Adaptive Clinical Trial Designs Research Articles

Adaptive biased-coin designs for clinical trials with several treatments

Adaptive biased-coin designs for clinical trials with several treatments

Lung Cancer Clinical Trials with a Seamless Phase II/III Design: Systematic Review

Current lung cancer clinical research focuses on biomarkers and personalized treatment strategies. Adaptive clinical trial designs have gained significant ground due to their increased flexibility, compared to the conventional model of drug development from phase I to phase IV trials. One such adaptive approach is the seamless phase II/III design, which has been used to reduce the total sample size and drug development time. In this context, an algorithmic systematic search was conducted in MEDLINE (PUBMED), SCOPUS, EMBASE and Cochrane Central Register of Controlled Trials until 31 June 2022 in order to identify lung cancer trials of systematic treatments that have employed the seamless phase II/III methodology and to describe their characteristics. The search strategy yielded a total of 1420 records that were screened through their title and abstract; 28 eligible trials were included in the systematic review. Based on the study endpoints, the most common subtype included phase II/III trials with inefficacy/futility analyses (61%; 17/28), followed by dose escalation phase II/III trials (18%; 5/28), one multi-arm multi stage trial and 5 trials with other design (18%). Most eligible trials were open-label (71%; 20/27), included patients with non-small cell lung cancer (82%; 23/28), evaluated targeted therapies and/or immunotherapies (82%; 23/28) and recruited patients with advanced disease (89.3%; 25/28). In conclusion, the seamless phase II/III design is a feasible and suitable approach in lung cancer research, with distinct design subcategories according to study endpoints.

A Comparative Tutorial of Bayesian Sequential Design and Reinforcement Learning

Reinforcement learning (RL) is a computational approach to reward-driven learning in sequential decision problems. It implements the discovery of optimal actions by learning from an agent interacting with an environment rather than from supervised data. We contrast and compare RL with traditional sequential design, focusing on simulation-based Bayesian sequential design (BSD). Recently, there has been an increasing interest in RL techniques for healthcare applications. We introduce two related applications as motivating examples. In both applications, the sequential nature of the decisions is restricted to sequential stopping. Rather than a comprehensive survey, the focus of the discussion is on solutions using standard tools for these two relatively simple sequential stopping problems. Both problems are inspired by adaptive clinical trial design. We use examples to explain the terminology and mathematical background that underlie each framework and map one to the other. The implementations and results illustrate the many similarities between RL and BSD. The results motivate the discussion of the potential strengths and limitations of each approach.

Multi-arm covariate adjusted response adaptive designs for ordinal outcome clinical trials.

Covariate adjusted response adaptive designs are developed with ordinal categorical responses for phase III clinical trial involving multiple treatments. Stochastic ordering principle is used to order the treatments according to effectiveness and consequently allocation functions are developed by combining the cumulative odds ratios suitably. The performance of the proposed designs is investigated through relevant exact as well as large sample measures. To investigate the performance in a real situation, a real clinical trial involving lung cancer patients is further redesigned using the proposed allocation design.

Adaptive clinical trial designs with blinded selection of binary composite endpoints and sample size reassessment.

For randomized clinical trials where a single, primary, binary endpoint would require unfeasibly large sample sizes, composite endpoints (CEs) are widely chosen as the primary endpoint. Despite being commonly used, CEs entail challenges in designing and interpreting results. Given that the components may be of different relevance and have different effect sizes, the choice of components must be made carefully. Especially, sample size calculations for composite binary endpoints depend not only on the anticipated effect sizes and event probabilities of the composite components but also on the correlation between them. However, information on the correlation between endpoints is usually not reported in the literature which can be an obstacle for designing future sound trials. We consider two-arm randomized controlled trials with a primary composite binary endpoint and an endpoint that consists only of the clinically more important component of the CE. We propose a trial design that allows an adaptive modification of the primary endpoint based on blinded information obtained at an interim analysis. Especially, we consider a decision rule to select between a CE and its most relevant component as primary endpoint. The decision rule chooses the endpoint with the lower estimated required sample size. Additionally, the sample size is reassessed using the estimated event probabilities and correlation, and the expected effect sizes of the composite components. We investigate the statistical power and significance level under the proposed design through simulations. We show that the adaptive design is equally or more powerful than designs without adaptive modification on the primary endpoint. Besides, the targeted power is achieved even if the correlation is misspecified at the planning stage while maintaining the type 1 error. All the computations are implemented in R and illustrated by means of a peritoneal dialysis trial.

