Decision-making in early clinical drug development

Effect of melatonin on postoperative cognitive function in elderly patients submitted to transurethral resection of the prostate under spinal anesthesia

BACKGROUND

Hospitalization for Transurethral Prostate Resection (TURP) involves circadian rhythm disturbance - a possible cause of Postoperative Neurocognitive Disorder (POCD) in elderly patients. This study investigated whether melatonin ameliorated this effect.

METHODS

A double-blind, randomized clinical trial used a battery of neuropsychological tests to evaluate cognitive performance of 118 patients aged ≥ 60, before TURP with spinal anesthesia, and at 21- and 180-days PO. Patients received 10 mg of melatonin, or a placebo, on the night before surgery and 1-, 2- and 3-days PO. Delayed neurocognitive recovery in the two groups at 21 days PO was compared using the Chi-Squared test; individual performances in each test at each date were compared using the General Mixed Model. Results with p < 0.05 were considered significant.

RESULTS

Pre-surgery, both groups had significant cognitive deficits. Delayed cognitive recovery at 21 days PO was the same in both. There were no cases of POCD at 180 days. The melatonin group performed better in the delayed-recall FOME, which assesses memory, and in the Digit Span test, which assesses attention and cognitive flexibility. Unexpectedly, global neurocognitive performance was improved at 180 PO in both groups.

CONCLUSIONS

Melatonin had no statistical effect on POCD, but a selective beneficial effect was observed in two cognitive areas. The high prevalence of preoperative cognitive impairment may be related to the lower urinary tract symptoms which were reasons for the surgery; the unexpected improvement of cognitive performance in all patients at 180 days PO may reflect alleviation of these symptoms.

From Plan to Pivot: How Model-Informed Drug Development Shaped the Dose Strategy of the Zibotentan/Dapagliflozin ZENITH Trials

Getting the dose right is a key challenge in drug development; model-informed drug development (MIDD) provides powerful tools to shape dose strategies and inform decision making. In this tutorial, the case study of the ZENITH trials showcases how a set of clinical pharmacology and MIDD approaches informed an impactful dose strategy. The endothelin A receptor antagonist zibotentan, combined with the sodium-glucose co-transporter-2 inhibitor dapagliflozin, has yielded a robust and significant albuminuria reduction in the Phase IIb trial ZENITH-CKD and is being investigated for reduction of kidney function decline in a high-risk chronic kidney disease population in the Phase III trial ZENITH High Proteinuria. Endothelin antagonist treatment has, until now, been limited by the class effect fluid retention. ZENITH-CKD investigated a wide range of zibotentan doses based on pharmacokinetics in renal impairment, competitor-data exposure-response modeling, and clinical trial simulations. Recruitment delays reduced interim analysis data availability; here, supportive dose-response modeling recovered decision-making confidence. At trial completion, the low-dose arm enabled Phase III dose selection between Phase IIb doses. Dose-response modeling of efficacy and Kaplan-Meier analyses of tolerability identified a kidney-function-based low-dose strategy of 0.25 or 0.75 mg zibotentan (with 10 mg dapagliflozin) to balance benefit/risk in ZENITH High Proteinuria. The applied clinical pharmacology and MIDD principles enabled successful Phase IIb dose finding, rationalized and built confidence in the innovative Phase III dosing strategy and identified a potential therapeutic window for zibotentan/dapagliflozin, providing the opportunity for a significant improvement in the treatment of chronic kidney disease with high proteinuria.

Incorporation of healthy volunteers data on receptor occupancy into a phase II proof-of-concept trial using a Bayesian dynamic borrowing design

Receptor occupancy in targeted tissues measures the proportion of receptors occupied by a drug at equilibrium and is sometimes used as a surrogate of drug efficacy to inform dose selection in clinical trials. We propose to incorporate data on receptor occupancy from a phase I study in healthy volunteers into a phase II proof-of-concept study in patients, with the objective of using all the available evidence to make informed decisions. A minimal physiologically based pharmacokinetic modeling is used to model receptor occupancy in healthy volunteers and to predict it in the patients of a phase II proof-of-concept study, taking into account the variability of the population parameters and the specific differences arising from the pathological condition compared to healthy volunteers. Then, given an estimated relationship between receptor occupancy and the clinical endpoint, an informative prior distribution is derived for the clinical endpoint in both the treatment and control arms of the phase II study. These distributions are incorporated into a Bayesian dynamic borrowing design to supplement concurrent phase II trial data. A simulation study in immuno-inflammation demonstrates that the proposed design increases the power of the study while maintaining a type I error at acceptable levels for realistic values of the clinical endpoint.

