Integrating predictive biomarkers and classifiers into oncology clinical development programmes

Aiding the Adoption of Master Protocols by Optimizing Patient Engagement

Master protocols (MPs) are an important addition to the clinical trial repertoire. As defined by the U.S. Food and Drug Administration (FDA), this term means "a protocol designed with multiple sub-studies, which may have different objectives (goals) and involve coordinated efforts to evaluate one or more investigational drugs in one or more disease subtypes within the overall trial structure." This means we now have a unique, scientifically based MP that describes how a clinical trial will be conducted using one or more potential candidate therapies to treat patients in one or more diseases. Patient engagement (PE) is also a critical factor that has been recognized by FDA through its Patient-Focused Drug Development (PFDD) initiative, and by the European Medicines Agency (EMA), which states on its website that it has been actively interacting with patients since the creation of the Agency in 1995. We propose that utilizing these PE principles in MPs can make them more successful for sponsors, providers, and patients. Potential benefits of MPs for patients awaiting treatment can include treatments that better fit a patient's needs; availability of more treatments; and faster access to treatments. These make it possible to develop innovative therapies (especially for rare diseases and/or unique subpopulations, e.g., pediatrics), to minimize untoward side effects through careful dose escalation practices and, by sharing a control arm, to lower the probability of being assigned to a placebo arm for clinical trial participants. This paper is authored by select members of the American Statistical Association (ASA)/DahShu Master Protocol Working Group (MPWG) People and Patient Engagement (PE) Subteam. DahShu is a 501(c)(3) non-profit organization, founded to promote research and education in data science. This manuscript does not include direct feedback from US or non-US regulators, though multiple regulatory-related references are cited to confirm our observation that improving patient engagement is supported by regulators. This manuscript represents the authors' independent perspective on the Master Protocol; it does not represent the official policy or viewpoint of FDA or any other regulatory organization or the views of the authors' employers. The objective of this manuscript is to provide drug developers, contract research organizations (CROs), third party capital investors, patient advocacy groups (PAGs), and biopharmaceutical executives with a better understanding of how including the patient voice throughout MP development and conduct creates more efficient clinical trials. The PE Subteam also plans to publish a Plain Language Summary (PLS) of this publication for clinical trial participants, patients, caregivers, and the public as they seek to understand the risks and benefits of MP clinical trial participation.

Emerging Biomarkers of Response to Systemic Therapies in Metastatic Clear Cell Renal Cell Carcinoma

Patients with metastatic clear cell renal cell carcinoma (mccRCC) experience highly heterogeneous outcomes when treated with standard-of-care systemic regimens. Therefore, valid biomarkers are needed to predict the clinical response to these therapies and help guide management. In this review, the authors outline relevant and promising biomarkers for patients with mccRCC receiving systemic therapies, with a focus on immunotherapy-based regimens.

The Effect of Biomarkers on Clinical Trial Risk in Gastric Cancer

OBJECTIVES

This study examined clinical trial success rates for new drug developments in gastric cancer since 1998. We also examined the clinical trial design features that may mitigate the risk of clinical trial failure.

MATERIALS AND METHODS

Clinical trial data was obtained from clinicaltrials.gov. Drugs were included if they entered testing between January 1, 1998 and January 1, 2022 and were excluded if they did not have a completed phase I trial or treated secondary effects of gastric cancer. Transition probabilities were calculated for each phase and compared with industry averages. Success rates were determined based on biomarker usage, drug target, type of therapy, and drug chemistry.

RESULTS

Upon screening 1990 trials, 219 drugs met our inclusion criteria. The probability of a drug completing all phases of testing and obtaining FDA approval was 7%, which is below the 11% industry average. The use of biomarkers in clinical development resulted in nearly a 2-fold increase in the cumulative success rate. Biologics also exhibited higher success rates (17%) as opposed to small molecules (1%). This was true even when we compared both drug types that shared the same target. When comparing only receptor-targeted therapies, biologics (62%) continued to outperform small molecules (18%). Similarly, when narrowed down to drugs targeting solely HER2 receptors, biologics continued to prevail (64% vs. 24%).

CONCLUSIONS

Implementing biomarkers, receptor-targeted therapies, and biologics in clinical development improves clinical trial success rates in gastric cancer. Thus, physicians should prioritize the enrollment of gastric cancer patients in clinical trials that incorporate the aforementioned features.

Antitumor pharmacological research in the era of personalized medicine

Anticancer drug discovery has yielded unprecedented progress in recent decades, resulting in the approval of innovative treatment options for patients and the successful implementation of personalized medicine in clinical practice. This remarkable progress has also reshaped the research scope of pharmacological research. This article, as a tribute to cancer research at Shanghai Institute of Materia Medica in celebration of the institute's 90th birthday, provides an overview of the conceptual revolution occurring in anticancer therapy, and summarizes our recent progress in the development of molecularly targeted therapeutics and exploration of new strategies in personalized medicine. With this review, we hope to provide a glimpse into how antitumor pharmacological researchers have embraced the new era of personalized medicine research and to propose a future path for anticancer drug discovery and pharmacological research.

Innovations in Clinical Development in Rare Diseases of Children and Adults: Small Populations and/or Small Patients

Many of the afflictions of children are rare diseases. This creates numerous drug development challenges related to small populations, including limited information about the disease state, enrollment challenges, and diminished incentives for pediatric development of novel therapies by pharmaceutical and biotechnology sponsors. We review selected innovations in clinical development that may partially mitigate some of these difficulties, starting with the concept of development efficiency for individual clinical trials, clinical programs (involving multiple trials for a single drug), and clinical portfolios of multiple drugs, and decision analysis as a tool to optimize efficiency. Development efficiency is defined as the ability to reach equally rigorous or more rigorous conclusions in less time, with fewer trial participants, or with fewer resources. We go on to discuss efficient methods for matching targeted therapies to biomarker-defined subgroups, methods for eliminating or reducing the need for natural history data to guide rare disease development, the use of basket trials to enhance efficiency by grouping multiple similar disease applications in a single clinical trial, and the use of alternative data sources including historical controls to augment or replace concurrent controls in clinical studies. Greater understanding and broader application of these methods could lead to improved therapies and/or more widespread and rapid access to novel therapies for rare diseases in both children and adults.

