Role of B Cell Recovery in Relapse Risk and CD19 Phenotype Following Real-World Use of Tisagenlecleucel for Pediatric B-Acute Lymphoblastic Leukemia: A Multi-Institutional Retrospective Study

Purpose:

Relapse is a fundamental challenge following tisagenlecleucel (CD19 CAR T cells; tisa-cel) in children and young adults (CAYA) with relapsed/refractory B-acute lymphoblastic leukemia (B-ALL). B cell aplasia (BCA) is a surrogate marker of tisa-cel persistence, and early B cell recovery (BCR <6 months) has been associated with an increased risk of relapse on Phase I-II clinical trials. The impact of late BCR (≥ 6 months) on relapse is less well defined, and the impact of early and late BCR on relapse risk and relapse immunophenotype with real-world data is not known.

Methods:

This multi-institutional retrospective study of CAYA infused with commercial tisa-cel included patients from Pediatric Real World CAR Consortium (PRWCC) centers (N=207), treated between 2017 and 2024. BCR was defined as return of normal CD19-positive (CD19+) B cells in the peripheral blood or bone marrow hematogones, with thresholds determined by institution. Loss of BCA was defined as the first occurrence of any CD19+ event (BCR or CD19+ relapse). Duration of BCA was defined as days from infusion without loss of BCA, censored for additional leukemia-directed therapy and CD19-negative (CD19-) or CD19 unknown relapse, with median duration of BCA determined by the Kaplan-Meier (KM) method. Cumulative incidence functions were separately performed for loss of BCA and CD19+ relapse, considering CD19- or CD19 unknown relapse and therapy in remission as competing events. Relapse-free survival (RFS) analyses via KM were performed for overall, CD19+, and CD19- relapses separately, censoring for therapy in remission and non-relapse death. Landmark analyses for RFS at landmark times of day + 90, 180, 270, and 365 days were performed, comparing groups with and without BCR by each landmark time. BCR was evaluated as a time-varying covariate (TVC) for CD19+ and CD19- relapse in a Cox proportional hazard model. Therapy for BCR while in remission, defined as tisa-cel re-infusion, alternative CAR T product, HSCT, or other cancer-directed therapy, was evaluated as a TVC for subsequent relapse. Event-free survival (EFS) was evaluated using the extended KM method in those who received versus did not receive treatment for BCR including death and relapse as events.

Results:

Of 207 patients, 178 (86%) had a complete response at 1-month with 94% of responders achieving flow MRD-negative status. The 1- and 2-year RFS in responders were 62% and 53%. Cumulative incidence of any relapse at 2 years was 45% (21% CD19+, 20% CD19-, and 4% CD19 unknown). Among 156 responders with reported B cell assessments, 75 experienced loss of BCA (including 9 who had CD19+ relapse without documented BCR) with median BCA duration of 262 days (95% CI 174, NA) with censoring and 396 days (95% CI 192, NA) without censoring for non-CD19+ relapse or further therapy. Cumulative incidence of loss of BCA at day 90, day 180, 1 year, and 2 years 22%, 38%, 45%, and 50% respectively. Of the 66 patients who experienced BCR, 80% (N=53) occurred early (<180 days) and 83% (N=55) occurred before day of relapse. In landmark analyses, CD19+ RFS was significantly worse for those with BCR prior to landmarks of day 90, 180, 270, and 365, compared to those without BCR by each timepoint (p<0.001 for all). No significant differences in CD19- RFS were observed across landmark times. Similarly, when evaluated as a TVC, BCR within 1 year of infusion increased risk of CD19+ relapse (HR 6.34; 95% CI 2.79, 14.45; p<0.001), but not CD19-relapse (HR 0.83; 0.20-3.68; p=0.80). Therapy given in remission for BCR, evaluated as a TVC, was associated with lower risk of CD19+ relapse [HR 0.27; 95% CI 0.10, 0.73, p=0.01] and significantly improved EFS compared to those who did not receive therapy (median EFS not reached vs. 181 days from BCR, p=0.007).

Conclusion:

In the real-world commercial tisa-cel setting, we show a relatively shorter duration of BCA when compared with the ELIANA trial, where the median time to BCR was 35.3 months from the time of remission. Our work confirms the prognostic significance of B cell recovery for CD19+ but not CD19- relapse and suggests BCR between 6 months to 1 year post-infusion continues to be associated with a risk of CD19+ relapse. Finally, we show that leukemia-directed therapy after BCR was associated with lower risk of CD19+ relapse and improved EFS. The continued risk of CD19+ relapse for patients with BCR between 6 months to 1 year raises questions of optimal management of BCR in this window.