AZD9567 versus prednisolone in patients with active rheumatoid arthritis: A phase IIa, randomized, double-blind, efficacy, and safety study

Oral corticosteroid use is limited by side effects, some caused by off-target actions on the mineralocorticoid receptor that disrupt electrolyte balance. AZD9567 is a selective, nonsteroidal glucocorticoid receptor modulator. The efficacy, safety, and tolerability of AZD9567 and prednisolone were assessed in a phase IIa study. Anti-inflammatory mechanism of action was also evaluated in vitro in monocytes from healthy donors. In this randomized, double-blind, parallel-group, multicenter study, patients with active rheumatoid arthritis were randomized 1:1 to AZD9567 40 mg or prednisolone 20 mg once daily orally for 14 days. The primary end point was change from baseline in DAS28-CRP at day 15. Secondary end points included components of DAS28-CRP, American College of Rheumatology (ACR) response criteria (ACR20, ACR50, and ACR70), and safety end points, including serum electrolytes. Overall, 21 patients were randomized to AZD9567 (n = 11) or prednisolone (n = 10), and all completed the study. As anticipated, AZD9567 had a similar efficacy profile to prednisolone, with no clinically meaningful (i.e., >1.0) difference in change from baseline to day 15 in DAS28-CRP between AZD9567 and prednisolone (least-squares mean difference: 0.47, 95% confidence interval: -0.49 to 1.43). Similar results were observed for the secondary efficacy end points. In vitro transcriptomic analysis showed that anti-inflammatory responses were similar for AZD9567, prednisolone, and dexamethasone. Unlike prednisolone, AZD9567 had no effect on the serum sodium:potassium ratio. The safety profile was not different from that of prednisolone. Larger studies of longer duration are required to determine whether AZD9567 40 mg may in the future be an alternative to prednisolone in patients with inflammatory disease.

Bayesian optimal phase II designs with dual-criterion decision making

The conventional phase II trial design paradigm is to make the go/no-go decision based on the hypothesis testing framework. Statistical significance itself alone, however, may not be sufficient to establish that the drug is clinically effective enough to warrant confirmatory phase III trials. We propose the Bayesian optimal phase II trial design with dual-criterion decision making (BOP2-DC), which incorporates both statistical significance and clinical relevance into decision making. Based on the posterior probability that the treatment effect reaches the lower reference value (statistical significance) and the clinically meaningful value (clinical significance), BOP2-DC allows for go/consider/no-go decisions, rather than a binary go/no-go decision. BOP2-DC is highly flexible and accommodates various types of endpoints, including binary, continuous, time-to-event, multiple, and coprimary endpoints, in single-arm and randomized trials. The decision rule of BOP2-DC is optimized to maximize the probability of a go decision when the treatment is effective or minimize the expected sample size when the treatment is futile. Simulation studies show that the BOP2-DC design yields desirable operating characteristics. The software to implement BOP2-DC is freely available at www.trialdesign.org.

A Bayesian three-tier quantitative decision-making framework for single arm studies in early phase oncology

In early phase oncology drug development, single arm proof-of-concept (POC) studies are increasingly being used to drive the early decisions for future development of the drug. Decision-makings based on such studies, typically involving small sample size and early surrogate efficacy endpoints, are extremely challenging. In particular, given the tremendous competition in the development of immunotherapies, expedition of the most promising programs is desired. To this end, we have proposed a Bayesian three-tier approach to facilitate the decision-making process, inheriting all the benefits of Bayesian decision-making approaches and formally allowing the option of acceleration. With pre-specified Bayesian decision criteria, three types of decisions regarding the future development of the drug can be made: (1) terminating the program, (2) further investigation, considering totality of evidence or additional POC studies, and (3) accelerating the program. We further proposed a Bayesian adaptive three-tier (BAT) design, extending the decision-making approach to incorporate adaptive thresholds and allow for continuous monitoring of the study. We compare the performance of the proposed methods with some other existing methods through simulations.

