Mario Torchia1, Niccolò Totis2, Fianne Sips2, Giuseppe Pasculli1, Pauline Bambury1, Jane Knöchel1, Maud Beneton1, Paolo Messina1, Daniel Röshammar1, Edward Coles3, Damien Bates3, Mike Fitzpatrick3
1InSilicoTrials Technologies S.p.A., 2InSilicoTrials Technologies B.V., 3Cellphire Therapeutics
Objectives: This work aimed to utilize population pharmacokinetic-pharmacodynamic (PKPD) modeling to support the development of a novel platelet-derived hemostatic agent, designed for intravenous administration in the management of perioperative bleeding. Specifically, the analysis was used to identify the optimal dosing regimens and optimal sample sizes for a clinical trial in patients undergoing emergency cardiothoracic surgery. Methods: A population PKPD model was developed using preclinical data from three animal species: mice, canines and non-human primates. The drug effect was defined as the time to hemostasis. Allometric scaling was applied to facilitate cross-species extrapolation. Inter-individual variability was accounted for in both PK and PD parameters. The initial dose administered in the first period was 3.2e+09, 5e+09, or 1e+10 cells/kg. In the subsequent two periods, the dose was either maintained at the initial level or reduced to half, following a ‘1-1-1’ strategy for consistent dosing across all periods or a ‘1-0.5-0.5’ strategy when the dose was halved after the first period. Inter-dose intervals of 10 or 20 minutes were also evaluated. The optimal dosing regimens were identified as the ones minimizing the overlap in predicted PK and PD endpoints, ensuring clear differentiation between treatment outcomes. By simulating endpoints across increasing sample sizes and re-estimating model parameters, the optimal sample size was determined as the one able to maintain the mean Relative Standard Error (RSE%) of PKPD parameter estimates below 25%. Results: A one-compartment PK model combined with a direct complete-response PD model, driven by total drug exposure, quantified by the area under the drug concentration curve (AUC) described the data. Goodness-of-fit plots demonstrated the model’s accuracy and PKPD parameters were estimated with good precision. Simulation results identified three dosing regimens with minimal overlap in drug exposure and clear differences in predicted therapeutic responses, while remaining feasible in terms of safety and administration. These regimens included a first regimen characterized by an initial dose of 3.2e+09 cells/kg, a 20-minute inter-dose interval and a ‘1-1-1’ dosing strategy, a second regimen differing from the first for a 10-minute inter-dose interval, and a third regimen characterized by an initial dose of 1e+10 cells/kg, a 10-minutes inter-dose interval and a 1-0.5-0.5 dosing strategy. Simulations indicated that the third regimen is expected to achieve near maximum efficacy, with the greatest reduction in bleed time achieved after 45 minutes of exposure. Notably, increasing follow-up doses in this regimen, as in the 1-1-1 dosing strategy, did not provide a significant benefit over the 1-0.5-0.5 strategy. Furthermore, doses exceeding 1e+10 cells/kg were predicted to translate into a saturated drug effect with no additional clinical benefit. These findings suggest that the identified regimens are strong candidates for clinical evaluation. Furthermore, sample size estimations indicated that enrolling eight subjects per treatment arm would provide sufficient data for a model-based analysis of clinical trial outcomes. Conclusions: Using preclinical data from three animal species, a population PKPD model was built to support the clinical development of a novel platelet-derived hemostatic agent. Simulations identified three dosing regimens with distinct therapeutic responses and minimal exposure overlap, making them strong alternatives for future clinical evaluation. Sample size estimations indicated that enrolling eight patients per treatment arm would provide sufficient data for a model-based analysis while controlling parameter uncertainty. Overall, these results provide valuable guidance to minimize risks and maximize insights in future early-phase clinical trials.
Reference: PAGE 33 (2025) Abstr 11704 [www.page-meeting.org/?abstract=11704]
Poster: Methodology - Study Design