An overview of statistical methods for biomarkers relevant to early clinical development of cancer immunotherapies.

Over the last decade, a new paradigm for cancer therapies has emerged which leverages the immune system to act against the tumor. The novel mechanism of action of these immunotherapies has also introduced new challenges to drug development. Biomarkers play a key role in several areas of early clinical development of immunotherapies including the demonstration of mechanism of action, dose finding and dose optimization, mitigation and prevention of adverse reactions, and patient enrichment and indication prioritization. We discuss statistical principles and methods for establishing the prognostic, predictive aspect of a (set of) biomarker and for linking the change in biomarkers to clinical efficacy in the context of early development studies. The methods discussed are meant to avoid bias and produce robust and reproducible conclusions. This review is targeted to drug developers and data scientists interested in the strategic usage and analysis of biomarkers in the context of immunotherapies.

Simulation-based evaluation of the Pharmpy Automatic Model Development tool for population pharmacokinetic modeling in early clinical drug development.

The Pharmpy Automatic Model Development (AMD) tool automates the building of population pharmacokinetic (popPK) models by utilizing a systematic stepwise process. In this study, the performance of the AMD tool was assessed using simulated datasets. Ten true models mimicking classical popPK models were created. From each true model, dataset replicates were simulated assuming a typical phase I study design-single and multiple ascending doses with/without dichotomous food effect, with rich PK sampling. For every dataset replicate, the AMD tool automatically built an AMD model utilizing NONMEM for parameter estimation. The AMD models were compared to the true and reference models (true model fitted to simulated datasets) based on their model components, predicted population and individual secondary PK parameters (SP) (AUC0-24, cmax, ctrough), and model quality metrics (e.g., model convergence, parameter relative standard errors (RSEs), Bayesian Information Criterion (BIC)). The models selected by the AMD tool closely resembled the true models, particularly in terms of distribution and elimination, although differences were observed in absorption and inter-individual variability components. Bias associated with the derived SP was low. In general, discrepancies between AMD and true SP were also observed for reference models and therefore were attributed to the inherent stochasticity in simulations. In summary, the AMD tool was found to be a valuable asset in automating repetitive modeling tasks, yielding reliable PK models in the scenarios assessed. This tool has the potential to save time during early clinical drug development that can be invested in more complex modeling activities within model-informed drug development.

PBPK Modeling of Entrectinib and Its Active Metabolite to Derive Dose Adjustments in Pediatric Populations Co-Administered with CYP3A4 Inhibitors.

Physiologically based pharmacokinetic (PBPK) models of entrectinib and its equipotent metabolite, M5, were established in healthy adult subjects and extrapolated to pediatric patients to predict increases in steady-state systemic exposure on co-administration of strong and moderate CYP3A4 inhibitors (itraconazole at 5 mg/kg, erythromycin at 7.5-12.5 mg/kg and fluconazole at 3-12 mg/kg, respectively). Adult model establishment involved the optimization of fraction metabolized by CYP3A4 (0.92 for entrectinib and 0.98 for M5) using data from an itraconazole DDI study. This model captured well the exposure changes of entrectinib and M5 seen in adults co-administered with the strong CYP3A4 inducer rifampicin. In pediatrics, reasonable prediction of entrectinib and M5 pharmacokinetics in ≧2 year olds was achieved when using the default models for physiological development and enzyme ontogenies. However, a two to threefold misprediction of entrectinib and M5 exposures was seen in <2 year olds which may be due to missing mechanistic understanding of gut physiology and/or protein binding in very young children. Model predictions for ≧2 year olds showed that entrectinib AUC(0-t) was increased by approximately sevenfold and five to threefold by strong and high-moderate and low-moderate CYP3A4 inhibitors, respectively. Based on these victim DDI predictions, dose adjustments for entrectinib when given concomitantly with strong and moderate CYP3A4 inhibitors in pediatric subjects were recommended. These simulations informed the approved entrectinib label without the need for additional clinical pharmacology studies.

Response-Based Dosing for Ponatinib: Model-Based Analyses of the Dose-Ranging OPTIC Study.

