Predictive biomarker validation in practice: lessons from real trials

Two-stage stratified designs with survival outcomes and adjustment for misclassification in predictive biomarkers

Biomarker stratified clinical trial designs are versatile tools to assess biomarker clinical utility and address its relationship with clinical endpoints. Due to imperfect assays and/or classification rules, biomarker status is prone to errors. To account for biomarker misclassification, we consider a two-stage stratified design for survival outcomes with an adjustment for misclassification in predictive biomarkers. Compared to continuous and/or binary outcomes, the test statistics for survival outcomes with an adjustment for biomarker misclassification is much more complicated and needs to take special care. We propose to use the information from the observed biomarker status strata to construct adjusted log-rank statistics for true biomarker status strata. These adjusted log-rank statistics are then used to develop sequential tests for the global (composite) hypothesis and component-wise hypothesis. We discuss the power analysis with the control of the type-I error rate by using the correlations between the adjusted log-rank statistics within and between the design stages. Our method is illustrated with examples of the recent successful development of immunotherapy in nonsmall-cell lung cancer.

Biomarkers in Polycystic Kidney Disease: Are We There?

This article describes the use of prognostic, predictive, and response biomarkers that have been developed for autosomal dominant polycystic kidney disease and their use in clinical care or drug development. We focus on biochemical markers that can be assayed in patients' blood and urine and their association with the outcome of decreased glomerular filtration rate. There have been several studies on prognostic biomarkers. The most promising ones have been markers of tubular injury, inflammation, metabolism, or the vasopressin-urinary concentration axis. So far, none have been shown to be superior to kidney volume-based biomarkers. Several biomarkers are additive to kidney volume and genotype in prognostic models, but there have been few direct comparisons between the biochemical markers to identify the best ones. Moreover, there is a lack of uniformity in the statistical tools used to assess and compare biomarkers. There have been few reports of predictive and response biomarkers, and none are suitable surrogate endpoints. The U.S. Food and Drug Administration's Biomarker Qualification Program provides a regulatory pathway to approve biomarkers for use across multiple drug-development programs.

Confirmatory prediction-driven RCTs in comparative effectiveness settings for cancer treatment

BACKGROUND

Medical advances in the treatment of cancer have allowed the development of multiple approved treatments and prognostic and predictive biomarkers for many types of cancer. Identifying improved treatment strategies among approved treatment options, the study of which is termed comparative effectiveness, using predictive biomarkers is becoming more common. RCTs that incorporate predictive biomarkers into the study design, called prediction-driven RCTs, are needed to rigorously evaluate these treatment strategies. Although researched extensively in the experimental treatment setting, literature is lacking in providing guidance about prediction-driven RCTs in the comparative effectiveness setting.

METHODS

Realistic simulations with time-to-event endpoints are used to compare contrasts of clinical utility and provide examples of simulated prediction-driven RCTs in the comparative effectiveness setting.

RESULTS

Our proposed contrast for clinical utility accurately estimates the true clinical utility in the comparative effectiveness setting while in some scenarios, the contrast used in current literature does not.

DISCUSSION

It is important to properly define contrasts of interest according to the treatment setting. Realistic simulations should be used to choose and evaluate the RCT design(s) able to directly estimate that contrast. In the comparative effectiveness setting, our proposed contrast for clinical utility should be used.

Pre- and on-treatment lactate dehydrogenase as a prognostic and predictive biomarker in advanced non-small cell lung cancer

BACKGROUND

The survival outcomes of patients with advanced non–small cell lung cancer (NSCLC) treated with immune checkpoint inhibitors (ICIs) are variable. This study investigated whether pre‐ and on‐treatment lactate dehydrogenase (LDH) could better prognosticate and select patients for ICI therapy.

METHODS

Using data from the POPLAR and OAK trials of atezolizumab versus docetaxel in previously treated advanced NSCLC, the authors assessed the prognostic and predictive value of pretreatment LDH (less than or equal to vs greater than the upper limit of normal). They further examined changes in on‐treatment LDH by performing landmark analyses and estimated overall survival (OS) distributions according to the LDH level stratified by the response category (complete response [CR]/partial response [PR] vs stable disease [SD]). They repeated pretreatment analyses in subgroups defined by the programmed death ligand 1 (PD‐L1) status.

RESULTS

This study included 1327 patients with available pretreatment LDH. Elevated pretreatment LDH was associated with an adverse prognosis regardless of treatment (hazard ratio [HR] for atezolizumab OS, 1.49; P = .0001; HR for docetaxel OS, 1.30; P = .004; P for treatment by LDH interaction = .28). Findings for elevated pretreatment LDH were similar for patients with positive PD‐L1 expression treated with atezolizumab. Persistently elevated on‐treatment LDH was associated with a 1.3‐ to 2.8‐fold increased risk of death at weeks 6, 12, 18, and 24 regardless of treatment. Elevated LDH at 6 weeks was associated with significantly shorter OS regardless of radiological response (HR for CR/PR, 2.10; P = .04; HR for SD, 1.50; P < .01), with similar findings observed at 12 weeks.

CONCLUSIONS

In previously treated advanced NSCLC, elevated pretreatment LDH is an independent adverse prognostic marker. There is no evidence that pretreatment LDH predicts ICI benefit. Persistently elevated on‐treatment LDH is associated with worse OS despite radiologic response.

This analysis of 1327 patients with advanced non–small cell lung cancer from the POPLAR and OAK randomized controlled trials has found that lactate dehydrogenase (LDH) is a useful pre‐ and on‐treatment prognostic marker that can assist clinicians in counselling patients undergoing second‐ or later‐line atezolizumab or docetaxel. However, the findings fail to support the use of LDH as a predictive biomarker for immune checkpoint inhibitor therapy and reinforce the importance of rigorous validation of promising predictive biomarkers using randomized data.

