Brain Tumor Resection Research Articles

Safety and Efficacy of Ketorolac After Craniotomy for Tumor Resection

ObjectivePostoperative pain is the most common undesirable outcome after neurosurgery. Ketorolac is a non-steroidal anti-inflammatory drug (NSAID) which is administered parenterally and carries a theoretical increased risk of bleeding. Our study aims to determine whether ketorolac after craniotomy for tumor resection significantly changes the rate of postoperative adverse events, adequately controls pain, and decreases concurrent narcotic usage. MethodsWe performed a retrospective chart review of all adult patients who underwent craniotomy for brain tumor resection from 2013 to 2022. Analysis of patients who received ketorolac and those who did not in the post-operative period were compared for adverse events associated with ketorolac use. Secondary outcomes included patient-reported pain scores and postoperative opioid usage. Results1,114 patients were included, of which 70 received ketorolac in the postoperative period. Patients who received ketorolac often received it after narcotics failed to provide adequate pain control. Patients receiving ketorolac were younger (p=0.001) and had a lower comorbidity index (p=0.041) compared to the non-ketorolac group. Patients receiving ketorolac did not experience a significantly increased rate of bleeding events (p=0.850). Patients recieving ketorolac had significantly higher baseline levels of pain (p=0.018) and opioid usage (p=0.047). When matched for chronic comorbidities including pain disorders, the ketorolac group only displayed higher levels of pain early in the postoperative course (POD 0-1) but not in latter part of the initial postoperative period. ConclusionsKetorolac is a safe and effective option for pain control after craniotomy for tumor resection. Prospective data is needed to better validate these retrospective observations.

SURGICAL OUTCOMES OF TUBULAR RETRACTOR SYSTEMS IN THE MANAGEMENT OF BRAIN TUMOURS: A SYSTEMATIC REVIEW AND META-ANALYSIS

Abstract AIMS Retraction is a critical step in surgical removal of brain tumours, enabling visualisation and resection. Conventional retraction applies focal pressure to overlying tissue, potentially resulting in retraction-induced injury. Tubular retractors (minimally invasive parafascicular surgery) circumferentially distribute pressure to theoretically minimise injury. Whilst previous reviews have provided insight into their potential, the literature has since significantly expanded, necessitating an update. The objective was to synthesise all current evidence regarding surgical outcomes of tubular retractors in resection of brain tumours. METHOD Studies reporting on use of tubular retractors to resect brain tumours in adult patients were eligible. PRISMA guidelines were followed. MEDLINE, Embase and the Cochrane Library were searched to 18th October 2023. Duplicate title/abstract screening, data extraction and risk of bias assessments were conducted. RESULTS Of the 2453 studies screened, 49 were included in the final synthesis, involving 690 patients. Tubular retractor systems assessed were: BrainPath (n=199), VBAS (n=198), METRx (n=63) and other (n=230). Pooled estimate of gross total resection rate was 75% (95% CI 0.65-0.84; I2=83%), whilst pooled estimate of complication rate was 17% (95% CI 0.11-0.24; I2=75%). Sub-group analysis of gross total resection rate was conducted for: glioma (150 patients, pooled prevalence 47%, 95% CI 0.36-0.57; I2=29%), colloid cyst (121 patients, pooled prevalence 100%, CI 0.99-1.00; I2=0%), and metastasis (114 patients, pooled prevalence 82%, 95% CI 0.67-0.94; I2=64%). There was no significant difference between tubular retractor systems with regards to gross total resection or complication rate (p>0.05). CONCLUSION There is mounting evidence for the utility of tubular retractors in resection of brain tumours. However, this evidence remains largely in the form of low-quality case-series, providing proof-of-concept, but limited in its ability to draw more robust comparisons to conventional retraction. Future studies with greater sample sizes and inclusion of control groups are now needed to consolidate the position of tubular retractors in neuro-oncology.