A Bayesian adaptive design for clinical trials of rare efficacy outcomes with multiple definitions.

Bayesian adaptive designs for clinical trials have gained popularity in the recent years due to the flexibility and efficiency that they offer. We consider the scenario where the outcome of interest comprises events with relatively low risk of occurrence and different case definitions resulting in varying control group risk assumptions. This is a scenario that occurs frequently for infectious diseases in global health research. We propose a Bayesian adaptive design that incorporates different case definitions of the outcome of interest that vary in stringency. A set of stopping rules are proposed where superiority and futility may be concluded with respect to different outcome definitions and therefore maintain a realistic probability of stopping in trials with low event rates. Through a simulation study, a variety of stopping rules and design configurations are compared. The simulation results are provided in an interactive web application that allows the user to explore and compare the design operating characteristics for a variety of assumptions and design parameters with respect to different outcome definitions. The results for select simulation scenarios are provided in the article. Bayesian adaptive designs offer the potential for maximizing the information learned from the data collected through clinical trials. The proposed design enables monitoring and utilizing multiple composite outcomes based on rare events to optimize the trial design operating characteristics.

CLRM-19 USING FUNCTIONAL PRECISION MEDICINE TO GUIDE CLINICAL TRIAL ENROLLMENT IN GBM

Interventional clinical trials in glioblastoma (GBM) have been consistently disappointing, attributable to various factors such as ineffective therapies, inadequate trial designs including lack of control arms, or enrollment criteria that do not represent real-world practice. Novel paradigms for clinical trial design(s) in GBM are desperately needed to produce clinically useful patient outcomes. KIYATEC has developed a patient- and tumor-specific technology platform to evaluate cellular response(s) to therapeutics using 3D cell culture methods that provide functional, patient-specific response predictions. Employing KIYATEC’s technology to screen compounds against both primary patient-, and PDX-derived specimens, enables clinical prioritization of early-stage assets most likely to have therapeutic response in vivo. In addition, KIYATEC’s 3D Predict™ Glioma test has shown clinical correlation of test-predicted response(s) and clinical outcomes in GBM patients. Incorporating KIYATEC’s 3D ex vivo technology into GBM therapeutic development is positioned to accelerate more successful trial results by 1) identifying early-stage compounds likely to possess clinical effects in vivo, and 2) prospectively identifying patients expected to have a clinical response to therapeutics in development. 3D Predict Glioma provides patient-specific responses within 7-10 days of tissue acquisition, providing an avenue for test integration into adaptive clinical trials, whereby functional characterization could provide gating information relating to trial execution. Specifically, functional response prediction may play a pivotal role in identifying newly diagnosed patients who might derive greater benefit from clinical trials compared to standard of care and by optimizing effective therapeutic selection in the recurrent setting. Therefore, a priori knowledge of an early-stage assets’ potential, combined with therapeutic sensitivity of individual patient tissue, may facilitate a new era for adaptive clinical trial design by assimilating KIYATEC’s analytically and clinically validated test into various steps of clinical trial execution such as randomization, stratification, therapy-switching, or compound addition/discontinuation.

Bayesian adaptive clinical trial designs for respiratory medicine.

The use of Bayesian adaptive designs for clinical trials has increased in recent years, particularly during the COVID‐19 pandemic. Bayesian adaptive designs offer a flexible and efficient framework for conducting clinical trials and may provide results that are more useful and natural to interpret for clinicians, compared to traditional approaches. In this review, we provide an introduction to Bayesian adaptive designs and discuss its use in recent clinical trials conducted in respiratory medicine. We illustrate this approach by constructing a Bayesian adaptive design for a multi‐arm trial that compares two non‐invasive ventilation treatments to standard oxygen therapy for patients with acute cardiogenic pulmonary oedema. We highlight the benefits and some of the challenges involved in designing and implementing Bayesian adaptive trials.