The Effect of Biomarker Use on the Speed and Duration of Clinical Trials for Cancer Drugs

Background

The purpose of this study was to explore the effects biomarkers have on the duration and speed of clinical trials in oncology.

Materials and Methods

Clinical trial data was pooled from www.clinicaltrials.gov within the 4 cancer indications of non-small cell lung cancer, breast cancer, melanoma, and colorectal cancer. Heatmaps of clinical timelines were used to display differences in the frequency and timing of clinical trials across trials that used or did not use biomarkers, for all 4 indications.

Results

Screening of 8630 clinical trials across the 4 indications yielded 671 unique drugs corresponding to 1224 eligible trials used in our analysis. The constructed heatmaps visually represented that biomarkers did not have an effect on the time gap between trial phases for non-small cell lung cancer and melanoma but did for colorectal and breast cancer trials, reducing the speed of trial timelines. It was also observed that biomarker trials were more often concurrent over shorter periods of time and began later in the timeline for non-small cell lung and colorectal cancers.

Conclusion

The novel visualization method revealed longer gaps between trial phases, later clinical trial start times, and shorter periods of concurrently run trials for drugs that used biomarkers. The study highlights that biomarker-driven trials might impact drug approval timelines and need to be considered carefully in clinical development plan.

Efficiency of a randomized confirmatory basket trial design constrained to control the family wise error rate by indication

Basket trials pool histologic indications sharing molecular pathophysiology, improving development efficiency. Currently, basket trials have been confirmatory only for exceptional therapies. Our previous randomized basket design may be generally suitable in the resource-intensive confirmatory phase, maintains high power even with modest effect sizes, and provides nearly k-fold increased efficiency for k indications, but controls false positives for the pooled result only. Since family wise error rate by indications may sometimes be required, we now simulate a variant of this basket design controlling family wise error rate at 0.025 k, the total family wise error rate of k separate randomized trials. We simulated this modified design under numerous scenarios varying design parameters. Only designs controlling family wise error rate and minimizing estimation bias were allowable. Optimal performance results when [Formula: see text]. We report efficiency (expected # true positives/expected sample size) relative to k parallel studies, at 90% power (“uncorrected”) or at the power achieved in the basket trial (“corrected,” because conventional designs could also increase efficiency by sacrificing power). Efficiency and power (percentage active indications identified) improve with a higher percentage of initial indications active. Up to 92% uncorrected and 38% corrected efficiency improvement is possible. Even under family wise error rate control, randomized confirmatory basket trials substantially improve development efficiency. Initial indication selection is critical.

Development of a model to demonstrate the impact of National Institute of Health and Care Excellence cost-effectiveness assessment on health utility for targeted medicines

Advances in medical technology have led to a better understanding of heterogeneity of diseases and patients, and to the development of targeted medicines. This development is beneficial to society but can come at an increased cost to pharmaceutical manufacturers due to the costs associated with developing and manufacturing a diagnostic test. For such medicines, the conventional pricing structure, where a therapy is approved if it is deemed cost-effective, may not appropriately incentivize targeted drug development. We model the decision-making processes for both the healthcare provider and the pharmaceutical manufacturer, capturing their main priorities, and populate it with information from a recent appraisal by the National Institute of Health and Care Excellence. Healthcare providers prefer a stratified drug to be developed for a subgroup of the population when the drug is on average effective in the subgroup but with a detrimental effect in the complement. Whilst pharmaceutical manufacturers' preferences are similar, regions of disagreement exist. We show how preferences can be aligned by either penalizing the development of a non-stratified drug or rewarding the development of a stratified drug. The cost and position of alignment depends on the true value of health to the healthcare provider, among other parameters.

Optimal one-stage design and analysis for efficacy expansion in Phase I oncology trials

BACKGROUND

Contemporary Phase I oncology trials often include efficacy expansion in various tumor indications post dose finding. Preliminary anti-tumor activity from efficacy expansion can aid Go/No-Go decision for Phase 2 or Phase 3 initiation. Tumor cohorts in efficacy expansion are commonly analyzed independently in practice, which are often underpowered due to small sample size. Pooled analysis is also sometimes conducted, but it ignores the heterogeneity of the anti-tumor activity across cohorts.

METHODS

We propose an optimal one-stage design and analysis strategy for the efficacy expansion to assess whether the treatment is effective. Allowing heterogeneous anti-tumor effects across tumor cohorts, inactive cohorts are pruned, and the potentially active cohorts are pooled together to gain study power. For a prospective design with a target power, the total sample size across all cohorts is minimized; or for an ad hoc analysis with pre-specified sample size for each cohort, the pruning criteria are optimized to achieve maximum power. The global type I error is controlled after proper multiplicity adjustment, and a penalty adjusted significance level is used for the pooled test.

RESULTS

Simulation studies show that the proposed optimal design has desirable operating characteristics in increasing the overall power and detecting more true positive tumor cohorts.

CONCLUSION

The proposed optimal design and analysis strategy provides a practical approach to design and analyze heterogeneous efficacy expansion cohorts in a basket setting with global type I and type II error being controlled.

Planning for the Next Pandemic: Ethics and Innovation Today for Improved Clinical Trials Tomorrow

The coronavirus pandemic has brought public attention to the steps required to produce valid scientific clinical research in drug development. Traditional ethical principles that guide clinical research remain the guiding compass for physicians, patients, public health officials, investigators, drug developers and the public. Accelerating the process of delivering safe and effective treatments and vaccines against COVID-19 is a moral imperative. The apparent clash between the regulated system of phased randomized clinical trials and urgent public health need requires leveraging innovation with ethical scientific rigor. We reflect on the Belmont principles of autonomy, beneficence and justice as the pandemic unfolds, and illustrate the role of innovative clinical trial designs in alleviating pandemic challenges. Our discussion highlights selected types of innovative trial design and correlates them with ethical parameters and public health benefits. Details are provided for platform trials and other innovative designs such as basket and umbrella trials, designs leveraging external data sources, multi-stage seamless trials, preplanned control arm data sharing between larger trials, and higher order systems of linked trials coordinated more broadly between individual trials and phases of development, recently introduced conceptually as "PIPELINEs."