Artificial intelligence, molecular subtyping, biomarkers, and precision oncology

A targeted cancer therapy is only useful if there is a way to accurately identify the tumors that are susceptible to that therapy. Thus rapid expansion in the number of available targeted cancer treatments has been accompanied by a robust effort to subdivide the traditional histological and anatomical tumor classifications into molecularly defined subtypes. This review highlights the history of the paired evolution of targeted therapies and biomarkers, reviews currently used methods for subtype identification, and discusses challenges to the implementation of precision oncology as well as possible solutions.

Chronic Lymphocytic Leukemia

Patients with chronic lymphocytic leukemia can be divided into three categories: those who are minimally affected by the problem, often never requiring therapy; those that initially follow an indolent course but subsequently progress and require therapy; and those that from the point of diagnosis exhibit an aggressive disease necessitating treatment. Likewise, such patients pass through three phases: development of the disease, diagnosis, and need for therapy. Finally, the leukemic clones of all patients appear to require continuous input from the exterior, most often through membrane receptors, to allow them to survive and grow. This review is presented according to the temporal course that the disease follows, focusing on those external influences from the tissue microenvironment (TME) that support the time lines as well as those internal influences that are inherited or develop as genetic and epigenetic changes occurring over the time line. Regarding the former, special emphasis is placed on the input provided via the B-cell receptor for antigen and the C-X-C-motif chemokine receptor-4 and the therapeutic agents that block these inputs. Regarding the latter, prominence is laid upon inherited susceptibility genes and the genetic and epigenetic abnormalities that lead to the developmental and progression of the disease.

Translational biomarkers in the era of precision medicine

In this chapter we discuss the past, present and future of clinical biomarker development. We explore the advent of new technologies, paving the way in which health, medicine and disease is understood. This review includes the identification of physicochemical assays, current regulations, the development and reproducibility of clinical trials, as well as, the revolution of omics technologies and state-of-the-art integration and analysis approaches.

MET inhibitors in cancer: pitfalls and challenges

Introduction: The HGF/MET axis is a key therapeutic pathway in cancer; it is aberrantly activated because of mutations, fusions, amplification or aberrant ligand production. Extensive efforts have been made to discover predictive factors of anti-MET therapeutic efficacy, but they have mostly unsuccessful. An understanding of the intrinsic and acquired mechanism of MET resistance will be fundamental for the development of new therapeutic interventions.Areas covered: This article provides a systematic review of phase II randomized and phase III clinical trials investigating the use of MET inhibitors in the treatment of cancer. We discuss preliminary findings on efficacy and methodologic design flaws in these trials.Expert opinion: MET inhibitors showed poor activity in unselected patients or patients selected by MET expression, p-MET or high HGF basal levels. The efficacy in advanced solid tumors is very modest and in phase III clinical trials, survival differences did not fulfill the stringent requirements of ESMO-Magnitude Clinical Benefit Score (MCBS). Prospective novel liquid biomarker-driven studies and novel trial designs such as Umbrella and Basket trials are necessary to progress MET inhibitor development.

Qualifying antibodies for image-based immune profiling and multiplexed tissue imaging

Multiplexed tissue imaging enables precise, spatially resolved enumeration and characterization of cell types and states in human resection specimens. A growing number of methods applicable to formalin-fixed, paraffin-embedded (FFPE) tissue sections have been described, the majority of which rely on antibodies for antigen detection and mapping. This protocol provides step-by-step procedures for confirming the selectivity and specificity of antibodies used in fluorescence-based tissue imaging and for the construction and validation of antibody panels. Although the protocol is implemented using tissue-based cyclic immunofluorescence (t-CyCIF) as an imaging platform, these antibody-testing methods are broadly applicable. We demonstrate assembly of a 16-antibody panel for enumerating and localizing T cells and B cells, macrophages, and cells expressing immune checkpoint regulators. The protocol is accessible to individuals with experience in microscopy and immunofluorescence; some experience in computation is required for data analysis. A typical 30-antibody dataset for 20 FFPE slides can be generated within 2 weeks.

Translating Immuno-oncology Biomarkers to Diagnostic Tests: A Regulatory Perspective

The rapid development of effective immunotherapy using immune-checkpoint inhibitors (ICIs) against many different cancer types opened a new front in cancer treatment. Immunotherapy is undoubtedly one of the biggest breakthroughs in cancer therapy within the past decade. The identification of predictive biomarkers to select the patients most likely to respond to ICI monotherapies or emerging combination therapies remains one of the major unmet needs for the oncology community.This chapter provides an overview of existing and emerging biomarkers associated with ICI response. Additionally, using several case studies of FDA approved or authorized in vitro diagnostic oncology devices, this chapter also provides an overview of analytical and clinical validation considerations of diagnostic tests for immuno-oncology biomarkers.

An alternative method to analyse the biomarker-strategy design

Recent developments in genomics and proteomics enable the discovery of biomarkers that allow identification of subgroups of patients responding well to a treatment. One currently used clinical trial design incorporating a predictive biomarker is the so‐called biomarker strategy design (or marker‐based strategy design). Conventionally, the results from this design are analysed by comparing the mean of the biomarker‐led arm with the mean of the randomised arm. Several problems regarding the analysis of the data obtained from this design have been identified in the literature. In this paper, we show how these problems can be resolved if the sample sizes in the subgroups fulfil the specified orthogonality condition. We also propose a different analysis strategy that allows definition of test statistics for the biomarker‐by‐treatment interaction effect as well as for the classical treatment effect and the biomarker effect. We derive equations for the sample size calculation for the case of perfect and imperfect biomarker assays. We also show that the often used 1:1 randomisation does not necessarily lead to the smallest sample size. In addition, we provide point estimators and confidence intervals for the treatment effects in the subgroups. Application of our method is illustrated using a real data example.