Targeted Therapy in Early Stage Non-small Cell Lung Cancer.

Tyrosine kinase inhibitors (TKIs) have dramatically improved tumor response rates and survival benefits in advanced oncogenic non-small-cell lung cancer (NSCLC). Given the impressive success, a renewed interest has been raised in the study of these agents in the perioperative setting. Preliminary data have shown dramatic effectiveness compared to conventional chemotherapy. Given the explicit need to induce durable responses and raise cure rates, we summarize the current progression, identify key challenges, and raise potential opportunities for perioperative targeted therapy that range from precise biomarkers to optimal adjuvant regimens for individual patients. As perioperative treatment indeed provides researchers with a unique platform to address the challenges mentioned above, investigators could obtain a comprehensive analysis of genomic profiling and trace resistance mechanisms. Multidisciplinary collaboration and adaptive clinical trial designs are warranted to integrate translational research into personalized perioperative TKI treatment paradigms.

An adaptive clinical trial design to identify the target dose of tenecteplase for treatment of acute pulmonary embolism.

Fibrinolytic therapy with tenecteplase has been proposed for patients with pulmonary embolism but the optimal dose is unknown. Higher-than-necessary dosing is likely to cause excess bleeding. We designed an adaptive clinical trial to identify the minimum and assumed safest dose of tenecteplase that maintains efficacy. We propose a Bayesian adaptive, placebo-controlled, group-sequential dose-finding trial using response-adaptive randomization to preferentially allocate subjects to the most promising doses, dual analyses strategies (continuous and dichotomized) using a gatekeeping approach to maximize clinical impact, and interim stopping rules to efficiently address competing trial objectives. The operating characteristics of the proposed design were evaluated using Monte Carlo simulation across multiple hypothetical efficacy scenarios. Simulation demonstrated response-adaptive randomization can preferentially allocate subjects to doses which appear to be performing well based on interim data. Interim decision-making, including the interim evaluation of both analysis strategies with gatekeeping, allows the trial to continue enrollment when success with the dichotomized analysis strategy appears sufficiently likely and to stop enrollment and declare superiority based on the continuous analysis strategy when there is little chance of ultimately declaring superiority with the dichotomized analysis. The proposed design allows evaluation of a greater number of dose levels than would be possible with a non-adaptive design and avoids the need to choose either the continuous or the dichotomized analysis strategy for the primary endpoint.

Two-stage re-estimation adaptive design: a simulation study


 
 
 Background: adaptive clinical trial design has been proposed as a promising new approach to improve the drug discovery process. Among the many options available, adaptive sample size re-estimation is of great interest mainly because of its ability to avoid a large ‘up-front’ commitment of resources. In this simulation study, we investigate the statistical properties of two-stage sample size re-estimation designs in terms of type I error control, study power and sample size, in comparison with the fixed-sample study.
 Methods: we simulated a balanced two-arm trial aimed at comparing two means of normally distributed data, using the inverse normal method to combine the results of each stage, and considering scenarios jointly defined by the following factors: the sample size re-estimation method, the information fraction, the type of group sequential boundaries and the use of futility stopping. Calculations were performed using the statistical software SAS™ (version 9.2).
 Results: under the null hypothesis, any type of adaptive design considered maintained the prefixed type I error rate, but futility stopping was required to avoid the unwanted increase in sample size. When deviating from the null hypothesis, the gain in power usually achieved with the adaptive design and its performance in terms of sample size were influenced by the specific design options considered.
 Conclusions: we show that adaptive designs incorporating futility stopping, a sufficiently high information fraction (50-70%) and the conditional power method for sample size re-estimation have good statistical properties, which include a gain in power when trial results are less favourable than anticipated. 
 

