Cohn, J. B.

8011 Background: Maintenance therapy is generally not used in treatment of DLBCL. However, older pts are a unique group in which to study this approach, due to poorer outcome. Methods: As part of intergroup trial E4494/C9793, 415 pts >= 60 years (yr) of age with DLBCL who responded to induction therapy with CHOP or R-CHOP were randomized to MR (rituximab, 375 mg/m2 weekly {wk} X 4 every 6 months for 2 yr) starting 4 wk after the last chemotherapy (n=207) or observation (OBS, n=208) in a prospective 2 × 2 randomized trial design (J Clin Oncol 2006;24:3121). Impact of MR on failure-free survival (FFS) was the primary objective, and impact on time to failure (TTF) and overall survival (OS) were secondary objectives. All p-values are two-sided. Results: Results are presented for 352 centrally reviewed evaluable pts, with a median follow-up after maintenance randomization of 5.5 yr. All results were similar for the intent-to-treat population. Baseline characteristics and response to induction therapy are balanced among the CHOP/R-CHOP pts. At 6 yr, FFS was 46% for those pt who received MR and 36% for the OBS cohort (p=0.005, hazard rate {HR} 0.64); 6-yr OS was 62% for MR and 64% for OBS (p=0.83, HR 0.96). Outcome was also examined in the four treatment subgroups. The 6-yr FFS were: CHOP+MR 44%, CHOP+OBS 35%, R-CHOP+MR 47%, and R- CHOP+OBS 40%. No differences in 6-yr OS were observed according to maintenance randomization following CHOP (57% MR vs 55% OBS) or R-CHOP (67% MR and 72% OBS). Median TTF after maintenance randomization following CHOP+MR was 5.2 yr vs 1.6 yr for CHOP-OBS (p=0.0004), and following R-CHOP+MR was 5.6 yr vs 5.4 yr with R-CHOP-OBS (p=0.50). Proportionately more of the treatment failures occurred within 2 yr after CHOP+OBS (85%) compared to CHOP+MR (61%), p=0.01. In contrast, the proportion of failures within 2 yr was similar for R- CHOP+OBS (47%) and R-CHOP+MR (53%), p=NS. Conclusions: MR after CHOP, but not after R-CHOP, significantly prolongs TTF, but fails to prolong OS, possibly due to a delayed pattern of relapse and/or the efficacy of rituximab in the salvage setting. As 6-yr FFS declined to <50% among R-CHOP responders, with or without MR, there is a need for more effective treatment strategies in older DLBCL pts. No significant financial relationships to disclose.