First-line consolidation with cemacabtagene ansegedleucel (cema-cel) in patients with large B-cell lymphoma (LBCL) and minimal residual disease (MRD) after response to standard therapy: The pivotal, randomized, open‑label Phase 2 ALPHA3 study

Abstract Background and Significance: R-CHOP as first-line (1L) therapy for LBCL has a cure rate of ~60%. However, ~10% of patients (pts) have refractory LBCL (Coiffier B et al, NEJM 2002) and ~30% of responders relapse within 2 years (Padala SA & Kallam A, In StatPearls. StatPearls Publishing; 2023). Autologous CAR T-cell therapies have been revolutionary in the treatment of relapsed/refractory (R/R) LBCL and they are considered standard second-line treatment but might not be an option due to aggressive disease, pt comorbidity, access barriers, or manufacturing issues/delays. Identifying responders to 1L therapy who are at high risk of relapse and rapidly administering an off-the-shelf CAR T-cell therapy for remission consolidation may improve outcomes. Presence of circulating tumor DNA (ctDNA)–based MRD, measured by an ultrasensitive MRD test at the end of 1L therapy, is highly prognostic for relapse. Cema-cel is an immediately available, off-the-shelf, HLA-unmatched allogeneic CD19 CAR T-cell product that utilizes Cellectis technologies has shown potent antitumor activity and manageable safety in phase 1 studies of pts with R/R LBCL and is a promising agent for consolidation in this treatment setting. The pivotal, randomized, open-label, phase 2 ALPHA3 study (NCT06500273) was designed to evaluate the efficacy and safety of consolidation with cema-cel compared with standard-of-care (SOC) observation in pts with LBCL who are in response after 1L immunochemotherapy but have detectable MRD by ctDNA-based testing. We report the updated study design.Study Design and Methods: Adults with histologically confirmed diffuse LBCL (DLBCL; includes DLBCL not otherwise specified, Epstein-Barr virus–positive DLBCL, DLBCL with IRF4/MUM1 rearrangement), high-grade B-cell lymphoma (HGBCL; includes HGBCL not otherwise specified, or with MYC and BCL2 and/or BCL6 rearrangements), or primary mediastinal B-cell lymphoma per WHO 2017 classification and ≥1 of the following clinical criteria at diagnosis—International Prognostic Index score of 2-5; Ann Arbor stage III/IV disease; and history of equivocal response at interim or end-of-therapy positron emission tomography (PET)/computed tomography (CT)—will be prescreened. Eligible pts must have completed a full course of standard 1L therapy that included an anthracycline and an anti-CD20 monoclonal antibody; achieved complete response (CR) or partial response (PR) suitable for observation at the end of 1L therapy per PET/CT evaluation by Lugano 2014 criteria and without evidence of progression by randomization; MRD positivity per ctDNA-based testing; Eastern Cooperative Oncology Group performance status score of 0 or 1; and adequate organ function. The study consists of a 2-part seamless design. In Part A, pts will be randomly assigned 1:1 to SOC observation or to cema-cel (120×106 CAR T cells) after a 3-day LD with fludarabine (30 mg/m2/day) and cyclophosphamide (300 mg/m2/day) (FC). Second cema-cel arm that utilized 3-day LD with FC plus the anti-CD52 monoclonal antibody ALLO-647 (30 mg/day) was closed in August 2025 because of a grade 5 hepatic failure event caused by a disseminated adenovirus infection that was attributed to ALLO-647. Part A will conclude with interim safety and surrogate biomarker–based efficacy analyses. In Part B, pts will continue to be randomly assigned 1:1 to cema-cel after FC or SOC observation. Randomization in Parts A and B will be stratified by best response to 1L therapy (CR vs PR). The primary endpoint is event-free survival per Lugano 2014 criteria by independent review committee (IRC), with hierarchical testing of key secondary end points of progression-free survival per Lugano 2014 criteria by IRC and overall survival. Other secondary end points include rate of MRD clearance and safety of cema-cel. Enrollment in Part A is ongoing. Pts randomly assigned to the treatment arm or followed up in observation during Part A will be included in the inferential testing in Part B. Approximately 110 pts will be enrolled in each arm across academic- and community-based centers. The study was initiated in June 2024.