Group-Sequential Methods for Adaptive Seamless Phase II/III Clinical Trials

Optimal sample size division in two-stage seamless designs

Inferentially seamless 2/3 designs are increasingly popular in clinical trials. It is important to understand their relative advantages compared with separate phase 2 and phase 3 trials, and to understand the consequences of design choices such as the proportion of patients included in the phase 2 portion of the design. Extending previous work in this area, we perform a simulation study across multiple numbers of arms and efficacy response curves. We consider a design space crossing the choice of a separate versus seamless design with the choice of allocating 0%-100% of available patients in phase 2, with the remainder in phase 3. The seamless designs achieve greater power than their separate trial counterparts. Importantly, the optimal seamless design is more robust than the optimal separate program, meaning that one range of values for the proportion of patients used in phase 2 (30%-50% of the total phase 2/3 sample size) is nearly optimal for a wide range of response scenarios. In contrast, a percentage of patients used in phase 2 for separate trials may be optimal for some alternative scenarios but decidedly inferior for other alternative scenarios. When operationally and scientifically viable, seamless trials provide superior performance compared with separate phase 2 and phase 3 trials. The results also provide guidance for the implementation of these trials in practice.

Using dichotomized survival data to construct a prior distribution for a Bayesian seamless Phase II/III clinical trial

Master protocol designs allow for simultaneous comparison of multiple treatments or disease subgroups. Master protocols can also be designed as seamless studies, in which two or more clinical phases are considered within the same trial. They can be divided into two categories: operationally seamless, in which the two phases are separated into two independent studies, and inferentially seamless, in which the interim analysis is considered an adaptation of the study. Bayesian designs are scarcely studied. Our aim is to propose and compare Bayesian operationally seamless Phase II/III designs using a binary endpoint for the first stage and a time-to-event endpoint for the second stage. At the end of Phase II, arm selection is based on posterior (futility) and predictive (selection) probabilities. The results of the first phase are then incorporated into prior distributions of a time-to-event model. Simulation studies showed that Bayesian operationally seamless designs can approach the inferentially seamless counterpart, allowing for an increasing simulated power with respect to the operationally frequentist design.

Point and interval estimation in two-stage adaptive designs with time to event data and biomarker-driven subpopulation selection

In personalized medicine, it is often desired to determine if all patients or only a subset of them benefit from a treatment. We consider estimation in two-stage adaptive designs that in stage 1 recruit patients from the full population. In stage 2, patient recruitment is restricted to the part of the population, which, based on stage 1 data, benefits from the experimental treatment. Existing estimators, which adjust for using stage 1 data for selecting the part of the population from which stage 2 patients are recruited, as well as for the confirmatory analysis after stage 2, do not consider time to event patient outcomes. In this work, for time to event data, we have derived a new asymptotically unbiased estimator for the log hazard ratio and a new interval estimator with good coverage probabilities and probabilities that the upper bounds are below the true values. The estimators are appropriate for several selection rules that are based on a single or multiple biomarkers, which can be categorical or continuous.

Seamless phase 2/3 oncology trial design with flexible sample size determination

Conventional seamless phase 2/3 design with fixed sample size determination (SSD) has gained its popularity in oncology drug development due to attractive features such as significantly shortening the development timeline, minimizing sample size, as well as early decision making. However, this design is not immune to inaccurate treatment effect assumption when only limited efficacy data are available at study design stage. We propose an innovative seamless phase 2/3 study design with flexible SSD for oncology trials, in which the trial is designed under a distribution of treatment effect instead of one single assumption due to huge uncertainty of treatment effect at design stage and the sample size for end of phase 3 analysis is not predetermined at design stage, but rather dynamically determined based on observed treatment effect at phase 2 portion. Some practical sample size determination rules for end of phase 3 analysis will be discussed. The proposed design can lead to reduced sample size or/and improved power compared with conventional seamless phase 2/3 design with fixed SSD. This innovative study design can be especially useful for programs with aggressive development strategy to expedite the process in delivering efficacious treatment to patients.

