Abstract 510: Impact of tisagenlecleucel product attributes on clinical outcomes in pediatric and young adult patients with relapsed or refractory acute lymphoblastic leukemia (r/r ALL)

Abstract Background Tisagenlecleucel CAR-T cell therapy showed durable clinical outcomes in pediatric/young adult patients with r/r ALL in ELIANA/ENSIGN trials. We examined the effect of tisagenlecleucel product attributes on clinical outcomes (Table). Methods Key product attributes of tisagenlecleucel were examined with univariate and multivariate (lasso, random forest) approaches (N=143) to evaluate the effects on efficacy (best overall response [BOR] ≤3 months, duration of remission, event-free survival [EFS], and safety (cytokine release syndrome [CRS] and neurological events). Results Univariate: High in vitro IFNγ release positively correlated with BOR and EFS. Relative and absolute number (no.) of early differentiated (ED) CD4+ and CD8+, activated (HLA-DR+CD38+) CD8+, and absolute no. of naïve CD4+ cells positively correlated with efficacy. Total CD4+ cells with central memory markers, LAG3+, CD25+ had weak positive correlation with grade 3/4 CRS. Relative no. of effector memory CD4+ and KLRG1+CD8+ and senescent CD4+ and CD8+ cells negatively correlated with efficacy. Median T-cell transduction efficiency by flow (23%; range, 1.5%-56%) and viability did not correlate with efficacy or safety. Also, the median CD4:CD8 cell ratio did not correlate with outcomes. Multivariate: ED CD4+ and CD8+, and activated CD8+ cells positively correlated with BOR. Naïve CD4+ cells positively correlated with survival, while KLRG1+CD8+ cells and effector memory CD4+ cells negatively correlated with survival. Conclusions Less mature T cells trended toward improved efficacy, while senescent and differentiated cells led to worse outcomes. No strong correlations to safety outcomes were observed. While phenotypes do not definitively distinguish outcomes, the analyses provide insight into the potential set of parameters to be used in predictive/causal modeling. Table. Univariate analysis of tisagenlecleucel product attributes and clinical outcomesProduct Attributes (CAR+ Cells)Correlation,a +/–BORDOREFSCRS G3/4NE G3/4IFNγ++++Transduction efficiency+Total CD4++++Total CD8+++++ED CD4+ cells+++++++ED CD8+ cells++++++TIM3+ CD4+ cellsTIM3+ CD8+ cells+++Activated CD4+ cells+++Activated CD8+ cells++++++HLA-DR+ CD4+ cells+HLA-DR+ CD8+ cells++++Naïve CD4+ cells++++++Naïve CD8+ cells+LAG3+ CD4+ cells++CD25+ CD4+ cells++Effector memory CD4+ cells– –– – –KLRG1+ CD4+ cells– –KLRG1+ CD8+ cells– – –– –– –Senescent CD4+ cells– – –Senescent CD8+ cells–Key: +++/– – –, P value <0.01; ++/– –, P value <0.05; +/–, P value <0.1 P values refer to Cox regression or Wilcoxon tests. Note: Central memory CD4+ and CD8+, LAG3+ CD8+, CD25+ CD8+ cells and CD4:CD8 ratio were analyzed but showed no significance. aComparison of ≤median to >median. BOR, best overall response; CAR, chimeric antigen receptor; CD, cluster of differentiation;CRS, cytokine release syndrome; DOR, duration of remission; ED, early differentiated;EFS, event-free survival; G, grade; HLA-DR, human leukocyte antigen-DR; IFN, interferon; KLRG1, Killer cell lectin-like receptor G1; LAG3, lymphocyte-activation gene 3; NE, neurological events; TIM3, T-cell immunoglobulin and mucin domain-3. Citation Format: Irina Gershgorin, Edward R. Waldron, Stephan A. Grupp, John E. Levine, Michael A. Pulsipher, Stella M. Davies, G. Doug Myers, Margit Jeschke, Boris Engels, Chris del Corral, Andre Baruchel. Impact of tisagenlecleucel product attributes on clinical outcomes in pediatric and young adult patients with relapsed or refractory acute lymphoblastic leukemia (r/r ALL) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 510.