Clofarabine-Based Consolidation Improves Relapse-Free Survival of Younger Adults with Non-Favorable Acute Myeloid Leukemia (AML) in First Remission: Results of the Randomized ALFA-0702/Clara Study (NCT 00932412)

Background: Clofarabine has been shown as effective as single-agent or in combination with cytarabine in various populations of AML patients. Nevertheless, two randomized studies failed to demonstrate prolonged survival when compared to low-dose cytarabine in older AML patients (Burnett, 2013) or intermediate-dose cytarabine (IDAC) in patients aged 55y or more with relapsed/refractory AML (Faderl, 2012). As this was likely due to excessive toxicity and poor compliance in these difficult-to-treat populations, we made the hypothesis that a more apparent benefit might be observed during post-remission therapy. We thus randomly assessed the role of CLARA combination (clofarabine + IDAC) versus high-dose cytarabine (HiDAC) as consolidation cycles in younger AML patients.

Methods: Between 2009 and 2013, 713 patients aged 18-60y old with de novo AML (CBF-AML excluded) were enrolled in the ALFA-0702 trial by 33 French centers. They received a timed-sequential induction with daunorubicin (60 mg/m², day 1-3; 35 mg/m², day 8-9), cytarabine (500 mg/m² CI, day 1-3; 1 g/m²/12h bolus, day 8-10) and G-CSF priming. A first salvage with idarubicin and HiDAC may be proposed to patients not achieving CR/CRp after this first course. Patients in CR/CRp with non-favorable AML (according to ELN classification) or those who needed the salvage (late CR/CRp) were eligible for: 1) allogeneic stem cell transplantation (SCT) if they had a sibling or fully matched unrelated donor; or 2) randomization for consolidation with three HiDAC (3 g/m²/12h cytarabine, day 1/3/5; G-CSF priming) or CLARA (30 mg/m² clofarabine, day 2-6; 1 g/m²/12h cytarabine, day 1-5; G-CSF priming) cycles if no identified donor. Primary endpoint was relapse-free survival (RFS), patients receiving allogeneic SCT in first CR/CRp after randomization (late donor identification) being censored at SCT time. Secondary endpoints were cumulative incidence of relapse (CIR) and death in first CR/CRp (CID), overall survival (OS), and safety. Anticipating a 35% SCT rate in randomized patients, 230 patients should be randomized to demonstrate a 20% gain in 2-year RFS. A sensitivity analysis without SCT censoring was planned.

Results: Among 468 CR/CRp patients of interest (465/552 non-favorable AML + 3/161 favorable AML with late CR/CRp), 221 were randomized between HiDAC (n=114) and CLARA (n=107) consolidation and 247 were not randomized for the following reasons: 181 had an identified donor for SCT; 19 had favorable ELN genotype even if no evidence of normal karyotype; and 47 were negatively selected (13 early relapses, 1 early death, 18 AEs, 15 drop-out). The ITT analysis was conducted in these 221 eligible patients (median age, 48y; 136 intermediate-risk, 85 unfavorable-risk; 18 late CR/CRp; without imbalances between randomization arms), the remaining 9 randomized patients having non-eligibility criteria (6 favorable-risk AML and 3 not in CR/CRp). Among these patients, the rate of SCT in first CR/CRp was higher than anticipated (50%; 55 patients in each arm). With a median follow-up of 37.4 months and SCT censoring, 2-year RFS was 52.1% (40-68) in the CLARA arm versus 30.5% (20-46) in the HiDAC arm (HR, 0.62 [0.39-0.99]; p= 0.042). This was due to a lower CIR in the CLARA arm (44.0 versus 67.7%; p= 0.023) with similar CID (3.9 versus 1.9%, p= 0.60). At 2 years, OS was 68.1% (56-82) in the CLARA arm versus 49.8% (38-66) in the HiDAC arm (p= 0.18). Interestingly, the gain in CIR and RFS was observed in both intermediate- and unfavorable-risk groups and persisted after adjustment on AML risk and age. As expected, these gains were lower when not censoring at SCT (2-year RFS, 57.8 versus 45.6%; p= 0.12). In the 110 randomized patients allografted in first CR/CRp, no differences in post-SCT outcome were observed according to randomization arm. In Mantel-Byar time-dependent analysis, longer RFS was observed after SCT in the HiDAC arm (HR, 0.45; p= 0.004), while not in the CLARA arm (HR, 0.71; p= 0.26). Finally, CLARA cycles were associated with higher hematologic toxicity, more infections and more liver toxicities than HiDAC cycles, especially after cycle 1.

Conclusions: As compared to HiDAC, clofarabine-based consolidation improved RFS in younger patients with intermediate- and unfavorable-risk AML in first CR/CRp. CLARA combination might thus become a new standard for post-remission therapy in these patients, especially in those who may not benefit from allogeneic SCT.