Optimizing subgroup selection in two-stage adaptive enrichment and umbrella designs

We design two‐stage confirmatory clinical trials that use adaptation to find the subgroup of patients who will benefit from a new treatment, testing for a treatment effect in each of two disjoint subgroups. Our proposal allows aspects of the trial, such as recruitment probabilities of each group, to be altered at an interim analysis. We use the conditional error rate approach to implement these adaptations with protection of overall error rates. Applying a Bayesian decision‐theoretic framework, we optimize design parameters by maximizing a utility function that takes the population prevalence of the subgroups into account. We show results for traditional trials with familywise error rate control (using a closed testing procedure) as well as for umbrella trials in which only the per‐comparison type 1 error rate is controlled. We present numerical examples to illustrate the optimization process and the effectiveness of the proposed designs.

Friction and Decision Rules in Portfolio Decision Analysis

In portfolio decision analysis, features comprise the objectives, alternatives, physics, and information that define a decision context. By modeling features, decision analysts forecast the expected utilities of the alternatives. A model is complete if it contains all the features. A model is well-calibrated if it correctly predicts the probability distributions of each alternative’s utility, whereas ill-calibrated models, like those that suffer the optimizer’s curse, do not. Friction identifies qualities of a situation that prevent decision analysts from creating complete, well-calibrated models. When friction is significant, can maximizing expected utility be a suboptimal decision rule? Is satisfying decision theory’s axioms a necessary or sufficient condition for good decision making? Can rules that violate the axioms outperform rules that satisfy them? A simulation study of how unbiased, imprecise forecasts of payoffs affect project selection finds that, for the example tested, the answers are yes, no, and yes, which suggests that further studies of friction may be worthwhile. Discussions of friction bookend the study, starting the paper by defining friction and concluding by presenting three frameworks, each one from a different field of study, that provide mathematical tools for studying friction.

Does biomarker use in oncology improve clinical trial failure risk? A large-scale analysis

Purpose

To date there has not been an extensive analysis of the outcomes of biomarker use in oncology.

Methods

Data were pooled across four indications in oncology drawing upon trial outcomes from www.clinicaltrials.gov: breast cancer, non‐small cell lung cancer (NSCLC), melanoma and colorectal cancer from 1998 to 2017. We compared the likelihood drugs would progress through the stages of clinical trial testing to approval based on biomarker status. This was done with multi‐state Markov models, tools that describe the stochastic process in which subjects move among a finite number of states.

Results

Over 10000 trials were screened, which yielded 745 drugs. The inclusion of biomarker status as a covariate significantly improved the fit of the Markov model in describing the drug trajectories through clinical trial testing stages. Hazard ratios based on the Markov models revealed the likelihood of drug approval with biomarkers having nearly a fivefold increase for all indications combined. A 12, 8 and 7‐fold hazard ratio was observed for breast cancer, melanoma and NSCLC, respectively. Markov models with exploratory biomarkers outperformed Markov models with no biomarkers.

Conclusion

This is the first systematic statistical evidence that biomarkers clearly increase clinical trial success rates in three different indications in oncology. Also, exploratory biomarkers, long before they are properly validated, appear to improve success rates in oncology. This supports early and aggressive adoption of biomarkers in oncology clinical trials.

Sample size re-estimation in adaptive enrichment design

Clinical trial participants are often heterogeneous, which is a fundamental problem in the rapidly developing field of precision medicine. Participants heterogeneity causes considerable difficulty in the current phase III trial designs. Adaptive enrichment designs provide a flexible and intuitive solution. At the interim analysis, we enrich the subgroup of trial participants who have a higher likelihood to benefit from the new treatment. However, it is critical to control the level of the test size and maintain adequate power after enrichment of certain subgroup of participants. We develop two adaptive enrichment strategies with sample size re-estimation and verify their feasibility and practicability through extensive simulations and sensitivity analyses. The simulation studies show that the proposed methods can control the overall type I error rate and exhibit competitive improvement in terms of statistical power and expected sample size. The proposed designs are exemplified with a real trial application.

Efficient two-stage sequential arrays of proof of concept studies for pharmaceutical portfolios

Previous work has shown that individual randomized "proof-of-concept" (PoC) studies may be designed to maximize cost-effectiveness, subject to an overall PoC budget constraint. Maximizing cost-effectiveness has also been considered for arrays of simultaneously executed PoC studies. Defining Type III error as the opportunity cost of not performing a PoC study, we evaluate the common pharmaceutical practice of allocating PoC study funds in two stages. Stage 1, or the first wave of PoC studies, screens drugs to identify those to be permitted additional PoC studies in Stage 2. We investigate if this strategy significantly improves efficiency, despite slowing development. We quantify the benefit, cost, benefit-cost ratio, and Type III error given the number of Stage 1 PoC studies. Relative to a single stage PoC strategy, significant cost-effective gains are seen when at least one of the drugs has a low probability of success (10%) and especially when there are either few drugs (2) with a large number of indications allowed per drug (10) or a large portfolio of drugs (4). In these cases, the recommended number of Stage 1 PoC studies ranges from 2 to 4, tracking approximately with an inflection point in the minimization curve of Type III error.

A Scoring Method for Immunohistochemical Staining on Ki67

An accurate interpretation of immunohistochemistry (IHC) staining results is crucial for precise disease diagnosis. In this study, we present a novel scoring method for interpreting and reporting of IHC staining assay results for the nuclear-type molecule. On the basis of the histologic characteristics, the samples were subdivided into 3 basic structural units and tissue subtypes including covered, mosaic, and mesenchymal subtypes. A cut-off of moderate-positive (2+) cells and 10% as the differential expression were applied to stratify the results into 11 grade scoring system (0 to X level). The observer can directly identify and count the number and percentage of positive cells from IHC staining data. Furthermore, Ki67 staining results in 88 carcinoma specimens were re-evaluated to determine the ease, reliability, reproducibility, and variance among different observers. The results indicated the consistency ratio of 68.0% for the mosaic subtype and 80% for the mesenchymal subtype, and 68.2% for the covered subtype by 5 experienced pathologists independently. Using 10% as the cut-off threshold, the consistency ratio of 92.5%, 96.8%, and 92.9% was noted for mosaic, mesenchymal, and covered subtypes, respectively. Besides, the correlation of counts revealed excellent agreement among the 5 independent pathologists. Overall, the proposed IHC scoring method is a novel, simple, reliable, and reproducible grading system for accurate interpretation of IHC staining data. Furthermore, the presented practical grading approach has the potential to improve the clinical evaluation of the IHC staining data for personalized therapy.

