International practice variation in perioperative laboratory testing in glioblastoma patients-a retrospective cohort study.

PURPOSE: Although standard-of-care has been defined for the treatment of glioblastoma patients, substantial practice variation exists in the day-to-day clinical management. This study aims to compare the use of laboratory tests in the perioperative care of glioblastoma patients between two tertiary academic centers-Brigham and Women's Hospital (BWH), Boston, USA, and University Medical Center Utrecht (UMCU), Utrecht, the Netherlands. METHODS: All glioblastoma patients treated according to standard-of-care between 2005 and 2013 were included. We compared the number of blood drawings and laboratory tests performed during the 70-day perioperative period using a Poisson regression model, as well as the estimated laboratory costs per patient. Additionally, we compared the likelihood of an abnormal test result using a generalized linear mixed effects model. RESULTS: After correction for age, sex, IDH1 status, postoperative KPS score, length of stay, and survival status, the number of blood drawings and laboratory tests during the perioperative period were 3.7-fold (p < 0.001) and 4.7-fold (p < 0.001) higher, respectively, in BWH compared to UMCU patients. The estimated median laboratory costs per patient were 82 euros in UMCU and 256 euros in BWH. Furthermore, the likelihood of an abnormal test result was lower in BWH (odds ratio [OR] 0.75, p < 0.001), except when the prior test result was abnormal as well (OR 2.09, p < 0.001). CONCLUSIONS: Our results suggest a substantially lower clinical threshold for ordering laboratory tests in BWH compared to UMCU. Further investigating the clinical consequences of laboratory testing could identify over and underuse, decrease healthcare costs, and reduce unnecessary discomfort that patients are exposed to.

Beta-blockers and glioma: a systematic review of preclinical studies and clinical results.

Given the median survival of 15 months after diagnosis, novel treatment strategies are needed for glioblastoma. Beta-blockers have been demonstrated to inhibit angiogenesis and tumor cell proliferation in various cancer types. The aim of this study was to systematically review the evidence on the effect of beta-blockers on glioma growth. A systematic literature search was performed in the PubMed, Embase, Google Scholar, Web of Science, and Cochrane Central to identify all relevant studies. Preclinical studies concerning the pharmacodynamic effects of beta-blockers on glioma growth and proliferation were included, as well as clinical studies that studied the effect of beta-blockers on patient outcomes according to PRISMA guidelines. Among the 980 citations, 10 preclinical studies and 1 clinical study were included after title/abstract and full-text screening. The following potential mechanisms were identified: reduction of glioma cell proliferation (n = 9), decrease of glioma cell migration (n = 2), increase of drug sensitivity (n = 1), induction of glioma cell death (n = 1). Beta-blockers affect glioma proliferation by inducing a brief reduction of cAMP and a temporary cell cycle arrest in vitro. Contrasting results were observed concerning glioma cell migration. The identified clinical study did not find an association between beta-blockers and survival in glioma patients. Although preclinical studies provide scarce evidence for the use of beta-blockers in glioma, they identified potential pathways for targeting glioma. Future studies are needed to clarify the effect of beta-blockers on clinical endpoints including survival outcomes in glioma patients to scrutinize the value of beta-blockers in glioma care.

Survival prediction of glioblastoma patients-are we there yet? A systematic review of prognostic modeling for glioblastoma and its clinical potential.

Glioblastoma is associated with a poor prognosis. Even though survival statistics are well-described at the population level, it remains challenging to predict the prognosis of an individual patient despite the increasing number of prognostic models. The aim of this study is to systematically review the literature on prognostic modeling in glioblastoma patients. A systematic literature search was performed to identify all relevant studies that developed a prognostic model for predicting overall survival in glioblastoma patients following the PRISMA guidelines. Participants, type of input, algorithm type, validation, and testing procedures were reviewed per prognostic model. Among 595 citations, 27 studies were included for qualitative review. The included studies developed and evaluated a total of 59 models, of which only seven were externally validated in a different patient cohort. The predictive performance among these studies varied widely according to the AUC (0.58-0.98), accuracy (0.69-0.98), and C-index (0.66-0.70). Three studies deployed their model as an online prediction tool, all of which were based on a statistical algorithm. The increasing performance of survival prediction models will aid personalized clinical decision-making in glioblastoma patients. The scientific realm is gravitating towards the use of machine learning models developed on high-dimensional data, often with promising results. However, none of these models has been implemented into clinical care. To facilitate the clinical implementation of high-performing survival prediction models, future efforts should focus on harmonizing data acquisition methods, improving model interpretability, and externally validating these models in multicentered, prospective fashion.

