Background: Characterization of product cellular attributes contributing to the in vivo expansion and toxicity of allogeneic, CD19-directed chimeric antigen receptor (CAR) T-cell therapy is necessary to optimize efficacy and safety. Using a single step process, the azer-cel CD19 CAR is knocked into the T-cell receptor alpha constant (TRAC) gene locus via a TRAC-specific ARCUS nuclease in activated T cells. Unlike autologous CAR T-cell therapy, all allogeneic CAR T products are cryopreserved. In this phase I study of the allogeneic, CD19-directed CAR T azer-cel, we analyzed post-thaw product attributes and cell composition of the infused product associated with pharmacokinetics (PK), pharmacodynamics (PD), and clinical outcomes (NCT03666000). Methods: Azer-cel was administered as a single infusion to 44 subjects across several dose levels and fludarabine/cyclophosphamide lymphodepletion regimens. CAR T cell peak expansion (Cmax), PK and area under the concentration-time curve (AUC) was assessed by flow cytometry and PD by multiplex cytokine assays. Azer-cel post-thaw cell composition was assessed by flow cytometry (minimally differentiated stem central memory [SCM], CC-chemokine receptor 7 (CCR7)+; differentiated effector memory, CCR7-). Effective cell number infused for any antigenically defined population was estimated using the product of the population percentage and the total post-thaw viable cell number. Overall tumor burden was measured as sum of the products of perpendicular diameters of target lesions and response was evaluated using Lugano 2016 criteria. Responders had either a complete or partial response at Day 28 or later. Relationships to safety events and efficacy were assessed using univariate correlative analyses (variables selected based on Pearson correlation coefficients ρ(rho)>0.35 unless specified, P-value <0.05). Results: Peak azer-cel levels and AUC varied by dose level with median Cmax ranging from 38 cells/mL at DL2 to 6.7x 104 cells/mL at DL4b, whereas median AUC0-28 ranged between 3 x 102 to 5.1x 105(cells/mL*days). Both azer-cel Cmax and AUC correlated strongly with SCM cell dose (P=0.009 & P=0.0026, respectively). Development of cytokine release syndrome (CRS) associated with effective SCM CD8+ cell dose (P<0.05) and total tumor burden (P=0.0087), in addition to Cmax of the cytokines interleukin (IL)-6, TNFα, MIP1α, IL-10 and C-X-C motif chemokine ligand 10 (CXCL10) or Interferon gamma-induced protein 10 (IP10) (P<0.01). Immune effector cell-associated neurotoxicity syndrome (ICANS) correlated with azer-cel Cmax (P=0.035), AUC (P=0.0057), total tumor burden (P=0.004), and effective SCM CAR T cell dose (P<0.05), as well as changes in a distinct set of cytokines that included IL-2, IL-8, IIL-15, IL-6, and IP10 (P<0.05). Grade >3 ICANS correlated strongly with effective SCM CD4+ cell dose (P=0.0015), differentiated CD4+CCR7- cell dose (ρ =0.34, P=0.029) and CD4:CD8 ratio (ρ = 0.32, P= 0.036), in addition to Cmax of IL-6, IL-2 and IP10 (P< 0.01). Overall response to azer-cel treatment negatively correlated with differentiated CD4+ CCR7- cell dose (P=0.04). Conclusions: Post-thaw product composition and effective cell dose are predictive for in vivo expansion potential, CAR-T-related adverse events and response to treatment with azer-cel. This is the first analysis of an allogeneic CD19 CAR T product composition to demonstrate that strategies intended to maximize stem central memory T-cell fraction (CCR7+) while limiting CD4+ CCR7- differentiated fraction may improve safety and efficacy of CAR T therapy.