Commentary on Hey and Kimmelman

A Bayesian adaptive randomized phase II multicenter trial of bevacizumab with or without vorinostat in adults with recurrent glioblastoma

BACKGROUND

Bevacizumab has promising activity against recurrent glioblastoma (GBM). However, acquired resistance to this agent results in tumor recurrence. We hypothesized that vorinostat, a histone deacetylase (HDAC) inhibitor with anti-angiogenic effects, would prevent acquired resistance to bevacizumab.

METHODS

This multicenter phase II trial used a Bayesian adaptive design to randomize patients with recurrent GBM to bevacizumab alone or bevacizumab plus vorinostat with the primary endpoint of progression-free survival (PFS) and secondary endpoints of overall survival (OS) and clinical outcomes assessment (MD Anderson Symptom Inventory Brain Tumor module [MDASI-BT]). Eligible patients were adults (≥18 y) with histologically confirmed GBM recurrent after prior radiation therapy, with adequate organ function, KPS ≥60, and no prior bevacizumab or HDAC inhibitors.

RESULTS

Ninety patients (bevacizumab + vorinostat: 49, bevacizumab: 41) were enrolled, of whom 74 were evaluable for PFS (bevacizumab + vorinostat: 44, bevacizumab: 30). Median PFS (3.7 vs 3.9 mo, P = 0.94, hazard ratio [HR] 0.63 [95% CI: 0.38, 1.06, P = 0.08]), median OS (7.8 vs 9.3 mo, P = 0.64, HR 0.93 [95% CI: 0.5, 1.6, P = 0.79]) and clinical benefit were similar between the 2 arms. Toxicity (grade ≥3) in 85 evaluable patients included hypertension (n = 37), neurological changes (n = 2), anorexia (n = 2), infections (n = 9), wound dehiscence (n = 2), deep vein thrombosis/pulmonary embolism (n = 2), and colonic perforation (n = 1).

CONCLUSIONS

Bevacizumab combined with vorinostat did not yield improvement in PFS or OS or clinical benefit compared with bevacizumab alone or a clinical benefit in adults with recurrent GBM. This trial is the first to test a Bayesian adaptive design with adaptive randomization and Bayesian continuous monitoring in patients with primary brain tumor and demonstrates the feasibility of using complex Bayesian adaptive design in a multicenter setting.

Comparison of methods for control allocation in multiple arm studies using response adaptive randomization

BACKGROUND/AIMS

Response adaptive randomization has many polarizing properties in two-arm settings comparing control to a single treatment. The generalization of these features to the multiple arm setting has been less explored, and existing comparisons in the literature reach disparate conclusions. We investigate several generalizations of two-arm response adaptive randomization methods relating to control allocation in multiple arm trials, exploring how critiques of response adaptive randomization generalize to the multiple arm setting.

METHODS

We perform a simulation study to investigate multiple control allocation schemes within response adaptive randomization, comparing the designs on metrics such as power, arm selection, mean square error, and the treatment of patients within the trial.

RESULTS

The results indicate that the generalization of two-arm response adaptive randomization concerns is variable and depends on the form of control allocation employed. The concerns are amplified when control allocation may be reduced over the course of the trial but are mitigated in the methods considered when control allocation is maintained or increased during the trial. In our chosen example, we find minimal advantage to increasing, as opposed to maintaining, control allocation; however, this result reflects an extremely limited exploration of methods for increasing control allocation.

CONCLUSION

Selection of control allocation in multiple arm response adaptive randomization has a large effect on the performance of the design. Some disparate comparisons of response adaptive randomization to alternative paradigms may be partially explained by these results. In future comparisons, control allocation for multiple arm response adaptive randomization should be chosen to keep in mind the appropriate match between control allocation in response adaptive randomization and the metric or metrics of interest.

Bias in retrospective analyses of biomarker effect using data from an outcome-adaptive randomized trial

BACKGROUND/AIMS

Biomarker-stratified outcome-adaptive randomization trials, in which randomization probabilities depend on both biomarker value and outcomes of previously treated patients, are receiving increased attention in oncology research. Data from these trials can also form the basis of investigation of additional biomarkers that may not have been incorporated into the original trial design. In this article, we investigate the validity of a standard analytical method that utilizes data from a biomarker-stratified outcome-adaptive randomization trial to assess the effect of a newly identified biomarker on patient outcomes.

METHODS

In the context of an ancillary biomarker study for a two-arm phase II trial with a response endpoint, we conduct analytic and simulation studies to investigate bias in estimated biomarker effects under outcome-adaptive randomization. Conditions under which bias arises and magnitude of the bias are examined in several settings. We then propose unbiased estimators of biomarker effects with appropriate variance estimators.