A roadmap to using historical controls in clinical trials - by Drug Information Association Adaptive Design Scientific Working Group (DIA-ADSWG)

Historical controls (HCs) can be used for model parameter estimation at the study design phase, adaptation within a study, or supplementation or replacement of a control arm. Currently on the latter, there is no practical roadmap from design to analysis of a clinical trial to address selection and inclusion of HCs, while maintaining scientific validity. This paper provides a comprehensive roadmap for planning, conducting, analyzing and reporting of studies using HCs, mainly when a randomized clinical trial is not possible. We review recent applications of HC in clinical trials, in which either predominantly a large treatment effect overcame concerns about bias, or the trial targeted a life-threatening disease with no treatment options. In contrast, we address how the evidentiary standard of a trial can be strengthened with optimized study designs and analysis strategies, emphasizing rare and pediatric indications. We highlight the importance of simulation and sensitivity analyses for estimating the range of uncertainties in the estimation of treatment effect when traditional randomization is not possible. Overall, the paper provides a roadmap for using HCs.

Oncology phase II proof-of-concept studies with multiple targets: Randomized controlled trial or single arm?

For oncology drug development, phase II proof-of-concept studies have played a key role in determining whether or not to advance to a confirmatory phase III trial. With the increasing number of immunotherapies, efficient design strategies are crucial in moving successful drugs quickly to market. Our research examines drug development decision making under the framework of maximizing resource investment, characterized by benefit cost ratios (BCRs). In general, benefit represents the likelihood that a drug is successful, and cost is characterized by the risk adjusted total sample size of the phases II and III studies. Phase III studies often include a futility interim analysis; this sequential component can also be incorporated into BCRs. Under this framework, multiple scenarios can be considered. For example, for a given drug and cancer indication, BCRs can yield insights into whether to use a randomized control trial or a single-arm study. Importantly, any uncertainty in historical control estimates that are used to benchmark single-arm studies can be explicitly incorporated into BCRs. More complex scenarios, such as restricted resources or multiple potential cancer indications, can also be examined. Overall, BCR analyses indicate that single-arm trials are favored for proof-of-concept trials when there is low uncertainty in historical control data and smaller phase III sample sizes. Otherwise, especially if the most likely to succeed tumor indication can be identified, randomized controlled trials may be a better option. While the findings are consistent with intuition, we provide a more objective approach.

Optimal two-stage designs for exploratory basket trials

The primary goal of an exploratory oncology clinical trial is to identify an effective drug for further development. To account for tumor indication selection error, multiple tumor indications are often selected for simultaneous testing in a basket trial. In this article, we propose optimal and minimax two-stage basket trial designs for exploratory clinical trials. Inactive tumor indications are pruned in stage 1 and the active tumor indications are pooled at end of stage 2 to assess overall effectiveness of the test drug. The proposed designs explicitly control the type I and type II error rates with closed-form sample size formula. They can be viewed as a natural extension of Simon's optimal and minimax two-stage designs for single arm trials to multi-arm basket trials. A simulation study shows that the proposed design method has desirable operating characteristics as compared to other commonly used design methods for exploratory basket trials.

The arrival of predictive biomarkers for monitoring therapy response to natural compounds in cancer drug discovery

Intrinsic or acquired drug resistance, adverse drug reactions and tumor heterogeneity between and within cancer patients limit the efficacy of clinical management of advanced cancers. To overcome these barriers, predictive biomarkers have recently emerged to guide medical oncologists in the selection of cancer patients who will respond to various anticancer treatments and to improve the toxicity to benefit ratio. Notably, targeted therapy has significantly benefited from these advances, but the application of predictive biomarkers have been a bit slower with some drugs derived from natural sources such as trabectedin, cabazitaxel and alvocidib. In this paper, we discuss some recent advances regarding the use of cancer biomarkers to predict efficacy of some selected natural compounds with a focus on human clinical studies.

Prospects and challenges for clinical decision support in the era of big data

Recently, there has been burgeoning interest in developing more effective and robust clinical decision support systems (CDSSs) for oncology. This has been primarily driven by the demands for more personalized and precise medical practice in oncology in the era of so-called Big Data (BD); an era that promises to harness the power of large-scale data flow to revolutionize cancer treatment. This interest in BD analytics has created new opportunities as well as new unmet challenges. These include: routine aggregation and standardization of clinical data; patient privacy; transformation of current analytical approaches to handle such noisy and heterogeneous data; and expanded use of advanced statistical learning methods based on confluence of modern statistical methods and machine learning algorithms. In this review, we present the current status of CDSSs in oncology, the prospects and current challenges of BD analytics, and the promising role of integrated modern statistics and machine learning algorithms in predicting complex clinical endpoints, individualizing treatment rules, and optimizing dynamic personalized treatment regimens. We discuss issues pertaining to these topics and present application examples from an aggregate of experiences. We also discuss the role of human factors in improving the utilization and acceptance of such enhanced CDSSs and how to mitigate possible sources of human error to achieve optimal performance and wider acceptance.

Prognostic and predictive signatures for treatment decisions

AIM

We develop a subgroup selection procedure using both prognostic and predictive biomarkers to identify four patient subpopulations: low- and high-risk responders, and low- and high-risk nonresponders.

METHODS

We utilize three regression models to identify three sets of biomarkers: S, prognostic biomarkers; T, predictive biomarkers; and U, prognostic and predictive biomarkers. The prognostic signature C(S) combines with a predictive signature, either C(T) or C(U), to develop two procedures C(S,T) and C(S,U) for identification of four subgroups.

RESULTS

Simulation experiment showed that proposed models for identifying the biomarker sets S and U performed well, as did the procedure C(S,U) for subgroup identification.

CONCLUSION

The proposed model provides more comprehensive characterization of patient subpopulations, and better accuracy in patient treatment assignment.

A transatlantic perspective on the integration of immuno-oncology prognostic and predictive biomarkers in innovative clinical trial design

Immuno-therapeutics aim to activate the body's own immune system against cancer and are one of the most promising cancer treatment strategies, but currently limited by a variable response rate. Biomarkers may help to distinguish those patients most likely to respond to therapy; they may also help guide clinical decision making for combination therapies, dosing schedules, and determining progression versus relapse. However, there is a need to confirm such biomarkers in preferably prospective clinical trials before they can be used in practice. Accordingly, it is essential that clinical trials for immuno-therapeutics incorporate biomarkers. Here, focusing on the specific setting of immune therapies, we discuss both the scientific and logistical hurdles to identifying potential biomarkers and testing them in clinical trials.