Estimating mean local posterior predictive benefit for biomarker-guided treatment strategies

Precision medicine has emerged from the awareness that many human diseases are intrinsically heterogeneous with respect to their pathogenesis and composition among patients as well as dynamic over the course therapy. Its successful application relies on our understanding of distinct molecular profiles and their biomarkers which can be used as targets to devise treatment strategies that exploit current understanding of the biological mechanisms of the disease. Precision medicine present challenges to traditional paradigms of clinical translational, however, for which estimates of population-averaged effects from large randomized trials are used as the basis for demonstrating improvements comparative effectiveness. A general approach for estimating the relative effectiveness of biomarker-guided therapeutic strategies is presented herein. The statistical procedure attempts to define the local benefit of a given biomarker-guided therapeutic strategy in consideration of the treatment response surfaces, selection rule, and inter-cohort balance of prognostic determinants. Theoretical and simulation results are provided. Additionally, the methodology is demonstrated through a proteomic study of lower grade glioma.

Novel biomarker-based model for the prediction of sorafenib response and overall survival in advanced hepatocellular carcinoma: a prospective cohort study

Background

Prediction of the outcome of sorafenib therapy using biomarkers is an unmet clinical need in patients with advanced hepatocellular carcinoma (HCC). The aim was to develop and validate a biomarker-based model for predicting sorafenib response and overall survival (OS).

Methods

This prospective cohort study included 124 consecutive HCC patients (44 with disease control, 80 with progression) with Child-Pugh class A liver function, who received sorafenib. Potential serum biomarkers (namely, hepatocyte growth factor [HGF], fibroblast growth factor [FGF], vascular endothelial growth factor receptor-1, CD117, and angiopoietin-2) were tested. After identifying independent predictors of tumor response, a risk scoring system for predicting OS was developed and 3-fold internal validation was conducted.

Results

A risk scoring system was developed with six covariates: etiology, platelet count, Barcelona Clinic Liver Cancer stage, protein induced by vitamin K absence-II, HGF, and FGF. When patients were stratified into low-risk (score ≤ 5), intermediate-risk (score 6), and high-risk (score ≥ 7) groups, the model provided good discriminant functions on tumor response (concordance [c]-index, 0.884) and 12-month survival (area under the curve [AUC], 0.825). The median OS was 19.0, 11.2, and 6.1 months in the low-, intermediate-, and high-risk group, respectively (P < 0.001). In internal validation, the model maintained good discriminant functions on tumor response (c-index, 0.825) and 12-month survival (AUC, 0.803), and good calibration functions (all P > 0.05 between expected and observed values).

Conclusions

This new model including serum FGF and HGF showed good performance in predicting the response to sorafenib and survival in patients with advanced HCC.

Mammographic density: a potential monitoring biomarker for adjuvant and preventative breast cancer endocrine therapies

Increased mammographic density (MD) has been shown beyond doubt to be a marker for increased breast cancer risk, though the underpinning pathobiology is yet to be fully elucidated. Estrogenic activity exerts a strong influence over MD, which consequently has been observed to change predictably in response to tamoxifen anti-estrogen therapy, although results for other selective estrogen receptor modulators and aromatase inhibitors are less consistent. In both primary and secondary prevention settings, tamoxifen-associated MD changes correlate with successful modulation of risk or outcome, particularly among pre-menopausal women; an observation that supports the potential use of MD change as a surrogate marker where short-term MD changes reflect longer-term anti-estrogen efficacy. Here we summarize endocrine therapy-induced MD changes and attendant outcomes and discuss both the need for outcome surrogates in such therapy, as well as make a case for MD as such a monitoring marker. We then discuss the process and steps required to validate and introduce MD into practice as a predictor or surrogate for endocrine therapy efficacy in preventive and adjuvant breast cancer treatment settings.

Strategies for power calculations in predictive biomarker studies in survival data

PURPOSE

Biomarkers and genomic signatures represent potentially predictive tools for precision medicine. Validation of predictive biomarkers in prospective or retrospective studies requires statistical justification of power and sample size. However, the design of these studies is complex and the statistical methods and associated software are limited, especially in survival data. Herein, we address common statistical design issues relevant to these two types of studies and provide guidance and a general template for analysis.

METHODS

A statistical interaction effect in the Cox proportional hazards model is used to describe predictive biomarkers. The analytic form by Peterson et al. and Lachin is utilized to calculate the statistical power for both prospective and retrospective studies.

RESULTS

We demonstrate that the common mistake of using only Hazard Ratio's Ratio (HRR) or two hazard ratios (HRs) can mislead power calculations. We establish that the appropriate parameter settings for prospective studies require median survival time (MST) in 4 subgroups (treatment and control in positive biomarker, treatment and control in negative biomarker). For the retrospective study which has fixed survival time and censored status, we develop a strategy to harmonize the hypothesized parameters and the study cohort. Moreover, we provide an easily-adapted R software application to generate a template of statistical plan for predictive biomarker validation so investigators can easily incorporate into their study proposals.

CONCLUSION

Our study provides guidance and software to help biostatisticians and clinicians design sound clinical studies for testing predictive biomarkers.

The development of CHAMP: a checklist for the appraisal of moderators and predictors

Background

Personalized healthcare relies on the identification of factors explaining why individuals respond differently to the same intervention. Analyses identifying such factors, so called predictors and moderators, have their own set of assumptions and limitations which, when violated, can result in misleading claims, and incorrect actions. The aim of this study was to develop a checklist for critically appraising the results of predictor and moderator analyses by combining recommendations from published guidelines and experts in the field.

Methods

Candidate criteria for the checklist were retrieved through systematic searches of the literature. These criteria were evaluated for appropriateness using a Delphi procedure. Two Delphi rounds yielded a pilot checklist, which was tested on a set of papers included in a systematic review on reinforced home-based palliative care. The results of the pilot informed a third Delphi round, which served to finalize the checklist.