From innovative thinking to pharmaceutical industry implementation: Some success stories

In industry, successful innovation involves not only developing new statistical methodology, but also ensuring that this methodology is implemented successfully. This includes enabling applied statisticians to understand the method, its benefits and limitations and empowering them to implement the new method. This will include advocacy, influencing in-house and external stakeholders, such that these stakeholders are receptive to the new methodology. In this paper, we describe some industry successes and focus on our colleague, Andy Grieve's role in these.

Bayesian modeling and simulation to inform rare disease drug development early decision-making: Application to Duchenne muscular dystrophy

Rare disease clinical trials are constrained to small sample sizes and may lack placebo-control, leading to challenges in drug development. This paper proposes a Bayesian model-based framework for early go/no-go decision making in rare disease drug development, using Duchenne muscular dystrophy (DMD) as an example. Early go/no-go decisions were based on projections of long-term functional outcomes from a Bayesian model-based analysis of short-term trial data informed by prior knowledge based on 6MWT natural history literature data in DMD patients. Frequentist hypothesis tests were also applied as a reference analysis method. A number of combinations of hypothetical trial designs, drug effects and cohort comparison methods were assessed. The proposed Bayesian model-based framework was superior to the frequentist method for making go/no-go decisions across all trial designs and cohort comparison methods in DMD. The average decision accuracy rates across all trial designs for the Bayesian and frequentist analysis methods were 45.8 and 8.98%, respectively. A decision accuracy rate of at least 50% was achieved for 42 and 7% of the trial designs under the Bayesian and frequentist analysis methods, respectively. The frequentist method was limited to the short-term trial data only, while the Bayesian methods were informed with both the short-term data and prior information. The specific results of the DMD case study were limited due to incomplete specification of individual-specific covariates in the natural history literature data and should be reevaluated using a full natural history dataset. These limitations aside, the framework presented provides a proof of concept for the utility of Bayesian model-based methods for decision making in rare disease trials.

Comparing Go/No-Go Decision-Making Properties Between Single Arm Phase II Trial Designs in Oncology

INTRODUCTION

Simon's design has been widely used in oncology to conduct single arm phase II trials and to make Go/No-Go development decision. Other authors have proposed designs with decision-making frameworks that include a third, "Consider" outcome. For results in the Consider zone, a final Go/No-Go development decision must still be made; however it is typically a subjective decision based on the totality of data and the development landscape. Under this framework, the probability of continuing development when the candidate therapy is truly ineffective or the probability of stopping development when the candidate therapy is truly effective is undefined.

METHODS

We use a motivating example to compare end of trial decision-making between Simon's two-stage approach and a Multilevel outcome approach. We present the minimum and maximum development decision error probabilities by varying whether candidates that end in the Consider zone would ultimately continue with development or not.

RESULTS

The Multilevel approach typically requires fewer patients, but the risk of making an incorrect drug development decision is inflated above the statistically defined Type I and Type II error rates. Compared to a Type I error rate of 20%, the Multilevel trial's maximum probability of moving forward with an ineffective therapy is 22%, 27%, and 36% for Consider zone sizes of 10%, 20%, and 30%, respectively.

CONCLUSION

The Multilevel approach provides flexibility in interpreting moderate efficacy results. However, the flexibility is accomplished with a lower sample size and corresponding uncertainty in the trial outcome that increases the risk of incorrect drug development decisions.

A New Comprehensive Approach to Assess the Probability of Success of Development Programs Before Pivotal Trials

The point at which clinical development programs transition from early phase to pivotal trials is a critical milestone. Substantial uncertainty about the outcome of pivotal trials may remain even after seeing positive early phase data, and companies may need to make difficult prioritization decisions for their portfolio. The probability of success (PoS) of a program, a single number expressed as a percentage reflecting the multitude of risks that may influence the final program outcome, is a key decision-making tool. Despite its importance, companies often rely on crude industry benchmarks that may be "adjusted" by experts based on undocumented criteria and which are typically misaligned with the definition of success used to drive commercial forecasts, leading to overly optimistic expected net present value calculations. We developed a new framework to assess the PoS of a program before pivotal trials begin. Our definition of success encompasses the successful outcome of pivotal trials, regulatory approval and meeting the requirements for market access as outlined in the target product profile. The proposed approach is organized in four steps and uses an innovative Bayesian approach to synthesize all relevant evidence. The new PoS framework is systematic and transparent. It will help organizations to make more informed decisions. In this paper, we outline the rationale and elaborate on the structure of the proposed framework, provide examples, and discuss the benefits and challenges associated with its adoption.