Optimized adaptive enrichment designs

Based on a Bayesian decision theoretic approach, we optimize frequentist single- and adaptive two-stage trial designs for the development of targeted therapies, where in addition to an overall population, a pre-defined subgroup is investigated. In such settings, the losses and gains of decisions can be quantified by utility functions that account for the preferences of different stakeholders. In particular, we optimize expected utilities from the perspectives both of a commercial sponsor, maximizing the net present value, and also of the society, maximizing cost-adjusted expected health benefits of a new treatment for a specific population. We consider single-stage and adaptive two-stage designs with partial enrichment, where the proportion of patients recruited from the subgroup is a design parameter. For the adaptive designs, we use a dynamic programming approach to derive optimal adaptation rules. The proposed designs are compared to trials which are non-enriched (i.e. the proportion of patients in the subgroup corresponds to the prevalence in the underlying population). We show that partial enrichment designs can substantially improve the expected utilities. Furthermore, adaptive partial enrichment designs are more robust than single-stage designs and retain high expected utilities even if the expected utilities are evaluated under a different prior than the one used in the optimization. In addition, we find that trials optimized for the sponsor utility function have smaller sample sizes compared to trials optimized under the societal view and may include the overall population (with patients from the complement of the subgroup) even if there is substantial evidence that the therapy is only effective in the subgroup.

Increasing the efficiency of oncology basket trials using a Bayesian approach

With the rapid growth of targeted and immune-oncology therapies, novel statistical design approaches are needed to increase the flexibility and efficiency of early phase oncology trials. Basket trials enroll patients with defined biological deficiencies, but with multiple histologic tumor types (or indications), to discover in which indications the drug is active. In such designs different indications are typically analyzed independently. This, however, ignores potential biological similarities among the indications. Our research provides a statistical methodology to enhance such basket trials by assessing the homogeneity of the response rates among indications at an interim analysis, and applying a Bayesian hierarchical modeling approach in the second stage if the efficacy is deemed reasonably homogenous across indications. This increases the power of the study by allowing indications with similar response rates to borrow information from each other. Via simulations, we quantify the efficiency gain of our proposed approach relative to the conventional parallel approach. The operating characteristics of our method depend on the similarity of the response rates between the different indications. If the response rates are comparable in most or all indications after treatment with the investigational drug, a substantial increase in efficiency as compared to the conventional approach can be obtained as fewer patients are required or a higher power is attained. We also demonstrate that efficacy again decreases if the response rates vary considerably among tumor types but it is still better than the conventional approach.

Enhancement of the adaptive signature design for learning and confirming in a single pivotal trial

Because of the complexity of cancer biology, often the target pathway is not well understood at the time that phase III trials are initiated. A 2-stage trial design was previously proposed for identifying a subgroup of interest in a learn stage, on the basis of 1 or more baseline biomarkers, and then subsequently confirming it in a confirmation stage. In this article, we discuss some practical aspects of this type of design and describe an enhancement to this approach that can be built into the study randomization to increase the robustness of the evaluation. Furthermore, we show via simulation studies how the proportion of patients allocated to the learn stage versus the confirm stage impacts the power and provide recommendations.

Real-Time Reverse-Transcription Quantitative Polymerase Chain Reaction Assay Is a Feasible Method for the Relative Quantification of Heregulin Expression in Non-Small Cell Lung Cancer Tissue

In preclinical studies, heregulin (HRG) expression was shown to be the most relevant predictive biomarker for response to patritumab, a fully human anti–epidermal growth factor receptor 3 monoclonal antibody. In support of a phase 2 study of erlotinib ± patritumab in non–small cell lung cancer (NSCLC), a reverse-transcription quantitative polymerase chain reaction (RT-qPCR) assay for relative quantification of HRG expression from formalin-fixed paraffin-embedded (FFPE) NSCLC tissue samples was developed and validated and described herein. Test specimens included matched FFPE normal lung and NSCLC and frozen NSCLC tissue, and HRG-positive and HRG-negative cell lines. Formalin-fixed paraffin-embedded tissue was examined for functional performance. Heregulin distribution was also analyzed across 200 NSCLC commercial samples. Applied Biosystems TaqMan Gene Expression Assays were run on the Bio-Rad CFX96 real-time PCR platform. Heregulin RT-qPCR assay specificity, PCR efficiency, PCR linearity, and reproducibility were demonstrated. The final assay parameters included the Qiagen FFPE RNA Extraction Kit for RNA extraction from FFPE NSCLC tissue, 50 ng of RNA input, and 3 reference (housekeeping) genes (HMBS, IPO8, and EIF2B1), which had expression levels similar to HRG expression levels and were stable among FFPE NSCLC samples. Using the validated assay, unimodal HRG distribution was confirmed across 185 evaluable FFPE NSCLC commercial samples. Feasibility of an RT-qPCR assay for the quantification of HRG expression in FFPE NSCLC specimens was demonstrated.

Biomarker-Guided Non-Adaptive Trial Designs in Phase II and Phase III: A Methodological Review

Biomarker-guided treatment is a rapidly developing area of medicine, where treatment choice is personalised according to one or more of an individual’s biomarker measurements. A number of biomarker-guided trial designs have been proposed in the past decade, including both adaptive and non-adaptive trial designs which test the effectiveness of a biomarker-guided approach to treatment with the aim of improving patient health. A better understanding of them is needed as challenges occur both in terms of trial design and analysis. We have undertaken a comprehensive literature review based on an in-depth search strategy with a view to providing the research community with clarity in definition, methodology and terminology of the various biomarker-guided trial designs (both adaptive and non-adaptive designs) from a total of 211 included papers. In the present paper, we focus on non-adaptive biomarker-guided trial designs for which we have identified five distinct main types mentioned in 100 papers. We have graphically displayed each non-adaptive trial design and provided an in-depth overview of their key characteristics. Substantial variability has been observed in terms of how trial designs are described and particularly in the terminology used by different authors. Our comprehensive review provides guidance for those designing biomarker-guided trials.

