Patterns of Immune Reconstitution and Their Clinical Significance in Aggressive Lymphoma Treated with CAR-T Therapy

Introduction

Chimeric antigen receptor T-cell (CAR-T) therapy is an effective treatment option for large B-cell lymphoma (LBCL). However, toxicities, including immune suppression, remain a significant concern. The dynamics of immune reconstitution (IR) post-CAR-T and its impact on treatment outcomes have been primarily studied in relatively small cohorts treated with axicabtagene ciloleucel (axi-cel). We systematically investigate the dynamics and patterns of IR and its influence on outcomes in a large cohort of patients with relapsed/refractory LBCL treated using three commercial CD19-CAR-T cell products.

Methods

Adult patients with LBCL treated with axi-cel, tisagenlecleucel (tisa-cel), and lisocabtagene maraleucel (liso-cel) between 2016 and 2024 were included in this single-center retrospective study. Immune cell levels were measured by flow cytometry at various timepoints up to a year post-CAR-T. Measurements after disease progression, relapse, or next-line treatment were excluded. Serial landmark multivariable Cox models stratified by CAR-T product and adjusted for age, lactate dehydrogenase level prior to CAR-T, and bridging were utilized to assess associations between IR and overall survival (OS) and progression-free survival (PFS). Finally, determinants of immune cell recovery were assessed using linear regression modeling.

Results

We included 263 patients with LBCL. Median age was 66 (range 20-86). Axi-cel (44.9%) was the predominant product, followed by liso-cel (30.4%) and tisa-cel (24.7%). Cyclophosphamide and fludarabine lymphodepletion was used in 86.7% of patients. Cytokine release syndrome and immune effector cell-associated neurotoxicity syndrome rates were 71.9% and 27%, respectively; 35.7% of patients received steroids and 45.6% received tocilizumab. Almost two-thirds of patients (63.3%) achieved a complete response following infusion. Median follow-up of the cohort was 26 months (interquartile range [IQR] 12-48).

CD3+ T cells were the most abundant circulating lymphocytes following CAR-T, followed by natural killer (NK) cells and B cells. By 30 days following CAR-T, median (IQR) values (cells/µL) of CD3+ T cells, NK cells, and B cells were 457 (180-790), 94 (56-148), and 0 (0-0), respectively. Median B cell levels remained at 0 through 365 days post-CAR-T.

By days 30, 180, and 365, CD4+ T cell levels were below 200 cells/µL in 64.3%, 51.5%, and 48.1% of patients, respectively. CD4+ T cell reconstitution was governed largely by recovery of CCR7-45RA- effector memory (EM) cells. Importantly, CD4+ T cell levels at all time points were not associated with infection risk, OS, or PFS in serial landmark analysis.

CD8+ T cells followed a similar trajectory to CD4+ T cells. Median (IQR) levels (cells/µL) of CD8+ T cells by days 30, 180, and 365 were 224 (91-445), 226 (108-434), and 227 (95-469), respectively. CCR7-45RA+ TEMRA cells and EM cells were the most abundant subpopulations. Notably, neither CD4+ nor CD8+ T cell recovery was influenced by lymphodepletion regimen or exposure to tocilizumab or corticosteroids.

When stratified by product, tisa-cel was consistently associated with greater levels of CD3+ T cells compared with axi-cel and liso-cel, an association driven by the CD4+ T cell compartment. This association was most prominent among the EM subpopulation. Notably, no differences were seen in other immune subsets.

In multivariable Cox models, early NK cell recovery (through 55 days post-CAR-T for OS and 65 days for PFS) was significantly associated with favorable OS (hazard ratio [HR]: 0.64; 95% confidence interval [CI], 0.44-0.92; P=0.0161) and PFS (HR: 0.65; 95% CI, 0.48-0.88; P=0.0054). Other immune subsets were not associated with outcome. In multivariable linear regression, levels of C-reactive protein, ferritin, tumor necrosis factor alpha, and interleukin 10 measured at time of CAR-T were inversely associated with NK cell recovery. Finally, progression and relapse were not linked to B cell recovery post-CAR-T across all products.

Conclusion

We present the largest and most comprehensive analysis of immune reconstitution following CAR-T. Among immune subsets considered, NK cell recovery emerges as a robust predictor of favorable clinical outcomes. In summary, this novel biomarker may serve as an important prognostic marker of disease outcome following CAR-T and also provide insights into the biology of post-CAR-T relapse.