A model-based approach for historical borrowing, with an application to neovascular age-related macular degeneration.

Bayesian historical borrowing has recently attracted growing interest due to the increasing availability of historical control data, as well as improved computational methodology and software. In this article, we argue that the statistical models used for borrowing may be suboptimal when they do not adjust for differing factors across historical studies such as covariates, dosing regimen, etc. We propose an alternative approach to address these shortcomings. We start by constructing a historical model based on subject-level historical data to accurately characterize the control treatment by adjusting for known between trials differences. This model is subsequently used to predict the control arm response in the current trial, enabling the derivation of a model-informed prior for the treatment effect parameter of another (potentially simpler) model used to analyze the trial efficacy (i.e. the trial model). Our approach is applied to neovascular age-related macular degeneration trials, employing a cross-sectional regression trial model, and a longitudinal non-linear mixed-effects drug-disease-trial historical model. The latter model characterizes the relationship between clinical response, drug exposure and baseline covariates so that the derived model-informed prior seamlessly adapts to the trial population and can be extrapolated to a different dosing regimen. This approach can yield a more accurate prior for borrowing, thus optimizing gains in efficiency (e.g. increasing power or reducing the sample size) in future trials.

A Baseline Score to Predict Response to Ranibizumab Treatment in Neovascular Age-Related Macular Degeneration.

Purpose: What are the patient characteristics predictive of response to ranibizumab treatment? Methods: Model-based characterization of best-corrected visual acuity (BCVA) time profiles of patients with neovascular age-related macular degeneration under ranibizumab or sham treatment based on 24-month observations of BCVA in 2419 patients from randomized multicenter phase 3 trials of ranibizumab: ANCHOR, MARINA, PIER, and HARBOR. Goodness-of-fit plots and precision of parameter estimates were used for measure of accuracy. Results: The model incorporates a long-term effect on disease progression and an additive and more potent short-term effect of ranibizumab. Response to ranibizumab treatment and progression of the disease were found to be a function of seven baseline characteristics (visual acuity, age, leakage size, central retinal lesion thickness, presence or absence of cyst, type of choroidal neovascularization (CNV), and size of pigment epithelium detachment). A composite score of these seven baseline characteristics was derived and used to categorize response to ranibizumab treatment. The ranibizumab treatment arms of two proof-of-concept studies held out from the model development were used to validate the methodology. Conclusions: A composite score based on seven patient characteristics prior to treatment could be used to discriminate patients with predicted insufficient response to anti-vascular endothelial growth factor treatment. Translational Relevance: The method could be used to create a virtual ranibizumab treatment arm in clinical trials or to reduce the size of a ranibizumab active control arm.

Population pharmacokinetic analysis of RO5459072, a low water-soluble drug exhibiting complex food-drug interactions.

AIMS: RO5459072, a cathepsin-S inhibitor, Biopharmaceutics Classification System class 2 and P-glycoprotein substrate, exhibited complex, nonlinear pharmacokinetics (PK) while fasted that seemed to impact both the absorption and the disposition phases. When given with food, all nonlinearities disappeared. Physiologically based PK (PBPK) modelling attributed those nonlinearities to dose-dependent solubilisation and colonic absorption. The objective of this population PK analysis was to complement the PBPK analysis. METHODS: PK profiles in 39 healthy volunteers after first oral dosing (1-600 mg) while fasted or fed in single and multiple ascending dose studies were analysed using population compartmental modelling. RESULTS: The PK of RO5459072 while fed was characterized by a 1-compartmental PK model with linear absorption and elimination. The nonlinearities while fasted were captured using dose dependent bioavailability and 2 sequential first-order absorption phases: one following drug administration and one occurring 11 hours later and only for doses >10 mg. The bioavailability in the first absorption phase increased between 1 and 10 mg and then decreased with dose, in agreement with in vitro dissolution and solubility studies. The remaining fraction of doses to be absorbed by the second absorption phase was found to have a bioavailability similar to that in the first absorption phase. CONCLUSION: The population PK model supported that dissolution- and solubility-limited absorption from the proximal and distal intestine alone explains the nonlinear PK of RO5459072 in fasted state and the linear PK in fed state. This work, together with the PBPK analysis, raised our confidence in the understanding of this complex PK.

PKPD and cardiac single cell modeling of a DDI study with a CYP3A4 substrate and itraconazole to quantify the effects on QT interval duration.