Optimizing Ponatinib Treatment in CP-CML (OPTIC) was a randomized, phase II dose-optimization trial of ponatinib in chronic phase-chronic myeloid leukemia (CP-CML) resistant to ≥ 2 tyrosine kinase inhibitors or with T315I mutation. Patients were randomized to starting doses of 45-, 30-, or 15-mg ponatinib once daily. Patients receiving 45- or 30-mg reduced to 15-mg upon achievement of ≤ 1% BCR::ABL1IS (≥ molecular response with 2-log reduction (MR2)). The exposure-molecular response relationship was described using a four-state, discrete-time Markov model. Time-to-event models were used to characterize the relationship between exposure and arterial occlusive events (AOEs), grade ≥ 3 neutropenia, and thrombocytopenia. Increasing systemic exposures were associated with increasing probability of transitioning from no response to ≥ MR1, and from MR1 to ≥ MR1, with odds ratios of 1.63 (95% confidence interval (CI), 1.06-2.73) and 2.05 (95% CI, 1.53-2.89) for a 15-mg dose increase, respectively. Ponatinib exposure was a significant predictor of AOEs (hazard ratio (HR) 2.05, 95% CI, 1.43-2.93, for a 15-mg dose increase). In the exposure-safety models for neutropenia and thrombocytopenia, exposure was a significant predictor of grade ≥ 3 thrombocytopenia (HR 1.31, 95% CI, 1.05-1.64, for a 15-mg dose increase). Model-based simulations predicted a clinically meaningful higher rate of ≥ MR2 response at 12 months for the 45-mg starting dose (40.4%) vs. 30-mg (34%) and 15-mg (25.2%). The exposure-response analyses supported a ponatinib starting dose of 45 mg with reduction to 15 mg at response for patients with CP-CML.

Population PK and Semimechanistic PK/PD Modeling and Simulation of Relugolix Effects on Testosterone Suppression in Men with Prostate Cancer.

Relugolix, the first orally active, nonpeptide gonadotropin-releasing hormone receptor antagonist, is approved in the United States and the European Union for the treatment of adult patients with advanced prostate cancer. The recommended dosing regimen is a 360-mg loading dose followed by a 120-mg daily dose. Relugolix and testosterone concentration data and clinical information from two phase I studies, two phase II studies, and the phase III safety and efficacy study (HERO) were used to develop a population pharmacokinetic (PopPK) model and a semimechanistic population pharmacokinetic/pharmacodynamic (PopPK/PD) model that characterized relugolix exposure and its relationship to testosterone concentrations. Age, body weight, and Black/African American race had at most minimal effects on relugolix exposure or testosterone concentrations with no clinical relevance. Simulations using the PopPK/PD model confirmed the recommended dosing regimen of relugolix, with the median simulated testosterone concentrations predicted to achieve castration levels (< 50 ng/dL) and profound castration levels (< 20 ng/dL) by day 2 and day 9, respectively, and demonstrated that 97.3% and 85.5% of the patients remained at castration levels (< 50 ng/dL) upon temporary interruption of treatment for 7 days and 14 days, respectively. Collectively, simulations based on the PopPK and PopPK/PD models were consistent with actual data from clinical studies, reflecting the high predictiveness of the models and supporting the reliability of model-based simulations. These models can be used to provide guidance regarding dosing recommendations under various circumstances (e.g., temporary interruption of treatment, if needed) for relugolix.

Forward and reverse translational approaches to predict efficacy of neutralizing respiratory syncytial virus (RSV) antibody prophylaxis.

BACKGROUND: Neutralizing mAbs can prevent communicable viral diseases. MK-1654 is a respiratory syncytial virus (RSV) F glycoprotein neutralizing monoclonal antibody (mAb) under development to prevent RSV infection in infants. Development and validation of methods to predict efficacious doses of neutralizing antibodies across patient populations exposed to a time-varying force of infection (i.e., seasonal variation) are necessary. METHODS: Five decades of clinical trial literature were leveraged to build a model-based meta-analysis (MBMA) describing the relationship between RSV serum neutralizing activity (SNA) and clinical endpoints. The MBMA was validated by backward translation to animal challenge experiments and forward translation to predict results of a recent RSV mAb trial. MBMA predictions were evaluated against a human trial of 70 participants who received either placebo or one of four dose-levels of MK-1654 and were challenged with RSV [NCT04086472]. The MBMA was used to perform clinical trial simulations and predict efficacy of MK-1654 in the infant target population. FINDINGS: The MBMA established a quantitative relationship between RSV SNA and clinical endpoints. This relationship was quantitatively consistent with animal model challenge experiments and results of a recently published clinical trial. Additionally, SNA elicited by increasing doses of MK-1654 in humans reduced RSV symptomatic infection rates with a quantitative relationship that approximated the MBMA. The MBMA indicated a high probability that a single dose of ≥ 75 mg of MK-1654 will result in prophylactic efficacy (> 75% for 5 months) in infants. INTERPRETATION: An MBMA approach can predict efficacy of neutralizing antibodies against RSV and potentially other respiratory pathogens.

A Pharmacometric Analysis of Patient-Reported Outcomes in Breast Cancer Patients Through Item Response Theory.

