Cardiovascular events and venous thromboembolism after primary malignant or non-malignant brain tumour diagnosis: a population matched cohort study in Wales (United Kingdom).

BACKGROUND: Elevated standardised mortality ratio of cardiovascular diseases (CVD) in patients with brain tumours may result from differences in the CVD incidences and cardiovascular risk factors. We compared the risk of CVD among patients with a primary malignant or non-malignant brain tumour to a matched general population cohort, accounting for other co-morbidities. METHODS: Using data from the Secured Anonymised Information Linkage (SAIL) Databank in Wales (United Kingdom), we identified all adults aged ≥ 18 years in the primary care database with first diagnosis of malignant or non-malignant brain tumour identified in the cancer registry in 2000-2014 and a matched cohort (case-to-control ratio 1:5) by age, sex and primary care provider from the general population without any cancer diagnosis. Outcomes included fatal and non-fatal major vascular events (stroke, ischaemic heart disease, aortic and peripheral vascular diseases) and venous thromboembolism (VTE). We used multivariable Cox models adjusted for clinical risk factors to compare risks, stratified by tumour behaviour (malignant or non-malignant) and follow-up period. RESULTS: There were 2869 and 3931 people diagnosed with malignant or non-malignant brain tumours, respectively, between 2000 and 2014 in Wales. They were matched to 33,785 controls. Within the first year of tumour diagnosis, malignant tumour was associated with a higher risk of VTE (hazard ratio [HR] 21.58, 95% confidence interval 16.12-28.88) and stroke (HR 3.32, 2.44-4.53). After the first year, the risks of VTE (HR 2.20, 1.52-3.18) and stroke (HR 1.45, 1.00-2.10) remained higher than controls. Patients with non-malignant tumours had higher risks of VTE (HR 3.72, 2.73-5.06), stroke (HR 4.06, 3.35-4.93) and aortic and peripheral arterial disease (HR 2.09, 1.26-3.48) within the first year of diagnosis compared with their controls. CONCLUSIONS: The elevated CVD and VTE risks suggested risk reduction may be a strategy to improve life quality and survival in people with a brain tumour.

Cauda equina syndrome.

Aims: Patients with cauda equina syndrome (CES) require emergency imaging and surgical decompression. The severity and type of symptoms may influence the timing of imaging and surgery, and help predict the patient's prognosis. Categories of CES attempt to group patients for management and prognostication purposes. We aimed in this study to assess the inter-rater reliability of dividing patients with CES into categories to assess whether they can be reliably applied in clinical practice and in research. Methods: A literature review was undertaken to identify published descriptions of categories of CES. A total of 100 real anonymized clinical vignettes of patients diagnosed with CES from the Understanding Cauda Equina Syndrome (UCES) study were reviewed by consultant spinal surgeons, neurosurgical registrars, and medical students. All were provided with published category definitions and asked to decide whether each patient had 'suspected CES'; 'early CES'; 'incomplete CES'; or 'CES with urinary retention'. Inter-rater agreement was assessed for all categories, for all raters, and for each group of raters using Fleiss's kappa. Results: Each of the 100 participants were rated by four medical students, five neurosurgical registrars, and four consultant spinal surgeons. No groups achieved reasonable inter-rater agreement for any of the categories. CES with retention versus all other categories had the highest inter-rater agreement (kappa 0.34 (95% confidence interval 0.27 to 0.31); minimal agreement). There was no improvement in inter-rater agreement with clinical experience. Across all categories, registrars agreed with each other most often (kappa 0.41), followed by medical students (kappa 0.39). Consultant spinal surgeons had the lowest inter-rater agreement (kappa 0.17). Conclusion: Inter-rater agreement for categorizing CES is low among clinicians who regularly manage these patients. CES categories should be used with caution in clinical practice and research studies, as groups may be heterogenous and not comparable.

Reporting of PPI and the MCID in phase III/IV randomised controlled trials-a systematic review.

BACKGROUND: Patient and public involvement (PPI) in clinical trial design contributes to ensuring the research objectives and outcome measures are relevant to patients. The minimal clinically important difference (MCID) in the primary outcome influences trial design and feasibility and should be predicated on PPI. We aimed to determine current practice of reporting PPI and the MCID in phase III/IV randomised controlled trials (RCTs). METHODS: Following a search of Medline, Embase, and the Cochrane Central Register of Controlled Trials, we included primary publications of phase III/IV RCTs, in English, inclusive of any medical specialty or type of intervention, that reported a health-related outcome. We excluded protocols and secondary publications of RCTs. We extracted RCT characteristics, the use of PPI, and use of the MCID. RESULTS: Between 1 July 2019 and 13 January 2020, 123 phase III/IV RCTs matched our eligibility criteria. Ninety percent evaluated a medical rather than surgical intervention. Oncology accounted for 21% of all included RCTs. Only 2.4% (n = 3) and 1.6% (n = 2) RCTs described PPI and the MCID respectively. CONCLUSIONS: PPI and the MCID are poorly reported, so it is uncertain how these contributed to trial design. Improvement in the reporting of these items would increase confidence that results are relevant and clinically significant to patients, contributing to improving the overall trial design. TRIAL REGISTRATION: Not registered.

Variation in postoperative outcomes of patients with intracranial tumors: insights from a prospective international cohort study during the COVID-19 pandemic.

BACKGROUND: This study assessed the international variation in surgical neuro-oncology practice and 30-day outcomes of patients who had surgery for an intracranial tumor during the COVID-19 pandemic. METHODS: We prospectively included adults aged ≥18 years who underwent surgery for a malignant or benign intracranial tumor across 55 international hospitals from 26 countries. Each participating hospital recorded cases for 3 consecutive months from the start of the pandemic. We categorized patients' location by World Bank income groups (high [HIC], upper-middle [UMIC], and low- and lower-middle [LLMIC]). Main outcomes were a change from routine management, SARS-CoV-2 infection, and 30-day mortality. We used a Bayesian multilevel logistic regression stratified by hospitals and adjusted for key confounders to estimate the association between income groups and mortality. RESULTS: Among 1016 patients, the number of patients in each income group was 765 (75.3%) in HIC, 142 (14.0%) in UMIC, and 109 (10.7%) in LLMIC. The management of 200 (19.8%) patients changed from usual care, most commonly delayed surgery. Within 30 days after surgery, 14 (1.4%) patients had a COVID-19 diagnosis and 39 (3.8%) patients died. In the multivariable model, LLMIC was associated with increased mortality (odds ratio 2.83, 95% credible interval 1.37-5.74) compared to HIC. CONCLUSIONS: The first wave of the pandemic had a significant impact on surgical decision-making. While the incidence of SARS-CoV-2 infection within 30 days after surgery was low, there was a disparity in mortality between countries and this warrants further examination to identify any modifiable factors.

Development of prediction models for one-year brain tumour survival using machine learning: a comparison of accuracy and interpretability.

BACKGROUND AND OBJECTIVE: Prediction of survival in patients diagnosed with a brain tumour is challenging because of heterogeneous tumour behaviours and treatment response. Advances in machine learning have led to the development of clinical prognostic models, but due to the lack of model interpretability, integration into clinical practice is almost non-existent. In this retrospective study, we compare five classification models with varying degrees of interpretability for the prediction of brain tumour survival greater than one year following diagnosis. METHODS: 1028 patients aged ≥16 years with a brain tumour diagnosis between April 2012 and April 2020 were included in our study. Three intrinsically interpretable 'glass box' classifiers (Bayesian Rule Lists [BRL], Explainable Boosting Machine [EBM], and Logistic Regression [LR]), and two 'black box' classifiers (Random Forest [RF] and Support Vector Machine [SVM]) were trained on electronic patients records for the prediction of one-year survival. All models were evaluated using balanced accuracy (BAC), F1-score, sensitivity, specificity, and receiver operating characteristics. Black box model interpretability and misclassified predictions were quantified using SHapley Additive exPlanations (SHAP) values and model feature importance was evaluated by clinical experts. RESULTS: The RF model achieved the highest BAC of 78.9%, closely followed by SVM (77.7%), LR (77.5%) and EBM (77.1%). Across all models, age, diagnosis (tumour type), functional features, and first treatment were top contributors to the prediction of one year survival. We used EBM and SHAP to explain model misclassifications and investigated the role of feature interactions in prognosis. CONCLUSION: Interpretable models are a natural choice for the domain of predictive medicine. Intrinsically interpretable models, such as EBMs, may provide an advantage over traditional clinical assessment of brain tumour prognosis by weighting potential risk factors and their interactions that may be unknown to clinicians. An agreement between model predictions and clinical knowledge is essential for establishing trust in the models decision making process, as well as trust that the model will make accurate predictions when applied to new data.

