(W-044) Population Cellular Kinetics of Orvacabtagene Autoleucel, an Autologous BCMA-Directed Chimeric Antigen Receptor T-Cell Product, in Patients with Relapsed/Refractory Multiple Myeloma

Yiming Cheng, PhD – Associate Director, BMS; Hongxiang Hu, PhD – Senior Research Investigator, BMS; Julia Piasecki, PhD – Senior Principal Scientist, BMS; Daniela Hosseyni, PhD – Senior Clinical Trial Physician, BMS; Ken Ogasawara, PhD – Director, BMS; Yan Li, PhD – Senior Director, BMS

Objectives: Orvacabtagene autoleucel (orva-cel) is a chimeric antigen receptor (CAR) T-cell product targeting B-cell maturation antigen (BCMA) which was investigated in subjects with relapsed/refractory multiple myeloma. To better characterize the kinetic profile of orva-cel transgene levels, a modified piecewise model was investigated, and the effects of different covariates on the cellular kinetics of orva-cel were further assessed.

Methods: Data from a Phase 1/2 study (NCT03430011, EVOLVE) of 159 subjects with relapsed and/or refractory multiple myeloma (RRMM) were used for the population cellular kinetic analysis. During the phase 1 dose-escalation, five dose levels of CD3⁺ CAR⁺ T cells, ranging from 50 to 600 million, were evaluated. The phase 2 expansion was conducted at the recommended phase 2 dose of 600 million. A modified piecewise model with the expansion rate described by saturation mechanism was used as the structural model. Unexplained between subject variability (BSV) was assumed to be log-normally distributed, and further explained by the addition of covariates including baseline demographics, dose level, prior or concomitant medication, baseline disease burden, anti-therapeutic antibody (ATA) status and other laboratory parameters. Model performance was evaluated through diagnostic plots, objective function value, standard goodness-of-fit criteria, and visual predictive check (VPC). Nonlinear mixed-effect modeling approach was implemented in the Monolix software system (Version 2019 R2, Lixoft, Antony, France).

Results: Traditional piecewise models, using a first-order expansion rate with or without lag time, inadequately captured the peak orva-cel transgene levels (Cmax) and underestimated time of Cmax (Tmax) when compared to observations (median = 10 days), suggesting potential model deficiency. A modified piecewise model incorporating a cell number-dependent expansion rate improved the model performance by 1) more accurately representing the cellular expansion phase as demonstrated in VPC plots, and 2) yielding a Tmax that aligns more closely with observed data. Population means of cellular kinetic parameters in a typical subject were: baseline (C0) 2.9 copies/µg; Tmax 7.86 days; maximum expansion rate (Vm) 31760 copies/µg/day; transgene levels at 1/2 of Vm (Km) 14039 copies/µg; α phase decline rate constant 0.058/day; β phase decline rate constant 0.002/day; and conversion rate constant from α phase to β phase 0.0000077/day. The covariate search demonstrated that dose level, baseline plasma cell percentage in bone marrow, the usage of tocilizumab and/or corticosteroids, and ATA status were associated with orva-cel kinetics and retained in the final model.

Conclusion: Orva-cel cellular kinetics were adequately described by the modified piecewise model that consisted of a cell-number dependent expansion phase, closely reflecting the physiology of cellular expansion, followed by a bi-exponential contraction phase in RRMM patients.

Citations: [1] De Boer, R.J., Homann, D. & Perelson, A.S. Different dynamics of CD4+ and CD8+ T cell responses during and after acute lymphocytic choriomeningitis virus infection. J Immunol 171, 3928-35 (2003).