PURPOSE: An item response theory (IRT) pharmacometric framework is presented to characterize Functional Assessment of Cancer Therapy-Breast (FACT-B) data in locally-advanced or metastatic breast cancer patients treated with ado-trastuzumab emtansine (T-DM1) or capecitabine-plus-lapatinib. METHODS: In the IRT model, four latent well-being variables, based on FACT-B general subscales, were used to describe the physical, social/family, emotional and functional well-being. Each breast cancer subscale item was reassigned to one of the other subscales. Longitudinal changes in FACT-B responses and covariate effects were investigated. RESULTS: The IRT model could describe both item-level and subscale-level FACT-B data. Non-Asian patients showed better baseline social/family and functional well-being than Asian patients. Moreover, patients with Eastern Cooperative Oncology Group performance status of 0 had better baseline physical and functional well-being. Well-being was described as initially increasing or decreasing before reaching a steady-state, which varied substantially between patients and subscales. T-DM1 exposure was not related to any of the latent variables. Physical well-being worsening was identified in capecitabine-plus-lapatinib-treated patients, whereas T-DM1-treated patients typically stayed stable. CONCLUSION: The developed framework provides a thorough description of FACT-B longitudinal data. It acknowledges the multi-dimensional nature of the questionnaire and allows covariate and exposure effects to be evaluated on responses.

A Minimal Continuous-Time Markov Pharmacometric Model.

In this work, an alternative model to discrete-time Markov model (DTMM) or standard continuous-time Markov model (CTMM) for analyzing ordered categorical data with Markov properties is presented: the minimal CTMM (mCTMM). Through a CTMM reparameterization and under the assumption that the transition rate between two consecutive states is independent on the state, the Markov property is expressed through a single parameter, the mean equilibration time, and the steady-state probabilities are described by a proportional odds (PO) model. The mCTMM performance was evaluated and compared to the PO model (ignoring Markov features) and to published Markov models using three real data examples: the four-state fatigue and hand-foot syndrome data in cancer patients initially described by DTMM and the 11-state Likert pain score data in diabetic patients previously analyzed with a count model including Markovian transition probability inflation. The mCTMM better described the data than the PO model, and adequately predicted the average number of transitions per patient and the maximum achieved scores in all examples. As expected, mCTMM could not describe the data as well as more flexible DTMM but required fewer estimated parameters. The mCTMM better fitted Likert data than the count model. The mCTMM enables to explore the effect of potential predictive factors such as drug exposure and covariates, on ordered categorical data, while accounting for Markov features, in cases where DTMM and/or standard CTMM is not applicable or conveniently implemented, e.g., non-uniform time intervals between observations or large number of categories.

A pharmacokinetic binding model for bevacizumab and VEGF165 in colorectal cancer patients.

PURPOSE: To characterize the population pharmacokinetics of bevacizumab, its binding properties to VEGF165 and the effect of demographic data and VEGF-A polymorphisms on the interplay between bevacizumab serum pharmacokinetics and VEGF165 serum concentrations in patients with colorectal cancer stage IV. METHODS: Bevacizumab and VEGF165 data were collected from 19 adult patients with metastatic colorectal cancer enrolled in an observational clinical study. Bevacizumab was administered with one of the following combinations: 5-FU/Leucovorin/Irinotecan, 5-FU/Leucovorin/Oxaliplatin, Capecitabine/Irinotecan at doses ranging from 5 to 10 mg/kg every 2 or 3 weeks. Data analysis was performed using nonlinear mixed-effects modeling implemented in NONMEM 7.3. RESULTS: A target-mediated drug disposition model adequately described bevacizumab concentration changes over time and its binding characteristics to VEGF165. The estimated clearance of bevacizumab was 0.18 L/day, the free VEGF165 levels at baseline were 212 ng/L, and the elimination rate constant of free VEGF165 was 0.401 day(-1). Body weight was allometrically included in all PK parameters. CONCLUSION: The final model adequately described the pre- and post-dose concentrations of total bevacizumab and free VEGF165 in patients with colorectal cancer. Model parameters were consistent with those previously reported for patients with solid tumors. Correlations between the binding affinity of bevacizumab and the VEGF-2578C/A and VEGF-634G/C polymorphisms were noticed.

Population pharmacokinetic-pharmacodynamic modelling in oncology: a tool for predicting clinical response.

In oncology trials, overall survival (OS) is considered the most reliable and preferred endpoint to evaluate the benefit of drug treatment. Other relevant variables are also collected from patients for a given drug and its indication, and it is important to characterize the dynamic effects and links between these variables in order to improve the speed and efficiency of clinical oncology drug development. However, the drug-induced effects and causal relationships are often difficult to interpret because of temporal differences. To address this, population pharmacokinetic-pharmacodynamic (PKPD) modelling and parametric time-to-event (TTE) models are becoming more frequently applied. Population PKPD and TTE models allow for exploration towards describing the data, understanding the disease and drug action over time, investigating relevance of biomarkers, quantifying patient variability and in designing successful trials. In addition, development of models characterizing both desired and adverse effects in a modelling framework support exploration of risk-benefit of different dosing schedules. In this review, we have summarized population PKPD modelling analyses describing tumour, tumour marker and biomarker responses, as well as adverse effects, from anticancer drug treatment data. Various model-based metrics used to drive PD response and predict OS for oncology drugs and their indications are also discussed.