Ignacio González-GarcÃa, Carlos Fernández-Teruel, Rubin Lubomirov, Salvador Fudio
PharmaMar
Objectives:
There are different strategies to validate a PKPD model. However, only when the model is able to describe and predict the data correctly and additionally has been validated, it is appropriate to perform simulations in order to explore a new dose range, new schedules, etc. Other types of simulations are also relevant as a way to measure the effect of the covariates either on the dependent value (e.g. concentrations, tumor size, absolute neutrophil count) or in the PK or PD parameters.
Nowadays, performing this type of simulation is not complicated, and there are many useful software available. However, these simulations have an inconvenient, especially when NONMEM is used. Simulation parameters (e.g. doses and times of administration, number of cycles and sampling times) have to be pre-established before simulations are run, and cannot be modified during the simulation exercise. Moreover, these simulation parameters can be modified only once the results of the simulation are analyzed. Such modifications may entail repeated changes in the database until the final result is achieved, which is highly time-consuming and not efficient.
PhM-TSim (PharmaMar Trial Simulator) is a new tool aimed at speeding up this type of simulations where the knowledge of the subject’s condition at present is needed to simulate the next subject’s condition. Evaluating subject’s condition after the last drug administration, PhM-TSim is capable to decide the dose regimen of the subsequent drug administration, which may be the same or a reduced one. Dose delays, omissions or administration of new therapies, such as growth colony stimulation factor (G-CSF), red blood cells and/or albumin transfusions, can also be considered.
The objective of this work is to present PhM-TSim as a new tool that implements changes in the database automatically, based on predefined per protocol re-treatment criteria. PhM-TSim combined with longitudinal drug exposure-response (DER) models is a valuable tool to assess scenarios for dose selection in future clinical trials.
Lurbinectedin is a drug under clinical development with several phase II and III trials ongoing/planned. The most relevant toxicity associated to lurbinectedin is the decay of absolute neutrophil counts (ANC), and is used as an example of the implementation of PhM-TSim.
Methods:
A DER model for ANC was developed and validated for the purpose of this work. Different simulation scenarios were planned in order to measure the capability of PhM-TSim when adjusting the dose (e.g. reduction, delay) to avoid serious toxicities during consecutive lurbinectedin administrations.
The lurbinectedin PKPD model was a 3-comparmental model with lineal kinetics (either distribution or elimination) combined with a DER model for ANC based on the one proposed by Quartino, and adapted by Fernandez-Teruel, with transit compartments and feedback effects on mean maturation time and proliferating cell pool, which was suitable for describing the time course of neutrophils.
NONMEM 7.3 was used to perform the simulations, while R (v. 3.4.1) and R Studio® (v. 1.01.143) were used to build the data sets, graphical analysis and dose adjustment.
Results: For each scenario, a total of 100 subjects and a maximum of ten cycles were simulated. A univariate analysis of the covariates included in the PKPD model showed the real impact of the covariate in the subject’s condition.
Conclusions: In all the scenarios simulated, PhM-TSim was able to first evaluate every subject’s condition, and then, if necessary, to delay, reduce lurbinectedin administration or even withdrawal from the clinical trial. The versatility of PhM-TSim also allows the use of G-CSF or any other strategies predefined per protocol at the re-treatment criteria. The impact of a due covariate on subject’s condition may not be the same after first drug administration than along the whole treatment, due to adjustments in dose regimen to manage toxicities. PhM-TSim shows a better description of the PD along the treatment than common simulations that do not consider dose adjustments, and ultimately can be implemented in simulations of DER efficacy models, to account for such dose adjustments that affect subject’s outcome in the clinical field.
Reference: PAGE 27 (2018) Abstr 8577 [www.page-meeting.org/?abstract=8577]
Poster: Methodology - Other topics