Do strict decision criteria hamper productivity in the pharmaceutical industry?

The discouragingly high rates of attrition in drug development, and in particular in Phase 2, warrant a closer look at the decision criteria applied for investment in the next phase (Phase 3). We have in this article evaluated Stop/Go criteria after Phase 2, based on a model encompassing both Phase 2 and 3, as well as the eventual outcome on the market. The results indicate that the value of a drug project is often maximized if rather liberal decision criteria are applied. The routine adherence to standard criteria, e.g. requiring significance at 5% level, may lead to an unduly high rate of false negative decisions. This might ultimately hamper the productivity of drug development and leading to potentially useful drugs not being taken forward to benefit the intended patients.

The detailed clinical objectives approach to designing clinical trials and choosing estimands

Objective setting is a necessary early step in the development of a clinical trial. ICH E9(R1) notes that the clinical objectives of a trial lead directly to the choice of estimands but barely discusses objectives themselves. Indeed, there is very little guidance anywhere in literature about objectives in clinical trials. This article identifies the substantial overlap between description of estimands and high quality definitions of objectives. It consequently shows that the estimand is decided by the precise choice of trial objective, and that therefore estimand decisions should be made at the objective level. The Detailed Clinical Objectives approach is proposed to support this. It emphasises clarity, specificity and a clinical focus when choosing and documenting objectives. Template text and examples are included to provide guidance on how it can be used in real trials. Finally, we describe objective-driven trial design, emphasising how strong objective setting establishes an important foundation for rigorous trial design discussions, logistical and operational decision-making during trial preparations, and clear communication of results and conclusions at the end of the trial. Highlighting the distinctions between objectives and estimands, we note how an objective-based framework can build on the ICH E9(R1) estimand framework to address many of its unanswered questions.

Selection bias, investment decisions and treatment effect distributions

When making decisions regarding the investment and design for a Phase 3 programme in the development of a new drug, the results from preceding Phase 2 trials are an important source of information. However, only projects in which the Phase 2 results show promising treatment effects will typically be considered for a Phase 3 investment decision. This implies that, for those projects where Phase 3 is pursued, the underlying Phase 2 estimates are subject to selection bias. We will in this article investigate the nature of this selection bias based on a selection of distributions for the treatment effect. We illustrate some properties of Bayesian estimates, providing shrinkage of the Phase 2 estimate to counteract the selection bias. We further give some empirical guidance regarding the choice of prior distribution and comment on the consequences for decision-making in investment and planning for Phase 3 programmes.

Conditional assurance: the answer to the questions that should be asked within drug development

In this paper, we extend the use of assurance for a single study to explore how meeting a study's pre-defined success criteria could update our beliefs about the true treatment effect and impact the assurance of subsequent studies. This concept of conditional assurance, the assurance of a subsequent study conditional on success in an initial study, can be used assess the de-risking potential of the study requiring immediate investment, to ensure it provides value within the overall development plan. If the planned study does not discharge sufficient later phase risk, alternative designs and/or success criteria should be explored. By transparently laying out the different design options and the risks associated, this allows for decision makers to make quantitative investment choices based on their risk tolerance levels and potential return on investment. This paper lays out the derivation of conditional assurance, discusses how changing the design of a planned study will impact the conditional assurance of a future study, as well as presenting a simple illustrative example of how this methodology could be used to transparently compare development plans to aid decision making within an organisation.

