Peripheral Blood CAR T-Cell Level Using a Clia-Validated Assay Is Prognostic for Treatment Response and Icans in Large B-Cell Lymphoma

Background: Chimeric antigen receptor T-cell (CAR-T) therapy targeting CD19 offers curative potential for relapsed or refractory large B-cell lymphoma (LBCL), but recurrence occurs in >50% of patients. CAR-T is also associated with serious toxicities including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). There have been discordant results across few earlier studies as to whether CAR-T expansion, as determined by longitudinal quantification of blood CAR-T levels, is a predictive biomarker for response to therapy and for therapy-associated complications (reviewed in Lionel & Neelapu, Blood Advances 2024). Differing results from these studies could be due to variation across CAR-T quantification assays and measurement time points as well as the combined use of different CAR-T products and lymphoma histologies in analyses. We analyzed a uniform cohort of LBCL patients treated with a single CAR-T product using a CLIA-validated flow cytometry assay at precise time points to establish the prognostic utility of CAR-T expansion and investigate the impact of different pre-treatment factors on CAR-T expansion kinetics.

Methods: The study included all patients who received standard-of-care axicabtagene ciloleucel (axi-cel) for LBCL at our institution starting in October 2022, when the flow cytometry assay was introduced, until January 2024, when pre-determined target of 100 patients was reached (54% second-line CAR-T vs. 46% later-line). Blood samples were obtained on alternate days from patients starting on day +1 after CAR-T infusion and continued for about two weeks. Subsequent measurements were performed weekly during the first month post-infusion. Blood CAR-T quantification was performed by a flow cytometry-based assay which was validated as per CLIA standards and utilized antibody targeting CD19-specific monoclonal antibody FMC63. CART cells were reported as absolute count using a dual platform (WBC by hematology counter). The primary response endpoint was disease status on PET/CT scan at 90 days after infusion. CRS and ICANS were graded as per the ASTCT system.

Results: CAR-T expansion data was available from 100 patients with a median number of 9 measurements per patient (Q1-Q3 range 7-10). In the 81 patients who had data available from day+90 PET/CT scans, CAR-T levels were found to be significantly higher in the 58 responders (56 complete and 2 partial responses) compared to the 23 non-responders at days +5, +7 and +9 after infusion. Among these days, day+7 CAR-T levels had the most significant difference (p<0.001) and highest area under the curve. Receiver operating characteristic (ROC) analysis found day+7 CAR-T level of 30 cells/mL to be differentiating responders from non-responders (specificity 75%, sensitivity 77.8%). The day+7 CAR-T level continued to be significantly associated with response in a multivariate analysis incorporating other parameters at time of conditioning therapy including disease stage, LDH, IPI risk score and refractoriness to prior treatment.

The day+7 CAR-T level was also significantly higher in patients with grade ≥3 ICANS (p<0.001) and correlated with increased use of systemic steroids for ICANS treatment (p<0.01) and need for critical care (p<0.01). By ROC analysis, day+7 CAR-T level of >134 cells/mL was associated with severe ICANS (specificity 81.7%, sensitivity 66.7%); day+7 level continued to be significantly associated with severe ICANS in a multivariate analysis. The association between grade ≥3 CRS and day+7 levels did not reach statistical significance. Several pre-treatment parameters were associated with higher day+7 CAR-T levels including fewer prior lines of therapy (second-line vs. later-line), absence of prior bendamustine exposure and the use of post-apheresis bridging therapy.

Conclusions: Our findings provide evidence for the prognostic utility of peripheral blood CAR-T levels measured by a CLIA-validated assay as an early biomarker to predict response to treatment and risk for severe ICANS after axi-cel therapy in patients with LBCL. This has implications for the future clinical use of real-time CAR-T kinetics using a flow cytometry-based assay that could be implemented at most centers to help identify patients who are likely to need additional intervention to improve efficacy of CAR-T therapy and optimization of management strategies for severe ICANS.