Results

Forty-nine appraisal criteria were identified in the literature. Feedback was obtained from fourteen experts from (bio)statistics, epidemiology and other associated fields elicited via three Delphi rounds. Additional feedback from other researchers was collected in a pilot test. The final version of our checklist included seventeen criteria, covering the design (e.g. a priori plausibility), analysis (e.g. use of interaction tests) and results (e.g. complete reporting) of moderator and predictor analysis, together with the transferability of the results (e.g. clinical importance). There are criteria both for individual papers and for bodies of evidence.

Conclusions

The proposed checklist can be used for critical appraisal of reported moderator and predictor effects, as assessed in randomized or non-randomized studies using individual participant or aggregate data. This checklist is accompanied by a user’s guide to facilitate implementation. Its future use across a wide variety of research domains and study types will provide insights about its usability and feasibility.

Electronic supplementary material

The online version of this article (10.1186/s12874-017-0451-0) contains supplementary material, which is available to authorized users.

Relative efficiency of precision medicine designs for clinical trials with predictive biomarkers

Prospective randomized clinical trials addressing biomarkers are time consuming and costly, but are necessary for regulatory agencies to approve new therapies with predictive biomarkers. For this reason, recently, there have been many discussions and proposals of various trial designs and comparisons of their efficiency in the literature. We compare statistical efficiencies between the marker-stratified design and the marker-based precision medicine design regarding testing/estimating 4 hypotheses/parameters of clinical interest, namely, treatment effects in each marker-positive and marker-negative cohorts, marker-by-treatment interaction, and the marker's clinical utility. As may be expected, the stratified design is more efficient than the precision medicine design. However, it is perhaps surprising to find out how low the relative efficiency can be for the precision medicine design. We quantify the relative efficiency as a function of design factors including the marker-positive prevalence rate, marker assay and classification sensitivity and specificity, and the treatment randomization ratio. It is interesting to examine the trends of the relative efficiency with these design parameters in testing different hypotheses. We advocate to use the stratified design over the precision medicine design in clinical trials with predictive biomarkers.

Current challenges of metastatic breast cancer

Metastasis is one of the most characteristic yet problematic behaviors of cancer cells. Stage IV breast cancer accounts for a large portion of breast cancer-related morbidity and mortality. Despite early detection and improvement in survival owing to advancements in biomedical research and overall improvement of the health system, 6-10% of patients present with stage IV disease in the developed world, with a higher incidence noted elsewhere. Despite advances in biomedical research into cancer, up to 70-80% of patients with stage IV breast cancer die of cancer in 5 years, a disproportionally higher mortality compared with non-metastatic breast cancer. In this article, we review the incidence, survival, heterogeneity, current practice, and challenges in stage IV breast cancer, and we finish by noting new research initiatives to improve poor survival and suggesting future directions. By doing so, we hope to set the basis of future directions for both treating physicians and translational researchers to relieve the suffering of patients with stage IV breast cancer and improve the survival of patients with this dismal disease.

On study designs and hypotheses for clinical trials with predictive biomarkers

Recent interest in conducting clinical trials with predictive biomarkers has generated research in comparing relative efficiency of different trial designs. We find these comparisons of efficiency mostly misleading since they are based on different hypotheses. In this paper, we discuss several commonly used trial designs and consider the hypotheses that each design is capable to address. We first consider the ideal situation of no classification errors, then the more realistic situation where marker assay's sensitivity, specificity and the rule of classification are imperfect. We pay special attention to the differences between treatment utility versus absolute treatment effect, and marker by treatment interaction versus marker utility.

Precision oncology: A new era of cancer clinical trials

Traditionally, site of disease and anatomic staging have been used to define patient populations to be studied in individual cancer clinical trials. In the past decade, however, oncology has become increasingly understood on a cellular and molecular level, with many cancer subtypes being described as a function of biomarkers or tumor genetic mutations. With these changes in the science of oncology have come changes to the way we design and perform clinical trials. Increasingly common are trials tailored to detect enhanced efficacy in a patient subpopulation, e.g. patients with a known biomarker value or whose tumors harbor a specific genetic mutation. Here, we provide an overview of traditional and newer biomarker-based trial designs, and highlight lessons learned through implementation of several ongoing and recently completed trials.

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.

Strategies to design clinical studies to identify predictive biomarkers in cancer research

The discovery of reliable biomarkers to predict efficacy and toxicity of anticancer drugs remains one of the key challenges in cancer research. Despite its relevance, no efficient study designs to identify promising candidate biomarkers have been established. This has led to the proliferation of a myriad of exploratory studies using dissimilar strategies, most of which fail to identify any promising targets and are seldom validated. The lack of a proper methodology also determines that many anti-cancer drugs are developed below their potential, due to failure to identify predictive biomarkers. While some drugs will be systematically administered to many patients who will not benefit from them, leading to unnecessary toxicities and costs, others will never reach registration due to our inability to identify the specific patient population in which they are active. Despite these drawbacks, a limited number of outstanding predictive biomarkers have been successfully identified and validated, and have changed the standard practice of oncology. In this manuscript, a multidisciplinary panel reviews how those key biomarkers were identified and, based on those experiences, proposes a methodological framework-the DESIGN guidelines-to standardize the clinical design of biomarker identification studies and to develop future research in this pivotal field.

Using prognostic factors from case series and cohort studies to identify individuals with poor long-term outcomes during periodontal maintenance

BACKGROUND

The accuracy of applying prognostic factors to individual patients is uncertain.

AIM/METHOD

The aim was to apply prognostic factors from several outcome studies (case series and cohort) to identify: (1) patients who lost a tooth/teeth during periodontal maintenance; (2) patients who were non-responding to treatment; (3) patients needing re-treatment during periodontal maintenance. In addition, tooth loss was related to initial prognosis and it was determined which of the prognostic factors were also risk factors. Chi squared analysis was carried out for the outcomes of patients with-, and without prognostic factors. Significance level was set at p ≤ 0.05. Sensitivity and specificity was calculated for patients with and without prognostic factors.

