Primary Analysis of ZUMA-12: A Phase 2 Study of Axicabtagene Ciloleucel (Axi-Cel) As First-Line Therapy in Patients with High-Risk Large B-Cell Lymphoma (LBCL)

Background: High-risk LBCL is associated with poor prognosis after first-line anti-CD20 mAb-containing regimens, highlighting the need for novel treatments. Axi-cel, an autologous anti-CD19 chimeric antigen receptor (CAR) T-cell therapy, is approved for treatment of relapsed/refractory (R/R) LBCL after ≥2 lines of systemic therapy. Here we report the primary analysis of ZUMA-12, a Phase 2, multicenter, single-arm study of axi-cel as part of first-line therapy in patients with high-risk LBCL.

Methods: Eligible adults had high-risk LBCL, defined by histology (double‑ or triple-hit status [MYC and BCL2 and/or BCL6 translocations] per investigator) or an IPI score ≥3, plus a positive interim PET per Lugano Classification (Deauville score [DS] 4/5) after 2 cycles of an anti-CD20 mAb and anthracycline-containing regimen. Patients underwent leukapheresis and received conditioning chemotherapy (cyclophosphamide and fludarabine) followed by a single axi-cel infusion at 2×10⁶ CAR T cells/kg. Non-chemotherapy bridging could be administered before conditioning per investigator discretion. The primary endpoint was investigator-assessed complete response (CR) rate per Lugano. Secondary endpoints included objective response rate (ORR; CR + partial response), duration of response (DOR), event-free survival (EFS), progression-free survival (PFS), overall survival (OS), incidence of adverse events (AEs), and levels of CAR T cells in blood and cytokines in serum. The primary analysis occurred after all treated patients had ≥6 months of follow-up.

Results: As of May 17, 2021, 42 patients were enrolled and 40 were treated with axi-cel. Median age was 61 years (range, 23–86); 68% of patients were male, 63% had ECOG 1, 95% had stage III/IV disease, 48% had DS4, 53% had DS5, 25% had double‑ or triple-hit status per central assessment, and 78% had IPI score ≥3.

A total of 37 patients had centrally confirmed double‑ or triple-hit histology or an IPI score ≥3 and were evaluable for response, with 15.9 months of median follow-up (range, 6.0–26.7). The CR rate was 78% (n=29; 95% CI, 62–90); 89% of patients had an objective response, and median time to initial response was 1 month. Among all 40 treated patients, 90% had an objective response (80% CR rate). At data cutoff, 73% of response-evaluable patients had ongoing responses. Medians for DOR, EFS, and PFS were not reached; 12-month estimates were 81%, 73%, and 75%, respectively. The estimated OS at 12 months was 91%.

All 40 treated patients had AEs of any grade; 85% of patients had Grade ≥3 AEs, most commonly cytopenias (68%). Grade ≥3 cytokine release syndrome (CRS) and neurologic events (NEs) occurred in 3 patients (8%) and 9 patients (23%), respectively. Median times to onset of CRS and NEs were 4 days (range, 1–10) and 9 days (range, 2–44), with median durations of 6 days and 7 days, respectively. All CRS and most NEs (28/29) of any grade resolved by data cutoff (1 ongoing Grade 1 tremor); 39/40 CRS events resolved by 14 days post-infusion and 19/29 NEs resolved by 21 days post-infusion. Tocilizumab was administered to 63% and 3% of patients for management of CRS or NEs, respectively; corticosteroids were administered to 35% and 33% of patients for CRS and NE management. One Grade 5 event of COVID-19 occurred (Day 350).

Median peak CAR T-cell level in all treated patients was 36 cells/µL (range, 7–560), and median expansion by AUC_0–28 was 495 cells/µL × days (range, 74–4288). CAR T-cell levels peaked at a median of 8 days post-infusion (range, 8–37). Higher frequency of CCR7+CD45RA+ T cells in axi-cel product, previously associated with greater expansion of CAR T cells (Locke et al. Blood Adv. 2020), was observed in ZUMA-12, compared with the ZUMA-1 study in R/R LBCL (Neelapu et al. New Engl J Med. 2017).

Conclusion: In the primary analysis of ZUMA-12, axi-cel demonstrated a high rate of rapid and complete responses in patients with high-risk LBCL, a population with high unmet need. With 15.9 months of median follow-up, responses were durable as medians for DOR, EFS, and PFS were not yet reached and over 70% of patients remained in response at data cutoff. No new safety signals were reported with axi-cel in an earlier line. Overall, axi-cel may benefit patients exposed to fewer prior therapies, and further trials in first-line high-risk LBCL are warranted to assess axi-cel in this setting.