Does Artificial Intelligence Outperform Natural Intelligence in Interpreting Musculoskeletal Radiological Studies? A Systematic Review.

BACKGROUND: Machine learning (ML) is a subdomain of artificial intelligence that enables computers to abstract patterns from data without explicit programming. A myriad of impactful ML applications already exists in orthopaedics ranging from predicting infections after surgery to diagnostic imaging. However, no systematic reviews that we know of have compared, in particular, the performance of ML models with that of clinicians in musculoskeletal imaging to provide an up-to-date summary regarding the extent of applying ML to imaging diagnoses. By doing so, this review delves into where current ML developments stand in aiding orthopaedists in assessing musculoskeletal images. QUESTIONS/PURPOSES: This systematic review aimed (1) to compare performance of ML models versus clinicians in detecting, differentiating, or classifying orthopaedic abnormalities on imaging by (A) accuracy, sensitivity, and specificity, (B) input features (for example, plain radiographs, MRI scans, ultrasound), (C) clinician specialties, and (2) to compare the performance of clinician-aided versus unaided ML models. METHODS: A systematic review was performed in PubMed, Embase, and the Cochrane Library for studies published up to October 1, 2019, using synonyms for machine learning and all potential orthopaedic specialties. We included all studies that compared ML models head-to-head against clinicians in the binary detection of abnormalities in musculoskeletal images. After screening 6531 studies, we ultimately included 12 studies. We conducted quality assessment using the Methodological Index for Non-randomized Studies (MINORS) checklist. All 12 studies were of comparable quality, and they all clearly included six of the eight critical appraisal items (study aim, input feature, ground truth, ML versus human comparison, performance metric, and ML model description). This justified summarizing the findings in a quantitative form by calculating the median absolute improvement of the ML models compared with clinicians for the following metrics of performance: accuracy, sensitivity, and specificity. RESULTS: ML models provided, in aggregate, only very slight improvements in diagnostic accuracy and sensitivity compared with clinicians working alone and were on par in specificity (3% (interquartile range [IQR] -2.0% to 7.5%), 0.06% (IQR -0.03 to 0.14), and 0.00 (IQR -0.048 to 0.048), respectively). Inputs used by the ML models were plain radiographs (n = 8), MRI scans (n = 3), and ultrasound examinations (n = 1). Overall, ML models outperformed clinicians more when interpreting plain radiographs than when interpreting MRIs (17 of 34 and 3 of 16 performance comparisons, respectively). Orthopaedists and radiologists performed similarly to ML models, while ML models mostly outperformed other clinicians (outperformance in 7 of 19, 7 of 23, and 6 of 10 performance comparisons, respectively). Two studies evaluated the performance of clinicians aided and unaided by ML models; both demonstrated considerable improvements in ML-aided clinician performance by reporting a 47% decrease of misinterpretation rate (95% confidence interval [CI] 37 to 54; p < 0.001) and a mean increase in specificity of 0.048 (95% CI 0.029 to 0.068; p < 0.001) in detecting abnormalities on musculoskeletal images. CONCLUSIONS: At present, ML models have comparable performance to clinicians in assessing musculoskeletal images. ML models may enhance the performance of clinicians as a technical supplement rather than as a replacement for clinical intelligence. Future ML-related studies should emphasize how ML models can complement clinicians, instead of determining the overall superiority of one versus the other. This can be accomplished by improving transparent reporting, diminishing bias, determining the feasibility of implantation in the clinical setting, and appropriately tempering conclusions. LEVEL OF EVIDENCE: Level III, diagnostic study.

Preoperative functional MRI use in neurooncology patients: a clinician survey.

OBJECTIVE: Functional MRI (fMRI) is increasingly being investigated for use in neurosurgical patient care. In the current study, the authors characterize the clinical use of fMRI by surveying neurosurgeons' use of and attitudes toward fMRI as a surgical planning tool in neurooncology patients. METHODS: A survey was developed to inquire about clinicians' use of and experiences with preoperative fMRI in the neurooncology patient population, including example case images. The survey was distributed to all neurosurgical departments with a residency program in the US. RESULTS: After excluding incomplete surveys and responders that do not use fMRI (n = 11), 50 complete responses were included in the final analysis. Responders were predominantly from academic programs (88%), with 20 years or more in practice (40%), with a main area of practice in neurooncology (48%) and treating an adult population (90%). All 50 responders currently use fMRI in neurooncology patients, mostly for low- (94%) and high-grade glioma (82%). The leading decision factors for ordering fMRI were location of mass in dominant hemisphere, location in a functional area, motor symptoms, and aphasia. Across 10 cases, language fMRI yielded the highest interrater reliability agreement (Fleiss' kappa 0.437). The most common reasons for ordering fMRI were to identify language laterality, plan extent of resection, and discuss neurological risks with patients. Clinicians reported that fMRI results were not obtained when ordered a median 10% of the time and were suboptimal a median 27% of the time. Of responders, 70% reported that they had ever resected an fMRI-positive functional site, of whom 77% did so because the site was "cleared" by cortical stimulation. Responders reported disagreement between fMRI and awake surgery 30% of the time. Overall, 98% of responders reported that if results of fMRI and intraoperative mapping disagreed, they would rely on intraoperative mapping. CONCLUSIONS: Although fMRI is increasingly being adopted as a practical preoperative planning tool for brain tumor resection, there remains a substantial degree of discrepancy with regard to its current use and presumed utility. There is a need for further research to evaluate the use of preoperative fMRI in neurooncology patients. As fMRI continues to gain prominence, it will be important for clinicians to collectively share best practices and develop guidelines for the use of fMRI in the preoperative planning phase of brain tumor patients.

Machine learning reveals multimodal MRI patterns predictive of isocitrate dehydrogenase and 1p/19q status in diffuse low- and high-grade gliomas.

PURPOSE: Isocitrate dehydrogenase (IDH) and 1p19q codeletion status are importantin providing prognostic information as well as prediction of treatment response in gliomas. Accurate determination of the IDH mutation status and 1p19q co-deletion prior to surgery may complement invasive tissue sampling and guide treatment decisions. METHODS: Preoperative MRIs of 538 glioma patients from three institutions were used as a training cohort. Histogram, shape, and texture features were extracted from preoperative MRIs of T1 contrast enhanced and T2-FLAIR sequences. The extracted features were then integrated with age using a random forest algorithm to generate a model predictive of IDH mutation status and 1p19q codeletion. The model was then validated using MRIs from glioma patients in the Cancer Imaging Archive. RESULTS: Our model predictive of IDH achieved an area under the receiver operating characteristic curve (AUC) of 0.921 in the training cohort and 0.919 in the validation cohort. Age offered the highest predictive value, followed by shape features. Based on the top 15 features, the AUC was 0.917 and 0.916 for the training and validation cohort, respectively. The overall accuracy for 3 group prediction (IDH-wild type, IDH-mutant and 1p19q co-deletion, IDH-mutant and 1p19q non-codeletion) was 78.2% (155 correctly predicted out of 198). CONCLUSION: Using machine-learning algorithms, high accuracy was achieved in the prediction of IDH genotype in gliomas and moderate accuracy in a three-group prediction including IDH genotype and 1p19q codeletion.

Treatment and survival of osteosarcoma and Ewing sarcoma of the skull: a SEER database analysis.

BACKGROUND: Common primary bone tumors include osteosarcomas (OSC) and Ewing sarcomas (EWS). The skull is a rare site, and literature about their treatment and survival is scarce. Using the Surveillance, Epidemiology, and End Results (SEER) database, this study aims to assess the treatment and survival of skull OSC and skull EWS, as well as predictors for survival. METHODS: Skull OSC and EWS cases were obtained from the SEER database. Patient and tumor characteristics, treatment modalities, and survival were extracted. Overall survival (OS) was assessed using multivariable Cox proportional hazard regression stratified by tumor histology. Kaplan-Meier curves were constructed for OS comparing OSC and EWS, as well as histological subtypes in OSC. RESULTS: A total of 321 skull OSC and 80 skull EWS patients were registered from 1973 to 2013. EWS was more common in younger patients (p < 0.001). Resection was the predominant treatment strategy (80.1%), frequently in combination with adjuvant radiotherapy (30.4%). The 5-year survival rate varied significantly between OSC and EWS (51.0% versus 68.5%, p = 0.02). Kaplan-Meier curves show that EWS had a significantly better survival compared to OSC. Comparing histological subtypes of skull OSC, chondroblastic OSC had the best OS, Paget OSC the worst. Older age, tumor advancement, no surgical treatment, and the use of radiotherapy were identified as independent predictors of decreased OS in skull OSC. CONCLUSION: Overall prognosis is better for EWS compared to OSC. Chondroblastic OSC have the best overall survival, while OSC associated with Paget's disease of the bone has the poorest overall survival.