Letter to the Editor Regarding “Incidence and Risk Factors of Delirium Following Brain Tumor Resection: A Retrospective National Inpatient Sample Database Study”

Letter to the Editor Regarding “Incidence and Risk Factors of Delirium Following Brain Tumor Resection: A Retrospective National Inpatient Sample Database Study”

Routine ICU Surveillance after Brain Tumor Surgery: Patient Selection Using Machine Learning.

Background/Objectives: Routine postoperative ICU admission following brain tumor surgery may not benefit selected patients. The objective of this study was to develop a risk prediction instrument for early (within 24 h) postoperative adverse events using machine learning techniques. Methods: Retrospective cohort of 1000 consecutive adult patients undergoing elective brain tumor resection. Nine events/interventions (CPR, reintubation, return to OR, mechanical ventilation, vasopressors, impaired consciousness, intracranial hypertension, swallowing disorders, and death) were chosen as target variables. Potential prognostic features (n = 27) from five categories were chosen and a gradient boosting algorithm (XGBoost) was trained and cross-validated in a 5 × 5 fashion. Prognostic performance, potential clinical impact, and relative feature importance were analyzed. Results: Adverse events requiring ICU intervention occurred in 9.2% of cases. Other events not requiring ICU treatment were more frequent (35% of cases). The boosted decision trees yielded a cross-validated ROC-AUC of 0.81 ± 0.02 (mean ± CI95) when using pre- and post-op data. Using only pre-op data (scheduling decisions), ROC-AUC was 0.76 ± 0.02. PR-AUC was 0.38 ± 0.04 and 0.27 ± 0.03 for pre- and post-op data, respectively, compared to a baseline value (random classifier) of 0.092. Targeting a NPV of at least 95% would require ICU admission in just 15% (pre- and post-op data) or 30% (only pre-op data) of cases when using the prediction algorithm. Conclusions: Adoption of a risk prediction instrument based on boosted trees can support decision-makers to optimize ICU resource utilization while maintaining adequate patient safety. This may lead to a relevant reduction in ICU admissions for surveillance purposes.

The Pediatric Physiatric Posterior Fossa Symptoms scale (3PFSs): Impairments and outcome in pediatric inpatient rehabilitation for posterior fossa brain tumors.

Clinical recognition of the post-operative neurologic sequelae of posterior fossa tumors is inconsistent. This study aimed to characterize functional impairments and recovery trajectories in pediatric patients admitted to inpatient rehabilitation following surgical resection of posterior fossa brain tumors. This study also introduces the Pediatric Physiatric Posterior Fossa Symptom scale (3PFSs) for serial assessment of post-operative symptoms in pediatric posterior fossa brain tumors. This retrospective cohort study included 49 patients aged 1.1 to 19.9 years admitted to a pediatric unit of a free-standing rehabilitation hospital following resection of a posterior fossa brain tumor. Functional Independence Measure for Children (WeeFIM) and 3PFSs scores at admission and discharge were the primary outcome measures. Across the group, WeeFIM score improved from 51.5±23.5 points at admission to 74.2±28.2 points at discharge (t=4.34, p<0.001). The 3PFSs score also showed improvement from 10[IQR=9-12] points at admission to 8[7-10] points at discharge (t=9.3, p<0.0001). While change in both the WeeFIM and 3PFSs captured statistically significant improvement in function, there was low inter-rating correlation (p>0.7). In addition, mortality was correlated with higher discharge 3PFSs score (p=0.007) but not discharge WeeFIM score. In pediatric patients with post-operative neurologic sequelae due to posterior fossa brain tumors, inpatient rehabilitation resulted in global and domain specific functional improvements. This initial application of the 3PFSs demonstrates potential applicability for stratifying patients to appropriate levels of rehabilitation, capturing functionally relevant response to rehabilitation treatment, and prognosticating long-term outcomes. These initial results are promising but require additional validation in a larger cohort.

Toward standardized brain tumor tissue processing protocols in neuro-oncology: a perspective for gliomas and beyond.