RESULTS

The prognostic factors only identified a small proportion of patients who lost teeth (34-38%). Combining the prognostic factors resulted in a lower accuracy. A higher proportion of patients with no prognostic factors lost teeth (53.8-96.2%). The chance of identifying a non-responding patient based on family history was 5.9%, for stress 32.4%, and for heavy smoking 8.7%. Significantly more patients (29/40 , χ² = 16.2 p < 0.05) with initial uncertain/poor prognosis and significantly fewer patients (11/40, χ² = 16.2, p < 0.05) with erratic/no compliance needing re-treatment were identified. 21 of 40 patients (52.5%) (p = 0.655) with family history needing retreatment were identified. Combining the prognostic factors identified 5-22% out of a total of 40% of patients needing re-treatment. six out of nine (67%) teeth with an initial hopeless prognosis were lost, 10/109 (9%) teeth with a poor prognosis were lost, 11/346 (3%) teeth with a moderate prognosis were lost and 9/1972 (0.46%) of teeth with a good prognosis were lost. None of the prognostic factors was found also to be a risk factor for developing periodontal diseases.

CONCLUSION

Applying prognostic factors to identify individual patients with poor long-term outcomes is associated with low accuracy.

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.

Clinical trial designs incorporating predictive biomarkers

Development of oncologic therapies has traditionally been performed in a sequence of clinical trials intended to assess safety (phase I), preliminary efficacy (phase II), and improvement over the standard of care (phase III) in homogeneous (in terms of tumor type and disease stage) patient populations. As cancer has become increasingly understood on the molecular level, newer "targeted" drugs that inhibit specific cancer cell growth and survival mechanisms have increased the need for new clinical trial designs, wherein pertinent questions on the relationship between patient biomarkers and response to treatment can be answered. Herein, we review the clinical trial design literature from initial to more recently proposed designs for targeted agents or those treatments hypothesized to have enhanced effectiveness within patient subgroups (e.g., those with a certain biomarker value or who harbor a certain genetic tumor mutation). We also describe a number of real clinical trials where biomarker-based designs have been utilized, including a discussion of their respective advantages and challenges. As cancers become further categorized and/or reclassified according to individual patient and tumor features, we anticipate a continued need for novel trial designs to keep pace with the changing frontier of clinical cancer research.

Designing a study to evaluate the benefit of a biomarker for selecting patient treatment

Biomarkers that predict the efficacy of treatment can potentially improve clinical outcomes and decrease medical costs by allowing treatment to be provided only to those most likely to benefit. We consider the design of a randomized clinical trial in which one objective is to evaluate a treatment selection marker. The marker may be measured prospectively or retrospectively using samples collected at baseline. We describe and contrast criteria around which the trial can be designed. An existing approach focuses on determining if there is a statistical interaction between the marker and treatment. We propose three alternative approaches based on estimating clinically relevant measures of improvement in outcomes with use of the marker. Importantly, our approaches accommodate the common scenario in which the marker-based rule for recommending treatment is developed with data from the trial. Sample sizes are calculated for powering a trial to assess these criteria in the context of adjuvant chemotherapy for the treatment of estrogen-receptor-positive, node-positive breast cancer. In this example, we find that larger sample sizes are generally required for assessing clinical impact than for simply evaluating if there is a statistical interaction between marker and treatment. We also find that retrospectively selecting a case-control subset of subjects for marker evaluation can lead to large efficiency gains, especially if cases and controls are matched on treatment assignment.

Clinical Neuropathology practice guide 5-2015: MGMT methylation pyrosequencing in glioblastoma: unresolved issues and open questions

O6-methylguanine-methyltransferase (MGMT) promoter methylation status has prognostic and, in the subpopulation of elderly patients, predictive value in newly diagnosed glioblastoma. Therefore, knowledge of the MGMT promoter methylation status is important for clinical decision-making. So far, MGMT testing has been limited by the lack of a robust test with sufficiently high analytical performance. Recently, one of several available pyrosequencing protocols has been shown to be an accurate and robust method for MGMT testing in an intra- and interlaboratory ring trial. However, some uncertainties remain with regard to methodological issues, cut-off definitions, and optimal use in the clinical setting. In this article, we highlight and discuss several of these open questions. The main unresolved issues are the definition of the most relevant CpG sites to analyze for clinical purposes and the determination of a cut-off value for dichotomization of quantitative MGMT pyrosequencing results into "MGMT methylated" and "MGMT unmethylated" patient subgroups as a basis for further treatment decisions.

Assessing the TP53 marker type in patients treated with or without neoadjuvant chemotherapy for resectable colorectal liver metastases: a p53 Research Group study

The type of a biomarker - whether it is prognostic or predictive - is frequently not known, although such information is crucial for assessing the clinical value of a marker. In order to evaluate the type of marker TP53 is, we identified a cohort of 76 patients with colorectal liver metastases (CLM), homogeneously staged as resectable, who had been treated either with or without fluorouracil-based neoadjuvant chemotherapy. The TP53 genotype was assessed retrospectively from paraffin-embedded, diagnostic tumour biopsies using a standardised, p53 gene-specific sequencing protocol (mark53(®) kit). The overall median survival was 44.2 months, and the overall TP53 mutation frequency was 55%. A significant interaction was observed between chemotherapy and TP53 status (P = 0.045). To illustrate this effect, the 51 patients with and the 25 patients without neoadjuvant chemotherapy were described separately. In patients with neoadjuvant chemotherapy, mutated TP53 was significantly associated with poor survival (P = 0.0025), resulting in five-year survival rates of 22%, compared to 60% in patients with normal TP53. The hazard ratio was 3.12 (95% confidence intervals (CI): 1.46-6.95) to the disadvantage of TP53-mutated patients and 5.49 (P = 0.0001; 95% CI: 2.28-13.24) after adjustment for known prognostic factors. In patients treated with surgery alone, a mutated TP53 did not have a negative effect on survival (P = 0.54). A mutated TP53 status independently predicted survival disadvantage in CLM patients in the presence, but not in the absence, of neoadjuvant chemotherapy. Our data suggest that TP53 might be a pure predictive marker.