PIPELINEs: Creating Comparable Clinical Knowledge Efficiently by Linking Trial Platforms

Adaptive, seamless, multisponsor, multitherapy clinical trial designs executed as large scale platforms, could create superior evidence more efficiently than single-sponsor, single-drug trials. These trial PIPELINEs also could diminish barriers to trial participation, increase the representation of real-world populations, and create systematic evidence development for learning throughout a therapeutic life cycle, to continually refine its use. Comparable evidence could arise from multiarm design, shared comparator arms, and standardized endpoints-aiding sponsors in demonstrating the distinct value of their innovative medicines; facilitating providers and patients in selecting the most appropriate treatments; assisting regulators in efficacy and safety determinations; helping payers make coverage and reimbursement decisions; and spurring scientists with translational insights. Reduced trial times and costs could enable more indications, reduced development cycle times, and improved system financial sustainability. Challenges to overcome range from statistical to operational to collaborative governance and data exchange.

Long range personalized cancer treatment strategies incorporating evolutionary dynamics

Background

Current cancer precision medicine strategies match therapies to static consensus molecular properties of an individual’s cancer, thus determining the next therapeutic maneuver. These strategies typically maintain a constant treatment while the cancer is not worsening. However, cancers feature complicated sub-clonal structure and dynamic evolution. We have recently shown, in a comprehensive simulation of two non-cross resistant therapies across a broad parameter space representing realistic tumors, that substantial improvement in cure rates and median survival can be obtained utilizing dynamic precision medicine strategies. These dynamic strategies explicitly consider intratumoral heterogeneity and evolutionary dynamics, including predicted future drug resistance states, and reevaluate optimal therapy every 45 days. However, the optimization is performed in single 45 day steps (“single-step optimization”).

Results

Herein we evaluate analogous strategies that think multiple therapeutic maneuvers ahead, considering potential outcomes at 5 steps ahead (“multi-step optimization”) or 40 steps ahead (“adaptive long term optimization (ALTO)”) when recommending the optimal therapy in each 45 day block, in simulations involving both 2 and 3 non-cross resistant therapies. We also evaluate an ALTO approach for situations where simultaneous combination therapy is not feasible (“Adaptive long term optimization: serial monotherapy only (ALTO-SMO)”). Simulations utilize populations of 764,000 and 1,700,000 virtual patients for 2 and 3 drug cases, respectively. Each virtual patient represents a unique clinical presentation including sizes of major and minor tumor subclones, growth rates, evolution rates, and drug sensitivities.

While multi-step optimization and ALTO provide no significant average survival benefit, cure rates are significantly increased by ALTO. Furthermore, in the subset of individual virtual patients demonstrating clinically significant difference in outcome between approaches, by far the majority show an advantage of multi-step or ALTO over single-step optimization. ALTO-SMO delivers cure rates superior or equal to those of single- or multi-step optimization, in 2 and 3 drug cases respectively.

Conclusion

In selected virtual patients incurable by dynamic precision medicine using single-step optimization, analogous strategies that “think ahead” can deliver long-term survival and cure without any disadvantage for non-responders. When therapies require dose reduction in combination (due to toxicity), optimal strategies feature complex patterns involving rapidly interleaved pulses of combinations and high dose monotherapy.

Reviewers

This article was reviewed by Wendy Cornell, Marek Kimmel, and Andrzej Swierniak. Wendy Cornell and Andrzej Swierniak are external reviewers (not members of the Biology Direct editorial board). Andrzej Swierniak was nominated by Marek Kimmel.

Electronic supplementary material

The online version of this article (doi:10.1186/s13062-016-0153-2) contains supplementary material, which is available to authorized users.

Optimizing Trial Designs for Targeted Therapies

An important objective in the development of targeted therapies is to identify the populations where the treatment under consideration has positive benefit risk balance. We consider pivotal clinical trials, where the efficacy of a treatment is tested in an overall population and/or in a pre-specified subpopulation. Based on a decision theoretic framework we derive optimized trial designs by maximizing utility functions. Features to be optimized include the sample size and the population in which the trial is performed (the full population or the targeted subgroup only) as well as the underlying multiple test procedure. The approach accounts for prior knowledge of the efficacy of the drug in the considered populations using a two dimensional prior distribution. The considered utility functions account for the costs of the clinical trial as well as the expected benefit when demonstrating efficacy in the different subpopulations. We model utility functions from a sponsor's as well as from a public health perspective, reflecting actual civil interests. Examples of optimized trial designs obtained by numerical optimization are presented for both perspectives.

Adaptive Design for a Confirmatory Basket Trial in Multiple Tumor Types Based on a Putative Predictive Biomarker

Increasingly, tumors are defined on a molecular basis rather than only on histology, and targeted agents, which address these molecular subtypes, are being approved. This profusion of molecular subtypes creates "rare" diseases as subsets of common cancers, leading to difficulties in enrolling sufficiently large cohorts for confirmatory trials. However, if the molecular subtype is shared across various histologies, these may be pooled into a basket trial. To date, basket trials have been primarily for exploratory early development. In this perspective, we consider qualitative designs for confirmatory basket trials. These confirmatory basket designs will provide patients in niche indications with enhanced access to novel therapies, facilitate development and full approval for niche indications, allow accelerated approval for indications within a basket based on a surrogate endpoint, reduce development cost by combining trials, and enhance the ability of regulatory authorities to evaluate risk and benefit in niche indications.

The Impact of Adaptive Design on Portfolio Optimization

BACKGROUND

Pharmaceutical portfolios are optimized by improved allocation of a fixed budget into individual trials that leads to an improved value of a portfolio. This paper investigates how flexibility of adaptive design contributes to portfolio optimization.

METHODS

An example portfolio was designed, and strategies that did or did not include trials with adaptive designs were specified. Operating characteristics of a traditional portfolio were compared to that of an adaptive portfolio. Adaptive portfolios offer potential advantages over traditional ones. Its flexibility largely increases the number of decision points, and as such it allows for a much more frequent reassessment of portfolios. Additionally, an adaptive portfolio can correct itself if initial decisions were made incorrectly.

RESULTS

Despite all these advantages, the adaptive portfolio did not outperform the traditional portfolio. The main reason is that in this case, adaptive designs allowed for increases in sample size to the point where improvements per unit increase were minimal, instead of allocating this budget to additional trials.

