Introduction: Treatment options for advanced MZL include the anti-CD20 antibody rituximab (R) with chemotherapy (chemo). In the randomised Phase III GALLIUM trial (NCT01332968), a preplanned efficacy interim analysis in 1202 previously untreated FL pts demonstrated that obinutuzumab (GA101; G) plus chemo prolonged PFS relative to R-chemo (primary endpoint; Marcus 2016). MZL pts were also enrolled to assess whether treatment effects are consistent with FL pts; the analysis (cut-off date, 31 January 2016) was not powered to detect PFS differences. Methods: Pts were aged ≥18 yrs with previously untreated MZL (prior antibiotic/antiviral, surgery or radiation allowed), advanced disease (stage III/IV or stage II with tumour diameter ≥ 7 cm), ECOG PS 0–2, and requiring treatment according to the investigator (INV). Chemo regimens (CHOP, CVP or bendamustine [B]) were allocated by pt. Pts were randomised 1:1 (stratified by chemo, IPI group and geographical region) to R 375 mg/m2 on day (D) 1 of each cycle (C) or G 1000 mg on D1, 8 and 15 of C1 and D1 of subsequent cycles, for 6 or 8 cycles depending on chemo. Pts with CR or PR at end of induction (EOI; modified Cheson 2007) continued to receive R or G every 2 months for 2 yrs or until progression. Results: 195 MZL (nodal, 66; extranodal, 61; splenic, 68) pts were randomised: G-chemo (n = 99; median age 63 yrs, 55% male), R-chemo (n = 96, 62 yrs, 46% male). Most pts (83%) were Ann Arbor stage IV at diagnosis; 49% were IPI high risk. At baseline, extranodal involvement, bone marrow involvement, bulky disease and B-symptoms were more common in G-chemo pts. In each arm, 88 pts received all induction cycles; chemo received was B (71% pts), CHOP (16%) and CVP (12%); 56 (G-chemo) and 57 (R-chemo) pts received 2 yrs' maintenance. After 38.4 months' median observation time, there was no clinically relevant difference in INV-assessed PFS between the study arms (HR, 0.82; 95% CL 0.45–1.46; p = 0.49); 3-year PFS rates were 75% in G-chemo pts and 78% in R-chemo pts. HRs for other time-to-event endpoints (including OS) were consistent with the INV-assessed PFS outcome (Table). CR and ORR at EOI were similar for the two arms (Table). More pts in the G-chemo than the R-chemo group had grade 3–5 AEs, SAEs and fatal AEs (Table). *From earlier of randomisation date or treatment start date to last date pt known to be alive. †Log rank test, stratified for chemotherapy regimen and IPI risk group. ‡Assessed by an independent radiology and oncology review committee (IRC). §Three pts randomised to R-chemo received G (n=2) or no antibody (n=1). ¶Grade ≥3 AEs with >5% incidence in either treatment arm. **All events in MedDRA System Organ Class ‘Infections and Infestations’. ††Standardised MedDRA query for malignant or unspecified tumours >6 months after treatment. Conclusions: The current analysis of this large RCT in previously untreated MZL pts did not show a clinically relevant difference in PFS between arms. G-chemo was associated with a higher frequency of grade 3–5 AEs, SAEs and fatal AEs. Safety by chemo will be presented. Keywords: marginal zone lymphoma (MZL); obinutuzumab; rituximab.