Presentation, management, and outcomes of cauda equina syndrome up to one year after surgery, using clinician and participant reporting: A multi-centre prospective cohort study.

Background: Cauda equina syndrome (CES) results from nerve root compression in the lumbosacral spine, usually due to a prolapsed intervertebral disc. Evidence for management of CES is limited by its infrequent occurrence and lack of standardised clinical definitions and outcome measures. Methods: This is a prospective multi-centre observational cohort study of adults with CES in the UK. We assessed presentation, investigation, management, and all Core Outcome Set domains up to one year post-operatively using clinician and participant reporting. Univariable and multivariable associations with the Oswestry Disability Index (ODI) and urinary outcomes were investigated. Findings: In 621 participants with CES, catheterisation for urinary retention was required pre-operatively in 31% (191/615). At discharge, only 13% (78/616) required a catheter. Median time to surgery from symptom onset was 3 days (IQR:1-8) with 32% (175/545) undergoing surgery within 48 h. Earlier surgery was associated with catheterisation (OR:2.2, 95%CI:1.5-3.3) but not with admission ODI or radiological compression. In multivariable analyses catheter requirement at discharge was associated with pre-operative catheterisation (OR:10.6, 95%CI:5.8-20.4) and one-year ODI was associated with presentation ODI (r = 0.3, 95%CI:0.2-0.4), but neither outcome was associated with time to surgery or radiological compression. Additional healthcare services were required by 65% (320/490) during one year follow up. Interpretation: Post-operative functional improvement occurred even in those presenting with urinary retention. There was no association between outcomes and time to surgery in this observational study. Significant healthcare needs remained post-operatively. Funding: DCN Endowment Fund funded study administration. Castor EDC provided database use. No other study funding was received.

Impact of tumour characteristics and cancer treatment on cerebrovascular mortality after glioma diagnosis: Evidence from a population-based cancer registry.

Objective: We aimed to examine brain tumour grade, a marker of biological aggressiveness, tumour size and cancer treatment are associated with cerebrovascular mortality among patients with malignant glioma, the most common and aggressive type of brain tumour. Methods: We conducted a retrospective, observational cohort study using the US National Cancer Institute's state and regional population-based cancer registries. We identified adult patients with glioma in 2000 to 2018 (N=72,916). The primary outcome was death from cerebrovascular disease. Cox regression modelling was used to estimate the associations with cerebrovascular mortality of tumour grade, tumour size and treatment (surgery, radiotherapy, chemotherapy), calculating hazard ratios (HR) adjusted for these factors as well as for age, sex, race, marital status and calendar year. Results: Higher grade (Grade IV vs Grade II: HR=2.47, 95% CI=1.69-3.61, p<0.001) and larger brain tumours (size 3 to <6 cm: HR=1.40, 95% CI=1.03 -1.89, p<0.05; size ≥ 6 cm: HR=1.47, 95% CI=1.02-2.13, p<0.05 compared to size < 3cm) were associated with increased cerebrovascular mortality. Cancer treatment was associated with decreased risk (surgery: HR= 0.60, p<0.001; chemotherapy: HR=0.42, p<0.001; radiation: HR= 0.69, p<0.05). However, among patents surviving five years or more from cancer diagnosis radiotherapy was associated with higher risk of cerebrovascular mortality (HR 2.73, 95% CI 1.49-4.99, p<0.01). Conclusion: More aggressive tumour characteristics are associated with increased cerebrovascular mortality. Radiotherapy increased risk of cerebrovascular mortality five-year after cancer diagnosis. Further research is needed to better understand the long-term cardiovascular consequences of radiation therapy, and whether the consequent risk can be mitigated.

Association of extent of resection and functional outcomes in diffuse low-grade glioma: systematic review & meta-analysis.