Bayesian adaptive design for pediatric clinical trials incorporating a community of prior beliefs

BackgroundPediatric population presents several barriers for clinical trial design and analysis, including ethical constraints on the sample size and slow accrual rate. Bayesian adaptive design methods could be considered to address these challenges in pediatric clinical trials.MethodsWe developed an innovative Bayesian adaptive design method and demonstrated the approach as a re-design of a published phase III pediatric trial. The innovative design used early success criteria based on skeptical prior and early futility criteria based on enthusiastic prior extrapolated from a historical adult trial, and the early and late stopping boundaries were calibrated to ensure a one-sided type I error of 2.5%. We also constructed several alternative designs which incorporated only one type of prior belief and the same stopping boundaries. To identify a preferred design, we compared operating characteristics including power, expected trial size and trial duration for all the candidate adaptive designs via simulation when performing an increasing number of equally spaced interim analyses.ResultsWhen performing an increasing number of equally spaced interim analyses, the innovative Bayesian adaptive trial design incorporating both skeptical and enthusiastic priors at both interim and final analyses outperforms alternative designs which only consider one type of prior belief, because it allows more reduction in sample size and trial duration while still offering good trial design properties including controlled type I error rate and sufficient power.ConclusionsDesigning a Bayesian adaptive pediatric trial with both skeptical and enthusiastic priors can be an efficient and robust approach for early trial stopping, thus potentially saving time and money for trial conduction.

A two-stage adaptive clinical trial design with data-driven subgroup identification at interim analysis.

In this paper, we consider randomized controlled clinical trials comparing two treatments in efficacy assessment using a time to event outcome. We assume a relatively small number of candidate biomarkers available in the beginning of the trial, which may help define an efficacy subgroup which shows differential treatment effect. The efficacy subgroup is to be defined by one or two biomarkers and cut-offs that are unknown to the investigator and must be learned from the data. We propose a two-stage adaptive design with a pre-planned interim analysis and a final analysis. At the interim, several subgroup-finding algorithms are evaluated to search for a subgroup with enhanced survival for treated versus placebo. Conditional powers computed based on the subgroup and the overall population are used to make decision at the interim to terminate the study for futility, continue the study as planned, or conduct sample size recalculation for the subgroup or the overall population. At the final analysis, combination tests together with closed testing procedures are used to determine efficacy in the subgroup or the overall population. We conducted simulation studies to compare our proposed procedures with several subgroup-identification methods in terms of a novel utility function and several other measures. This research demonstrated the benefit of incorporating data-driven subgroup selection into adaptive clinical trial designs.

Glioblastoma: The Current State of Biology and Therapeutic Strategies.

Over the past two decades, there have been advances in surgical technologies and chemoradiation strategies for glioblastoma, yet durable remissions are rarely seen. As the biological challenges and genetic basis of glioblastoma have become more understood, new therapeutic strategies may lead to more durable clinical responses and long-term remissions. We believe specialized academic centers that form meaningful corporate partnerships to complement basic science infrastructure and use adaptive clinical trial designs will achieve more rapid translation of innovative approaches to glioblastoma. Here we outline the core biological challenges to be overcome in the management of glioblastoma.

Semi-blinded design in clinical trials

The semi-blinded design can be applied for the adaptive design in clinical trials. According the the FDA guidance, an adaptive clinical design study would modify the study design and hypotheses based on the analysis of accrued data. An independent Data Monitoring Committee analyzes the unblinded data in the middle of a clinical trial and makes recommendations to the sponsor if any action is needed. The unblinded interim analysis could potentially introduce bias and put the integrity of the trial at risk. The proposed methodology is technically blinded for the interim analysis. In fact, actual treatment labels are not released. It is based on a treatment re-labeling approach in which the dummy treatments associated with the actual treatments via a pre-specified transition probability matrix are released in the dataset. The semi-blinded test statistic depends on this transition probability matrix. However, it can be evaluated without knowing the transition probability matrix if this matrix is a Latin square. As the stagnation probability approach to 0.5, the more information we will lose. It will lead to obtain a lower power of testing hypothesis, and may not be able to show the utility or efficacy of study treatment if this probability is close to 0.5. The higher the stagnation probability, the higher disclosure risk we have. The randomization codes may be broken as the stagnation probability approach to 1. The sponsor should strike a balance between the statistical power and disclosure risk.

A comparison of estimation methods adjusting for selection bias in adaptive enrichment designs with time-to-event endpoints.

