Harmonizing flow-based quantification of CD19-directed CAR-T cells: Results from a multi-center interlaboratory comparison

CD19-directed CAR-T cell therapy has become the standard of care for the treatment of relapsed/refractory LBCL, MCL, FL and BCP-ALL. Expansion of CAR-T cells in the peripheral blood (PB) of patients in the first weeks after infusion is an important biomarker for the prediction of therapy response and occurrence of toxicities. Importantly, meaningful comparison and aggregation of CAR-T cell expansion data across studies and centers requires harmonized quantification approaches. We have therefore jointly developed and harmonized a flow cytometry-based quantification workflow for CD19-directed CAR-T cells and established an external quality assessment (EQA) scheme for centers based in Germany, Austria and Switzerland. CAR-T cells were detected using the dual-platform approach integrating a PE-conjugated FMC63 idiotype antibody. The assay was validated using PB from CAR-T cell patients and healthy donors spiked with CAR-T cells. Accuracy was evaluated in three interlaboratory comparisons using the harmonized workflow. Furthermore, we established an EQA scheme for CD19-CAR T cell quantification including also other quantification approaches. Using our workflow, we found CAR-T cell quantification to be highly sensitive (LOD/LLOQ: 0.02/0.03% of CD45⁺ events or 0.9/1.5 CAR-T cells/μL) and observed a high intra-assay precision with a coefficient of variation (CV) of typically less than 10%. Quantification was challenged by CAR-T cell stability resulting in about 4-fold underestimation of the CAR-T cell concentration when EDTA blood was analyzed 3 days after sample collection. However, this could be markedly improved by sample stabilization with Streck Cell Preservative. Interlaboratory comparisons underlined the accuracy of our workflow relying on a harmonized gating strategy and the following formula, which integrates absolute leukocyte counts: $\mathrm{CAR}-T\text{cells}\left(/\mathrm{\mu L}\right)=\text{Leukocytes}\left(/\mathrm{\mu L}\right)x\left(\text{Events}\left(\text{viable}\mathrm{CD}{45}^{+}{\text{CD3}}^{+}{\mathrm{CAR}}^{+}\right)/\text{Events}\left(\text{viable}\mathrm{CD}{45}^{+}\right)\right)$ . Results from four EQA tests including seven samples and 31 participating centers and multiple quantification methods showed measurement uncertainties in CAR-T cell quantification, with %CVs between 16 and 92% depending on the absolute CAR-T cell level. Importantly, laboratories applying the harmonized workflow described in this paper showed a markedly improved concordance of quantification results (4.9-69 %CV) compared to other quantification approaches (26-158 %CV). Taken together, we report a simple and reproducible workflow for routine quantification of CD19-CAR T cells that can be easily adapted for other target antigens. In contrast to the variability observed in the EQA tests, interlaboratory testing using the described harmonized workflow demonstrated improved concordance. This emphasizes the importance of thorough harmonization of wet and dry lab procedures in order to achieve comparable CAR-T cell quantification results.