Randomized controlled trials comparing surgery to non-operative management in neurosurgery: a systematic review.

BACKGROUND: A randomized controlled trial (RCT) remains the pinnacle of clinical research design. However, RCTs in neurosurgery, especially those comparing surgery to non-operative treatment, are rare and their relevance and applicability have been questioned. This study set out to assess trial design and quality and identify their influence on outcomes in recent neurosurgical trials that compare surgery to non-operative treatment. METHODS: From 2000 to 2017, PubMed and Embase databases and four trial registries were searched. RCTs were evaluated for study design, funding, adjustments to reported outcome measures, accrual of patients, and academic impact. RESULTS: Eighty-two neurosurgical RCTs were identified, 40 in spine disorders, 19 neurovascular and neurotrauma, 11 functional neurosurgery, ten peripheral nerve, and two pituitary surgery. Eighty-four RCTs were registered, of which some are ongoing. Trial registration rate differed per subspecialty. Funding was mostly from non-industry institutions (58.5%), but 25.6% of RCTs did not report funding sources. 36.4% of RCTs did not report a difference between surgical and non-operative treatment, 3.7% favored non-operative management. Primary and secondary outcome measures were changed in 13.2% and 34.2% of RCTs respectively and varied by subspecialty. 41.9% of RCTs subtracted ≥ 10% of the anticipated accrual and 12.9% of RCTs added ≥ 10%. 7.3% of registered RCTs were terminated, mostly due to too slow recruitment. Subspecialty, registration, funding, masking, population size, and changing outcome measures were not significantly associated with a reported benefit of surgery. High Jadad scores (≥ 4) were negatively associated with a demonstration of surgical benefit (P < 0.05). CONCLUSIONS: Neurosurgical RCTs comparing surgical to non-operative treatment often find a benefit for surgical treatment. Changes to outcome measurements and anticipated accrual are common and funding sources are not always reported.

A nationwide analysis of 30-day adverse events, unplanned readmission, and length of hospital stay after peripheral nerve surgery in extremities and the brachial plexus.

BACKGROUND: Little is known on adverse events and their timing after peripheral nerve surgery in extremities. The aim of this study is to identify predictors and typical timing of complications, unplanned readmission, and length of hospital stay for patients undergoing peripheral nerve surgery in the extremities. METHODS: Data were extracted from the National Surgical Quality Improvement Program (NSQIP) registry from 2005 to 2015. Adult patients undergoing peripheral nerve surgery in the extremities were included. A subgroup analysis was performed for brachial plexus operations. Multivariable logistic regression was performed to identify predictors of any complication, surgical site infection, unplanned readmission, and reoperation. RESULTS: A total of 2,840 patients were identified; 628 were brachial plexus operations. Overall complications were 4.4% and 7.0%, respectively. Median time for occurrence of any complication was 8 days. The most common complications were wound-related (1.7%), which occurred at a median of 15 days postoperatively. Reoperation occurred in 1.8% of all cases; most commonly for musculoskeletal repair (16.7%). Unplanned readmissions occurred in 2.3% and were most often due to wound-related problems (24.1%). Preoperatively contaminated wounds, inpatient procedures, and longer operative time seemed to have the most influence on all adverse events. In brachial plexus pathology, insulin-dependent diabetes and emergency cases also negatively affected outcomes. CONCLUSIONS: Complications usually occur one to two weeks postoperatively. Preoperatively contaminated wounds, inpatient procedures, and longer operative times influence outcome. Anatomical level of operation results in significantly different lengths of hospital stay; brachial plexus pathology has the longest length of stay.

Venous thromboembolism and intracranial hemorrhage after craniotomy for primary malignant brain tumors: a National Surgical Quality Improvement Program analysis.