Implementation of standardized protocols in neurooncology during the surgical resection of brain tumors is needed to advance the clinical treatment paradigms that use tissue for diagnosis, prognosis, bio-banking, and treatment. Currently recommendations on intraoperative tissue procurement only exist for diffuse gliomas but management of other brain tumor subtypes can also benefit from these protocols. Fresh tissue from surgical resection can now be used for intraoperative diagnostics and functional precision medicine assays. A multidisciplinary neuro-oncology perspective is critical to develop the best avenues for practical standardization. This perspective from the multidisciplinary Oncology Tissue Advisory Board (OTAB) discusses current advances, future directions, and the imperative of adopting standardized protocols for diverse brain tumor entities. There is a growing need for consistent operating room practices to enhance patient care, streamline research efforts, and optimize outcomes.

Procedural volume is linearly associated with mortality, major complications, and readmissions in patients undergoing malignant brain tumor resection.

Improved outcomes have been noted in patients undergoing malignant brain tumor resection at high-volume centers. Studies have arbitrarily chosen high-volume dichotomous cutoffs and have not evaluated volume-outcome associations at specific institutional procedural volumes. We sought to establish the continuous association of volume with patient outcomes and identify cutoffs significantly associated with mortality, major complications, and readmissions. We hypothesized that a linear volume-outcome relationship can estimate likelihood of adverse outcomes when comparing any two volumes. The patient cohort was identified with ICD-10 coding in the Nationwide Readmissions Database(NRD). The association of volume and mortality, major complications, and 30-/90-day readmissions were evaluated in multivariate analyses. Volume was used as a continuous variable with two/three-piece splines, with various knot positions to reflect the best model performance, based on the Quasi Information Criterion(QIC). From 2016 to 2018, 34,486 patients with malignant brain tumors underwent resection. When volume was analyzed as a continuous variable, mortality risk decreased at a steady rate of OR 0.988 per each additional procedure increase for hospitals with 1-65 cases/year(95% CI 0.982-0.993, p < 0.0001). Risk of major complications decreased from 1 to 41 cases/year(OR 0.983, 95% CI 0.979-0.988, p < 0.0001), 30-day readmissions from 1 to 24 cases/year(OR 0.987, 95% CI 0.979-0.995, p = 0.001) and 90-day readmissions from 1 to 23 cases/year(OR 0.989, 95% CI 0.983-0.995, p = 0.0003) and 24-349 cases/year(OR 0.9994, 95% CI 0.999-1, p = 0.01). In multivariate analyses, institutional procedural volume remains linearly associated with mortality, major complications, and 30-/90-day readmission up to specific cutoffs. The resulting linear association can be used to calculate relative likelihood of adverse outcomes between any two volumes.

Advanced Imaging Techniques in Preoperative Planning for Brain Tumor Resection: Evaluating the Impact on Surgical Precision and Neurological Outcomes

This study investigates advanced imaging techniques and their impact on surgical precision and neurological outcomes in preoperative planning for brain tumor resections. Selected studies include large human sample sizes, peer-reviewed research, systematic reviews, and meta-analyses focusing on advanced imaging techniques and their impact on surgical precision and neurological outcomes. Latest imaging modalities experiments on animal studies are considered. Systematic review of recent literature on advanced imaging modalities—such as MRI, fMRI, PET, and DTI—and their application in preoperative planning. Our findings suggest advanced imaging techniques, including Functional MRI, Ultra-High Field MRI, Diffusion Tensor Imaging (DTI), PET/CT and Deuterium Magnetic Resonance Spectroscopy (2H MRS) improve surgical precision and neurological outcomes in brain tumor resections by enhancing tumor targeting, reducing morbidity, and improving resection accuracy. Future advancements should focus on integrating and optimizing these technologies to further improve preoperative planning and patient-specific treatment strategies.