RESULTS

We demonstrate that use of biomarker-stratified outcome-adaptive randomization perturbs the patient population and treatment assignments. Consequently, application of standard analysis methods to data from an outcome-adaptive randomization trial either to estimate prognostic effect of a new biomarker in uniformly treated patients or to estimate effect of treatment in relation to the new biomarker can lead to substantially biased estimates. The proposed adjusted estimators are asymptotically unbiased, and the proposed variance estimators correctly reflect the sample variability in the estimators.

CONCLUSION

This article demonstrates existence of bias when standard, naïve statistical methods are utilized to assess biomarker effects using data from a biomarker-stratified outcome-adaptive randomization trial, and hence that results from naïve analyses must be interpreted with great caution. These findings highlight that, in an era where data and specimens are increasingly being shared for biomarker studies, care must be taken to document and understand implications of the study design under which specimens or data have been obtained.

Using Adaptive Designs to Avoid Selecting the Wrong Arms in Multiarm Comparative Effectiveness Trials

Limited resources are a challenge when planning comparative effectiveness studies of multiple promising treatments, often prompting study planners to reduce the sample size to meet the financial constraints. The practical solution is often to increase the efficiency of this sample size by selecting a pair of treatments among the pool of promising treatments before the clinical trial begins. The problem with this approach is that the investigator may inadvertently leave out the most beneficial treatment. This paper demonstrates a possible solution to this problem by using Bayesian adaptive designs. We use a planned comparative effectiveness clinical trial of treatments for sialorrhea in amyotrophic lateral sclerosis as an example of the approach. Rather than having to guess at the two best treatments to compare based on limited data, we suggest putting more arms in the trial and letting response adaptive randomization (RAR) determine better arms. To ground this study relative to previous literature we first compare RAR, adaptive equal randomization (ER), arm(s) dropping, and a fixed design. Given the goals of this trial we demonstrate that we may avoid 'type III errors' - inadvertently leaving out the best treatment - with little loss in power compared to a two-arm design, even when choosing the correct two arms for the two-armed design. There are appreciable gains in power when the two arms are prescreened at random.

Outcome-adaptive randomization in clinical trials: issues of participant welfare and autonomy

Outcome-adaptive randomization (OAR) has been proposed as a corrective to certain ethical difficulties inherent in the traditional randomized clinical trial (RCT) using fixed-ratio randomization. In particular, it has been suggested that OAR redresses the balance between individual and collective ethics in favour of the former. In this paper, I examine issues of welfare and autonomy arising in relation to OAR. A central issue in discussions of welfare in OAR is equipoise, and the moral status of OAR is crucially influenced by the way in which this concept is construed. If OAR is based on a model of equipoise that demands strict indifference between competing interventions throughout the trial, such equipoise is disturbed by accruing data favouring one treatment over another; OAR seeks to redress this by weighting randomization to the seemingly superior treatment. However, this is a partial response, as patients continue to be allocated to the inferior therapy. Moreover, it rests upon considerations of aggregate harms and benefits, and does not therefore uphold individual ethics. Issues of fairness also arise, as early and late enrollees are randomized on a different basis. Fixed-ratio randomization represents a fuller and more consistent response to a loss of equipoise, as so construed. With regard to consent, the complexity of OAR poses challenges to adequate disclosure and comprehension. Additionally, OAR does not offer a remedy to the therapeutic misconception-participants' tendency to attribute treatment allocation in an RCT to individual clinical judgments, rather than to scientific considerations-and, if anything, accentuates rather than alleviates this misconception. In relation to these issues, OAR fails to offer ethical advantages over fixed-ratio randomization. More broadly, the ethical basis of OAR can be seen to lie more in collective than in individual ethics, and overall it fares worse in this territory than fixed-ratio randomization.

Impact of adaptation algorithm, timing, and stopping boundaries on the performance of Bayesian response adaptive randomization in confirmative trials with a binary endpoint

Despite the concerns of time trend in subject profiles, the use of Bayesian response adaptive randomization (BRAR) in large multicenter phase 3 confirmative trials has been reported in recent years, motivated by the potential benefits in subject ethics and/or trial efficiency. However three issues remain unclear to investigators: 1) among several BRAR algorithms, how to choose one for the specific trial setting; 2) when to start and how frequently to update the allocation ratio; and 3) how to choose the interim analyses stopping boundaries to preserve the type 1 error. In this paper, three commonly used BRAR algorithms are evaluated based on type 1 error, power, sample size, the proportion of subjects assigned to the better performing arm, and the total number of failures, under two specific trial settings and different allocation ratio update timing and frequencies. Simulation studies show that for two-arm superiority trials, none of the three BRAR algorithms has predominant benefits in both patient ethics and trial efficiency when compared to fixed equal allocation design. For a specific trial aiming to identify the best or the worst among three treatments, a properly selected BRAR algorithm and its implementation parameters are able to gain ethical and efficiency benefits simultaneously. Although the simulation results come from a specific trial setting, the methods described in this paper are generally applicable to other trials.

