Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints

G-computation, propensity score-based methods, and targeted maximum likelihood estimator for causal inference with different covariates sets: a comparative simulation study

Controlling for confounding bias is crucial in causal inference. Distinct methods are currently employed to mitigate the effects of confounding bias. Each requires the introduction of a set of covariates, which remains difficult to choose, especially regarding the different methods. We conduct a simulation study to compare the relative performance results obtained by using four different sets of covariates (those causing the outcome, those causing the treatment allocation, those causing both the outcome and the treatment allocation, and all the covariates) and four methods: g-computation, inverse probability of treatment weighting, full matching and targeted maximum likelihood estimator. Our simulations are in the context of a binary treatment, a binary outcome and baseline confounders. The simulations suggest that considering all the covariates causing the outcome led to the lowest bias and variance, particularly for g-computation. The consideration of all the covariates did not decrease the bias but significantly reduced the power. We apply these methods to two real-world examples that have clinical relevance, thereby illustrating the real-world importance of using these methods. We propose an R package RISCA to encourage the use of g-computation in causal inference.

Revisiting the optical bandgap of semiconductors and the proposal of a unified methodology to its determination

Along the last two centuries, the story of semiconductor materials ranged from a mix of disbelief and frustration to one of the most successful technological achievements ever seen. Such a progress comprised the development of materials and models that, allied to the knowledge provided by spectroscopic techniques, resulted in the (nowadays) omnipresent electronic gadgets. Within this context, optically-based methods were of special importance since, amongst others, they presented details about the electronic states and energy bandgap E_gap of semiconductors which, ultimately, decided about their application in devices. Stimulated by these aspects, this work investigated the semiconductors silicon, germanium, and gallium-arsenide in the crystalline (bulk and powder) and amorphous (film) forms. The detailed analysis of the experimental results indicates that accurate E_gap values can be obtained by fitting a sigmoid (Boltzmann) function to their corresponding optical absorption spectra. The method is straightforward and, contrary to the traditional approaches to determine E_gap, it is exempt from errors due to experimental spectra acquisition and data processing. Additionally, it complies with the requirements of direct, indirect, and amorphous bandgap semiconductors, and it is able to probe the (dis)order of the material as well. In view of these characteristics, a new−unified methodology based on the fitting of the absorption spectrum with a Boltzmann function is being proposed to efficiently determine the optical bandgap of semiconductor materials.

Sex-specific glioma genome-wide association study identifies new risk locus at 3p21.31 in females, and finds sex-differences in risk at 8q24.21

Incidence of glioma is approximately 50% higher in males. Previous analyses have examined exposures related to sex hormones in women as potential protective factors for these tumors, with inconsistent results. Previous glioma genome-wide association studies (GWAS) have not stratified by sex. Potential sex-specific genetic effects were assessed in autosomal SNPs and sex chromosome variants for all glioma, GBM and non-GBM patients using data from four previous glioma GWAS. Datasets were analyzed using sex-stratified logistic regression models and combined using meta-analysis. There were 4,831 male cases, 5,216 male controls, 3,206 female cases and 5,470 female controls. A significant association was detected at rs11979158 (7p11.2) in males only. Association at rs55705857 (8q24.21) was stronger in females than in males. A large region on 3p21.31 was identified with significant association in females only. The identified differences in effect of risk variants do not fully explain the observed incidence difference in glioma by sex.

The Value of lncRNA NEAT1 as a Prognostic Factor for Survival of Cancer Outcome: A Meta-Analysis

The present meta-analysis aimed to analyze available data to identify the prognostic role of NEAT1 in multiple carcinomas. A systematic search was performed by using several computerized databases from inception to June 7, 2017. The quantity of the publications was assessed according to MOOSE checklist. Pooled HRs with 95% CI was calculated to summarize the effect. A total of 12 studies with 3,262 cancer patients were pooled in the analysis to evaluate the prognostic value of NEAT1 in multiple tumors. High expression levels of NEAT1 were demonstrated to be associated with poor OS (HR = 1.71, 95%CI: 1.37-2.14, P < 0.001) and tumor progression (III/IV vs. I/II: HR 1.76, 95%CI: 1.40-2.21, P < 0.00001). Subgroup analysis showed that NEAT1 detection method (qRT-PCR) and sample size (more or less than 100) did not alter the predictive value of NEAT1 on OS in various cancers. According to the meta-regression results, the large heterogeneity of meta-analysis may be attributed to the differences of NEAT1 detection method. Furthermore, elevated NEAT1 expression significantly predicted lymph node metastasis (HR: 2.10, 95%CI: 1.32-3.33, P = 0.002) and distant metastasis (HR: 2.80, 95%CI: 1.60-4.91, P = 0.0003) respectively. The results indicate that NEAT1 expression level is a prognostic biomarker for OS and metastasis in general tumors.

