Predicting Survival for Chimeric Antigen Receptor T-Cell Therapy: A Validation of Survival Models Using Follow-Up Data From ZUMA-1

ObjectivesSurvival extrapolation for chimeric antigen receptor T-cell therapies is challenging, owing to their unique mechanistic properties that translate to complex hazard functions. Axicabtagene ciloleucel is indicated for the treatment of relapse or refractory diffuse large B-cell lymphoma after 2 or more lines of therapy based on the ZUMA-1 trial. Four data snapshots are available, with minimum follow-up of 12, 24, 36, and 48 months. This analysis explores how survival extrapolations for axicabtagene ciloleucel using ZUMA-1 data can be validated and compared.MethodsThree different parametric modeling approaches were applied: standard parametric, spline-based, and cure-based models. Models were compared using a range of metrics, across the 4 data snapshot, including visual fit, plausibility of long-term estimates, statistical goodness of fit, inspection of hazard plots, point-estimate accuracy, and conditional survival estimates.ResultsStandard and spline-based parametric extrapolations were generally incapable of fitting the ZUMA-1 data well. Cure-based models provided the best fit based on the earliest data snapshot, with extrapolations remaining consistent as data matured. At 48 months, the maximum survival overestimate was 8.3% (Gompertz mixture-cure model) versus the maximum underestimate of 33.5% (Weibull standard parametric model).ConclusionsWhere a plateau in the survival curve is clinically plausible, cure-based models may be helpful in making accurate predictions based on immature data. The ability to reliably extrapolate from maturing data may reduce delays in patient access to potentially lifesaving treatments. Additional research is required to understand how models compare in broader contexts, including different treatments and therapeutic areas.     

Recent advances in preclinical in vitro approaches towards quantitative prediction of hepatic clearance and drug-drug interactions involving organic anion transporting polypeptide (OATP) 1B transporters

Hepatic uptake mediated by organic anion transporting polypeptide (OATP) 1B1 and 1B3 can serve as a major elimination pathway for various anionic drugs and as a site of drug-drug interactions (DDIs). This article provides an overview of the in vitro approaches used to predict human hepatic clearance (CL_h) and the risk of DDIs involving OATP1Bs. On the basis of the so-called extended clearance concept, in vitro–in vivo extrapolation methods using human hepatocytes as in vitro systems have been used to predict the CL_h involving OATP1B-mediated hepatic uptake. CL_h can be quantitatively predicted using human donor lots possessing adequate OATP1B activities. The contribution of OATP1Bs to hepatic uptake can be estimated by the relative activity factor, the relative expression factor, or selective inhibitor approaches, which offer generally consistent outcomes. In OATP1B1 inhibition assays, substantial substrate dependency was observed. The time-dependent inhibition of OATP1B1 was also noted and may be a mechanism underlying the in vitro–in vivo differences in the inhibition constant of cyclosporine A. Although it is still challenging to quantitatively predict CL_h and DDIs involving OATP1Bs from only preclinical data, understanding the utility and limitation of the current in vitro methods will pave the way for better prediction.     

In Vitro-In Vivo Extrapolation and Hepatic Clearance-Dependent Underprediction

Accurately predicting the hepatic clearance of compounds using in vitro to in vivo extrapolation (IVIVE) is crucial within the pharmaceutical industry. However, several groups have recently highlighted the serious error in the process. Although empirical or regression-based scaling factors may be used to mitigate the common underprediction, they provide unsatisfying solutions because the reasoning behind the underlying error has yet to be determined. One previously noted trend was intrinsic clearance-dependent underprediction, highlighting the limitations of current in vitro systems. When applying these generated in vitro intrinsic clearance values during drug development and making first-in-human dose predictions for new chemical entities though, hepatic clearance is the parameter that must be estimated using a model of hepatic disposition, such as the well-stirred model. Here, we examine error across hepatic clearance ranges and find a similar hepatic clearance-dependent trend, with high clearance compounds not predicted to be so, demonstrating another gap in the field.     