CONCLUSIONS

It is critical to minimize missed opportunities to initiate new promising trials, and to increase sample size only in regions that promise meaningful improvements in power.

Biomarker-Guided Adaptive Trial Designs in Phase II and Phase III: A Methodological Review

BACKGROUND

Personalized medicine is a growing area of research which aims to tailor the treatment given to a patient according to one or more personal characteristics. These characteristics can be demographic such as age or gender, or biological such as a genetic or other biomarker. Prior to utilizing a patient's biomarker information in clinical practice, robust testing in terms of analytical validity, clinical validity and clinical utility is necessary. A number of clinical trial designs have been proposed for testing a biomarker's clinical utility, including Phase II and Phase III clinical trials which aim to test the effectiveness of a biomarker-guided approach to treatment; these designs can be broadly classified into adaptive and non-adaptive. While adaptive designs allow planned modifications based on accumulating information during a trial, non-adaptive designs are typically simpler but less flexible.

METHODS AND FINDINGS

We have undertaken a comprehensive review of biomarker-guided adaptive trial designs proposed in the past decade. We have identified eight distinct biomarker-guided adaptive designs and nine variations from 107 studies. Substantial variability has been observed in terms of how trial designs are described and particularly in the terminology used by different authors. We have graphically displayed the current biomarker-guided adaptive trial designs and summarised the characteristics of each design.

CONCLUSIONS

Our in-depth overview provides future researchers with clarity in definition, methodology and terminology for biomarker-guided adaptive trial designs.

Methods for identification and confirmation of targeted subgroups in clinical trials: A systematic review

Important objectives in the development of stratified medicines include the identification and confirmation of subgroups of patients with a beneficial treatment effect and a positive benefit-risk balance. We report the results of a literature review on methodological approaches to the design and analysis of clinical trials investigating a potential heterogeneity of treatment effects across subgroups. The identified approaches are classified based on certain characteristics of the proposed trial designs and analysis methods. We distinguish between exploratory and confirmatory subgroup analysis, frequentist, Bayesian and decision-theoretic approaches and, last, fixed-sample, group-sequential, and adaptive designs and illustrate the available trial designs and analysis strategies with published case studies.

Predictive biomarkers for treatment selection: statistical considerations

Predictive biomarkers are developed for treatment selection to identify patients who are likely to benefit from a particular therapy. This review describes statistical methods and discusses issues in the development of predictive biomarkers to enhance study efficiency for detection of treatment effect on the selected responder patients in clinical studies. The statistical procedure for treatment selection consists of three components: biomarker identification, subgroup selection and clinical utility assessment. Major statistical issues discussed include biomarker designs, procedures to identify predictive biomarkers, classification models for subgroup selection, subgroup analysis and multiple testing for clinical utility assessment and evaluation.

Translating predictive biomarkers within oncology clinical development programs

Predictive biomarkers provide essential information to enable personalized medicine, and hold the promise for enhancing the effectiveness and value of cancer therapies. However, they do not always work. This review provides a framework for managing the risk of predictive biomarkers and maximally harvesting their benefit. Methods are provided which permit data-driven, adaptive decision making about the use of predictive biomarkers during clinical development, applying them to the extent they are validated by the clinical data. Techniques for optimizing overall development efficiency, measured as the number of successful drug indications approved per patient utilized, are also presented.

Clinical utility estimation for assay cutoffs in early phase oncology enrichment trials

Predictive enrichment strategies use biomarkers to selectively enroll oncology patients into clinical trials to more efficiently demonstrate therapeutic benefit. Because the enriched population differs from the patient population eligible for screening with the biomarker assay, there is potential for bias when estimating clinical utility for the screening eligible population if the selection process is ignored. We write estimators of clinical utility as integrals averaging regression model predictions over the conditional distribution of the biomarker scores defined by the assay cutoff and discuss the conditions under which consistent estimation can be achieved while accounting for some nuances that may arise as the biomarker assay progresses toward a companion diagnostic. We outline and implement a Bayesian approach in estimating these clinical utility measures and use simulations to illustrate performance and the potential biases when estimation naively ignores enrichment. Results suggest that the proposed integral representation of clinical utility in combination with Bayesian methods provide a practical strategy to facilitate cutoff decision-making in this setting.

Clinical Translation and Validation of a Predictive Biomarker for Patritumab, an Anti-human Epidermal Growth Factor Receptor 3 (HER3) Monoclonal Antibody, in Patients With Advanced Non-small Cell Lung Cancer

Background

During early clinical development, prospective identification of a predictive biomarker and validation of an assay method may not always be feasible. Dichotomizing a continuous biomarker measure to classify responders also leads to challenges. We present a case study of a prospective–retrospective approach for a continuous biomarker identified after patient enrollment but defined prospectively before the unblinding of data. An analysis of the strengths and weaknesses of this approach and the challenges encountered in its practical application are also provided.

Methods

HERALD (NCT02134015) was a double-blind, phase 2 study in patients with non-small cell lung cancer (NSCLC) randomized to erlotinib with placebo or with high or low doses of patritumab, a monoclonal antibody targeted against human epidermal growth factor receptor 3 (HER3). While the primary objective was to assess safety and progression-free survival (PFS), a secondary objective was to determine a single predictive biomarker hypothesis to identify subjects most likely to benefit from the addition of patritumab. Although not identified as the primary biomarker in the study protocol, on the basis of preclinical results from 2 independent laboratories, expression levels of the HER3 ligand heregulin (HRG) were prospectively declared the predictive biomarker before data unblinding but after subject enrollment. An assay to measure HRG mRNA was developed and validated. Other biomarkers, such as epidermal growth factor receptor (EGFR) mutation status, were also evaluated in an exploratory fashion. The cutoff value for high vs. low HRG mRNA levels was set at the median delta threshold cycle. A maximum likelihood analysis was performed to evaluate the provisional cutoff. The relationship of HRG values to PFS hazard ratios (HRs) was assessed as a measure of internal validation. Additional NSCLC samples were analyzed to characterize HRG mRNA distribution.