Plasma drug concentration and electrocardiogram (ECG) data from a drug-drug interaction (DDI) study employing the metabolic inhibitor itraconazole have been used as part of a prospectively defined pharmacokinetic/pharmacodynamic modelling strategy to quantify the potential for QT interval prolongation from basmisanil, an investigational compound. ECG data were collected on multiple days during repeat dosing treatment regimens, thereby allowing the capture of QT data across a wide range of drug concentrations in each study participant and encompassing both "therapeutic" and "supra-therapeutic" exposures. The data were used to develop a non-linear mixed effect concentration-QT (C-QT) model that differentiated drug-induced QT prolongation from other factors altering QT interval duration. Food effects were accounted by quantitating their influences on the parameters describing the diurnal variation of QT. The final model demonstrated that itraconazole does not cause QT prolongation, while for basmisanil, the 1-sided upper 95% CI of the QT interval at 240 mg (the highest dose tested in ongoing phase 2 studies) with DDI, was below the 10 ms threshold considered to be of clinical significance by regulatory authorities. The empirical modelling was complemented with a human mechanistic cardiac single cell model that was used to simulate the change in action potential duration as a function of drug concentration. The results of the two approaches were in agreement, suggesting that the effect of basmisanil on QT interval duration can be attributed to the effect on hERG alone. The C-QT model for basmisanil can be used to derive the QT interval corrected changes in heart rate (QTc) and thus inform cardiac safety strategy in later development without the need for a separate, dedicated study.

Ocular Half-Life of Intravitreal Biologics in Humans and Other Species: Meta-Analysis and Model-Based Prediction.

Therapeutic antibodies administered intravitreally are the current standard of care to treat retinal diseases. The ocular half-life (t1/2) is a key determinant of the duration of target suppression. To support the development of novel, longer-acting drugs, a reliable determination of t1/2 is needed together with an improved understanding of the factors that influence it. A model-based meta-analysis was conducted in humans and nonclinical species (rat, rabbit, monkey, and pig) to determine consensus values for the ocular t1/2 of IgG antibodies and Fab fragments. Results from multiple literature and in-house pharmacokinetic studies are presented within a mechanistic framework that assumes diffusion-controlled drug elimination from the vitreous. Our analysis shows, both theoretically and experimentally, that the ocular t1/2 increases in direct proportion to the product of the hydrodynamic radius of the macromolecule (3.0 nm for Fab and 5.0 nm for IgG) and the square of the radius of the vitreous globe, which varies approximately 24-fold from the rat to the human. Interspecies differences in the proportionality factors are observed and discussed in mechanistic terms. In addition, mathematical formulae are presented that allow prediction of the ocular t1/2 for molecules of interest. The utility of these formulae is successfully demonstrated in case studies of aflibercept, brolucizumab, and PEGylated Fabs, where the predicted ocular t1/2 values are found to be in reasonable agreement with the experimental data available for these molecules.

Time-to-event modelling of effect of codrituzumab on overall survival in patients with hepatocellular carcinoma.

AIMS: Codrituzumab (GC33) is a recombinant, humanized mAb that binds to glypican-3 (GPC3), an oncofetal protein highly expressed in hepatocellular carcinoma (HCC). This investigation aimed to identify clinically relevant factors that may affect the overall survival (OS) in HCC patients treated with codrituzumab and to quantitatively annotate their effects. METHODS: Codrituzumab exposure was estimated by a population pharmacokinetics model with a nonlinear elimination pathway. Analysis of OS was performed using a time-to-event model in 181 patients with advanced HCC. The model was tested with the addition of various covariates, including levels of immune biomarkers, such as CD16 (measured in terms of molecules of equivalent soluble fluorophore; CD16MESF ) and CD4, codrituzumab exposure and potential prognostic biomarkers of HCC such as baseline tumour size and soluble GPC3. RESULTS: The time-to-event model estimated a prolonged OS (>3 months) in patients with codrituzumab exposure of ≥230 µg ml-1 and high CD16MESF level (>5.26 × 105 MESF at least). The Weibull model was selected as the base hazard model. The baseline tumour size was included in the hazard model as a parameter independent of the drug effect. A logistic model was applied to explain the effects of drug exposure and CD16MESF level. CONCLUSIONS: The final model indicates that adequate drug exposure plus a favourable immune environment are associated with prolonged OS. This quantitative model should be further validated with emerging data so as to guide study design in future clinical trials.

Bayesian adaptive dose-escalation designs for simultaneously estimating the optimal and maximum safe dose based on safety and efficacy.

The main purpose of dose-escalation trials is to identify the dose(s) that is/are safe and efficacious for further investigations in later studies. In this paper, we introduce dose-escalation designs that incorporate both the dose-limiting events and dose-limiting toxicities (DLTs) and indicative responses of efficacy into the procedure. A flexible nonparametric model is used for modelling the continuous efficacy responses while a logistic model is used for the binary DLTs. Escalation decisions are based on the combination of the probabilities of DLTs and expected efficacy through a gain function. On the basis of this setup, we then introduce 2 types of Bayesian adaptive dose-escalation strategies. The first type of procedures, called "single objective," aims to identify and recommend a single dose, either the maximum tolerated dose, the highest dose that is considered as safe, or the optimal dose, a safe dose that gives optimum benefit risk. The second type, called "dual objective," aims to jointly estimate both the maximum tolerated dose and the optimal dose accurately. The recommended doses obtained under these dose-escalation procedures provide information about the safety and efficacy profile of the novel drug to facilitate later studies. We evaluate different strategies via simulations based on an example constructed from a real trial on patients with type 2 diabetes, and the use of stopping rules is assessed. We find that the nonparametric model estimates the efficacy responses well for different underlying true shapes. The dual-objective designs give better results in terms of identifying the 2 real target doses compared to the single-objective designs.