Congenital Brain Tumors: Surgical Outcomes and Long-Term Prognostic Factors

ObjectiveTo evaluate long-term outcomes of surgical resection for congenital brain tumors (CBTs) in infants under one year of age and to identify factors related to survival. MethodsOur retrospective study analyzed infants who underwent gross total (GTR) or subtotal resection (STR) for CBTs between 2001 and 2019. Data were obtained from medical records, including demographics, clinical presentation, diagnosis, tumor characteristics, and presence of hydrocephalus. Additional factors such as pre- and/or postoperative ventriculoperitoneal shunt (VPS) placement and adjuvant chemotherapy or radiotherapy were also reviewed. Cox regression analysis was used to identify factors associated with survival. ResultsThe study included 70 patients, with median age at surgery of 198.5 days, and 28 (40%) were girls. Seizures (31.4%) and vomiting (24.3%) were the most common presenting symptoms. High-grade tumors were present in 29 (41.4%) patients. GTR was achieved in 64.3% of cases, with surgical mortality rate of 7.1%. Overall survival rates at 5 and 10 years were 78% and 63%, respectively. Long-term follow-up data were available for 61 patients (87%), with median follow-up of 74.2 months. Among 45 long-term survivors, 55.5% had neurological sequelae. Factors associated with reduced survival included high-grade, preoperative hydrocephalus, larger tumor size, and VPS placement. The extent of resection improved survival only in low-grade tumor cases. Multivariable Cox regression analysis identified tumor grade and size as independent predictors of poor prognosis. ConclusionsSurgical resection remains crucial for treating CBTs in infants under one year, yet the aggressive nature of malignant tumors results in suboptimal outcomes regarding prognosis.

Is postoperative high dependency care really needed for children undergoing supratentorial brain tumour surgery?

Purpose We present our analysis of the existing Paediatric High Dependency Unit (HDU) admission policy at our institution and discuss our thoughts for its revision in the context of paediatric supratentorial tumour surgery. Materials and methods We screened our prospectively maintained database of all children undergoing supratentorial craniotomy for resection of paediatric brain tumours over a fifteen-year period. The post-operative course of each patient was reviewed, assessing the number of patients who had true HDU needs in the immediate post-operative period, and the relative depth of input from paediatric HDU specialists that each patient received. Results Forty-three patients underwent craniotomy for supratentorial tumour resections during the study period. The median age of the children was 8 years old. Forty-two patients in the study cohort did not require any HDU-level monitoring or treatment post-operatively; all these patients were able to be discharged from HDU to a standard ward bed very rapidly post-operatively. Only one patient (2%) from the study cohort had any tangible HDU needs in the acute post-operative period, comprising of invasive cardiovascular monitoring and repeated blood transfusions. This child’s tumour was known to be large, highly vascular, and invasive pre-operatively. Conclusions We would advocate a rational and nuanced approach with regards to predicting which children are most likely need paediatric HDU care following supratentorial craniotomy for resection of a brain tumour. This rationalisation could improve resource availability and reduce financial burdens upon paediatric neurosurgical units.

Association Between Postoperative Decrease of Albumin and Outcomes in Patients Undergoing Craniotomy for Brain Tumors