Translational research: precision medicine, personalized medicine, targeted therapies: marketing or science?

Personalized medicine is based on: 1) improved clinical or non-clinical methods (including biomarkers) for a more discriminating and precise diagnosis of diseases; 2) targeted therapies of the choice or the best drug for each patient among those available; 3) dose adjustment methods to optimize the benefit-risk ratio of the drugs chosen; 4) biomarkers of efficacy, toxicity, treatment discontinuation, relapse, etc. Unfortunately, it is still too often a theoretical concept because of the lack of convenient diagnostic methods or treatments, particularly of drugs corresponding to each subtype of pathology, hence to each patient. Stratified medicine is a component of personalized medicine employing biomarkers and companion diagnostics to target the patients likely to present the best benefit-risk balance for a given active compound. The concept of targeted therapy, mostly used in cancer treatment, relies on the existence of a defined molecular target, involved or not in the pathological process, and/or on the existence of a biomarker able to identify the target population, which should logically be small as compared to the population presenting the disease considered. Targeted therapies and biomarkers represent important stakes for the pharmaceutical industry, in terms of market access, of return on investment and of image among the prescribers. At the same time, they probably represent only the first generation of products resulting from the combination of clinical, pathophysiological and molecular research, i.e. of translational research.

The direct assignment option as a modular design component: an example for the setting of two predefined subgroups

Background. A phase II design with an option for direct assignment (stop randomization and assign all patients to experimental treatment based on interim analysis, IA) for a predefined subgroup was previously proposed. Here, we illustrate the modularity of the direct assignment option by applying it to the setting of two predefined subgroups and testing for separate subgroup main effects. Methods. We power the 2-subgroup direct assignment option design with 1 IA (DAD-1) to test for separate subgroup main effects, with assessment of power to detect an interaction in a post-hoc test. Simulations assessed the statistical properties of this design compared to the 2-subgroup balanced randomized design with 1 IA, BRD-1. Different response rates for treatment/control in subgroup 1 (0.4/0.2) and in subgroup 2 (0.1/0.2, 0.4/0.2) were considered. Results. The 2-subgroup DAD-1 preserves power and type I error rate compared to the 2-subgroup BRD-1, while exhibiting reasonable power in a post-hoc test for interaction. Conclusion. The direct assignment option is a flexible design component that can be incorporated into broader design frameworks, while maintaining desirable statistical properties, clinical appeal, and logistical simplicity.

Between pathways and networks lies context: implications for precision medicine

Precision medicine, broadly defined as considering individual variability in genes, environment, and lifestyle for each person in disease prevention and selection of suitable medical intervention, shows strong promise in the treatment of cancer Selecting therapies is complicated by multiple routes to gene dysregulation, which manifest in the individual patient within the many different types of genomic measurements. Additionally, multiple mutations exist in patients, aphenomenon known as oncogenic collaboration, which further complicates the selection of therapy. In this article, we discuss current approaches using biological pathways and networks to unify the many types of OMICs data. We argue that a contextual approach combining cancer pathways and networks could lead to a proper understanding of the biology of this significant disease.

Evaluation of a chemoresponse assay as a predictive marker in the treatment of recurrent ovarian cancer: further analysis of a prospective study

BACKGROUND

Recently, a prospective study reported improved clinical outcomes for recurrent ovarian cancer patients treated with chemotherapies indicated to be sensitive by a chemoresponse assay, compared with those patients treated with non-sensitive therapies, thereby demonstrating the assay's prognostic properties. Due to cross-drug response over different treatments and possible association of in vitro chemosensitivity of a tumour with its inherent biology, further analysis is required to ascertain whether the assay performs as a predictive marker as well.

METHODS

Women with persistent or recurrent epithelial ovarian cancer (n=262) were empirically treated with one of 15 therapies, blinded to assay results. Each patient's tumour was assayed for responsiveness to the 15 therapies. The assay's ability to predict progression-free survival (PFS) was assessed by comparing the association when the assayed therapy matches the administered therapy (match) with the association when the assayed therapy is randomly selected, not necessarily matching the administered therapy (mismatch).

RESULTS

Patients treated with assay-sensitive therapies had improved PFS vs patients treated with non-sensitive therapies, with the assay result for match significantly associated with PFS (hazard ratio (HR)=0.67, 95% confidence interval (CI)=0.50-0.91, P=0.009). On the basis of 3000 simulations, the mean HR for mismatch was 0.81 (95% range=0.66-0.99), with 3.4% of HRs less than 0.67, indicating that HR for match is lower than for mismatch. While 47% of tumours were non-sensitive to all assayed therapies and 9% were sensitive to all, 44% displayed heterogeneity in assay results. Improved outcome was associated with the administration of an assay-sensitive therapy, regardless of homogeneous or heterogeneous assay responses across all of the assayed therapies.

CONCLUSIONS

These analyses provide supportive evidence that this chemoresponse assay is a predictive marker, demonstrating its ability to discern specific therapies that are likely to be more effective among multiple alternatives.

