Abstract Anti-CD19 Chimeric Antigen Receptor (CAR) T-cell therapy is a promising option for relapsed or refractory lymphoma patients, yet our mechanistic understanding of response heterogeneity remains incomplete. Using retrospective longitudinal data from 21 patients treated with axicabtagene ciloleucel (including CAR T counts, absolute lymphocyte counts, tumor burden, and survival), we developed a computational modeling approach to distinguish between two expansion mechanisms: homeostasis- and antigen-driven. For each patient, we trained and compared three-compartment models to describe the dynamics of normal T-, CAR T-, and tumor cells. Comparisons revealed two distinct patient groups: patients who can be exclusively characterized by homeostatic proliferation as the CAR T expansion mechanism (homeostatic expanders, 10/21) and those of mixed type (mixed homeostatic/antigen-driven expanders, 11/21). These groups were distinguished by differences in the relationship between baseline metabolic tumor volume (taken one to two weeks before CAR T) and the inferred tumor burden at dosing. Notably, homeostatic expanders demonstrated significantly better overall survival (60% vs. 10% beyond day 180, p = 0.011), with 6 of 7 patients in this group achieving long-term responses. Our findings highlight how the ability of CAR T-cells to function in a homeostatic reconstitution context and the ability to quantify tumor burden at CAR T dosing influence the predictability of long-term responses. Our personalized mathematical modeling approach provides novel insights into optimizing CAR T-cell therapy and understanding the dynamics of cellular immunotherapy. Importantly, the absence of antigen-driven expansion increases the negative impact of high tumor burden at the time of infusion. Citation Format: Philipp Martin Altrock, Álvaro Martinez-Rubio, María Rosa, Arne Traulsen, Michael D. Jain, Frederick L. Locke. Modeling the role of homeostatic T-cell reconstitution in durable response to CAR T-cell therapy abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6834.
Altrock et al. (Fri,) studied this question.