Introduction: Rituximab (R) plus chemotherapy (chemo) is standard of care for patients (pts) with diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL). A subcutaneous (SC) formulation of R (R-SC) that has a non-inferior pharmacokinetic profile to the intravenous (IV) formulation (R-IV) and similar efficacy and safety in DLBCL and FL pts is now available for clinical use. The randomized, open-label, crossover, Phase 3b PrefMab study (NCT01724021) aimed to explore pt preference and treatment satisfaction for the SC vs IV formulation of R. In the primary analysis (data cut-off, January 19, 2015; Rummel et al. Ann Oncol 2017), a strong pt preference for R-SC over R-IV administration was demonstrated (81% vs 11% at Cycle 8; primary endpoint). In addition, on the Rituximab Administration Satisfaction Questionnaire, median scores were higher for R-SC vs R-IV for ‘Impact on Activities of Daily Living’, ‘Convenience of Therapy’, ‘Satisfaction with Therapy’, and ‘Psychological Impact’. No new safety signals were reported. Here, we present the final analysis of the PrefMab study, which was performed after all pts had completed at least 24 months' follow-up (data cut-off, December 20, 2016) with a focus on long-term pt outcomes and safety.Methods: Pts aged 18-80 yrs with previously untreated CD20+ DLBCL or FL were randomized 1:1 to receive one of two 8-cycle schedules of R in combination with 6-8 cycles of chemo (CHOP, CVP, or bendamustine, as per local practice): Sequence A (n=372), 1 cycle R-IV (375mg/m2) then 3 cycles R-SC (1400mg) then 4 cycles R-IV; Sequence B (n=371), 4 cycles R-IV then 4 cycles R-SC. Maintenance therapy was not mandated by the study protocol; however, pts with FL could receive maintenance at the investigator's discretion. Secondary endpoints reported here include overall (OR) and complete response (CR) rates (assessed according to Cheson 1999 criteria) and investigator-assessed progression-free (PFS), event-free (EFS), disease-free, and overall survival. Safety was also assessed by treatment-emergent adverse event (AE) reporting. Efficacy was analyzed in all randomized pts, safety in all pts who received ≥1 dose of R.Results: Of 743 randomized pts, 740 received study treatment (465 pts with DLBCL [63%], 273 with FL [37%], and 2 other [0.3%]). Demographics and baseline characteristics are presented in Table 1. The median age of pts with DLBCL was 61 yrs and of pts with FL was 59 yrs. The majority of pts with DLBCL and FL received CHOP chemo (Table 1). OR rates were >90% in both DLBCL and FL pts (Table 2). CR rates were 55.1% in pts with DLBCL and 44.4% in pts with FL (Table 2). PFS rates at 3 yrs were 79% in DLBCL pts and 91% in FL pts. Other time-to-event endpoints are presented in Table 2. EFS was shorter in FL pts than DLBCL pts; however, because R maintenance was considered a non-protocol-specified anti-lymphoma therapy, FL pts were censored at the time of starting maintenance therapy. Treatment sequence had no impact on response rates or time-to-event endpoints. Pharmacokinetic serum-level data will be presented.Of pts who received ≥2 cycles of treatment, 653/712 (92%) experienced ≥1 AE on or after Cycle 2. AEs were generally balanced according to administration route and diagnosis, although a higher rate of grade ≥3 AEs and serious AEs occurred in DLBCL pts compared with FL (Table 3). Most AEs occurred in Cycle 1, during which all pts received R-IV (pts with AE in Cycle 1: Arm A, 71%; Arm B, 69%). No administration-related reactions (ARRs) occurred at a frequency of ≥2% in pts receivingR-SC or R-IV on or after Cycle 2. There were no differences in rates of ARRs according to IV or SC administration. FL pts were more likely to experience an ARR than DLBCL pts. At the end of the study, 104 pts had died (85 DLBCL; 19 FL). Fatal AEs related to R, according to the investigator's judgment, included oropharyngeal candidiasis, pneumonia, septic shock, febrile neutropenia, and myocardial infarction (1 case each).Conclusions: In addition to the strong pt preference for SC dosing and improved satisfaction observed in the primary analysis, these findings confirm that R-SC is an efficacious and well-tolerated treatment option in DLBCL and FL pts. Overall, outcomes did not appear to be impacted by treatment sequence. [Display omitted] DisclosuresRummel:Roche: Honoraria, Other: PrefMab is sponsored by F. Hoffmann-La Roche Ltd. Third-party Medical Writing assistance, under the direction of Mathias Rummel, was provided by Lynda McEvoy of Gardiner-Caldwell Communications, and was funded by F. Hoffmann-La Roche Ltd, Research Funding. Brugger:AstraZeneca: Employment, Equity Ownership. Re:Gilead: Membership on an entity's Board of Directors or advisory committees. Task:Genentech: Employment, Equity Ownership. McIntyre:Roche: Employment, Equity Ownership. Meier:Roche: Employment. Osborne:Roche: Employment. Smith:Roche: Employment. Grigg:Roche: Other: Funding to attend ASH 2016.