Serum albumin reflects nutritional status and is associated with postoperative complications and mortality. Delta albumin (ΔAlb), defined as the difference between preoperative and lowest postoperative levels, could predict complications and mortality, even with postoperative levels above 30g/L prompting albumin infusions. This study aimed to assess how ΔAlb relates to outcomes in craniotomy patients with brain tumors. This retrospective study screened patients diagnosed with a brain tumor who underwent cerebral surgery from a single Chinese hospital between December 2010 and April 2021. Patients were divided into 4 groups based on their ΔAlb levels: <5g/L (normal), 5-9.9g/L (mild ΔAlb), 10-14.9g/L (moderate ΔAlb), and ≥15g/L (severe ΔAlb). The primary outcome was postoperative 30-day mortality. Among the 9660 patients undergoing craniotomy for brain tumors, the median ΔAlb level after craniotomy was 7.3g/L. ΔAlb was associated with increased postoperative 30-day mortality; odds ratios for mild, moderate, and severe ΔAlb were 1.93 (95% confidence interval [CI], 1.17-3.18, P= 0.01), 2.21 (95% CI, 1.28-3.79, P= 0.004), and 7.26 (95% CI, 4.19-12.58, P<0.01), respectively. Significantly, ΔAlb >5g/L was found to have a strong association with a higher risk of mortality, even when the nadir Alb remained greater than 30g/L (odds ratio, 1.84; 95% CI, 1.13-3.00, P= 0.014). Among patients undergoing craniotomy for brain tumor resection, a mild degree of ΔAlb was associated with increased 30-day mortality, even if the nadir Alb remained greater than 30g/L. Moreover, ΔAlb was associated with postoperative complications and longer lengths of stay.

Open access segmentations of intraoperative brain tumor ultrasound images.

Registration and segmentation of magnetic resonance (MR) and ultrasound (US) images could play an essential role in surgical planning and resectioning brain tumors. However, validating these techniques is challenging due to the scarcity of publicly accessible sources with high-quality ground truth information. To this end, we propose a unique set of segmentations (RESECT-SEG) of cerebral structures from the previously published RESECT dataset to encourage a more rigorous development and assessment of image-processing techniques forneurosurgery. The RESECT database consists of MR and intraoperative US (iUS) images of 23 patients who underwent brain tumor resection surgeries. The proposed RESECT-SEG dataset contains segmentations of tumor tissues, sulci, falx cerebri, and resection cavity of the RESECT iUS images. Two highly experienced neurosurgeons validated the quality of the segmentations. Segmentations are provided in 3D NIFTI format in the OSF open-science platform: https://osf.io/jv8bk. The proposed RESECT-SEG dataset includes segmentations of real-world clinical US brain images that could be used to develop and evaluate segmentation and registration methods. Eventually, this dataset could further improve the quality of image guidance inneurosurgery.

Intraoperative goal-directed fluid management and postoperative brain edema in patients having high-grade gliomas resections: a randomized trial.

Patients with high-grade gliomas often have severe brain edema. Goal-directed fluid management protects neurological function, but whether reduces postoperative brain edema remains unknown. Patients having elective resection of supratentorial malignant gliomas were randomly assigned to goal-directed versus routine fluid management. Patients assigned to goal-directed management group were given 3mL kg-1 hydroxyethyl starch solution when stroke volume variation exceeded 15% for 5 minutes. Fluid was managed per routine by attending anesthesiologists in reference patients. The primary outcome was cerebral edema volume after surgery as assessed by computerized tomography. A total of 480 eligible patients were randomly assigned to the goal-directed (n = 240) or the routine fluid management group (n = 240). The amounts of crystalloid (5.4 vs. 7.0ml kg-1 hour-1, P < 0.001), colloid (1.1 vs. 1.7ml kg-1 hour-1, P < 0.001), and overall fluid balance (0.3 vs. 1.9ml kg-1 hour-1, P < 0.001) were significantly lower in goal-directed fluid management. There was no significant difference in postoperative brain edema volume between groups (36.0cm3 vs. 38.9cm3, mean difference: 0.18cm3, 95% CI: -5.7 to 5.9). Goal-directed patients had lower intraoperative dural tension (risk ratio: 0.63, 95% CI: 0.50 to 0.80, P < 0.001). There was no significant difference in Karnofsky Performance Status between the two groups at 30 days after surgery. Goal-directed fluid therapy substantially reduced intravenous fluid volumes, but did not reduce postoperative brain edema in patients having brain tumor resections.