Subgroup-Based Adaptive (SUBA) Designs for Multi-Arm Biomarker Trials

Targeted therapies based on biomarker profiling are becoming a mainstream direction of cancer research and treatment. Depending on the expression of specific prognostic biomarkers, targeted therapies assign different cancer drugs to subgroups of patients even if they are diagnosed with the same type of cancer by traditional means, such as tumor location. For example, Herceptin is only indicated for the subgroup of patients with HER2+ breast cancer, but not other types of breast cancer. However, subgroups like HER2+ breast cancer with effective targeted therapies are rare and most cancer drugs are still being applied to large patient populations that include many patients who might not respond or benefit. Also, the response to targeted agents in humans is usually unpredictable. To address these issues, we propose SUBA, subgroup-based adaptive designs that simultaneously search for prognostic subgroups and allocate patients adaptively to the best subgroup-specific treatments throughout the course of the trial. The main features of SUBA include the continuous reclassification of patient subgroups based on a random partition model and the adaptive allocation of patients to the best treatment arm based on posterior predictive probabilities. We compare the SUBA design with three alternative designs including equal randomization, outcome-adaptive randomization and a design based on a probit regression. In simulation studies we find that SUBA compares favorably against the alternatives.

Prognostic factors in chronic lymphocytic leukemia: a conceptual approach

SUMMARY 

Prognostic assessment is an essential component in the management of patients with chronic lymphocytic leukemia. Prognostic factors (e.g., clinical stage, lymphocyte doubling time and IGHV mutational status/ZAP70 expression), allow predicting time to disease progression and need of therapy and also provide a rough estimate of the overall survival. The most important predictor of survival in patients requiring intervention is response to therapy and its degree, patients with undetectable minimal residual disease following therapy having a much better outcome than those with an inferior response. Given the increasing number of treatment modalities for CLL, the identification of predictive factors is important. Unfortunately, with the only exception of del(17p)/TP53 mutations that predict a poor response to purine analogs-based therapy and del(11q) that correlates with inferior response to fludarabine given as single agent, there are not reliable predictors of response. Well-conducted studies aimed at identifying prognostic and, particularly, predictive factors are needed. New prognostic and predictive parameters should demonstrate superiority over already well validated markers and be helpful in the management of patients with CLL.

Stratified medicine in European Medicines Agency licensing: a systematic review of predictive biomarkers

OBJECTIVES

Stratified medicine is often heralded as the future of clinical practice. Key part of stratified medicine is the use of predictive biomarkers, which identify patient subgroups most likely to benefit (or least likely to experience harm) from an intervention. We investigated how many and what predictive biomarkers are currently included in European Medicines Agency (EMA) licensing.

SETTING

EMA licensing.

PARTICIPANTS

Indications and contraindications of all drugs considered by the EMA and published in 883 European Public Assessment Reports and Pending Decisions.

PRIMARY AND SECONDARY OUTCOME MEASURES

Data were collected on: the type of the biomarker, whether it selected a subgroup of patients based on efficacy or toxicity, therapeutic area, marketing status, date of licensing decision, date of inclusion of the biomarker in the indication or contraindication and on orphan designation.

RESULTS

49 biomarker-indication-drug (B-I-D) combinations were identified over 16 years, which included 37 biomarkers and 41 different drugs. All identified biomarkers were molecular. Six drugs (relating to 10 B-I-D combinations) had an orphan designation at the time of licensing. The identified B-I-D combinations were mainly used in cancer and HIV treatment, and also in hepatitis C and three other indications (cystic fibrosis, hyperlipoproteinaemia type I and methemoglobinaemia). In 45 B-I-D combinations, biomarkers were used as predictive of drug efficacy and in four of drug toxicity. It appeared that there was an increase in the number of B-I-D combinations introduced each year; however, the numbers were too small to identify any trends.

CONCLUSIONS

Given the large body of literature documenting research into potential predictive biomarkers and extensive investment into stratified medicine, we identified relatively few predictive biomarkers included in licensing. These were also limited to a small number of clinical areas. This might suggest a need for improvement in methods of translation from laboratory findings to clinical practice.

Drug-diagnostics co-development in oncology

Developments in genomics are providing a biological basis for the heterogeneity of clinical course and response to treatment that have long been apparent to clinicians. The ability to molecularly characterize human diseases presents new opportunities to develop more effective treatments and new challenges for the design and analysis of clinical trials. In oncology, treatment of broad populations with regimens that benefit a minority of patients is less economically sustainable with expensive molecularly targeted therapeutics. The established molecular heterogeneity of human diseases requires the development of new paradigms for the design and analysis of randomized clinical trials as a reliable basis for predictive medicine. We review prospective designs for the development of new therapeutics and predictive biomarkers to inform their use. We cover designs for a wide range of settings. At one extreme is the development of a new drug with a single candidate biomarker and strong biological evidence that marker negative patients are unlikely to benefit from the new drug. At the other extreme are phase III clinical trials involving both genome-wide discovery of a predictive classifier and internal validation of that classifier. We have outlined a prediction based approach to the analysis of randomized clinical trials that both preserves the type I error and provides a reliable internally validated basis for predicting which patients are most likely or unlikely to benefit from a new regimen.

Evaluating many treatments and biomarkers in oncology: a new design

There is a pressing need for more-efficient trial designs for biomarker-stratified clinical trials. We suggest a new approach to trial design that links novel treatment evaluation with the concurrent evaluation of a biomarker within a confirmatory phase II/III trial setting. We describe a new protocol using this approach in advanced colorectal cancer called FOCUS4. The protocol will ultimately answer three research questions for a number of treatments and biomarkers: (1) After a period of first-line chemotherapy, do targeted novel therapies provide signals of activity in different biomarker-defined populations? (2) If so, do these definitively improve outcomes? (3) Is evidence of activity restricted to the biomarker-defined groups? The protocol randomizes novel agents against placebo concurrently across a number of different biomarker-defined population-enriched cohorts: BRAF mutation; activated AKT pathway: PI3K mutation/absolute PTEN loss tumors; KRAS and NRAS mutations; and wild type at all the mentioned genes. Within each biomarker-defined population, the trial uses a multistaged approach with flexibility to adapt in response to planned interim analyses for lack of activity. FOCUS4 is the first test of a protocol that assigns all patients with metastatic colorectal cancer to one of a number of parallel population-enriched, biomarker-stratified randomized trials. Using this approach allows questions regarding efficacy and safety of multiple novel therapies to be answered in a relatively quick and efficient manner, while also allowing for the assessment of biomarkers to help target treatment.