BACKGROUND: Surgical resection offers survival benefits in patients with diffuse low-grade glioma (DLGG) but its association with functional outcomes is uncertain. This systematic review assessed functional outcomes associated with extent of resection (EoR) in adults with DLGG. METHODS: We searched Medline, Embase and CENTRAL on the 19th of February 2021 for observational studies reporting functional outcomes after surgical resection for patients aged ≥ 18 years with a new diagnosis of supratentorial DLGG according to any World Health Organization classification of primary brain tumors. The Newcastle-Ottawa Scale (NOS) informed our risk of bias assessments. The proportion of patients returning to work within 12 months entered a random-effects meta-analysis. PROSPERO registration number CRD42021238387. RESULTS: There were seven eligible moderate to high-quality (NOS > 6) observational studies identified from 1,183 records involving 234 patients with DLGG. Functional outcomes reported included neurocognition (n = 2 studies), performance status (n = 3), quality of life (QoL) (n = 1) and return to work (n = 6). The proportion of patients who returned to work within 12 months of surgery was 84% (95% confidence interval [CI] 50-96%, I-squared = 38%, 5 studies) for gross total resection, 66% (95% CI 14-96%, I2 = 57%, 5 studies) for subtotal resection, and 31% (95% CI 4-82%, I2 = 0%, 4 studies) for partial resection. There was insufficient data on other functional outcomes for quantitative synthesis. CONCLUSION: A higher proportion of DLGG patients returned to work following gross total resection compared with those who had a subtotal or partial resection. Further studies with standardized assessments can clarify the association between EoR and different functional outcomes.

Demographics of Cauda Equina Syndrome: A Population-Based Incidence Study.

INTRODUCTION: Cauda equina syndrome (CES) has significant medical, social, and legal consequences. Understanding the number of people presenting with CES and their demographic features is essential for planning healthcare services to ensure timely and appropriate management. We aimed to establish the incidence of CES in a single country and stratify incidence by age, gender, and socioeconomic status. As no consensus clinical definition of CES exists, we compared incidence using different diagnostic criteria. METHODS: All patients presenting with radiological compression of the cauda equina due to degenerative disc disease and clinical CES requiring emergency surgical decompression during a 1-year period were identified at all centres performing emergency spinal surgery across Scotland. Initial patient identification occurred during the emergency hospital admission, and case ascertainment was checked using ICD-10 diagnostic coding. Clinical information was reviewed, and incidence rates for all demographic and clinical groups were calculated. RESULTS: We identified 149 patients with CES in 1 year from a total population of 5.4 million, giving a crude incidence of 2.7 (95% CI: 2.3-3.2) per 100,000 per year. CES occurred more commonly in females and in the 30-49 years age range, with an incidence per year of 7.2 (95% CI: 4.7-10.6) per 100,000 females age 30-39. There was no association between CES and socioeconomic status. CES requiring catheterization had an incidence of 1.1 (95% CI: 0.8-1.5) per 100,000 adults per year. The use of ICD-10 codes alone to identify cases gave much higher incidence rates, but was inaccurate, with 55% (117/211) of patients with a new ICD-10 code for CES found not to have CES on clinical notes review. CONCLUSION: CES occurred more commonly in females and in those between 30 and 49 years and had no association with socioeconomic status. The incidence of CES in Scotland is at least four times higher than previous European estimates of 0.3-0.6 per 100,000 population per year. Incidence varies with clinical diagnostic criteria. To enable comparison of rates of CES across populations, we recommend using standardized clinical and radiological criteria and standardization for population structure.

Imaging timing after glioblastoma surgery (INTERVAL-GB): protocol for a UK and Ireland, multicentre retrospective cohort study.

INTRODUCTION: Glioblastoma is the most common malignant primary brain tumour with a median overall survival of 12-15 months (range 6-17 months), even with maximal treatment involving debulking neurosurgery and adjuvant concomitant chemoradiotherapy. The use of postoperative imaging to detect progression is of high importance to clinicians and patients, but currently, the optimal follow-up schedule is yet to be defined. It is also unclear how adhering to National Institute for Health and Care Excellence (NICE) guidelines-which are based on general consensus rather than evidence-affects patient outcomes such as progression-free and overall survival. The primary aim of this study is to assess MRI monitoring practice after surgery for glioblastoma, and to evaluate its association with patient outcomes. METHODS AND ANALYSIS: ImagiNg Timing aftER surgery for glioblastoma: an eVALuation of practice in Great Britain and Ireland is a retrospective multicentre study that will include 450 patients with an operated glioblastoma, treated with any adjuvant therapy regimen in the UK and Ireland. Adult patients ≥18 years diagnosed with glioblastoma and undergoing surgery between 1 August 2018 and 1 February 2019 will be included. Clinical and radiological scanning data will be collected until the date of death or date of last known follow-up. Anonymised data will be uploaded to an online Castor database. Adherence to NICE guidelines and the effect of being concordant with NICE guidelines will be identified using descriptive statistics and Kaplan-Meier survival analysis. ETHICS AND DISSEMINATION: Each participating centre is required to gain local institutional approval for data collection and sharing. Formal ethical approval is not required since this is a service evaluation. Results of the study will be reported through peer-reviewed presentations and articles, and will be disseminated to participating centres, patients and the public.