Venous thromboembolism (VTE), including deep venous thrombosis (DVT) and pulmonary embolism (PE), frequently complicates the postoperative course of primary malignant brain tumor patients. Thromboprophylactic anticoagulation is commonly used to prevent VTE at the risk of intracranial hemorrhage (ICH). We extracted all patients who underwent craniotomy for a primary malignant brain tumor from the National Surgical Quality Improvement Program (NSQIP) registry (2005-2015) to perform a time-to-event analysis and identify relevant predictors of DVT, PE, and ICH within 30 days after surgery. Among the 7376 identified patients, the complication rates were 2.6, 1.5, and 1.3% for DVT, PE, and ICH, respectively. VTE was the second-most common major complication and third-most common reason for readmission. ICH was the most common reason for reoperation. The increased risk of VTE extends beyond the period of hospitalization, especially for PE, whereas ICH occurred predominantly within the first days after surgery. Older age and higher BMI were overall predictors of VTE. Dependent functional status and longer operative times were predictive for VTE during hospitalization, but not for post-discharge events. Admission two or more days before surgery was predictive for DVT, but not for PE. Preoperative steroid usage and male gender were predictive for post-discharge DVT and PE, respectively. ICH was associated with various comorbidities and longer operative times. This multicenter study demonstrates distinct critical time periods for the development of thrombotic and hemorrhagic events after craniotomy. Furthermore, the VTE risk profile depends on the type of VTE (DVT vs. PE) and clinical setting (hospitalized vs. post-discharge patients).

An introduction and overview of machine learning in neurosurgical care.

BACKGROUND: Machine learning (ML) is a branch of artificial intelligence that allows computers to learn from large complex datasets without being explicitly programmed. Although ML is already widely manifest in our daily lives in various forms, the considerable potential of ML has yet to find its way into mainstream medical research and day-to-day clinical care. The complex diagnostic and therapeutic modalities used in neurosurgery provide a vast amount of data that is ideally suited for ML models. This systematic review explores ML's potential to assist and improve neurosurgical care. METHOD: A systematic literature search was performed in the PubMed and Embase databases to identify all potentially relevant studies up to January 1, 2017. All studies were included that evaluated ML models assisting neurosurgical treatment. RESULTS: Of the 6,402 citations identified, 221 studies were selected after subsequent title/abstract and full-text screening. In these studies, ML was used to assist surgical treatment of patients with epilepsy, brain tumors, spinal lesions, neurovascular pathology, Parkinson's disease, traumatic brain injury, and hydrocephalus. Across multiple paradigms, ML was found to be a valuable tool for presurgical planning, intraoperative guidance, neurophysiological monitoring, and neurosurgical outcome prediction. CONCLUSIONS: ML has started to find applications aimed at improving neurosurgical care by increasing the efficiency and precision of perioperative decision-making. A thorough validation of specific ML models is essential before implementation in clinical neurosurgical care. To bridge the gap between research and clinical care, practical and ethical issues should be considered parallel to the development of these techniques.

Visual outcomes after endoscopic endonasal pituitary adenoma resection: a systematic review and meta-analysis.

PURPOSE: Patients with pituitary adenomas often present with visual deficits. While the aim of endoscopic endonasal transsphenoidal surgery (EETS) is to improve these deficits, permanent worsening is a possible outcome. The aim of this meta-analysis was to evaluate the effect of EETS for pituitary adenomas on visual outcomes. METHODS: A meta-analysis was conducted according to the PRISMA guidelines. Pooled prevalence was calculated for complete recovery, improvement, and deterioration of visual field deficits, visual acuity and unspecified visual function in fixed- and random-effect models, including assessment of heterogeneity (I2) and publication bias (Begg's test). RESULTS: Out of 2636 articles, 35 case series were included in the meta-analysis. Results are described for fixed-effect models. For patients with impaired visual acuity, only one study reported complete recovery (27.2%). Pooled prevalence for improvement was 67.5% (95% CI = 59.1-75.0%), but with considerable heterogeneity (I2: 86.0%), and 4.50% (95% CI = 1.80-10.8%) for patients experiencing deterioration. For patients with visual field deficits, the prevalence was 40.4% (95% CI = 34.8-46.3%) for complete recovery, 80.8% (95% CI = 77.7-83.6%) for improvement, and 2.3% (95% CI = 1.1-4.7%) for deterioration. For the unspecified visual outcomes, pooled prevalence of complete recovery was 32.9% (95% CI: 28.5-37.7%), but with considerable heterogeneity (I2 = 84.2%). The prevalence was 80.9% (95% CI = 77.9-83.6) for improvement and 2.00% (95% CI = 1.10-3.40%) for deterioration. Random-effect models yielded similar results. Publication bias was non-significant for all the outcomes. CONCLUSION: While visual deficits improved after EETS in the majority of patients, complete recovery was only achieved in less than half of the patients and some patients even suffered from visual deterioration.