A stochastically curtailed two-arm randomised phase II trial design for binary outcomes

Randomised controlled trials are considered the gold standard in trial design. However, phase II oncology trials with a binary outcome are often single-arm. Although a number of reasons exist for choosing a single-arm trial, the primary reason is that single-arm designs require fewer participants than their randomised equivalents. Therefore, the development of novel methodology that makes randomised designs more efficient is of value to the trials community. This article introduces a randomised two-arm binary outcome trial design that includes stochastic curtailment (SC), allowing for the possibility of stopping a trial before the final conclusions are known with certainty. In addition to SC, the proposed design involves the use of a randomised block design, which allows investigators to control the number of interim analyses. This approach is compared with existing designs that also use early stopping, through the use of a loss function comprised of a weighted sum of design characteristics. Comparisons are also made using an example from a real trial. The comparisons show that for many possible loss functions, the proposed design is superior to existing designs. Further, the proposed design may be more practical, by allowing a flexible number of interim analyses. One existing design produces superior design realisations when the anticipated response rate is low. However, when using this design, the probability of rejecting the null hypothesis is sensitive to misspecification of the null response rate. Therefore, when considering randomised designs in phase II, we recommend the proposed approach be preferred over other sequential designs.

Blood-Brain Barrier Disruption in Neuro-Oncology: Strategies, Failures, and Challenges to Overcome

The blood-brain barrier (BBB) presents a formidable challenge in the development of effective therapeutics in neuro-oncology. This has fueled several decades of efforts to develop strategies for disrupting the BBB, but progress has not been satisfactory. As such, numerous drug- and device-based methods are currently being investigated in humans. Through a focused assessment of completed, active, and pending clinical trials, our first aim in this review is to outline the scientific foundation, successes, and limitations of the BBBD strategies developed to date. Among 35 registered trials relevant to BBBD in neuro-oncology in the ClinicalTrials.gov database, mannitol was the most common drug-based method, followed by RMP-7 and regadenoson. MR-guided focused ultrasound was the most common device-based method, followed by MR-guided laser ablation, ultrasound, and transcranial magnetic stimulation. While most early-phase studies focusing on safety and tolerability have met stated objectives, advanced-phase studies focusing on survival differences and objective tumor response have been limited by heterogeneous populations and tumors, along with a lack of control arms. Based on shared challenges among all methods, our second objective is to discuss strategies for confirmation of BBBD, choice of systemic agent and drug design, alignment of BBBD method with real-world clinical workflow, and consideration of inadvertent toxicity associated with disrupting an evolutionarily-refined barrier. Finally, we conclude with a strategic proposal to approach future studies assessing BBBD.

"In-House" Data on the Outside-A Mobile Health Approach

Mobile health (mHealth) technologies have the potential to capture dense patient data on the background of real-life behavior. Merck & Co., Inc. (Kenilworth, NJ), in collaboration with Koneksa Health, conducted a phase I clinical trial to validate cardiovascular mHealth technologies for concordance with traditional approaches and to establish sensitivity to detect effects of pharmacological intervention. This two-part study enrolled 18 healthy male subjects. Part I, a 5-day study, compared mHealth measures of heart rate (HR) and blood pressure (BP) to those from traditional methods. Hypotheses of similarity, in the clinic and at home, were tested individually for HR, systolic BP, and diastolic BP, at a 2-sided 0.05 alpha level, with a prespecified criterion for similarity being the percentage differences between the 2 measurements within 15%. Part II, a 7-day, 3-period randomized balanced crossover study, evaluated the mHealth technology's ability to detect effects of bisoprolol and salbutamol. Hypotheses that the changes from baseline in HR were greater in the bisoprolol (reduction in HR) and salbutamol (increase in HR) groups compared with no treatment were tested, at a 1-sided 0.05 alpha level. Linear mixed-effects models, Pearson's correlation coefficients, summary statistics, and exploratory plots were applied to analyze the data. The mHealth measures of HR and BP were demonstrated to be similar to those from traditional methods, and sensitive to changes in cardiovascular parameters induced by bisoprolol and salbutamol.

Beyond p-values: A phase II dual-criterion design with statistical significance and clinical relevance

Background

Well-designed phase II trials must have acceptable error rates relative to a pre-specified success criterion, usually a statistically significant p-value. Such standard designs may not always suffice from a clinical perspective because clinical relevance may call for more. For example, proof-of-concept in phase II often requires not only statistical significance but also a sufficiently large effect estimate.

Purpose

We propose dual-criterion designs to complement statistical significance with clinical relevance, discuss their methodology, and illustrate their implementation in phase II.