Perioperative Management of Antiplatelet and Anticoagulation in Brain Tumor Surgery: A Survey of International Practices

Perioperative management of antithrombotic therapy is a necessary preprocedural consideration for patients prescribed direct oral anticoagulants (DOACs), vitamin K antagonists, or antiplatelet medications. There is a lack of evidence-based guidelines to help inform decision-making in managing antiplatelet and anticoagulation medications in the perioperative period around brain tumor resection. The objective of this study was to provide an example of the heterogeneity in practice and raise awareness for the need to create standardized guidelines for managing these medications. A survey was sent to a list of over 800 international neurosurgeons who are members of the Neurosurgery Research Listserv. The survey comprised 70 questions assessing individual practices for managing thromboprophylaxis, antiplatelet medications, and anticoagulation in the perioperative period. The survey was sent via e-mail invitation between March 2021 and June2021. A total of 72 surgeons responded to the survey. There was no difference in medication management preoperatively or postoperatively when comparing intra- and extra-axial tumor resections. Cessation of antiplatelet medications varied between 3 and 11 days while restart varied between 1 and 14 days. Preoperative management of vitamin K antagonists varied between indication for use (P < 0.001) while DOAC management did not. In our group of respondents, 90% started heparin products within 5 days of surgery, while the same fraction restarted DOAC within 14 days. Respondents demonstrated significant heterogeneity in their perioperative management of antiplatelet and anticoagulation medication in brain tumor resection. This may lead to an unacceptable level of heterogeneity in practice that has the potential to cause patient harm due to errors in medication management.

Tumor-Specific Alterations in Motor Cortex Excitability and Tractography of the Corticospinal Tract-A Navigated Transcranial Magnetic Stimulation Study.

Non-invasive brain mapping using navigated transcranial magnetic stimulation (nTMS) is a valuable tool prior to resection of malignant brain tumors. With nTMS motor mapping, it is additionally possible to analyze the function of the motor system and to evaluate tumor-induced neuroplasticity. Distinct changes in motor cortex excitability induced by certain malignant brain tumors are a focal point of research. A retrospective single-center study was conducted involving patients with malignant brain tumors. Clinical data, resting motor threshold (rMT), and nTMS-based tractography were evaluated. The interhemispheric rMT-ratio (rMTTumor/rMTControl) was calculated for each extremity and considered pathological if it was >110% or <90%. Distances between the corticospinal tract and the tumor (lesion-to-tract-distance - LTD) were measured. 49 patients were evaluated. 16 patients (32.7%) had a preoperative motor deficit. The cohort comprised 22 glioblastomas (44.9%), 5 gliomas of Classification of Tumors of the Central Nervous System (CNS WHO) grade 3 (10.2%), 6 gliomas of CNS WHO grade 2 (12.2%) and 16 cerebral metastases (32.7%). 26 (53.1%) had a pathological rMT-ratio for the upper extremity and 35 (71.4%) for the lower extremity. All patients with tumor-induced motor deficits had pathological interhemispheric rMT-ratios, and presence of tumor-induced motor deficits was associated with infiltration of the tumor to the nTMS-positive cortex (p = 0.04) and shorter LTDs (all p < 0.021). Pathological interhemispheric rMT-ratio for the upper extremity was associated with cerebral metastases, but not with gliomas (p = 0.002). Our study underlines the diagnostic potential of nTMS motor mapping to go beyond surgical risk stratification. Pathological alterations in motor cortex excitability can be measured with nTMS mapping. Pathological cortical excitability was more frequent in cerebral metastases than in gliomas.

