Summary
We conducted a prospective randomized phase II study to evaluate two chemotherapy regimens: (i) rituximab plus hyperfractionated cyclophosphamide, vincristine, doxorubicin and dexamethasone (R-HCVAD) alternating with rituximab, high-dose methotrexate, and cytarabine (R-MA) and (ii) rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R-CHOP) in diffuse large B-cell lymphoma (DLBCL). This study randomized patients aged ≤60 years with DLBCL and an age-adjusted international prognostic index ≥2 to R-HCVAD/R-MA or R-CHOP based on a Bayesian adaptive algorithm. Interim analysis of the first 26 eligible patients showed that the complete response rate (CRR) was higher with R-HCVAD/R-MA than R-CHOP (P = 0·03); thus, R-CHOP arm was closed. In the final analysis, 49 and 10 eligible patients were treated in R-HCVAD/R-MA and R-CHOP arms respectively; CRR were 82% and 60% respectively (P = 0·13); 3-year progression-free survival (PFS) rates were 75·7% and 77·8% respectively (P = 0·53). In the R-HCVAD/R-MA arm, 3-year PFS rates in patients aged 46–60 years and ≤45 years were 70·3% and 87·1% respectively (P = 0·13), and the treatment-associated early mortality rate in patients >45 years was 12%. In conclusion, R-HCVAD/R-MA is associated with excellent outcome in patients ≤45 years old. However, in patients >45 years old, R-HCVAD/R-MA is associated with unacceptable mortality rates.