Impact of Inflammation, Tumor and Product Attributes on Clinical Outcomes in Patients with Relapsed/Refractory Follicular Lymphoma Treated with Axicabtagene Ciloleucel

Background: The ZUMA-5 trial, investigating axicabtagene ciloleucel (axi-cel), an anti-CD19 CAR T cell therapy, in third-line indolent-NHL, demonstrated high rate of durable response with a manageable safety profile (Jacobson et al., Lancet Oncology 2022). Results from this trial led to the approval of axi-cel in relapsed/refractory (r/r) Follicular Lymphoma (FL) patients. While clinical outcomes were remarkable, biomarkers associated with efficacy and high-grade toxicity in r/r FL patients treated with axi-cel are not well established. Methods: This report presents univariate and multivariate analyses of product attributes, serum biomarkers, clinical features, patient characteristics, and tumor characteristics to identify pre- (230 covariates were assessed) and post-treatment biomarkers (66 covariates were assessed) that associate with efficacy and/or high-grade (grade >= 3) cytokine release syndrome (CRS) or neurologic events (NE) in 124 r/r FL patients from the ZUMA-5 trial (NCT03105336). To investigate factors in the tumor microenvironment associated with outcome, we conducted gene expression profiling via Nanostring analysis of pre-treatment tumor biopsies from a subset of patients (N=34) using pre-defined PanCancer IO360™ signatures. This analysis was complemented by bulk RNA sequencing (N=35, including 30 patients overlapping with the Nanostring dataset). CD19 protein expression was measured using immunohistochemistry (IHC). Results: Univariate and multivariate analyses of 296 covariates (pre-treatment and post-treatment) identified pre-treatment inflammatory peripheral blood markers, including TNFa and IL12p40, as well as total metabolic tumor volume (TMTV) as associated with disease progression. Moreover, TNFa and other inflammatory markers showed strong correlation with TMTV. Conversely, T-naïve-like product phenotype (CCR7+ CD45RA+) associated with improved progression-free survival (PFS), especially in high TMTV patients. These covariates improved risk stratification when combined with the FL International Prognostic Index (FLIPI) score. Multivariate analysis also showed positive correlation of T-cells with grade≥3 CRS and NE. Post-treatment, CAR T-cell expansion associated with improved PFS, while serum inflammatory and immuno-modulatory markers, including TNFa associated with disease progression and occurrence of high-grade CRS or NE. This presents possible targets for improving the therapeutic index of CAR T cells in r/r FL. Investigation of pre-treatment tumor gene expression with IO360™ signatures revealed that tumor IFN signaling, indicated by IFNg IO360 score associated with disease progression. This was supported by bulk RNAseq analysis where IFN stimulated genes (ISGs) and IFN-related gene signatures were higher in relapsed and non-responders. Further, IFN signaling was also associated with higher expression of T cell exhaustion markers, including TIM3 and LAG3, indicative of a pre-treatment TME with dysfunctional T cells. Unlike the findings in second-line diffuse large B cell lymphoma (LBCL) (ZUMA-7 study), B-cell signature and CD19 expression in pre-treatment tumor biopsies were not associated with clinical outcomes in r/r FL, possibly due to significantly higher levels in r/r FL compared to second-line LBCL. Post-treatment loss of tumor CD19, measured by IHC did not emerge as a major mechanism of relapse. Conclusions: These findings underscore the significance of pre-treatment systemic inflammation (e.g. serum TNFa levels), product T cell phenotype, TMTV, CAR T-cell expansion, tumor IFN signaling and FLIPI as independent factors associated with clinical outcomes in patients with r/r FL treated with axi-cel. These results offer insights into mechanisms of resistance and toxicity, risk stratification, combination therapy and development of next-generation CAR-T therapy for patients with r/r FL.