Agents for fluorescence-guided glioma surgery: a systematic review of preclinical and clinical results.

BACKGROUND: Fluorescence-guided surgery (FGS) is a technique used to enhance visualization of tumor margins in order to increase the extent of tumor resection in glioma surgery. In this paper, we systematically review all clinically tested fluorescent agents for application in FGS for glioma and all preclinically tested agents with the potential for FGS for glioma. METHODS: We searched the PubMed and Embase databases for all potentially relevant studies through March 2016. We assessed fluorescent agents by the following outcomes: rate of gross total resection (GTR), overall and progression-free survival, sensitivity and specificity in discriminating tumor and healthy brain tissue, tumor-to-normal ratio of fluorescent signal, and incidence of adverse events. RESULTS: The search strategy resulted in 2155 articles that were screened by titles and abstracts. After full-text screening, 105 articles fulfilled the inclusion criteria evaluating the following fluorescent agents: 5-aminolevulinic acid (5-ALA) (44 studies, including three randomized control trials), fluorescein (11), indocyanine green (five), hypericin (two), 5-aminofluorescein-human serum albumin (one), endogenous fluorophores (nine) and fluorescent agents in a pre-clinical testing phase (30). Three meta-analyses were also identified. CONCLUSIONS: 5-ALA is the only fluorescent agent that has been tested in a randomized controlled trial and results in an improvement of GTR and progression-free survival in high-grade gliomas. Observational cohort studies and case series suggest similar outcomes for FGS using fluorescein. Molecular targeting agents (e.g., fluorophore/nanoparticle labeled with anti-EGFR antibodies) are still in the pre-clinical phase, but offer promising results and may be valuable future alternatives.

An Online Calculator for the Prediction of Survival in Glioblastoma Patients Using Classical Statistics and Machine Learning.

BACKGROUND: Although survival statistics in patients with glioblastoma multiforme (GBM) are well-defined at the group level, predicting individual patient survival remains challenging because of significant variation within strata. OBJECTIVE: To compare statistical and machine learning algorithms in their ability to predict survival in GBM patients and deploy the best performing model as an online survival calculator. METHODS: Patients undergoing an operation for a histopathologically confirmed GBM were extracted from the Surveillance Epidemiology and End Results (SEER) database (2005-2015) and split into a training and hold-out test set in an 80/20 ratio. Fifteen statistical and machine learning algorithms were trained based on 13 demographic, socioeconomic, clinical, and radiographic features to predict overall survival, 1-yr survival status, and compute personalized survival curves. RESULTS: In total, 20 821 patients met our inclusion criteria. The accelerated failure time model demonstrated superior performance in terms of discrimination (concordance index = 0.70), calibration, interpretability, predictive applicability, and computational efficiency compared to Cox proportional hazards regression and other machine learning algorithms. This model was deployed through a free, publicly available software interface (https://cnoc-bwh.shinyapps.io/gbmsurvivalpredictor/). CONCLUSION: The development and deployment of survival prediction tools require a multimodal assessment rather than a single metric comparison. This study provides a framework for the development of prediction tools in cancer patients, as well as an online survival calculator for patients with GBM. Future efforts should improve the interpretability, predictive applicability, and computational efficiency of existing machine learning algorithms, increase the granularity of population-based registries, and externally validate the proposed prediction tool.

Orally Administered 5-aminolevulinic Acid for Isolation and Characterization of Circulating Tumor-Derived Extracellular Vesicles in Glioblastoma Patients.

Background: In glioblastoma (GB), tissue is required for accurate diagnosis and subtyping. Tissue can be obtained through resection or (stereotactic) biopsy, but these invasive procedures provide risks for patients. Extracellular vesicles (EVs) are small, cell-derived vesicles that contain miRNAs, proteins, and lipids, and possible candidates for liquid biopsies. GB-derived EVs can be found in the blood of patients, but it is difficult to distinguish them from circulating non-tumor EVs. 5-aminolevulinic acid (5-ALA) is orally administered to GB patients to facilitate tumor visualization and maximal resection, as it is metabolized to fluorescent protoporphyrin IX (PpIX) that accumulates in glioma cells. In this study, we assessed whether PpIX accumulates in GB-derived EVs and whether these EVs could be isolated and characterized to enable a liquid biopsy in GB. Methods: EVs were isolated from the conditioned media of U87 cells treated with 5-ALA by differential ultracentrifugation. Blood samples were collected and processed from healthy controls and patients undergoing 5-ALA guided surgery for GB. High-resolution flow cytometry (hFC) enabled detection and sorting of PpIX-positive EVs, which were subsequently analyzed by digital droplet PCR (ddPCR). Results: PpIX-positive EVs could be detected in conditioned cell culture media as well as in patient samples after administration of 5-ALA. By using hFC, we could sort the PpIX-positive EVs for further analysis with ddPCR, which indicated the presence of EVs and GB-associated miRNAs. Conclusion: GB-derived EVs can be isolated from the plasma of GB patients by using 5-ALA induced fluorescence. Although many challenges remain, our findings show new possibilities for the development of blood-based liquid biopsies in GB patients.

