CD19/22 CAR T-cells in children and young adults with B-ALL: Final Phase 1 trial results

Abstract

Abstract Introduction: Initial experience from the dose-escalation portion of our phase 1 bivalent CD19.22.BBζ in children and young adults (CAYA) with B-cell malignancies illustrated safety, identified a recommended phase 2 expansion dose [RP2D], and highlighted limitations in CD22 targeting compared to single antigen constructs. (Shalabi et al, Blood 2022) We now report on results for patients treated at the expansion dose (NCT03448393). Methods: This was a single center phase 1 trial of autologous CD19.22.BBζ CAR T-cells in CAYA with relapsed/refractory CD19/CD22 positive B-cell malignancies. All patients treated at the RP2D (3x106 transduced CAR T-cells/kg) were included in the efficacy analysis (inclusive of those in dose-escalation) and stratified as either CAR naïve (or had interval hematopoietic stem cell transplant [HSCT] after prior CAR) or CAR pretreated. CAR naïve patients received fludarabine 75mg/m2 and cyclophosphamide 900mg/m2 for lymphodepletion (LD), whereas CAR pretreated patients received intensified LD with fludarabine 120mg/m2 and cyclophosphamide 1200mg/m2. During the expansion phase, eligibility was broadened to include an isolated CNS (iCNS) disease cohort and to test the efficacy of siltuximab as first line therapy for cytokine release syndrome (CRS). Data cut off was February 15, 2025. Results: Thirty CAYA were treated at the RP2D; 28 with B-ALL and 2 with Burkitt lymphoma. With a primary analysis on B-ALL, the median age was 21.3 years (range 6.3-38.1) and 54% (n=15) were male. Patients had received a median of 4 (range 2-9) prior lines of therapy; 50% (n=14) had received prior blinatumomab, 18% (n=5) received prior CAR T-cells, and 43% (n=12) had prior HSCT. The CAR naïve cohort comprised the majority (n=25), 2 patients were CAR pretreated, and 1 had iCNS disease. At infusion 39% (n=11) of B-ALL patients had high bone marrow (BM) disease (> 5% marrow blasts), all were CNS1, and 7 (25%) had non-CNS extramedullary disease (EMD). A single patient with isolated CNS2 disease at enrollment was CNS1 at infusion. Twenty patients (71%) developed CRS at a median of 4 days (range 0-17) post infusion; only 2 (10%) were grade > 3. Seven (25%) required both tocilizumab and steroids. One patient received siltuximab as first-line for CRS with full resolution after 2 doses. Grade 3 ICANS occurred in 3 (11%) patients, all of whom received intrathecal and systemic steroids with full resolution. Immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome was not seen. At one-month restaging, 25 (89%) achieved an MRD negative complete remission (CR), 6 without experiencing any CRS or ICANS. Prior CAR had no impact on outcomes; all 5 with and without an interval HSCT achieved an MRD negative CR. EMD fully cleared in 4 of 7 (57%). The only 3 non-responders had persistent EMD, including one who developed lineage switch with B-ALL eradication. Twenty-three of 25 (92%) achieving CR proceeded to consolidative HSCT at a median of 51 days (range: 45-68 days) post CAR infusion; among 9 this was a second HSCT. Median relapse free survival for those achieving CR (n=25) and OS (n=28) were both 34 months (95% CI 13-infinity and 14-infinity, respectively). With a median follow up of 30.5 months (range 8-65 months), 16 remain alive. Of the 2 patients with Burkitt lymphoma one experienced progressive disease and the other had a transient CR. CAR T-cell expansion was detected in all 28 B-ALL patients by peripheral blood (PB) flow cytometry, with peak expansion at a median of 7 days (range 6-15). Expansion was higher in patients with CRS (median 75 CAR/mcL vs 15.9 CAR/mcL, p=0.016) and in patients with high BM disease (median 92.2 CAR/mcL in patients with >5% vs median 21.3 CAR/mcL in patients with <5%, p=0.002). CAR T-cells were detected in the BM at day 28 in 19 of 27 patients (70.4%). At day 28, 24 of 25 responders had PB flow available, and all demonstrated B-cell aplasia (CD19<10/mcl). Conclusions: The extended experience with this construct highlights safety and efficacy warranting further exploration of these properties in future constructs. Non-response occurred only in patients with non-CNS EMD, highlighting ongoing challenges in targeting EMD. Institutional efforts have shifted to test a bicistronic CD19.28ζ /CD22.BBζ construct with initial results demonstrating enhanced efficacy, CD22 targeting, and persistence (NCT05442515).