Interventional neurorehabilitation for glioma patients: A systematic review

Abstract Harnessing the neuroplastic potential of the human brain is being increasingly recognized as an important neuro-oncological paradigm to facilitate safe resection of brain tumors while preserving neurological function and quality of life. Interventional neurorehabilitation, employing both invasive and noninvasive neuromodulation techniques, represents an important emerging therapeutic strategy to induce or enhance neural plasticity to promote functional recovery in brain tumor patients. This study aimed to conduct a comprehensive review of interventional neurorehabilitation techniques for glioma patients. Methods: In accordance with PRISMA guidelines, searches of Medline, Embase, Web of Science, APA PsycINFO, and Cochrane were undertaken from database inception to November 28, 2023. Studies reporting on neuromodulation applied to glioma patients were included. Results: Seven studies reporting findings from 118 patients met the inclusion criteria. Three neuromodulation techniques were identified and included transcranial magnetic stimulation (TMS) reported in 5 out of 7 (71.4%) studies; transcranial direct current stimulation (tDCS); and continuous cortical electrical stimulation (cCES) using grid electrodes, reported in one study each. All studies applying noninvasive stimulation to ameliorate postoperative deficits demonstrated an improvement on at least one outcome measure. The 2 studies applying tDCS and cCES to induce plasticity reported evidence of functional reorganization. Conclusions: There is emerging evidence of benefits of neuromodulation to improve postoperative outcome in glioma patients. In the current literature, noninvasive stimulation has shown to have a favorable safety profile. Large-scale, double-blind, sham-controlled trials are warranted to further investigate the effectiveness of these interventions for modulating different cognitive networks in patients undergoing glioma surgery.

Utilizing connectome fingerprinting functional MRI models for motor activity prediction in presurgical planning: A feasibility study.

Presurgical planning prior to brain tumor resection is critical for the preservation of neurologic function post-operatively. Neurosurgeons increasingly use advanced brain mapping techniques pre- and intra-operatively to delineate brain regions which are "eloquent" and should be spared during resection. Functional MRI (fMRI) has emerged as a commonly used non-invasive modality for individual patient mapping of critical cortical regions such as motor, language, and visual cortices. To map motor function, patients are scanned using fMRI while they perform various motor tasks to identify brain networks critical for motor performance, but it may be difficult for some patients to perform tasks in the scanner due to pre-existing deficits. Connectome fingerprinting (CF) is a machine-learning approach that learns associations between resting-state functional networks of a brain region and the activations in the region for specific tasks; once a CF model is constructed, individualized predictions of task activation can be generated from resting-state data. Here we utilized CF to train models on high-quality data from 208 subjects in the Human Connectome Project (HCP) and used this to predict task activations in our cohort of healthy control subjects (n = 15) and presurgical patients (n = 16) using resting-state fMRI (rs-fMRI) data. The prediction quality was validated with task fMRI data in the healthy controls and patients. We found that the task predictions for motor areas are on par with actual task activations in most healthy subjects (model accuracy around 90%-100% of task stability) and some patients suggesting the CF models can be reliably substituted where task data is either not possible to collect or hard for subjects to perform. We were also able to make robust predictions in cases in which there were no task-related activations elicited. The findings demonstrate the utility of the CF approach for predicting activations in out-of-sample subjects, across sites and scanners, and in patient populations. This work supports the feasibility of the application of CF models to presurgical planning, while also revealing challenges to be addressed in future developments. PRACTITIONER POINTS: Precision motor network prediction using connectome fingerprinting. Carefully trained models' performance limited by stability of task-fMRI data. Successful cross-scanner predictions and motor network mapping in patients with tumor.

Value of the Acute Physiology and Chronic Health Evaluation II (APACHE II) score in predicting hospital mortality for postoperative brain tumor patients in intensive care units in Japan: A retrospective case-control study

