Worldwide Prevalence of Epidermal Growth Factor Receptor Mutations in Non-Small Cell Lung Cancer: A Meta-Analysis.

BACKGROUND: Identification of variable epidermal growth factor receptor (EGFR) gene mutations in non-small cell lung cancer (NSCLC) is important for the selection of appropriate targeted therapies. This meta-analysis was conducted to provide a worldwide overview of EGFR mutation and submutation (specifically exon 19 deletions, exon 21 L858R substitutions, and others) prevalence, and identify important covariates that influence EGFR mutation status in patients with advanced NSCLC to address this clinical data gap. METHODS: Embase® and MEDLINE® in Ovid were searched for studies published between 2004 and 2019 with cohorts of ≥ 50 adults with EGFR mutations, focusing on stage III/IV NSCLC (≤ 20% of patients with stage I/II NSCLC). Linear mixed-effects models were fitted to EGFR mutation endpoints using logistic transformation (logit), assuming a binomial distribution. The model included terms for an intercept reflecting European studies and further additive terms for other continents. EGFR submutations examined were exon 19 deletions, exon 21 L858R substitutions, and others. RESULTS: Of 3969 abstracts screened, 57 studies were included in the overall EGFR mutation analysis and 74 were included in the submutation analysis relative to the overall EGFR mutation population (Europe, n = 12; Asia, n = 51; North America, n = 5; Central America, n = 1; South America, n = 1; Oceania, n = 1; Global, n = 3). The final overall EGFR mutations model estimated Asian and European prevalence of 49.1% and 12.8%, respectively, and included an additive covariate for the proportion of male patients in a study. There were no significant covariates in the submutation analyses. Most submutations were actionable: exon 19 deletions (49.2% [Asia]; 48.4% [Europe]); exon 21 L858R substitutions (41.1% [Asia]; 29.9% [Europe]). CONCLUSIONS: Although EGFR mutation prevalence was higher in Asian than Western countries, data support worldwide testing for EGFR overall and submutations to inform appropriate targeted treatment decisions.

Performance of model-based network meta-analysis (MBNMA) of time-course relationships: A simulation study.

Time-course model-based network meta-analysis (MBNMA) has been proposed as a framework to combine treatment comparisons from a network of randomized controlled trials reporting outcomes at multiple time-points. This can explain heterogeneity/inconsistency that arises by pooling studies with different follow-up times and allow inclusion of studies from earlier in drug development. The aim of this study is to explore using simulation: (a) how MBNMA model parameters are affected by the quantity/location of observed time-points across studies/comparisons, (b) how reliably an appropriate MBNMA model can be identified, (c) the robustness of model estimates and predictions under different dataset characteristics. Our results indicate that model parameters for a given treatment comparison are estimated with low mean bias even when no direct evidence was available, provided there was sufficient indirect evidence to estimate the time-course. A staged model selection strategy that selects time-course function, then heterogeneity, then covariance structure, identified the true model most reliably and efficiently. Predictions and parameter estimates from selected models had low mean bias even in the presence of high heterogeneity/correlation between time-points. However, failure to properly account for heterogeneity/correlation could lead to high error in precision of the estimates. Time-course MBNMA provides a statistically robust framework for synthesizing direct and indirect evidence to estimate relative effects and predicted mean responses whilst accounting for time-course and incorporating correlation and heterogeneity. This supports the use of MBNMA in evidence synthesis, particularly when additional studies are available with follow-up times that would otherwise prohibit their inclusion by conventional meta-analysis.

Modelling time-course relationships with multiple treatments: Model-based network meta-analysis for continuous summary outcomes.

BACKGROUND: Model-based meta-analysis (MBMA) is increasingly used to inform drug-development decisions by synthesising results from multiple studies to estimate treatment, dose-response, and time-course characteristics. Network meta-analysis (NMA) is used in Health Technology Appraisals for simultaneously comparing effects of multiple treatments, to inform reimbursement decisions. Recently, a framework for dose-response model-based network meta-analysis (MBNMA) has been proposed that combines, often nonlinear, MBMA modelling with the statistically robust properties of NMA. Here, we aim to extend this framework to time-course models. METHODS: We propose a Bayesian time-course MBNMA modelling framework for continuous summary outcomes that allows for nonlinear modelling of multiparameter time-course functions, accounts for residual correlation between observations, preserves randomisation by modelling relative effects, and allows for testing of inconsistency between direct and indirect evidence on the time-course parameters. We demonstrate our modelling framework using an illustrative dataset of 23 trials investigating treatments for pain in osteoarthritis. RESULTS: Of the time-course functions that we explored, the Emax model gave the best fit to the data and has biological plausibility. Some simplifying assumptions were needed to identify the ET50 , due to few observations at early follow-up times. Treatment estimates were robust to the inclusion of correlations in the likelihood. CONCLUSIONS: Time-course MBNMA provides a statistically robust framework for synthesising evidence on multiple treatments at multiple time points. The use of placebo-controlled studies in drug-development means there is limited potential for inconsistency. The methods can inform drug-development decisions and provide the rigour needed in the reimbursement decision-making process.