Phase III Randomized Study of R-CHOP Versus DA-EPOCH-R and Molecular Analysis of Untreated Diffuse Large B-Cell Lymphoma: CALGB/Alliance 50303

Background: Diffuse large B-cell lymphoma (DLBCL) is comprised of multiple diseases with different outcomes, cell of origin and molecular pathogenesis. DLBCL derived from germinal center B-cells (GCB) and activated B-cells (ABC) constitutes over 80% of DLBCL. A small number of DLBCL tumors harbor MYC translocations, which are associated with a poorer prognosis. Primary mediastinal B-cell lymphoma (PMBL) occurs in younger patients and is typically treated with intensive or combined modality therapy. R-CHOP is the standard of care for DLBCL but has not been prospectively studied within DLBCL molecular subtypes. DA-EPOCH-R is an alternative and more dose-intensive regimen that showed a PFS of 81% at 4-years in a phase II multicenter CALGB study (Haematologica 2012; 97:758). CALGB/Alliance undertook a phase III randomized study to compare R-CHOP and DA-EPOCH-R in DLBCL and specifically within the GCB and ABC subtypes.

Methods: Patients had stage II or higher newly diagnosed DLBCL, age ≥ 18 years and were HIV negative. Subjects were randomized 1:1 to receive R-CHOP or DA-EPOCH-R as previously published. Central nervous system prophylaxis in at-risk patients was 12 mg intrathecal methotrexate in cycles 3-6 (4 doses total). All subjects with accessible tumor were required to have a biopsy or waiver. Subjects received a total of 6 treatment cycles. The primary study endpoints were EFS of R-CHOP (Arm A) versus DA-EPOCH-R (Arm B) and to develop a molecular predictor of outcome based on GCB and ABC DLBCL. Additional analyses include toxicity, pharmacogenomics and analysis of tumor genetics by next generation sequencing, and.

Results: 524 patients registered (262 on each arm) between May 2005 and May 2013. Efficacy analysis includes R-CHOP (n=233) and R-EPOCH (n=231) assigned patients with confirmed eligibility who received any treatment. Characteristics of patients registered to Arm A and B did not differ with median age (58; 56), male sex (53.2%; 53.7%), high-intermediate/high IPI (33.6%; 38.2%) and primary mediastinal B-cell lymphoma histology (6.9%; 5.2%). Therapy was completed per protocol in 89% and 83%, respectively, for Arm A and B, and disease progression on therapy was 2.6% and 1.7%. Adverse side effects leading to treatment discontinuation were 1.7% and 5.6%, respectively, for Arm A and B. Arm B compared to Arm A was associated with more grade 4 neutropenia (90%; 56%), grade 4 thrombocytopenia (35%; 6%), grade 3/4 febrile neutropenia (37%; 19%) and grade 3 neuropathy (motor: 8%;1% and sensory: 15%; 3%). Grade 5 events on study were the same in both arms. Preliminary analysis of EFS in confirmed eligible subjects shows no difference between Arm A and B with hazard ratio of 1.02 and p=0.89 at a median follow-up of 4.9 years. Overall survival is also similar with hazard ratio of 1.19 and p=0.40 at median 5.0 years. Additional analyses of EFS and OS by age (<60 and >=60) and by GCB versus ABC DLBCL are in process. Conclusions: There was no difference in EFS or OS between R-CHOP and DA-EPOCH-R when considering all patients. DA-EPOCH-R showed increased toxicity consistent with higher dose-intensity but not increased grade 5 toxicity. Compared to R-CHOP, more patients on DA-EPOCH-R did not complete treatment, which may reflect patterns of care or toxicity. Due to the clinical and genetic diversity of DLBCL, subset analyses are necessary to determine the effect of CNS relapse, GCB and ABC subtypes, age and IPI on outcomes of the two arms. These data do not address the efficacy of these regimens in PMBL or MYC+ DLBCL due to their low frequency, and where more dose-intense regimens appear to be important. Full molecular analyses are ongoing. Support: U10CA180821, U10CA180882, U10CA180888, CA180799, CA180820, CA180833, CA21076, CA21115.ClinicalTrials.gov Identifier: NCT00118209.