Exploring genetic and non-genetic risk factors for delayed graft function, acute and subclinical rejection in renal transplant recipients.

AIMS: This study aimed at identifying pharmacological factors such as pharmacogenetics and drug exposure as new predictive biomarkers for delayed graft function (DGF), acute rejection (AR) and/or subclinical rejection (SCR). METHODS: Adult renal transplant recipients (n = 361) on cyclosporine-based immunosuppression were followed for the first 6 months after transplantation. The incidence of DGF and AR were documented as well as the prevalence of SCR at 6 months in surveillance biopsies. Demographic, transplant-related factors, pharmacological and pharmacogenetic factors (ABCB1, CYP3A5, CYP3A4, CYP2C8, NR1I2, PPP3CA and PPP3CB) were analysed in a combined approach in relation to the occurrence of DGF, AR and prevalence of SCR at month 6 using a proportional odds model and time to event model. RESULTS: Fourteen per cent of the patients experienced at least one clinical rejection episode and only DGF showed a significant effect on the time to AR. The incidence of DGF correlated with a deceased donor kidney transplant (27% vs. 0.6% of living donors). Pharmacogenetic factors were not associated with risk for DGF, AR or SCR. A deceased donor kidney and acute rejection history were the most important determinants for SCR, resulting in a 52% risk of SCR at 6 months (vs. 11% average). In a sub-analysis of the patients with AR, those treated with rejection treatment including ATG, significantly less frequent SCR was found in the 6-month biopsy (13% vs. 50%). CONCLUSIONS: Transplant-related factors remain the most important determinants of DGF, AR and SCR. Furthermore, rejection treatment with depleting antibodies effectively prevented SCR in 6-month surveillance biopsies.

Bayesian adaptive dose-escalation procedures for binary and continuous responses utilizing a gain function.

One of the main aims of early phase clinical trials is to identify a safe dose with an indication of therapeutic benefit to administer to subjects in further studies. Ideally therefore, dose-limiting events (DLEs) and responses indicative of efficacy should be considered in the dose-escalation procedure. Several methods have been suggested for incorporating both DLEs and efficacy responses in early phase dose-escalation trials. In this paper, we describe and evaluate a Bayesian adaptive approach based on one binary response (occurrence of a DLE) and one continuous response (a measure of potential efficacy) per subject. A logistic regression and a linear log-log relationship are used respectively to model the binary DLEs and the continuous efficacy responses. A gain function concerning both the DLEs and efficacy responses is used to determine the dose to administer to the next cohort of subjects. Stopping rules are proposed to enable efficient decision making. Simulation results shows that our approach performs better than taking account of DLE responses alone. To assess the robustness of the approach, scenarios where the efficacy responses of subjects are generated from an Emax model, but modelled by the linear log-log model are also considered. This evaluation shows that the simpler log-log model leads to robust recommendations even under this model showing that it is a useful approximation to the difficulty in estimating Emax model. Additionally, we find comparable performance to alternative approaches using efficacy and safety for dose-finding.

Bayesian adaptive designs in single ascending dose trials in healthy volunteers.

AIM: Recent publications indicate a strong interest in applying Bayesian adaptive designs in first time in humans (FTIH) studies outside of oncology. The objective of the present work was to assess the performance of a new approach that includes Bayesian adaptive design in single ascending dose (SAD) trials conducted in healthy volunteers, in comparison with a more traditional approach. METHODS: A trial simulation approach was used and seven different scenarios of dose-response were tested. RESULTS: The new approach provided less biased estimates of maximum tolerated dose (MTD). In all scenarios, the number of subjects needed to define a MTD was lower with the new approach than with the traditional approach. With respect to duration of the trials, the two approaches were comparable. In all scenarios, the number of subjects exposed to a dose greater than the actual MTD was lower with the new approach than with the traditional approach. CONCLUSIONS: The new approach with Bayesian adaptive design shows a very good performance in the estimation of MTD and in reducing the total number of healthy subjects. It also reduces the number of subjects exposed to doses greater than the actual MTD.

Pharmacokinetic-pharmacodynamic models for categorical toxicity data.

We propose a pharmacokinetic-pharmacodynamic (PK/PD) model (with possibly different choices for the PD link) for categorical toxicity data analysis. This is extension of the one-comportment model that applies to toxic endpoints categorised by grades (e.g., benign, mild, severe, and very severe). The model assumes that the area under the curve (AUC) of the internal quantity of the chemical substance is the critical dose-metric that drives the acute toxic phenomenon. That model handles time-varying concentrations and takes into account follow-up time, i.e., time at which effects are observed. Moreover the model bridges mechanistically based dose-response models and standard dose-response models, retaining the advantages of both. We use Markov chain-Monte Carlo (MCMC) simulations to fit the model to mortality data for mice exposed to chlorine, rats exposed to ammonia, and categorical data (different severity levels) from acute exposures of rats and humans to hydrogen sulfide.