Automating Clinical Chart Review: An Open-Source Natural Language Processing Pipeline Developed on Free-Text Radiology Reports From Patients With Glioblastoma.

PURPOSE: The aim of this study was to develop an open-source natural language processing (NLP) pipeline for text mining of medical information from clinical reports. We also aimed to provide insight into why certain variables or reports are more suitable for clinical text mining than others. MATERIALS AND METHODS: Various NLP models were developed to extract 15 radiologic characteristics from free-text radiology reports for patients with glioblastoma. Ten-fold cross-validation was used to optimize the hyperparameter settings and estimate model performance. We examined how model performance was associated with quantitative attributes of the radiologic characteristics and reports. RESULTS: In total, 562 unique brain magnetic resonance imaging reports were retrieved. NLP extracted 15 radiologic characteristics with high to excellent discrimination (area under the curve, 0.82 to 0.98) and accuracy (78.6% to 96.6%). Model performance was correlated with the inter-rater agreement of the manually provided labels (ρ = 0.904; P < .001) but not with the frequency distribution of the variables of interest (ρ = 0.179; P = .52). All variables labeled with a near perfect inter-rater agreement were classified with excellent performance (area under the curve > 0.95). Excellent performance could be achieved for variables with only 50 to 100 observations in the minority group and class imbalances up to a 9:1 ratio. Report-level classification accuracy was not associated with the number of words or the vocabulary size in the distinct text documents. CONCLUSION: This study provides an open-source NLP pipeline that allows for text mining of narratively written clinical reports. Small sample sizes and class imbalance should not be considered as absolute contraindications for text mining in clinical research. However, future studies should report measures of inter-rater agreement whenever ground truth is based on a consensus label and use this measure to identify clinical variables eligible for text mining.

Trends in High-Impact Neurosurgical Randomized Controlled Trials Published in General Medical Journals: A Systematic Review.

BACKGROUND: The neurosurgery literature lacks a comprehensive report of neurosurgical randomized controlled trials (RCTs) published in general medical journals. RCTs published in these journals have high visibility and impact on decision-making by general medical practitioners and health care policymakers. METHODS: A systematic review of neurosurgical RCTs in the New England Journal of Medicine, The Lancet, Journal of the American Medical Association, The BMJ, and Annals of Internal Medicine was completed. RESULTS: There were 78 neurosurgical RCTs published in the selected high-impact journals from 2000 to 2017. The most common study topics were neurovascular (n = 39, 50%) and spine (n = 24, 30.8%). Of these RCTs, 44 (56.4%) compared operative with nonoperative management. For studies published before 2017, the mean number of citations was 899. Approximately half of the studies showed superiority of operative management over nonoperative management in the intent to treat primary outcome of interest (n = 24, 54.5%). However, stratified by subsubspecialty, 7 (87.5%) of the functional RCTs, 9 (50%) of the neurovascular RCTs, 1 (50%) of the trauma RCTs, and 7 (43.8%) of the spinal RCTs demonstrated superiority of operative management over nonoperative management. Additionally, there were large subspecialty differences in study characteristics, such as rate of double blinding, proportion of patient enrollment from patients screened, and proportion of crossover from nonsurgical to surgical arm. CONCLUSIONS: Neurosurgical RCTs in general medical journals have large subspecialty differences in characteristics such as crossovers from nonsurgical to surgical treatment arms and the proportion of studies demonstrating benefit of operative intervention over nonoperative management.

Natural Language Processing for Automated Quantification of Brain Metastases Reported in Free-Text Radiology Reports.