Adaptive Clinical Trials: Advantages and Disadvantages of Various Adaptive Design Elements

There is a wide range of adaptive elements of clinical trial design (some old and some new), with differing advantages and disadvantages. Classical interim monitoring, which adapts the design based on early evidence of superiority or futility of a treatment arm, has long been known to be extremely useful. A more recent application of interim monitoring is in the use of phase II/III designs, which can be very effective (especially in the setting of multiple experimental treatments and a reliable intermediate end point) but do have the cost of having to commit earlier to the phase III question than if separate phase II and phase III trials were performed. Outcome-adaptive randomization is an older technique that has recently regained attention; it increases trial complexity and duration without offering substantial benefits to the patients in the trial. The use of adaptive trials with biomarkers is new and has great potential for efficiently identifying patients who will be helped most by specific treatments. Master protocols in which trial arms and treatment questions are added to an ongoing trial can be especially efficient in the biomarker setting, where patients are screened for entry into different subtrials based on evolving knowledge about targeted therapies. A discussion of three recent adaptive clinical trials (BATTLE-2, I-SPY 2, and FOCUS4) highlights the issues.

Adaptive Clinical Trials: Advantages and Disadvantages of Various Adaptive Design Elements

There is a wide range of adaptive elements of clinical trial design (some old and some new), with differing advantages and disadvantages. Classical interim monitoring, which adapts the design based on early evidence of superiority or futility of a treatment arm, has long been known to be extremely useful. A more recent application of interim monitoring is in the use of phase II/III designs, which can be very effective (especially in the setting of multiple experimental treatments and a reliable intermediate end point) but do have the cost of having to commit earlier to the phase III question than if separate phase II and phase III trials were performed. Outcome-adaptive randomization is an older technique that has recently regained attention; it increases trial complexity and duration without offering substantial benefits to the patients in the trial. The use of adaptive trials with biomarkers is new and has great potential for efficiently identifying patients who will be helped most by specific treatments. Master protocols in which trial arms and treatment questions are added to an ongoing trial can be especially efficient in the biomarker setting, where patients are screened for entry into different subtrials based on evolving knowledge about targeted therapies. A discussion of three recent adaptive clinical trials (BATTLE-2, I-SPY 2, and FOCUS4) highlights the issues.

Adaptive Clinical Trials: Advantages and Disadvantages of Various Adaptive Design Elements

There is a wide range of adaptive elements of clinical trial design (some old and some new), with differing advantages and disadvantages. Classical interim monitoring, which adapts the design based on early evidence of superiority or futility of a treatment arm, has long been known to be extremely useful. A more recent application of interim monitoring is in the use of phase II/III designs, which can be very effective (especially in the setting of multiple experimental treatments and a reliable intermediate end point) but do have the cost of having to commit earlier to the phase III question than if separate phase II and phase III trials were performed. Outcome-adaptive randomization is an older technique that has recently regained attention; it increases trial complexity and duration without offering substantial benefits to the patients in the trial. The use of adaptive trials with biomarkers is new and has great potential for efficiently identifying patients who will be helped most by specific treatments. Master protocols in which trial arms and treatment questions are added to an ongoing trial can be especially efficient in the biomarker setting, where patients are screened for entry into different subtrials based on evolving knowledge about targeted therapies. A discussion of three recent adaptive clinical trials (BATTLE-2, I-SPY 2, and FOCUS4) highlights the issues.

Comparing three regularization methods to avoid extreme allocation probability in response-adaptive randomization

We examine three variations of the regularization methods for response-adaptive randomization (RAR) and compare their operating characteristics. A power transformation (PT) is applied to refine the randomization probability. The clip method is used to bound the randomization probability within specified limits. A burn-in period of equal randomization (ER) can be added before adaptive randomization (AR). For each method, more patients are assigned to the superior arm and overall response rate increase as the scheme approximates simple AR, while statistical power increases as it approximates ER. We evaluate the performance of the three methods by varying the tuning parameter to control the extent of AR to achieve the same statistical power. When there is no early stopping rule, PT method generally performed the best in yielding higher proportion to the superior arm and higher overall response rate, but with larger variability. The burn-in method showed smallest variability compared with the clip method and the PT method. With the efficacy early stopping rule, all three methods performed more similarly. The PT and clip methods are better than the burn-in method in achieving higher proportion randomized to the superior arm and higher overall response rate but burn-in method required fewer patients in the trial. By carefully choosing the method and the tuning parameter, RAR methods can be tailored to strike a balance between achieving the desired statistical power and enhancing the overall response rate.