Work

Car‐hematotox: a Discriminative Model for Car T‐cell Related Hematotoxicity in Relapsed/refractory Large B‐cell Lymphoma

Year: 2021

Type: article

Abstract: Background: Hematotoxicity represents a frequent CAR T-cell related adverse event that remains poorly understood. Prolonged cytopenia and aplastic courses can predispose for severe infectious complications, which represent the most common cause of non-relapse mortality after CAR T-cell therapy. Methods: In this multicenter retrospective analysis, we studied patterns of hematopoietic reconstitution and evaluated potential predictive markers in 258 patients receiving Axicabtagene ciloleucel (Axi-cel, n = 170) or Tisagenlecleucel (Tisa-cel, n = 88) for relapsed/refractory large B-cell lymphoma. Cytopenia was graded according to CTCAE criteria with further sub-characterization of neutropenia as severe (ANC < 500/µl), profound (ANC < 100/µl), protracted (≥7 days) and prolonged (ANC < 1000/µl after day 21). We studied the impact of baseline ( = prior to lymphodepletion) clinical and laboratory markers on hematotoxicity by uni- & multivariate analysis using a binary logistic regression analysis for the outcome of severe neutropenia ≥14 days ( = Test group) vs. <14 days ( = Control group). A discriminative model termed CAR-HEMATOTOX was developed. Receiver operator characteristic (ROC) analyses were used to assess measures of performance (sensitivity, specificity, AUC). Results: Across all studied patients, we observed a high incidence of profound and prolonged neutropenia of 72 and 64% respectively (see Table). The incidence of severe neutropenia was 91% with a median duration of 9 days (95% CI: 8 – 10 days). Uni- and multivariate analysis revealed an association between markers associated with hematopoietic reserve (e.g. platelet count, ANC, hemoglobin), baseline inflammatory state (e.g. ferritin, CRP), and longer duration of neutropenia. In contrast, neither CRS/ICANS grade nor peak cytokine levels were associated with long neutropenia. To develop the CAR-HEMATOTOX model in our training cohort (n = 58), the discriminatory threshold for each variable significantly associated with hematotoxicity was determined on the basis of optimizing the respective Youden J statistic. Additional weighting of factors identified by stepwise multivariate analysis using a backward elimination was applied. We validated the model in two independent cohorts from Europe (n = 91) and the USA (n = 109) and identified patients with severe neutropenia ≥14 days in both (AUC = 0.77-0.91, P < 0.001, Sensitivity 84-100%, Specificity 67-69%). A high CAR-HEMATOTOX score resulted in a longer duration of neutropenia (12 vs. 5.5 days, P < 0.001) and discriminated for patients with aplastic features. Conclusion: These findings underline the strikingly high incidence of hematotoxicity after CAR T-cell therapy. The CAR-HEMATOTOX will be useful for risk-adapted management of hematotoxicity, to guide prophylactic antimicrobial strategies, and to identify patients for potential autologous stem cell back-up. Keywords: Risk Models, Cellular therapies, Aggressive B-cell non-Hodgkin lymphoma No conflicts of interests pertinent to the abstract.

Source: Hematological Oncology

DOI: 10.1002/hon.82_2879