Meta-CART: A tool to identify interactions between moderators in meta-analysis

In the framework of meta-analysis, moderator analysis is usually performed only univariately. When several study characteristics are available that may account for treatment effect, standard meta-regression has difficulties in identifying interactions between them. To overcome this problem, meta-CART has been proposed: an approach that applies classification and regression trees (CART) to identify interactions, and then subgroup meta-analysis to test the significance of moderator effects. The previous version of meta-CART has its shortcomings: when applying CART, the sample sizes of studies are not taken into account, and the effect sizes are dichotomized around the median value. Therefore, this article proposes new meta-CART extensions, weighting study effect sizes by their accuracy, and using a regression tree to avoid dichotomization. In addition, new pruning rules are proposed. The performance of all versions of meta-CART was evaluated via a Monte Carlo simulation study. The simulation results revealed that meta-regression trees with random-effects weights and a 0.5-standard-error pruning rule perform best. The required sample size for meta-CART to achieve satisfactory performance depends on the number of study characteristics, the magnitude of the interactions, and the residual heterogeneity.

Prognostic value of c-Met in colorectal cancer: A meta-analysis

AIM: To assess the prognostic value of c-Met status in colorectal cancer.

METHODS: We conducted a search in PubMed, Web of Science, and the Cochrane Library covering all published papers up to July 2014. Only studies assessing survival in colorectal cancer by c-Met status were included. This meta-analysis was performed by using STATA11.0.

RESULTS: Ultimately, 11 studies were included in this analysis. Meta-analysis of the hazard ratios (HR) indicated that patients with high c-Met expression have a significantly poorer overall survival (OR) (HR = 1.33, 95%CI: 1.06-1.59) and progression-free survival (PFS) (HR = 1.47, 95%CI: 1.03-1.91). Subgroup analysis showed a significant association between high c-Met expression and poorer overall survival in the hazard ratio reported (HR = 1.41, 95%CI: 1.08-1.74).

CONCLUSION: The present meta-analysis indicated that high c-Met expression was associated with poor prognosis in patients with colorectal cancer.

S-1-based combination therapy vs S-1 monotherapy in advanced gastric cancer: a meta-analysis

AIM

To assess the efficacy and safety of combination therapy based on S-1, a novel oral fluoropyrimidine, vs S-1 monotherapy in advanced gastric cancer (AGC).

METHODS

We searched PubMed, EMBASE and the Cochrane Library for eligible studies published before March 2013. Our analysis identified four randomized controlled trials involving 790 participants with AGC. The outcome measures were overall survival (OS), progression-free survival (PFS), overall response rate (ORR) and grade 3-4 adverse events.

RESULTS

Meta-analysis showed that S-1-based combination therapy significantly improved OS (HR = 0.77, 95%CI: 0.66-0.91, P = 0.002), PFS (HR = 0.58, 95%CI: 0.46-0.72, P = 0.000) and ORR (OR = 2.23, 95%CI: 1.54-3.21, P = 0.000). Sensitivity analysis further confirmed this association. Lower incidence of grade 3-4 leucopenia (OR = 4.06, 95%CI: 2.11-7.81), neutropenia (OR = 3.94, 95%CI: 2.1-7.81) and diarrhea (OR = 2.41, 95%CI: 1.31-4.44) was observed in patients with S-1 monotherapy.

CONCLUSION

S-1-based combination therapy is superior to S-1 monotherapy in terms of OS, PFS and ORR. S-1 monotherapy is associated with less toxicity.

Highlights for α-fetoprotein in determining prognosis and treatment monitoring for hepatocellular carcinoma

AIM: To explore the prognostic value in the monitoring of treatment efficacy of serial α-fetoprotein (AFP) in hepatocellular carcinoma (HCC) patients.

