Statistics in Oncology Series
Time-To-Event Data: An Overview and Analysis ConsiderationsIn oncology, overall survival and progression-free survival are common time-to-event end points used to measure treatment efficacy. Analyses of this type of data rely on a complex statistical framework and the analysis results are only valid when the data meet certain assumptions. This article provides an overview of time-to-event data, the basic mechanics of common analysis methods, and issues often encountered when analyzing such data. Our goal is to provide clinicians and other lung cancer researchers with the knowledge to choose the appropriate time-to-event analysis methods and to interpret the outcomes of such analyses appropriately.
Statistical Models in Clinical StudiesAlthough statistical models serve as the foundation of data analysis in clinical studies, their interpretation requires sufficient understanding of the underlying statistical framework. Statistical modeling is inherently a difficult task because of the general lack of information of the nature of observable data. In this article, we aim to provide some guidance when using regression models to aid clinical researchers to better interpret results from their statistical models and to encourage investigators to collaborate with a statistician to ensure that their studies are designed and analyzed appropriately.
Biomarker Discovery and Validation: Statistical ConsiderationsBiomarkers have various applications including disease detection, diagnosis, prognosis, prediction of response to intervention, and disease monitoring. In this era of precision medicine, having validated biomarkers to inform clinical decision making is more important than ever. In this article, we discuss best the practices and potential issues in biomarker discovery and validation. We encourage team science partnerships to bring cutting-edge discovery from bench to bedside, leading to improved patient care and outcomes.
Statistical Considerations for Subgroup AnalysesRandomized clinical trials (RCTs) are conducted to evaluate the effect of an experimental treatment on outcomes of a target patient population. Eligibility criteria for large trials are often broad to ensure that the trial results can be generalized to a larger patient population. Subgroup analyses, either specified a priori or post hoc, are perfo rmed to evaluate the treatment effect specific to a subgroup of treated patients. Regardless of whether a subgroup analysis is specified a priori or post hoc, investigators must consider inflated false-positive rates, chance differences in observed treatment effects, low power for the comparisons of interest, and interpretation of the subgroup results.
Clinical Versus Statistical Significance in Studies of Thoracic MalignanciesThe scientific synergy between statistics and medical research often leads to controversy when considering statistical versus clinical significance of findings from studies that are neither overwhelmingly practice-changing nor clearly negative. Studies can reach statistical significance but provide evidence that is not clinically meaningful, or results could not be statistically significant but very clinically relevant. To fully appreciate the debate about studies that fall in this gray area, one must understand how to interpret several features of statistical design and the interpretation of results.
Principles of Good Clinical Trial DesignClinical trials are a fundamental component of medical research and serve as the main route to obtain evidence of the safety and efficacy of treatment before its approval. A trial’s ability to provide the intended evidence hinges on appropriate design, background knowledge, trial rationale to sample size, and interim monitoring rules. In this article, we present some general design principles for investigators and their research teams to consider when planning to conduct a trial.
Random Survival ForestsIn the article by Chen et al,1 the authors used Random Survival Forests (RSF) as part of their approach for analyzing the data. In this note, we will explain RSF in a nontechnical way; precise details of the RSF method are described in the article by Ishwaran et al.2 RSF are an adaptation of Random Forests (RF)3 designed to be used for survival data. Software to run RSF is described in the article by Ishwaran and Kogalur.4 RSF differs from RF in that the response data are a survival time, which may be censored.
Systematic Reviews and Meta-Analysis of Published Studies: An Overview and Best PracticesSystematic reviews and meta-analytic approaches are widely used in the clinical arena to integrate outcome data from published studies in a patient population that address a set of related research hypotheses. The credibility of this line of research is dependent on how the studies are chosen, how the data are assembled, and how the results are reported. In this brief report, we provide an overview of the minimum set of reporting requirements for systematic reviews and meta-analyses based on the Preferred Reporting Items of Systematic reviews and Meta-Analyses guidelines.
Biostatistics Primer: What a Clinician Ought to Know: Hazard RatiosHazard ratios (HRs) are used commonly to report results from randomized clinical trials in oncology. However, they remain one of the most perplexing concepts for clinicians. A good understanding of HRs is needed to effectively interpret the medical literature to make important treatment decisions. This article provides clear guidelines to clinicians about how to appropriately interpret HRs. While this article focuses on the commonly used methods, the authors acknowledge that other statistical methods exist for analyzing survival data.
All-Comers versus Enrichment Design Strategy in Phase II TrialsDesigns for biomarker validation have been proposed and used in the phase III oncology clinical trial setting. Broadly, these designs follow either an enrichment (i.e., targeted) strategy or an all-comers (i.e., unselected) strategy. An enrichment design screens patients for the presence or absence of a marker or a panel of markers and then only includes patients who either have or do not have a certain marker characteristic or profile. In contrast, all patients meeting the eligibility criteria (regardless of a particular biomarker status) are entered into an all-comers design.
Measures of Interrater AgreementKappa statistics is used for the assessment of agreement between two or more raters when the measurement scale is categorical. In this short summary, we discuss and interpret the key features of the kappa statistics, the impact of prevalence on the kappa statistics, and its utility in clinical research. We also introduce the weighted kappa when the outcome is ordinal and the intraclass correlation to assess agreement in an event the data are measured on a continuous scale.
Receiver Operating Characteristic Curve in Diagnostic Test AssessmentThe performance of a diagnostic test in the case of a binary predictor can be evaluated using the measures of sensitivity and specificity. However, in many instances, we encounter predictors that are measured on a continuous or ordinal scale. In such cases, it is desirable to assess performance of a diagnostic test over the range of possible cutpoints for the predictor variable. This is achieved by a receiver operating characteristic (ROC) curve that includes all the possible decision thresholds from a diagnostic test result.
Randomized Phase II Trials: Time for a New Era in Clinical Trial DesignThe classic single-arm oncology phase II trial designs for evaluating an experimental regimen/agent are limited by multiple sources of bias arising from the inability to separate trial effects (such as patient selection, trial eligibility, imaging techniques and assessment schedule, and treatment locations) from treatment effect on clinical outcomes. Changes in patient population based on biologic subsetting, newer imaging technologies, the use of alternative end points, constrained resources, and the multitude of promising therapies for a given disease make randomized phase II designs, with a concurrent control arm where necessary, attractive.
Simple Statistical Measures for Diagnostic Accuracy AssessmentThe aim of diagnostic medicine research is to estimate and compare the accuracy of diagnostic tests to provide reliable information about a patient's disease status and thereby influencing patient care. When developing screening tools, researchers evaluate the discriminating power of the screening test by using simple measures such as the sensitivity and specificity of the test, as well as the positive and negative predictive values. In this brief report, we discuss these simple statistical measures that are used to quantify the diagnostic ability of a test.