Methods

Clinical relevance requires the effect estimate to pass a clinically motivated threshold (the decision value (DV)). In contrast to standard designs, the required effect estimate is an explicit design input, whereas study power is implicit. The sample size for a dual-criterion design needs careful considerations of the study’s operating characteristics (type I error, power).

Results

Dual-criterion designs are discussed for a randomized controlled and a single-arm phase II trial, including decision criteria, sample size calculations, decisions under various data scenarios, and operating characteristics. The designs facilitate GO/NO-GO decisions due to their complementary statistical–clinical criterion.

Limitations

While conceptually simple, implementing a dual-criterion design needs care. The clinical DV must be elicited carefully in collaboration with clinicians, and understanding similarities and differences to a standard design is crucial.

Conclusion

To improve evidence-based decision-making, a formal yet transparent quantitative framework is important. Dual-criterion designs offer an appealing statistical–clinical compromise, which may be preferable to standard designs if evidence against the null hypothesis alone does not suffice for an efficacy claim.

Quantitative decision-making in randomized Phase II studies with a time-to-event endpoint

One of the most critical decision points in clinical development is Go/No-Go decision-making after a proof-of-concept study. Traditional decision-making relies on a formal hypothesis testing with control of type I and type II error rates, which is limited by assessing the strength of efficacy evidence in a small isolated trial. In this article, we propose a quantitative Bayesian/frequentist decision framework for Go/No-Go criteria and sample size evaluation in Phase II randomized studies with a time-to-event endpoint. By taking the uncertainty of treatment effect into consideration, we propose an integrated quantitative approach for a program when both the Phase II and Phase III trials share a common endpoint while allowing a discount of the observed Phase II data. Our results confirm the argument that an increase in the sample size of a Phase II trial will result in greater increase in the probability of success of a Phase III trial than increasing the Phase III trial sample size by equal amount. We illustrate the steps in quantitative decision-making with a real example of a randomized Phase II study in metastatic pancreatic cancer.

Integrating dose estimation into a decision-making framework for model-based drug development

Model-informed drug discovery and development offers the promise of more efficient clinical development, with increased productivity and reduced cost through scientific decision making and risk management. Go/no-go development decisions in the pharmaceutical industry are often driven by effect size estimates, with the goal of meeting commercially generated target profiles. Sufficient efficacy is critical for eventual success, but the decision to advance development phase is also dependent on adequate knowledge of appropriate dose and dose-response. Doses which are too high or low pose risk of clinical or commercial failure. This paper addresses this issue and continues the evolution of formal decision frameworks in drug development. Here, we consider the integration of both efficacy and dose-response estimation accuracy into the go/no-go decision process, using a model-based approach. Using prespecified target and lower reference values associated with both efficacy and dose accuracy, we build a decision framework to more completely characterize development risk. Given the limited knowledge of dose response in early development, our approach incorporates a set of dose-response models and uses model averaging. The approach and its operating characteristics are illustrated through simulation. Finally, we demonstrate the decision approach on a post hoc analysis of the phase 2 data for naloxegol (a drug approved for opioid-induced constipation).

Bayesian Phase II optimization for time-to-event data based on historical information

After exploratory drug development, companies face the decision whether to initiate confirmatory trials based on limited efficacy information. This proof-of-concept decision is typically performed after a Phase II trial studying a novel treatment versus either placebo or an active comparator. The article aims to optimize the design of such a proof-of-concept trial with respect to decision making. We incorporate historical information and develop pre-specified decision criteria accounting for the uncertainty of the observed treatment effect. We optimize these criteria based on sensitivity and specificity, given the historical information. Specifically, time-to-event data are considered in a randomized 2-arm trial with additional prior information on the control treatment. The proof-of-concept criterion uses treatment effect size, rather than significance. Criteria are defined on the posterior distribution of the hazard ratio given the Phase II data and the historical control information. Event times are exponentially modeled within groups, allowing for group-specific conjugate prior-to-posterior calculation. While a non-informative prior is placed on the investigational treatment, the control prior is constructed via the meta-analytic-predictive approach. The design parameters including sample size and allocation ratio are then optimized, maximizing the probability of taking the right decision. The approach is illustrated with an example in lung cancer.