Adaptive enrichment designs in clinical trials have been developed to enhance drug developments. They permit, at interim analyses during the trial, to select the sub-populations that benefits the most from the treatment. Because of this selection, the naive maximum likelihood estimation of the treatment effect, commonly used in classical randomized controlled trials, is biased. In the literature, several methods have been proposed to obtain a better estimation of the treatments' effects in such contexts. To date, most of the works have focused on normally distributed endpoints, and some estimators have been proposed for time-to-event endpoints but they have not all been compared side-by-side. In this work, we conduct an extensive simulation study, inspired by a real case-study in heart failure, to compare the maximum-likelihood estimator (MLE) with an unbiased estimator, shrinkage estimators, and bias-adjusted estimators for the estimation of the treatment effect with time-to-event data. The performances of the estimators are evaluated in terms of bias, variance, and mean squared error. Based on the results, along with the MLE, we recommend to provide the unbiased estimator and the single-iteration bias-adjusted estimator: the former completely eradicates the selection bias, but is highly variable with respect to a naive estimator; the latter is less biased than the MLE estimator and only slightly more variable.

Adaptive Clinical Trial Designs with Surrogates: When Should We Bother?

The success of a new drug is assessed within a clinical trial using a primary endpoint, which is typically the true outcome of interest—for example, overall survival. However, regulators sometimes approve drugs using a surrogate outcome—an intermediate indicator that is faster or easier to measure than the true outcome of interest—for example, progression-free survival—as the primary endpoint when there is demonstrable medical need. Although using a surrogate outcome (instead of the true outcome) as the primary endpoint can substantially speed up clinical trials and lower costs, it can also result in poor drug-approval decisions because the surrogate is not a perfect predictor of the true outcome. In this paper, we propose combining data from both surrogate and true outcomes to improve decision making within a late-phase clinical trial. In contrast to broadly used clinical trial designs that rely on a single primary endpoint, we propose a Bayesian adaptive clinical trial design that simultaneously leverages both observed outcomes to inform trial decisions. We perform comparative statics on the relative benefit of our approach, illustrating the types of diseases and surrogates for which our proposed design is particularly advantageous. Finally, we illustrate our proposed design on metastatic breast cancer. We use a large-scale clinical trial database to construct a Bayesian prior and simulate our design on a subset of clinical trials. We estimate that our design would yield a 16% decrease in trial costs relative to existing clinical trial designs, while maintaining the same Type I/II error rates. This paper was accepted by J. George Shanthikumar, Management Science Special Section on Data-Driven Prescriptive Analytics.

Adaptive designs for clinical trials have potential advantages, but statistical challenges lurk!

Linked Article: Cro et al. Br J Dermatol 2022; 186:245–256.

Innovative Two-Stage Seamless Adaptive Clinical Trial Designs

Innovative Two-Stage Seamless Adaptive Clinical Trial Designs

An optimal response adaptive design for multi-treatment clinical trials with ordinal categorical outcomes

ABSTRACT In clinical trials, fixed randomizations in a prefixed proportion (e.g. 1:1 or 2:1 for two treatment trials) may be adopted to allocate the entering patients among the competing treatments. However, such an allocation procedure ignores the knowledge obtained from the accrued information on the performance of the treatments until that point. However, while allocating, a fixed randomization may favor the most and the least effective treatments in a prefixed manner, and hence becomes instrumental to induce a conflict with the “individual ethics” requirement. Adaptive allocation designs are considered instead, for their ability to dynamically settle the issue of running randomization towards the treatment doing better – all using the available data but with a scope to compromise in statistical precision. Although most of the developments are pertinent to binary, continuous and survival responses, ordinal categorical responses are natural outcomes in many disciplines of clinical trials like Orthopedics and Ophthalmology. Therefore, to balance between ethics and precision in the context of a multi-treatment clinical trial producing ordinal categorical responses, an optimal response adaptive design is derived by minimizing a measure of “precision” subject to constrained number of “failures” ensuring higher number of assignments to the “best” treatment. Related design and inference-based characteristics are extensively studied – both theoretically and empirically. Further, the practical applicability of the developed design is envisaged through re-designing of a real clinical trial, where the responses are immediate and are measured in ordinal categorical scale.