A systematic review identifying outcome measures used in evaluating adults sustaining cervical spine fractures.

OBJECTIVE: To assess the outcome measures used in studies investigating cervical spine fractures in adults, with or without associated spinal cord injury, to inform development of a core outcome set. METHODS: Medline, Embase and Scopus were searched for relevant studies until May 28, 2022, without a historic limit on study date. Study characteristics, population characteristics and outcomes reported were extracted and analyzed. RESULTS: Our literature search identified 536 studies that met criteria for inclusion, involving 393,266 patients. Most studies were single center (87.3%), retrospective studies (88.9%) and involved a median of 40 patients (range 6-167,278). Treatments assessed included: surgery (55.2%), conservative (6.2%), halo immobilization (4.9%), or a mixture (33.2%). Median study duration was 84 months (range 3-564 months); the timing of clinical and/or radiological follow-up assessment after injury was reported in 56.7%. There was significant heterogeneity in outcomes used, with 79 different reported outcomes measures. Differences in use were identified between smaller/larger, retro-/prospective and single/multicenter cohorts. Over time, the use of radiological outcomes has declined with greater emphasis on patient-reported outcome measures (PROMs). Studies of conservative management were more likely to detail PROMs and mortality, whereas surgical studies reported Frankel/ASIA grade, radiological fusion, complication rates, duration of hospital stay and re-operation rates more frequently. In studies assessing the elderly population (> 65 years), use of PROMs, mortality, hospital stay and discharge destination were more common, whereas fusion was reported less often. Response rates for outcome assessments were lower in studies assessing elderly patients, and studies using PROMs. CONCLUSIONS: We have classified the various outcome measures used for patients with cervical spine fractures based on the COMET outcome taxonomy. We also described the contexts in which different outcomes are more commonly employed to help guide decision-making when designing future research endeavors.

Utility of 5-ALA for resection of CNS tumours other than high-grade gliomas: a protocol for a systematic review.

INTRODUCTION: 5-aminolevulinic acid (5-ALA) is a proagent developed for fluorescent-guided surgery for high-grade glioma patients associated with a significant increase in resection conferring survival. 5-ALA was shown to penetrate the blood-brain barrier accumulating in malignant glioma cells with high selectivity, sensitivity and positive predictive value. However, those have yet to be explored aiding diagnosis for tumours of the central nervous system (CNS) other than high-grade gliomas (HGG). No up-to-date systematic review exists reporting the major surgical outcomes and diagnostic accuracy. We sought to conduct a systematic review of the literature summarising surgical outcomes, evaluate the quality of diagnostic accuracy reported in the literature and qualitatively assess the evidence to inform future studies. METHODS AND ANALYSIS: We will search electronic databases (Medline, Embase) with subsequent interrogation of references lists of articles reporting the use of 5-ALA for brain tumours other than high-grade glioma adult patients, which also report the extent of resection and/or survival. Prospective and retrospective cohort and case-control studies with more than five patients will be included. Two independent reviewers will screen the abstracts and full articles, with a third reviewer resolving any conflicts. The data will be extracted in a standardised template and outcomes will be reported using descriptive statists. The quality of non-randomised studies will be appraised. ETHICS AND DISSEMINATION: The study will summarise the available evidence on the effect of the clinical utility of 5-ALA in achieving resection and improving survival and its diagnostic accuracy for tumours of the CNS other than HGG. The data will be presented nationally and internationally and the manuscript will be published in a peer-reviewed journal. No ethical approvals were needed. The aim is to inform prospective studies minimising reporting bias allowing for more reliable, reproducible and generalisable results. The study has been registered in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.PROSPERO registration numberCRD42021260542.

Socioeconomic deprivation is associated with reduced efficacy of an insulin adjustment education program for people with type 1 diabetes.