Impact of Autonomic Regulation Therapy in Patients with Heart Failure

Background:

The ANTHEM-HFrEF (Autonomic Regulation Therapy to Enhance Myocardial Function and Reduce Progression of Heart Failure with Reduced Ejection Fraction) pivotal study is an adaptive, open-label, randomized, controlled study evaluating whether autonomic regulation therapy will benefit patients with advanced HFrEF. While early-phase studies have supported potential use of vagus nerve stimulation to deliver autonomic regulation therapy for HFrEF, results of larger clinical trials have been inconsistent. The ANTHEM-HFrEF study uses a novel design, with adaptive sample size selection, evaluating effects on morbidity and mortality as well as symptoms and function.

Methods:

The ANTHEM-HFrEF study will randomize patients (2:1) to autonomic regulation therapy plus guideline-directed medical therapy, or guideline-directed medical therapy alone. The morbidity and mortality trial utilizes a conventional frequentist approach for analysis of the primary outcome end point—reduction in the composite of cardiovascular death or first HF hospitalization—and a Bayesian adaptive approach toward sample size selection. Embedded within the ANTHEM-HFrEF study is a second trial evaluating improvement in symptoms and function. Symptom/function success will require meeting 2 risk-related conditions (trend for reduced cardiovascular death/HF hospitalization and sufficient freedom from device-related serious adverse events) and 3 efficacy end point components (changes in left ventricular EF, 6-minute walk distance, and Kansas City Cardiomyopathy Questionnaire overall score).

Conclusions:

Vagus nerve stimulation remains a promising, yet unproven treatment in HFrEF. A successful ANTHEM-HFrEF pivotal study would provide an important advance in HFrEF treatment and offer a model for expediting evaluation of new therapies.

Clinical Trial Registration:

URL: http://www.clinicaltrials.gov . Unique identifier: NCT03425422.

Group sequential designs with prospectively planned rules for subpopulation enrichment

We propose a class of randomized trial designs aimed at gaining the advantages of wider generalizability and faster recruitment while mitigating the risks of including a population for which there is greater a priori uncertainty. We focus on testing null hypotheses for the overall population and a predefined subpopulation. Our designs have preplanned rules for modifying enrollment criteria based on data accrued at interim analyses. For example, enrollment can be restricted if the participants from a predefined subpopulation are not benefiting from the new treatment. Our designs have the following features: the multiple testing procedure fully leverages the correlation among statistics for different populations; the asymptotic familywise Type I error rate is strongly controlled; for outcomes that are binary or normally distributed, the decision rule and multiple testing procedure are functions of the data only through minimal sufficient statistics. Our designs incorporate standard group sequential boundaries for each population of interest; this may be helpful in communicating the designs, because many clinical investigators are familiar with such boundaries, which can be summarized succinctly in a single table or graph. We demonstrate these designs through simulations of a Phase III trial of a new treatment for stroke. User-friendly, free software implementing these designs is described. Copyright © 2016 John Wiley & Sons, Ltd.

Modeling the relationship between progression-free survival and overall survival: the phase II/III trial

The standard phase II trial design has changed dramatically over the past decade. Randomized phase II studies have essentially become the standard phase II design in oncology for a variety of reasons. The use of these designs is motivated by concerns about the use of historical data to determine if a new agent or regimen shows promise of activity. However, randomized phase II designs come with the cost of increased study duration and patient resources. Progression-free survival (PFS) is an important endpoint used in many phase II designs. In many clinical settings, changes in PFS with the introduction of a new treatment may represent true benefit in terms of the gold standard outcome, overall survival (OS). The phase II/III design has been proposed as an approach to shorten the time of discovery of an active regimen. In this article, design considerations for a phase II/III trial are discussed and presented in terms of a model defining the relationship between OS and PFS. The design is also evaluated using 15 phase III trials completed in the Southwest Oncology Group (SWOG) between 1990 and 2005. The model provides a framework to evaluate the validity and properties of using a phase II/III design. In the evaluation of SWOG trials, three of four positive studies would have also proceeded to the final analysis and 10 of 11 negative studies would have stopped at the phase II analysis if a phase II/III design had been used. Through careful consideration and thorough evaluation of design properties, substantial gains could occur using this approach.