Résumé
Background: Glofitamab monotherapy induces durable remission in patients with relapsed or refractory diffuse large B-cell lymphoma after two or more previous therapies, but has not previously been assessed as a second-line therapy. We investigated the efficacy and safety of glofitamab plus gemcitabine-oxaliplatin (Glofit-GemOx) versus rituximab (R)-GemOx in patients with relapsed or refractory diffuse large B-cell lymphoma.Methods: The phase 3, randomised, open-label STARGLO trial was done at 62 centres in 13 countries in Asia and Australia, Europe, and North America. We recruited transplant-ineligible patients (aged ≥18 years) with histologically confirmed relapsed or refractory diffuse large B-cell lymphoma after one or more previous therapies. Patients were randomly assigned in permuted blocks (block size of six) via an interactive voice or web response system (2:1; stratified by 1 vs ≥2 previous lines of therapy and relapsed vs refractory status) to Glofit-GemOx (intravenous gemcitabine 1000 mg/m2 and oxaliplatin 100 mg/m2 plus glofitamab step-up dosing to 30 mg; for a total of eight cycles, plus four additional cycles of glofitamab monotherapy) or R-GemOx (intravenous gemcitabine 1000 mg/m2 and oxaliplatin 100 mg/m2 plus rituximab 375 mg/m2; for a total of eight cycles). The trial independent review committee, which evaluated all response-based endpoints, was masked to treatment assignment. The primary endpoint was overall survival. Efficacy analyses were by intention to treat in all randomly assigned patients. We present results from both the primary analysis (cutoff: March 29, 2023) and updated analysis after all patients had completed study therapy (cutoff: Feb 16, 2024). Safety analyses included all patients who received any study treatment. This study is registered with ClinicalTrials.gov, NCT04408638, and is ongoing (closed to recruitment).Findings: From Feb 23, 2021, to March 14, 2023, 274 patients were enrolled and randomly assigned to receive Glofit-GemOx (n=183) or R-GemOx (n=91). 158 (58%) patients were male and 116 (42%) were female; median age was 68 years (IQR 58-74). At the primary analysis after a median follow-up of 11·3 months (95% CI 9·6-12·7), overall survival was significantly improved with Glofit-GemOx versus R-GemOx (median not estimable [NE; 95% CI 13·8 months-NE] vs 9·0 months [7·3-14·4]; hazard ratio [HR] 0·59 [95% CI 0·40-0·89]; p=0·011). At the updated analysis after a median follow-up of 20·7 months (19·9-23·3), a consistent improvement in overall survival was observed with Glofit-GemOx versus R-GemOx (median 25·5 months [18·3-NE] vs 12·9 months [7·9-18·5]; HR 0·62 [0·43-0·88]). In the safety sets, 180 (100%) patients in the Glofit-GemOx group and 84 (96%) of 88 patients in the R-GemOx group had at least one adverse event during the study period. Cytokine release syndrome occurred in 76 (44%) of 172 glofitamab-exposed patients and was predominantly low grade. Deaths related to glofitamab or rituximab occurred in five (3%) patients in the Glofit-GemOx group and in one (1%) patient in the R-GemOx group.Interpretation: Glofit-GemOx had a significant overall survival benefit compared with R-GemOx, supporting its use in transplant-ineligible patients with relapsed or refractory diffuse large B-cell lymphoma after one or more previous lines of therapy.