Dulanermin-a non-polyhistidine-tagged soluble recombinant human apoptosis ligand 2 (Apo2L) or tumour-necrosis-factor-related apoptosis-inducing-ligand (TRAIL)-has pro-apoptotic activity in a range of cancers and synergistic preclinical activity with rituximab against lymphoma in vivo. We aimed to assess the safety, pharmacokinetics, and efficacy of dulanermin and rituximab in patients with relapsed indolent B-cell non-Hodgkin lymphoma. We did an open-label phase 1b/2 randomised study. Four study centres in the USA enrolled patients into phase 1b, and 27 study centres in the USA, Italy, Australia, France, Czech Republic, New Zealand, and Poland enrolled patients into phase 2. In phase 1b, patients (age ≥18 years) with indolent B-cell non-Hodgkin lymphoma with stable disease or better lasting at least 6 months after the most recent rituximab-containing regimen were included. In phase 2, patients (age ≥18 years) with follicular lymphoma grades 1-3a were included. In phase 1b, patients received 4 mg/kg or 8 mg/kg intravenous dulanermin on days 1-5 of up to four 21-day cycles and intravenous rituximab 375 mg/m(2) weekly for up to eight doses. In phase 2, patients were randomly assigned (1:1:1) centrally by an interactive voice response system to dulanermin (8 mg/kg for a maximum of four 21-day cycles), rituximab (375 mg/m(2) weekly for up to eight doses), or both in combination, stratified by baseline follicular lymphoma International Prognostic Index (0-3 vs 4-5) and geographic site (USA vs non-USA). The primary endpoints of the phase 1b study were the safety, tolerability, and pharmacokinetics of dulanermin with rituximab. The primary endpoint of phase 2 was the proportion of patients who achieved an objective response. All patients who received any dose of study drug were included in safety analyses. Efficacy analyses were per protocol. Treatment was open label; all patients and investigators were unmasked to treatment allocation. This study is registered with ClinicalTrials.gov, NCT00400764. Between June 6, 2006, and Feb 15, 2007, 12 patients were enrolled in phase 1b, and between April 4, 2007, and April 20, 2009, 60 patients were enrolled in phase 2, of whom 59 were included in safety analyses and 58 in efficacy analyses. No dose-limiting toxic effects were noted in phase 1b. The most common grade 1-2 adverse events in phase 1b were fatigue (nine; 75%), rash (five; 42%), and chills, decreased appetite, diarrhoea, and nausea (four each; 33%). 19 grade 3 or higher adverse effects were noted in five (42%) patients, with 14 occurring in one patient. After treatment with 8 mg/kg of dulanermin, in six patients the mean serum peak concentration was 80 μg/mL, dropping below the minimum detectable concentration (2 ng/mL) within 24 h after the dose. The mean steady state peak and trough concentrations of rituximab were 461 μg/mL (SD 97.5) and 303 μg/mL (92.8), respectively. In phase 2, eight (14%) of 59 patients experienced 12 grade 3 or higher adverse events. In phase 2, objective responses were noted in 14 of 22 (63.6%, 95% CI 41.8-81.3) patients treated with rituximab only, 16 of 25 (64.0%, 43.1-81.5) treated with dulanermin and rituximab, and one of 11 (9.1%, 0.5-39.0) treated with dulanermin only. The study was terminated early, on May 5, 2010, because of an absence of efficacy in the combination group. The addition of dulanermin to rituximab in patients with indolent B-cell non-Hodgkin lymphoma was tolerable but did not lead to increased objective responses. This combination is not being developed further in non-Hodgkin lymphoma. Genentech and Amgen.