BACKGROUND: The Dose Adjustment for Normal Eating (DAFNE) course teaches insulin dose adjustment to match dietary carbohydrates and improve glycaemic control in participants with type 1 diabetes mellitus (T1DM). We investigated the association between socioeconomic deprivation and reduction in HbA1c as a marker of sustained glycaemic control, after attending DAFNE education. METHODS: This retrospective observational study identified adults with T1DM who attended DAFNE training in NHS Lothian, South East Scotland. We extracted age, sex, postcode-based Scottish Index of Multiple Deprivation (SIMD) quintiles and annual HbA1c measurements available four years before and after course attendance. We calculated mean HbA1c before (baseline) and after attendance at DAFNE, across four annual measurements. Change in mean HbA1c (mmol/mol) was categorised into three groups: decrease (≥ - 2.5), no change (<±2.5), increase (≥ + 2.5). We used multivariable ordinal logistic regression, with baseline mean HbA1c as a covariate, to investigate the association of SIMD quintile with reduction in mean HbA1c. RESULTS: 335 participants were included. Age and sex distribution were similar across SIMD quintiles (Mean age = 45, range 21-91, 59% women). Lower SIMD quintiles (greater deprivation) had higher baseline mean HbA1c (SIMD 1: 76.0, SIMD 5: 69.0). Higher SIMD quintiles (lower deprivation) were associated with lower odds of no change/increase in mean HbA1c (SIMD 5, odds ratio = 0.25, 95% confidence interval 0.10, 0.58, p = 0.001, multivariable analysis). CONCLUSION: Socioeconomic deprivation was associated with higher baseline mean HbA1c and lower reduction in HbA1c following DAFNE education. Future research could explore causes and how best to support participants from deprived areas. PREVIOUS SUBMISSIONS: This work has not been previously submitted to a journal. This work was presented as a poster at The ABCD Conference 2021 and the abstract (of no more than 300 words) from the meeting has been published: Innes CWD, Henshall DE, Wilson B, Poon M, Morley SD, Ritchie SA. Socioeconomic deprivation is associated with reduced efficacy of an insulin adjustment education programme for people with type 1 diabetes. Br J Diabetes. 2021; 21: 293-296.

Analytic approaches to clinical validation of results from preclinical models of glioblastoma: A systematic review.

INTRODUCTION: Analytic approaches to clinical validation of results from preclinical models are important in assessment of their relevance to human disease. This systematic review examined consistency in reporting of glioblastoma cohorts from The Cancer Genome Atlas (TCGA) or Chinese Glioma Genome Atlas (CGGA) and assessed whether studies included patient characteristics in their survival analyses. METHODS: We searched Embase and Medline on 02Feb21 for studies using preclinical models of glioblastoma published after Jan2008 that used data from TCGA or CGGA to validate the association between at least one molecular marker and overall survival in adult patients with glioblastoma. Main data items included cohort characteristics, statistical significance of the survival analysis, and model covariates. RESULTS: There were 58 eligible studies from 1,751 non-duplicate records investigating 126 individual molecular markers. In 14 studies published between 2017 and 2020 using TCGA RNA microarray data that should have the same cohort, the median number of patients was 464.5 (interquartile range 220.5-525). Of the 15 molecular markers that underwent more than one univariable or multivariable survival analyses, five had discrepancies between studies. Covariates used in the 17 studies that used multivariable survival analyses were age (76.5%), pre-operative functional status (35.3%), sex (29.4%) MGMT promoter methylation (29.4%), radiotherapy (23.5%), chemotherapy (17.6%), IDH mutation (17.6%) and extent of resection (5.9%). CONCLUSION: Preclinical glioblastoma studies that used TCGA for validation did not provide sufficient information about their cohort selection and there were inconsistent results. Transparency in reporting and the use of analytic approaches that adjust for clinical variables can improve the reproducibility between studies.

Temozolomide sensitivity of malignant glioma cell lines - a systematic review assessing consistencies between in vitro studies.