Generation and validation of a normative, age-specific reference curve for lumbar spine trabecular bone score (TBS) in French women

Age-related changes in lumbar vertebral microarchitecture are evaluated, as assessed by trabecular bone score (TBS), in a cohort of 5,942 French women. The magnitude of TBS decline between 45 and 85 years of age is piecewise linear in the spine and averaged 14.5%. TBS decline rate increases after 65 years by 50%.

INTRODUCTION

This study aimed to evaluate age-related changes in lumbar vertebral microarchitecture, as assessed by TBS, in a cohort of French women aged 45-85 years.

METHODS

An all-comers cohort of French Caucasian women was selected from two clinical centers. Data obtained from these centers were cross-calibrated for TBS and bone mineral density (BMD). BMD and TBS were evaluated at L1-L4 and for all lumbar vertebrae combined using GE-Lunar Prodigy densitometer images. Weight, height, and body mass index (BMI) also were determined. To validate our all-comers cohort, the BMD normative data of our cohort and French Prodigy data were compared.

RESULTS

A cohort of 5,942 French women aged 45 to 85 years was created. Dual-energy X-ray absorptiometry normative data obtained for BMD from this cohort were not significantly different from French prodigy normative data (p = 0.15). TBS values at L1-L4 were poorly correlated with BMI (r = -0.17) and weight (r = -0.14) and not correlated with height. TBS values obtained for all lumbar vertebra combined (L1, L2, L3, L4) decreased with age. The magnitude of TBS decline at L1-L4 between 45 and 85 years of age was piecewise linear in the spine and averaged 14.5%, but this rate increased after 65 years by 50%. Similar results were obtained for other region of interest in the lumbar spine. As opposed to BMD, TBS was not affected by spinal osteoarthrosis.

CONCLUSION

The age-specific reference curve for TBS generated here could therefore be used to help clinicians to improve osteoporosis patient management and to monitor microarchitectural changes related to treatment or other diseases in routine clinical practice.

Multicenter trials using ¹⁸F-fluorodeoxyglucose (FDG) PET to predict chemotherapy response: effects of differential measurement error and bias on power calculations for unselected and enrichment designs

BACKGROUND

Clinical validation of a predictive biomarker is especially difficult when the biomarker cannot be assessed retrospectively. A cost-effective, prospective multicenter replication study with rapid accrual is warranted prior to further validation studies such as a marker-based strategy for treatment selection. However, it is often unknown how measurement error and bias in a multicenter trial will differ from that in single-institution studies.

PURPOSE

Power calculations using simulated data may inform the efficient design of a multicenter study to replicate single-institution findings. This case study used serial standardized uptake value (SUV) measures from (18)F-fluorodeoxyglucose (FDG) positron emission tomography (PET) to predict early response to breast cancer neoadjuvant chemotherapy. We examined the impact of accelerating accrual through increased inclusion of secondary sites with greater levels of measurement error and bias. We also examined whether enrichment designs based on breast cancer initial uptake could increase the study power for a fixed budget (200 total scans).

METHODS

Reference FDG PET SUV data were selected with replacement from a single-institution trial; pathologic complete response (pCR) data were simulated using a logistic regression model predicting response by mid-therapy percent change in SUV. The impact of increased error for SUV measurements in multicenter trials was simulated by sampling from error and bias distributions: 20%-40% measurement error, 0%-40% bias, and fixed error/bias values. The proportion of patients recruited from secondary sites (with higher additional error/bias compared to primary sites) varied from 25% to 75%.

RESULTS

Reference power (from source data with no added error) was 0.92 for N = 100 to detect an association between percentage change in SUV and response. With moderate (20%) simulated measurement error for 3/4, 1/2, and 1/4 of measurements and 40% for the remainder, power was 0.70, 0.61, and 0.53, respectively. Reduction of study power was similar for other manifestations of measurement error (bias as a percentage of true value, absolute error, and absolute bias). Enrichment designs, which recruit additional patients by not conducting a second scan in patients with unsuitable pre-therapy uptake (low baseline SUV), did not lead to greater power for studies constrained to the same total cost.

LIMITATIONS

Simulation parameters could be incorrect, or not generalizable. Under a different logistic regression model relating mid-therapy percent change in SUV to pCR (with no relationship for patients with low baseline SUV, rather than the modest point estimate from reference data), the enrichment design did have somewhat greater power than the unselected design.

CONCLUSION

Even moderate additional measurement error substantially reduced study power under both unselected and enrichment designs.

Development and validation of risk models and molecular diagnostics to permit personalized management of cancer

Despite the advances made in cancer management over the past few decades, improvements in cancer diagnosis and prognosis are still poor, highlighting the need for individualized strategies. Toward this goal, risk prediction models and molecular diagnostic tools have been developed, tailoring each step of risk assessment from diagnosis to treatment and clinical outcomes based on the individual's clinical, epidemiological, and molecular profiles. These approaches hold increasing promise for delivering a new paradigm to maximize the efficiency of cancer surveillance and efficacy of treatment. However, they require stringent study design, methodology development, comprehensive assessment of biomarkers and risk factors, and extensive validation to ensure their overall usefulness for clinical translation. In the current study, the authors conducted a systematic review using breast cancer as an example and provide general guidelines for risk prediction models and molecular diagnostic tools, including development, assessment, and validation.

Group sequential designs for developing and testing biomarker-guided personalized therapies in comparative effectiveness research

Biomarker-guided personalized therapies offer great promise to improve drug development and improve patient care, but also pose difficult challenges in designing clinical trials for the development and validation of these therapies. We first give a review of the existing approaches, briefly for clinical trials in new drug development and in more detail for comparative effectiveness trials involving approved treatments. We then introduce new group sequential designs to develop and test personalized treatment strategies involving approved treatments.