METHODS: We searched MEDLINE, EMBASE and COCHRANE LIBRARY through April 21, 2012, to find qualifying articles. Our overall search strategy included terms for HCC, AFP, treatment response, and prognosis. Literature was limited to English-language, human studies. Studies reporting cumulative survival rates were summarized qualitatively. For the prognostic meta-analysis, we undertook a series of meta-analyses that summarised the Cox proportional hazard ratios (HRs) by assuming a random effects model. With regards to the correlation of AFP change with radiologic response, the categorical dichotomous variables were assessed using Poisson relative risks (RRs), which were incorporated into the random effects model meta-analysis of accuracy prediction. Between-study heterogeneity was estimated by use of the I² statistic. Publication bias was evaluated using the Begg funnel plot and Egger plot. Sensitivity analyses were conducted first by separating systemic treatment estimates from locoregional therapy estimates, evaluating different AFP response cut-off point effects, and exploring the impact of different study sizes.

RESULTS: Of 142 titles identified in our original search, 11 articles (12 clinical studies) met our criteria. Six studies investigated outcome in a total of 464 cases who underwent systemic treatment, and six studies investigated outcome in a total of 510 patients who received locoregional therapy. A random-effects model meta-analysis showed that AFP response was associated with an mortality HR of 0.55 (95%CI, 0.47-0.65) across HCC in overall survival (OS) and 0.50 (95%CI, 0.38-0.65) in progression-free survival. Restricting analysis to the six eligible analyses of systemic treatment, the pooled HRs were 0.64 (95%CI, 0.53-0.77) for OS. Limiting analysis to the six analyses of locoregional therapy, the pooled HRs for OS was 0.39 (95%CI, 0.29-0.53). We showed a larger pooled HR in the 50% definition studies (HR, 0.67, 95%CI, 0.55-0.83) compared with that from the 20% definition studies (HR, 0.41, 95%CI, 0.32-0.53). Restricting analysis to the four studies including over 100 patients individually, the pooled HR was 0.65 (95%CI, 0.54-0.79), with a pooled HR for OS of 0.35 (95%CI, 0.23-0.46) in the studies of less than 100 patients. As to radiological imaging, 43.1% (155/360) of the patients in the AFP response group presented with a radiological overall response, while the response rate decreased to 11.5% (36/313) in the patients from the AFP nonresponse group. The RR of having no overall response was significantly lower in the AFP response group than the AFP nonresponse group (RR, 0.67; 95%CI, 0.61-0.75). In terms of disease control rate, 86.9% (287/330) in the AFP response group and 51.0% (153/300) in the AFP nonresponse group showed successful disease control, respectively. The RR of disease control failure, similarly, was significantly lower in the AFP response group (RR, 0.37; 95%CI, 0.23-0.58). But these findings could be overestimates because of publication and reporting bias.

CONCLUSION: HCC patients presenting with an AFP response are at decreased risk of mortality. In addition, patients with an AFP response also present with a higher overall response rate and disease control rate.

Even Highly Correlated Measures Can Add Incrementally to Predicting Recidivism Among Sex Offenders

Criterion-referenced measures, such as those used in the assessment of crime and violence, prioritize predictive accuracy (discrimination) at the expense of construct validity. In this article, we compared the discrimination and incremental validity of three commonly used criterion-referenced measures for sex offenders (Rapid Risk Assessment for Sex Offence Recidivism [RRASOR], Static-99R, and Static-2002R). In a meta-analysis of 20 samples ( n = 7,491), Static-99R and Static-2002R provided similar discrimination but outperformed the RRASOR in the prediction of sexual, violent, and any recidivism. Remarkably, despite large correlations between them ( rs ranging from .70 to .92), these risk scales consistently added incremental validity to one another. The direction of the incremental effects, however, was not consistently positive. When controlling for the other measures, high scores on the RRASOR were associated with lower risk for violent and any recidivism. We also examined different methods of combining risk scales and found that the averaging approach produced better discrimination than choosing the highest score and produced better calibration than either choosing the lowest or highest risk score. The findings reinforce the importance of understanding the psychological content of criterion-referenced measures, even when the sole purpose is to predict a particular outcome and provide some direction concerning the best methods for combining risk scales.