Background: Obinutuzumab with polatuzumab vedotin or lenalidomide showed tolerability and activity in phase 1b/2 trials that recruited patients with relapsed or refractory follicular lymphoma. We aimed to examine whether the novel polatuzumab vedotin-obinutuzumab-lenalidomide (Pola-G-Len) combination might enhance antitumour response in patients with relapsed or refractory follicular lymphoma. Methods: This multicentre, single-arm phase 1b/2 study tested Pola-G-Len in patients with relapsed or refractory follicular lymphoma, and polatuzumab vedotin in combination with rituximab and lenalidomide in patients with relapsed or refractory diffuse large B-cell lymphoma. Here we report the results from the cohort of patients with follicular lymphoma. The trial was done in 18 cancer centres across three countries (Spain, UK, and USA). Patients (≥18 years old) with CD20-positive relapsed or refractory follicular lymphoma (excluding grade 3b) and Eastern Cooperative Oncology Group performance status of 2 or less who had previously received anti-CD20-containing chemotherapy were eligible for inclusion. During the dose-escalation phase, patients received six 28-day cycles of induction treatment with intravenous obinutuzumab 1000 mg (all cohorts), and intravenous polatuzumab vedotin and oral lenalidomide (Celgene, Summit, NJ, USA) in the following doses: 1·4 mg/kg polatuzumab vedotin and 10 mg lenalidomide (cohort 1); 1·8 mg/kg polatuzumab vedotin and 10 mg lenalidomide (cohort 2); 1·4 mg/kg polatuzumab vedotin and 15 mg lenalidomide (cohort 3); 1·8 mg/kg polatuzumab vedotin and 15 mg lenalidomide (cohort 4); 1·4 mg/kg polatuzumab vedotin and 20 mg lenalidomide (cohort 5); and 1·8 mg/kg polatuzumab vedotin and 20 mg lenalidomide (cohort 6). Polatuzumab vedotin was administered on day 1, lenalidomide on days 1–21, and obinutuzumab on days 1, 8, and 15 of cycle one and day 1 of cycles two to six of each 28-day cycle. During phase 2 (dose expansion phase), patients received six cycles of induction with Pola-G-Len at the recommended phase 2 dose established during dose-escalation. Patients who had a response or stable disease at the end of induction were eligible to enter the maintenance phase, in which they received obinutuzumab for 24 months at 1000 mg on day 1 of every other 28-day cycle for a total of 12 doses, and lenalidomide for 12 months at 10 mg on days 1–21 of each 28-day cycle for a maximum of 12 cycles. The primary activity endpoint was complete response at the end of induction. Adverse events were monitored throughout the study. The primary safety objective was to determine the maximum tolerated dose of Pola-G-Len. Analyses were in the safety population, which included all patients that received at least one dose of any of the component drugs (ie, all patients who entered the induction phases in both the escalation and expansion phases), and activity-evaluable population, which included all patients who received at least one dose of any of the component drugs at the recommended phase 2 dose (ie, all patients who received the recommended phase 2 dose in the dose escalation investigation and all patients who entered induction in the dose expansion investigation). This ongoing trial is registered at ClinicalTrials.gov, NCT02600897. Findings: Between March 24, 2016, and August 23, 2018, 56 patients (33 [59%] men and 23 [41%] women; 49 [88%] non-Hispanic or Latino) were enrolled. Two of four patients in cohort 2 reported dose-limiting toxicity events during dose escalation (one patient had grade 4 amylase and lipase elevation and one patient had grade 4 neutropenia and grade 3 thrombocytopenia), and there were no dose-limiting toxicities observed in cohorts 3 or 5; therefore, the recommended phase 2 dose for the dose-expansion was 1·4 mg/kg polatuzumab vedotin plus 20 mg lenalidomide. 46 (82%) patients were included in the activity-evaluable population. After a median follow up of 26·7 months (IQR 22·2–31·3) the objective response rate was 76% (90% CI 64–86) and complete response rate was 63% (90 CI 50–75). After a median follow-up of 27·0 months (IQR 18·7–34·0), the most common grade 3–4 adverse events were neutropenia (31 [55%] of 56 patients) and thrombocytopenia (14 [25%] patients). 61 serious adverse events were reported in 35 (63%) patients; the most common of which were febrile neutropenia (five [9%] patients; a sixth patient had febrile neutropenia, but this was not considered serious by the investigator), pneumonia (four [7%] patients), and pyrexia (four [7%] patients). One fatal adverse event (grade 5 septic shock) occurred in a patient who had discontinued study treatment due to disease progression and had initiated a new anti-lymphoma tyrosine kinase inhibitor treatment. This was not considered related to study treatment by the investigator. Interpretation: Pola-G-Len showed high complete response rates, although it did not reached the prespecified threshold for activity, in patients who were heavily pretreated with refractory follicular lymphoma. Our findings compare favourably with available therapies and support future investigation of Pola-G-Len in a larger patient population. Funding: Genentech/F Hoffmann-La Roche.