Genetic aspects of animal reasoning

This paper reviews the investigations of Prof. L. V. Krushinsky and his colleagues into the genetics of complex behaviors in mammals. The ability of animals to extrapolate the direction of a food stimulus movement was investigated in wild and domesticated foxes (including different fur-color mutants), wild brown rats, and laboratory rats and mice. Wild animals (raised in the laboratory) were shown to be superior to their respective domesticated forms on performance of the extrapolation task, especially in their scores for the first presentation, in which no previous experience could be used. Laboratory rats and mice demonstrated a low level of extrapolation performance. This means that only a few laboratory animals were capable of solving the task, i.e., the percentage of correct solutions was equivalent to chance. The brain weight selection program resulted in two mice strains with a 20% (90-mg) difference in brain weight. Ability to solve the extrapolation task was present in low-brain weight mice in generations 7–11 but declined with further selection. Investigation of extrapolation ability in mice with different chromosomal anomalies demonstrated that animals with Robertsonian translocations Rb(8,17) 1lem and Rb(8,17) 6Sic were capable of solving this task in a statistically significant majority of cases, while mice with fusion of other chromosomes, as well as CBA normal karyotype mice, performed no better than expected by chance. Mice with two types of partial trisomies and animals homo- and heterozygous for translocations were also tested. Although mice with T6 trisomy performed no better than expected by chance, animals with trisomy for a chromosome 17 fragment solved the task successfully. Thus, a genetic component underlying the ability to solve the extrapolation task was demonstrated in three animal species. The extrapolation task in animals is considered to reveal a general capacity for elementary reasoning. The genetic basis of this capacity is very complex.     

A comparison of the toxicity to laboratory mice and pipistrelle bats Pipistrellus pipistrellus of exposure to remedially-treated timber

Many species of bats roost in roof spaces and can come into contact with pesticides used to treat roof timbers against rot. As part of a study to develop a test with laboratory mice which could be used to predict the likely toxicity of pesticides to bats, the toxic effects on pipistrelle bats Pipistrellus pipistrellus and laboratory mice (CFLP outbred-strain) of exposure to dieldrin-treated timber were compared. Dose (amount of dieldrin applied to the wood surface) — response (mortality) studies demonstrated that pipistrelle bats were 30 times more sensitive to dieldrin-treated wood than laboratory mice. However, dieldrin residues in the brain and liver of animals which had been poisoned were significantly higher in bats than mice. These residue data suggest that bats may be inherently less sensitive to dieldrin poisoning than mice and so it is likely that the 30-fold greater sensitivity of bats than mice to dieldrin-treated timber was a result of greater rate of uptake, per gram body weight, of the active ingredient from the wood surface and/or slower clearance from the body. The importance of differences in exposure and the reliability of residue data as an indicator of differences in inherent sensitivity to a chemical are discussed in terms of their importance when considering inter-species variation in toxic response.     

Comparative numerical modeling of inhaled micron-sized particle deposition in human and rat nasal cavities

Abstract

Micron-sized particle deposition in anatomically realistic models of a rat and human nasal cavity was numerically investigated. A steady laminar inhalation flow rate was applied and particles were released from the outside air. Particles showing equivalent total particle deposition fractions were classified into low, medium and high inertial particle. Typical particle sizes are 2.5, 9 and 20 μm for the human model and 1, 2 and 3 μm for the rat model, respectively. Using a surface-mapping technique the 3D nasal cavity surface was “unwrapped” into a 2D domain and the particle deposition locations were plotted for complete visual coverage of the domain surface. The total surface area comparison showed that the surface area of the human nasal model was about ten times the size of the rat model. In contrast, the regional surface area percentage analysis revealed the olfactory region of the rat model was significantly larger than all other regions making up ～55.6% of the total surface area, while that of the human nasal model only occupying 10.5%. Flow pattern comparisons showed rapid airflow acceleration was found at the nasopharynx region and the nostril region for the human and rat model, respectively. For the human model, the main passage is the major deposition region for micro-particles. While for the rat model, it is the vestibule. Through comparing the regional deposition flux between human and rat models, this study can contribute towards better extrapolation approach of inhalation exposure data between inter-subject species.     