BACKGROUND: Malignant glioma cell line models are integral to pre-clinical testing of novel potential therapies. Accurate prediction of likely efficacy in the clinic requires that these models are reliable and consistent. We assessed this by examining the reporting of experimental conditions and sensitivity to temozolomide in glioma cells lines. METHODS: We searched Medline and Embase (Jan 1994-Jan 2021) for studies evaluating the effect of temozolomide monotherapy on cell viability of at least one malignant glioma cell line. Key data items included type of cell lines, temozolomide exposure duration in hours (hr), and cell viability measure (IC50). RESULTS: We included 212 studies from 2789 non-duplicate records that reported 248 distinct cell lines. The commonest cell line was U87 (60.4%). Only 10.4% studies used a patient-derived cell line. The proportion of studies not reporting each experimental condition ranged from 8.0-27.4%, including base medium (8.0%), serum supplementation (9.9%) and number of replicates (27.4%). In studies reporting IC50, the median value for U87 at 24 h, 48 h and 72 h was 123.9 µM (IQR 75.3-277.7 µM), 223.1 µM (IQR 92.0-590.1 µM) and 230.0 µM (IQR 34.1-650.0 µM), respectively. The median IC50 at 72 h for patient-derived cell lines was 220 µM (IQR 81.1-800.0 µM). CONCLUSION: Temozolomide sensitivity reported in comparable studies was not consistent between or within malignant glioma cell lines. Drug discovery science performed on these models cannot reliably inform clinical translation. A consensus model of reporting can maximise reproducibility and consistency among in vitro studies.

Raised cardiovascular disease mortality after central nervous system tumor diagnosis: analysis of 171,926 patients from UK and USA.

BACKGROUND: Patients with central nervous system (CNS) tumors may be at risk of dying from cardiovascular disease (CVD). We examined CVD mortality risk in patients with different histological subtypes of CNS tumors. METHODS: We analyzed UK(Wales)-based Secure Anonymized Information Linkage (SAIL) for 8743 CNS tumors patients diagnosed in 2000-2015, and US-based National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) for 163,183 patients in 2005-2015. We calculated age-, sex-, and calendar-year-adjusted standardized mortality ratios (SMRs) for CVD comparing CNS tumor patients to Wales and US residents. We used Cox regression models to examine factors associated with CVD mortality among CNS tumor patients. RESULTS: CVD was the second leading cause of death for CNS tumor patients in SAIL (UK) and SEER (US). Patients with CNS tumors had higher CVD mortality than the general population (SAIL SMR = 2.64, 95% CI = 2.39-2.90, SEER SMR = 1.38, 95% CI = 1.35-1.42). Malignant CNS tumor patients had over 2-fold higher mortality risk in US and UK cohorts. SMRs for nonmalignant tumors were almost 2-fold higher in SAIL than in SEER. CVD mortality risk particularly cerebrovascular disease was substantially greater in patients diagnosed at age younger than 50 years, and within the first year after their cancer diagnosis (SAIL SMR = 2.98, 95% CI = 2.39-3.66, SEER SMR = 2.14, 95% CI = 2.03-2.25). Age, sex, race/ethnicity in USA, deprivation in UK and no surgery were associated with CVD mortality. CONCLUSIONS: Patients with CNS tumors had higher risk for CVD mortality, particularly from cerebrovascular disease compared to the general population, supporting further research to improve mortality outcomes.

The reporting quality of natural language processing studies: systematic review of studies of radiology reports.

BACKGROUND: Automated language analysis of radiology reports using natural language processing (NLP) can provide valuable information on patients' health and disease. With its rapid development, NLP studies should have transparent methodology to allow comparison of approaches and reproducibility. This systematic review aims to summarise the characteristics and reporting quality of studies applying NLP to radiology reports. METHODS: We searched Google Scholar for studies published in English that applied NLP to radiology reports of any imaging modality between January 2015 and October 2019. At least two reviewers independently performed screening and completed data extraction. We specified 15 criteria relating to data source, datasets, ground truth, outcomes, and reproducibility for quality assessment. The primary NLP performance measures were precision, recall and F1 score. RESULTS: Of the 4,836 records retrieved, we included 164 studies that used NLP on radiology reports. The commonest clinical applications of NLP were disease information or classification (28%) and diagnostic surveillance (27.4%). Most studies used English radiology reports (86%). Reports from mixed imaging modalities were used in 28% of the studies. Oncology (24%) was the most frequent disease area. Most studies had dataset size > 200 (85.4%) but the proportion of studies that described their annotated, training, validation, and test set were 67.1%, 63.4%, 45.7%, and 67.7% respectively. About half of the studies reported precision (48.8%) and recall (53.7%). Few studies reported external validation performed (10.8%), data availability (8.5%) and code availability (9.1%). There was no pattern of performance associated with the overall reporting quality. CONCLUSIONS: There is a range of potential clinical applications for NLP of radiology reports in health services and research. However, we found suboptimal reporting quality that precludes comparison, reproducibility, and replication. Our results support the need for development of reporting standards specific to clinical NLP studies.