Early Real Time Quantification of CD19. CAR T Cells By Flow Cytometry As a Predictive Biomarker for Response in Patients with R/R LBCL

Abstract

Introduction: Despite the clinical success of CD19-directed CAR T-cell therapy in relapsed / refractory (r/r) LBCL (large B-cell lymphoma), the majority of patients do not respond or relapse early after treatment. Biomarker analysis within pivotal trials showed the prognostic value of CAR T-cell expansion over time during the early post-transfusion period. We set up a longitudinal immunomonitoring to investigate whether distinct CAR T-cell levels can predict treatment response. Methods: 63 patients with r/r LBCL were treated in third line with the commercially available CD19-directed CAR T-cell products Axicabtagene ciloleucel (Axi-cel) or Tisagenlecleucel (Tisa-cel) between January 2019 and November 2021 at the University Hospitals of Erlangen and the LMU Munich. Aliquots of CAR T-cell products and EDTA-anticoagulated peripheral blood (15cc) were collected at day 4, 7, 14, 28, 60 and 90 post CAR T-cell transfusion. Clinical metadata were assembled prior and post transfusion. CAR T-cells were analyzed with flow cytometry utilizing a two step-staining with a biotinylated CD19 protein. CAR T-cells were detected with a sensitivity of 1:2000 and reported as cells/µl based on differential blood counts. Responders (R, complete or partial remission) were compared to non-responders (NR, stable or progressive disease) according to (PET-) CT scans three months after transfusion. Results: Surprisingly, the CAR T-cell products contained more untransduced T-cells than anti-CD19 expressing CAR T-cells with a median cell number of 5.75x108 Non-CAR T-cells and 1.76x108 CAR T-cells (p<0.0001). Even though a large inter-patient variability was observed, CAR T-cell peak expansion occurred between day 7 to 14 in the majority of patients. Interestingly, differences in the number of CAR T transduced T-cells were observed in Tisa-cel versus Axi-cel: Tisa-cel products contained higher numbers of Non-CAR T-cells compared to Axi-cel products (p<0.0001) followed by greater expansion of Non-CAR T-cells over time after Tisa-cel transfusion compared to Axi-cel transfusion (p=0.0155). However, the absolute number of CAR T transduced T-cells within the product did not impact response rate. More importantly, higher CAR T-cell expansion over time (AUC p=0.0007) and peak expansion (p=0.0009) were associated with treatment response. Moreover, higher CAR T-cell peak expansion in relation to tumor burden at baseline was associated with treatment response (p=0.0011). Comparing the absolute CAR T-cell numbers at the different time points, we identified the CAR T-cell levels at day 7 to be strongly associated with treatment response (p=0.0005). Next, through ROC analysis, we were able to identify 20/µl CAR T-cells on day 7 as a cut-off value, that was associated with treatment response (AUC=0.7656 and p=0.0007) and improved survival (PFS p=0.0001 and OS p=0.0002). Next, we hypothesized, that higher CAR T-cell numbers on day 7 reflect the superior capacity of CAR T-cells to proliferate and expand. This was indeed supported by our findings, that CAR T-cell numbers on day 7 correlated with the absolute number of CAR T-cells in the CAR product of R (r=0.4928, p=0.0144) but not of NR (r=-0.152). Finally, we found patient- (ECOG) and disease-associated (tumor volume, LDH) as well as baseline inflammatory markers (CRP and Ferritin) to be negatively correlated with CAR T-cell kinetics. Conclusion: Longitudinal quantification of CAR-T cells by multiparameter flow cytometry is feasible. Responding patients with r/r LBCL showed greater CAR T-cell expansion over the first 90 days post transfusion. Notably, a cut-off value of 20/µl CAR T-cells at day 7 was associated with both improved radiographic response and survival. Currently the cut-off value is being validated in an external control of 100 patients. Our data suggest that quantification of CAR T-cell levels might serve as a biomarker to predict treatment failure and open up the possibility to modify treatment concepts in these patients at an early time point. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal