Fianne L. P. Sips (1), Alberto Palazzin (1), Chiara Nicolò (1), Roberta Bursi (1)
1. InSilicoTrials Technologies, Trieste, Italy
Introduction:
In the last decades, Relapsing-remitting Multiple Sclerosis (RRMS) diagnostic methodology and treatment options have undergone great improvements [1–3], positively affecting patient care. However, these developments have also impacted clinical trial design. For example, the average annualized relapse rates characterizing typical clinical trial populations have declined over recent decades [4], which results in smaller expected relapse rates, and larger required groups sizes or longer required trial durations.
To support design of clinical trials in RRMS, we developed a clinical trial simulator (MS TreatSim, [5]). The simulator was designed to create heterogeneous subpopulations of virtual RRMS patients, apply common inclusion and exclusion criteria to populate trial arms, and finally perform simulations of all virtual patients in a clinical trial scenario. We then applied MS TreatSim to reproduce a phase III trial of natalizumab first published 15 years ago [6].
Objectives:
- Develop a trial simulation framework for in silico trials in Relapsing-remitting Multiple Sclerosis
- Reproduce the control group characteristics and treatment effect of a historical clinical trial of natalizumab with the simulator
Methods:
The trial simulator (MS TreatSim, [5]) consists of a simulation framework to set up and simulate an in silico trial and a web-based implementation of this simulator with a user-friendly interface, and is built on an agent-based model of the immune system and RRMS [7]. This underlying mechanistic modelling framework incorporates the innate and adaptive immune system and includes basic immune behaviours and components such as B cells, T helper cells, T regulatory cells, antigens, and cytokine signalling. For application to RRMS, the model was expanded with the autoimmune response and the brain’s oligodendrocytes. The model incorporates treatments, such as natalizumab, via their mechanism of action at the cellular level [7]. As a case study, we reproduced a phase III clinical trial for natalizumab (AFFIRM, [6]) with MS TreatSim, by setting the base characteristics, inclusion criteria, timelines, group compositions and treatment schedules to mirror those reported in [6].
Results:
MS TreatSim was developed to generate heterogeneous virtual patient groups through random sampling of base characteristics, customizable through probabilities that can be set by the user. The virtual patients are then selected and placed into trial groups based on customizable inclusion and exclusion criteria, and finally they are simulated according to the trial design specified by the user.
To evaluate MS TreatSim’s performance, we recreated a historical clinical trial of natalizumab. The results, expressed in percentages of relapse free subjects after two years, show that the simulator is able to reproduce the population itself well (control group: 41 % in [6], versus 40% in silico). In the treatment group, the treatment effect was clear in both clinical and in silico studies (67 % vs. > 90%). The treatment effects therefore seems somewhat overestimated, although other outcome metrics in [6] show a similarly large effect size to the difference in in silico relapse free subjects. An explanation for this overestimation may lie in the fact that complex disease histories are known to induce resistance to natalizumab treatment [8], and these resistance mechanisms are currently underrepresented by the in silico trial setup of this study.
Conclusion:
MS TreatSim synergistically combines a user-friendly web-based interface with the simulation framework developed here, as well as an underlying a mechanistic model. MS TreatSim can be used to explore clinical trial design scenarios, aiding optimization of clinical trial design.
Acknowledgements:
The cloud-based tool MS TreatSim is the result of a collaboration between InSilicoTrials Technologies and Mimesis Srl. The authors acknowledge Prof. Pappalardo and Dr. Russo for their contribution to this work.
References:
[1] A. J. Thompson et al. Lancet Neurol. (2018) 17, 162.
[2] M. Bross et al. Int. J. Mol. Sci. (2020) 21, 1.
[3] R. Dobson and G. Giovannoni, Eur. J. Neurol. (2019) 26, 27.
[4] Y. Zhang etal., Ther. Adv. Neurol. Disord. (2019) 12.
[5] MS TreatSim, mstreat.insilicotrials.com, InSilicoTrials Technologies SpA
[6] C. H. Polman et al. N Engl j Med (2006) 354.
[7] F. Pappalardo et al, Cells (2020) 9, 586.
[8] van Pesch et al. Clin. Neurol. Neurosurg. (2016) 149:55–63.
Reference: PAGE 30 (2022) Abstr 10156 [www.page-meeting.org/?abstract=10156]
Poster: Drug/Disease Modelling - CNS