Asymmetric tumor-related alterations of network‐specific intrinsic functional connectivity in glioma patients. Hum Brain Mapp 41(16):4549\u0026ndash;4561. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hbm.25140\u003c/span\u003e\u003cspan address=\"10.1002/hbm.25140\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArbula S, Ambrosini E, Della Puppa A et al (2020) Focal left prefrontal lesions and cognitive impairment: A multivariate lesion-symptom mapping approach. Neuropsychologia 136:107253. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuropsychologia.2019.107253\u003c/span\u003e\u003cspan address=\"10.1016/j.neuropsychologia.2019.107253\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Cognition, Glioma, Neuropsychological screening, Surgery","lastPublishedDoi":"10.21203/rs.3.rs-9163008/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9163008/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose\u003c/h2\u003e \u003cp\u003eGlioblastoma patients suffer from cognitive deficits across domains. We used principal component analysis and gaussian mixture modelling to identify preoperative cognitive phenotypes which are at risk of postoperative cognitive impairment.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003e44 patients underwent neuropsychological screening before and after surgery (within 1 week and 4\u0026ndash;6 weeks post-surgery). Patient scores were standardised to a normative reference population and compared using one-sided z-scores. Reliable change indices were calculated with correction for hospitalization effects using a cohort of biopsy patients. Gaussian mixture models were fit after principal component analysis of standardised preoperative scores. Linear mixed effects models were performed to assess predictors of postoperative principal component scores using demographic, clinical and imaging variables.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThree principal components accounted for 55% of the variance in cognitive scores: The first component loaded on Language, Executive Function and Perception (25%). The second, on Prospective and Retrospective Memory (15%). The third component on Recognition Memory and Working Memory (15%). The optimal model identified four preoperative cognitive phenotypes: Minimal (n\u0026thinsp;=\u0026thinsp;23), Mild (n\u0026thinsp;=\u0026thinsp;12), Moderate (n\u0026thinsp;=\u0026thinsp;5) and Severe (n\u0026thinsp;=\u0026thinsp;4) cognition. Post-surgery, preoperative cognition limited recovery and few patients returned to their preoperative baseline. Moderate and Severe phenotypes were significantly associated with worse postoperative Language, Executive Function as well as Prospective and Retrospective Memory.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eWe identified four preoperative cognitive phenotypes which stratify patients into those at risk of postoperative cognitive deficit. Better preoperative cognition was associated with improved postoperative cognition.\u003c/p\u003e","manuscriptTitle":"Multidimensional preoperative cognitive phenotypes predict postoperative cognition in glioblastoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-20 07:16:06","doi":"10.21203/rs.3.rs-9163008/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"708f9279-9573-48c2-9e37-a0bcf700f8f5","owner":[],"postedDate":"March 20th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-21T12:09:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-20 07:16:06","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9163008","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9163008","identity":"rs-9163008","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}