Results

The subgroup of patients with high HRG mRNA levels (“HRG-high”) demonstrated clinical benefit from patritumab treatment with HRs of 0.37 (P = 0.0283) and 0.29 (P = 0.0027) in the high- and low-dose patritumab arms, respectively. However, only 102 of the 215 randomized patients (47.4%) had sufficient tumor samples for HRG mRNA measurement. Maximum likelihood analysis showed that the provisional cutoff was within the optimal range. In the placebo arm, the HRG-high subgroup demonstrated worse prognosis compared with HRG-low. A continuous relationship was observed between increased HRG mRNA levels and lower HR. Additional NSCLC samples (N = 300) demonstrated a similar unimodal distribution to that observed in this study, suggesting that the defined cutoff may be applicable to future NSCLC studies.

Conclusions

The prospective–retrospective approach was successful in clinically validating a probable predictive biomarker. Post hoc in vitro studies and statistical analyses permitted further testing of the underlying assumptions. However, limitations of this analysis include the incomplete collection of adequate tumor tissue and a lack of stratification. In a phase 3 study, findings are being confirmed, and the HRG cutoff value is being further refined.

ClinicalTrials.gov Number

NCT02134015.

•

High heregulin levels predict benefit from patritumab treatment in patients with NSCLC.

•

A prospective–retrospective approach provisionally validated a predictive biomarker.

•

Post hoc analyses can be used to test underlying assumptions in biomarker validation.

•

The median may be a reasonable initial cutoff for a unimodal continuous biomarker.

Adaptive designs for subpopulation analysis optimizing utility functions

If the response to treatment depends on genetic biomarkers, it is important to identify predictive biomarkers that define (sub-)populations where the treatment has a positive benefit risk balance. One approach to determine relevant subpopulations are subgroup analyses where the treatment effect is estimated in biomarker positive and biomarker negative groups. Subgroup analyses are challenging because several types of risks are associated with inference on subgroups. On the one hand, by disregarding a relevant subpopulation a treatment option may be missed due to a dilution of the treatment effect in the full population. Furthermore, even if the diluted treatment effect can be demonstrated in an overall population, it is not ethical to treat patients that do not benefit from the treatment when they can be identified in advance. On the other hand, selecting a spurious subpopulation increases the risk to restrict an efficacious treatment to a too narrow fraction of a potential benefiting population. We propose to quantify these risks with utility functions and investigate nonadaptive study designs that allow for inference on subgroups using multiple testing procedures as well as adaptive designs, where subgroups may be selected in an interim analysis. The characteristics of such adaptive and nonadaptive designs are compared for a range of scenarios.

Cancer genomic research at the crossroads: realizing the changing genetic landscape as intratumoral spatial and temporal heterogeneity becomes a confounding factor

The US National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI) created the Cancer Genome Atlas (TCGA) Project in 2006. The TCGA’s goal was to sequence the genomes of 10,000 tumors to identify common genetic changes among different types of tumors for developing genetic-based treatments. TCGA offered great potential for cancer patients, but in reality has little impact on clinical applications. Recent reports place the past TCGA approach of testing a small tumor mass at a single time-point at a crossroads. This crossroads presents us with the conundrum of whether we should sequence more tumors or obtain multiple biopsies from each individual tumor at different time points. Sequencing more tumors with the past TCGA approach of single time-point sampling can neither capture the heterogeneity between different parts of the same tumor nor catch the heterogeneity that occurs as a function of time, error rates, and random drift. Obtaining multiple biopsies from each individual tumor presents multiple logistical and financial challenges. Here, we review current literature and rethink the utility and application of the TCGA approach. We discuss that the TCGA-led catalogue may provide insights into studying the functional significance of oncogenic genes in reference to non-cancer genetic background. Different methods to enhance identifying cancer targets, such as single cell technology, real time imaging of cancer cells with a biological global positioning system, and cross-referencing big data sets, are offered as ways to address sampling discrepancies in the face of tumor heterogeneity. We predict that TCGA landmarks may prove far more useful for cancer prevention than for cancer diagnosis and treatment when considering the effect of non-cancer genes and the normal genetic background on tumor microenvironment. Cancer prevention can be better realized once we understand how therapy affects the genetic makeup of cancer over time in a clinical setting. This may help create novel therapies for gene mutations that arise during a tumor’s evolution from the selection pressure of treatment.

Predictive biomarkers and companion diagnostics. The future of immunohistochemistry: "in situ proteomics," or just a "stain"?

A ‘companion diagnostic’ is a test for a predictive biomarker, that classifies patients (tumors) into responders and non-responders, for a specified therapeutic agent. Companion diagnostics are designated as Class III medical devices by the FDA, because the test result equates directly to administration of a drug.

Testing for HER2 expression was approved by the FDA in 1998, and served as the prototype for using immunohistochemistry (IHC) as the basis for a companion diagnostic. However, over four decades IHC has primarily been employed in a broad range of ‘special stains’, for identification and classification cells and tumors in FFPE (formalin fixed paraffin embedded) tissues. During the long use of IHC as a ‘special stain’ we have acquired some very bad habits, changing protocols, concentrations, incubation times, retrieval methods, or reagents, to achieve the perception of a ‘good’ stain, that ‘pleases the eye’ of the user pathologist. While this approach may be acceptable for IHC stains, it is a recipe for disaster when transferred to companion diagnostics, where quantification and absolute reproducibility are required. In the context of companion diagnostics the IHC method should be regarded as an assay, not simply a stain. Elevating IHC to a true immunoassay will necessitate a much more rigorous approach to performance, reproducibility and control. The ultimate goal is to supplement morphologic judgment with precise measurement of proteins in tissues and in individual cells, ‘in situ proteomics’ as it were.

Nonstandard personalized medicine strategies for cancer may lead to improved patient outcomes

Cancer is an evolutionary process that is driven by mutation and selection. Tumors are genetically unstable, and research has shown that this is the most efficient way for cancers to evolve. Genetic instability leads to genetic heterogeneity and dynamic change within a single individual's tumor, in turn leading to therapeutic resistance. Cancer treatment has also evolved from an empirical science of killing dividing cells to the current era of 'personalized medicine', exquisitely targeting the molecular features of individual cancers. However, current personalized medicine regards a single individual's cancer as largely uniform and static. Moreover, from a strategic perspective, current personalized medicine thinks primarily of the immediate therapy selection. Ongoing research suggests that new, nonstandard personalized treatment strategies that plan further ahead and consider intratumoral heterogeneity and the evolving nature of cancer (due to genetic instability) may lead to the next level of therapeutic benefit beyond current personalized medicine.