Abstract
The primary objective in multiply relapsed hairy cell leukemia and variant (HCL/HCLv) was to determine whether pentostatin-rituximab (DCFR) and bendamustine-rituximab (BR) each achieve an overall response rate (ORR) exceeding that historically achieved by rituximab alone (∼40%) in favor of 65%. Prospective data were unreported for either regimen. Fifty-six patients received 6 28-day cycles of rituximab (375 mg/m2, days 1 and 15) with either bendamustine (90 mg/m2, days 1 and 2) or pentostatin (4 mg/m2, days 1 and 15). Eligibility required ≥2 purine analogs, or 1 purine analog plus rituximab for response of <1 year to the initial purine analog. Although patients were assigned to either regimen through randomization to increase homogeneity of the 2 treatment groups, the DCFR arm had fewer previous purine analogs (P = .021) and lower baseline marrow HCL/HCLv infiltration (P = .013). ORRs for DCFR and BR were 93% (95% confidence intervals [CI], 83-102) and 86%, (95% CI, 73-99), respectively, exceeding 40% (P< .0001) for each group. Rates for complete remission (CR) and minimal residual disease-free CR and median progression-free survival (141 vs 50 months; HR, 0.63; 95% CI, 0.32-1.25) numerically favored DCFR, but that arm was significantly enriched with less previous purine analogs and marrow infiltration, each of which was associated post hoc with better response. Post hoc subgroup analysis, particularly for 41 patients with classic HCL, suggested any superiority of DCFR vs BR might apply to patients with more favorable disease. DCFR and BR were highly effective in multiply relapsed HCL/HCLv. Possible DCFR superiority was hypothesis-generating, given uneven baseline risks and trial design. This trial was registered at www.clinicaltrials.gov as #NCT01059786.