PURPOSE: Although the bulk of patient-generated health data are increasing exponentially, their use is impeded because most data come in unstructured format, namely as free-text clinical reports. A variety of natural language processing (NLP) methods have emerged to automate the processing of free text ranging from statistical to deep learning-based models; however, the optimal approach for medical text analysis remains to be determined. The aim of this study was to provide a head-to-head comparison of novel NLP techniques and inform future studies about their utility for automated medical text analysis. PATIENTS AND METHODS: Magnetic resonance imaging reports of patients with brain metastases treated in two tertiary centers were retrieved and manually annotated using a binary classification (single metastasis v two or more metastases). Multiple bag-of-words and sequence-based NLP models were developed and compared after randomly splitting the annotated reports into training and test sets in an 80:20 ratio. RESULTS: A total of 1,479 radiology reports of patients diagnosed with brain metastases were retrieved. The least absolute shrinkage and selection operator (LASSO) regression model demonstrated the best overall performance on the hold-out test set with an area under the receiver operating characteristic curve of 0.92 (95% CI, 0.89 to 0.94), accuracy of 83% (95% CI, 80% to 87%), calibration intercept of -0.06 (95% CI, -0.14 to 0.01), and calibration slope of 1.06 (95% CI, 0.95 to 1.17). CONCLUSION: Among various NLP techniques, the bag-of-words approach combined with a LASSO regression model demonstrated the best overall performance in extracting binary outcomes from free-text clinical reports. This study provides a framework for the development of machine learning-based NLP models as well as a clinical vignette of patients diagnosed with brain metastases.

Passive data collection and use in healthcare: A systematic review of ethical issues.

INTRODUCTION: Passive data refers to data generated without the active participation of the subject. This includes data from global positioning systems and accelerometers or metadata on phone call and text activity. Although the potential healthcare applications are far-reaching, passive data raises numerous ethical challenges. MATERIALS AND METHODS: We performed a systematic review to identify all ethical concerns, normative standpoints, and underlying arguments related to the use of passive data in healthcare. RESULTS: Among the various challenges discussed in the ethical literature, informational privacy, informed consent, and data security were the primary focus of the current debate. Other topics of discussion were the evaluation and regulation of products, equity in access, vulnerable patient groups, ownership, and secondary use. CONCLUSION: No clear ethical framework has been established that stimulates passive data-driven innovation while protecting patient integrity. The consensus in the ethical literature, as well as the parallels with similar concerns and solutions in other fields, can lay a foundation for the construction of an ethical framework. The future debate should focus on conflicts between two or more ethical, technical, or clinical values to ensure a safe and effective implementation of passive data in healthcare.

Routine Blood Tests Do Not Predict Survival in Patients with Glioblastoma-Multivariable Analysis of 497 Patients.

BACKGROUND: Multiple reports have attributed a prognostic value to routine blood tests results for patients with glioblastoma. However, these studies have reported conflicting results and have often had small sample sizes. We sought to validate the prognostic value of the described tests in an independent glioblastoma patient population. METHODS: We performed a retrospective single-center multivariable analysis of 497 patients with glioblastoma who had postoperatively undergone radiotherapy and/or chemotherapy to identify the prognostic value for median overall survival of hemoglobin, white blood cell, monocyte, neutrophil, leukocyte, and platelet counts, neutrophil/lymphocyte ratio, C-reactive protein, erythrocyte sedimentation rate, activated partial thromboplastin time, prothrombin time, and lactate dehydrogenase. We also evaluated known prognostic factors for survival such as patient age, intervention type, IDH1 status, Karnofsky clinical performance status, and postoperative treatment modality. RESULTS: In a multivariable model, after correcting for multiple testing bias, biopsy alone (hazard ratio, 0.35; 95% confidence interval, 0.26-0.49; false discovery rate-adjusted P < 0.001) and monotherapy after surgery (hazard ratio, 0.46; 95% confidence interval, 0.33-0.66; false discovery rate-adjusted P < 0.001) remained significantly associated with worse median overall survival. Patient age and Karnofsky performance status score ≥70 did not significantly influence survival in the multivariable model. No routine blood test included in the multivariable analysis was significantly associated with survival. CONCLUSIONS: In the present study, hemoglobin, white blood cell, monocyte, neutrophil, leukocyte, and platelet counts, neutrophil/lymphocyte ratio, C-reactive protein, erythrocyte sedimentation rate, activated partial thromboplastin time, prothrombin time, and lactate dehydrogenase levels did not independently predict for overall survival in patients with glioblastoma.