ObjectiveAcute Physiology and Chronic Health Evaluation II (APACHE II) is based on the data of intensive care unit (ICU) patients and often correlates with disease severity and prognosis. However, no prognostic predictors exist based on ICU admission data for patients with brain tumors, and no studies have reported an association between APACHE II and prognosis in patients with brain tumors. The Japanese Intensive Care Patients Database (JIPAD) was established to improve the quality of care delivered in intensive care medicine in Japan. We used JIPAD to examine factors associated with in-hospital mortality based on available data of postoperative patients with brain tumors admitted to the ICU. MethodsPatients aged ≥16 years enrolled in JIPAD between April 2015 and March 2018 after surgical brain tumor resection or biopsy of brain tumors. We examined factors related to outcomes at discharge based on blood tests and medical procedures performed during ICU admission, tumor type, and APACHE II score. ResultsAmong the 1454 patients (male:female ratio: 1:1.1, mean age: 62 years) in the study, 32 (2.2 %) died during hospital stay. In multivariate analysis, male sex (odds ratio [OR] 2.70, [95 % confidence interval, CI 1.22–6.00]), malignant tumor (OR 2.51 [95 % CI 1.13–5.55]), and APACHE II score ≥15 (OR 2.51 [95 % CI 3.08–14.3]) were significantly associated with in-hospital mortality. ConclusionBy picking up cases with a high risk of in-hospital death at an early stage, it is possible to improve methods of treatment and support for the patient's family.

Towards machine learning-based quantitative hyperspectral image guidance for brain tumor resection

BackgroundComplete resection of malignant gliomas is hampered by the difficulty in distinguishing tumor cells at the infiltration zone. Fluorescence guidance with 5-ALA assists in reaching this goal. Using hyperspectral imaging, previous work characterized five fluorophores’ emission spectra in most human brain tumors.MethodsIn this paper, the effectiveness of these five spectra was explored for different tumor and tissue classification tasks in 184 patients (891 hyperspectral measurements) harboring low- (n = 30) and high-grade gliomas (n = 115), non-glial primary brain tumors (n = 19), radiation necrosis (n = 2), miscellaneous (n = 10) and metastases (n = 8). Four machine-learning models were trained to classify tumor type, grade, glioma margins, and IDH mutation.ResultsUsing random forests and multilayer perceptrons, the classifiers achieve average test accuracies of 84–87%, 96.1%, 86%, and 91% respectively. All five fluorophore abundances vary between tumor margin types and tumor grades (p < 0.01). For tissue type, at least four of the five fluorophore abundances are significantly different (p < 0.01) between all classes.ConclusionsThese results demonstrate the fluorophores’ differing abundances in different tissue classes and the value of the five fluorophores as potential optical biomarkers, opening new opportunities for intraoperative classification systems in fluorescence-guided neurosurgery.

Spectral library and method for sparse unmixing of hyperspectral images in fluorescence guided resection of brain tumors.

Through spectral unmixing, hyperspectral imaging (HSI) in fluorescence-guided brain tumor surgery has enabled the detection and classification of tumor regions invisible to the human eye. Prior unmixing work has focused on determining a minimal set of viable fluorophore spectra known to be present in the brain and effectively reconstructing human data without overfitting. With these endmembers, non-negative least squares regression (NNLS) was commonly used to compute the abundances. However, HSI images are heterogeneous, so one small set of endmember spectra may not fit all pixels well. Additionally, NNLS is the maximum likelihood estimator only if the measurement is normally distributed, and it does not enforce sparsity, which leads to overfitting and unphysical results. In this paper, we analyzed 555666 HSI fluorescence spectra from 891 ex vivo measurements of patients with various brain tumors to show that a Poisson distribution indeed models the measured data 82% better than a Gaussian in terms of the Kullback-Leibler divergence, and that the endmember abundance vectors are sparse. With this knowledge, we introduce (1) a library of 9 endmember spectra, including PpIX (620 nm and 634 nm photostates), NADH, FAD, flavins, lipofuscin, melanin, elastin, and collagen, (2) a sparse, non-negative Poisson regression algorithm to perform physics-informed unmixing with this library without overfitting, and (3) a highly realistic spectral measurement simulation with known endmember abundances. The new unmixing method was then tested on the human and simulated data and compared to four other candidate methods. It outperforms previous methods with 25% lower error in the computed abundances on the simulated data than NNLS, lower reconstruction error on human data, better sparsity, and 31 times faster runtime than state-of-the-art Poisson regression. This method and library of endmember spectra can enable more accurate spectral unmixing to aid the surgeon better during brain tumor resection.