N1 Latencies of the Slow Auditory Evoked Potential

Several studies have been published concerning the relation between stimulus level and amplitude of the slow auditory evoked response. Little attention has, however, been given to the relation between level and latency of the N₁ component (100-200 ms). A mathematical model of such a relation is proposed based on the assumption of an exponential relationship. The model has been tested on normally hearing subjects as well as on patients with sensorineural lesions, cochlear as well as retrocochlear. The results indicate that the proposed method of evaluating the slow evoked responses offers a way of threshold extrapolation with rather good accuracy. It also appears to be useful in determining the type of hearing loss     

Commentary: Dermal penetration of bisphenol A—Consequences for risk assessment

With this comment we would raise awareness for applying appropriate procedures in route-to-route extrapolation. The paper of Demierre et al. (2012) prompted us to comment on the simple approach for route-to-route extrapolation and to explain some short comings. For the risk assessment of exposures resulting from a non-oral route, route-to-route extrapolation is often done by correcting the non-oral route exposure by the route specific absorption into the systemic circulation and comparing the result with the (oral) threshold value. Making use of this procedure means that an internal dose obtained from the non-oral route is compared with an external dose of the oral route. This procedure would be appropriate only if the absorption on the oral route is 100%. If the absorption on the oral route is less than 100% the procedure may underestimate the risk of the exposure of the non-oral route. For some chemicals with a high first pass metabolism in the liver, e.g. BPA, the situation is even more complex and in addition, the target organ for toxicity has to be taken into consideration.     

Comparative functional and pharmacological characterization of Sandoz proposed biosimilar adalimumab (GP2017): rationale for extrapolation across indications

Background: Biosimilars are approved biologics that match reference medicine in quality, safety, and efficacy. The development of Sandoz proposed biosimilar adalimumab (SPBA; GP2017) involved a target-directed, iterative state-of-the-art quality-by-design development program. Here, we describe the functional and pharmacological characterization of SPBA and its proposed mechanism of action in immune-mediated inflammatory diseases.

Methods: Sensitive in vitro binding and functional characterization studies, and nonclinical evaluations (pharmacokinetics, pharmacodynamics, and safety/toxicology) were performed as part of a stepwise approach to confirm the biosimilarity of SPBA with reference adalimumab.

Results: Matching values were reported for SPBA and reference adalimumab in binding assays involving tumor necrosis factor (TNF)-α, complement 1q and human immune effector cell Fcγ receptor subtypes in cell-based bioassays for Fc receptor function (complement- and antibody-dependent cytotoxicity), and in apoptosis inhibition. Furthermore, SPBA and reference adalimumab were equivalent in terms of membrane TNF binding and induction of reverse signaling. Pharmacokinetics of SPBA and reference adalimumab were comparable in rabbits, and the two biologics were equally effective in a human TNF transgenic mouse model of polyarthritis.

Conclusion: SPBA matches reference adalimumab with regards to target binding, functional, pharmacokinetic, and pharmacodynamic properties at the nonclinical level supporting its approval in all indications of the reference adalimumab.     

Interpreting in vitro developmental toxicity test battery results: The consideration of toxicokinetics

In the EU collaborative project ChemScreen an alternative, in vitro assay-based test strategy was developed to screen compounds for reproductive toxicity. A toxicokinetic modeling approach was used to allow quantitative comparison between effective concentrations in the in vitro test battery and observations of developmental toxicity in vivo. This modeling strategy is based on (1) the definition of relevant observations of toxicity in vivo, (2) simulation of the corresponding systemic concentrations in vivo by toxicokinetic modeling, and (3) correction for differences in protein binding and lipid partitioning between plasma and in vitro test media. The test results of a feasibility study with a number of known reproductive toxicants has been described previously (Piersma et al. [15]). In the present paper, we take a more detailed look at the toxicokinetics of these compounds, and add the analysis of some compounds from subsequent studies. We discuss how the consideration of toxicokinetics allowed comparison between test systems with differing test medium composition, has helped to interpret the in vitro findings in light of in vivo observations, and to gain confidence in the predictive value of the test battery outcomes. The same toxicokinetic modeling strategy, in reverse order, can now be used for risk assessment purposes to predict toxic doses in vivo from effective concentrations in vitro.