Prognostic Impact of Cytogenetics and Early Response and Efficacy of Sequential High-Dose AraC and Idarubicin (S-HAI) in Patients with Refractory and Relapsed Acute Myeloid Leukemia (AML): Results of a Prospective Study.

Management of patients with refractory and relapsed AML needs optimization. We performed a prospective study in these patients aiming at 1) the definition of the anti-leukemic efficacy of the S-HAI regimen; and 2) the evaluation of the prognostic impact of cytogenetic aberrations at relapse in the context of other prognostic parameters. Treatment consisted of AraC 1 g/sqm q 12 h days 1, 2, 8, and 9 and idarubicin 10 mg/sqm days 3, 4, 10, and 11. AraC was given at 3 g/sqm in patients under age 60 with refractory AML or relapse after CR1 <6 months. Fludarabine was given according to randomization at 15 mg/sqm 4 h before each dose of AraC. Between May 1996 and February 2004 306 patients were randomized, 261 are fully evaluable. The patients′ characteristics were median age 55 years (range, 18-83); refractory AML/relapse with CR1<6 months/relapse with CR1 >6 months 13%/25%/62%; cytogenetics at relapse favorable/intermediate/unfavorable/not available 7%/44%/24%/25%; secondary AML 7%. Median duration of neutropenia <1000/μl was 37 days. Non-hematologic side effects III°/IV° included diarrhea (21%), mucositis (19%), nausea/vomiting (17%), hyperbilirubinemia (12%), and bleeding (8%). Encontered infections were pneumonia 51%, FUO 41%, bacteremia 28%, abdominal 23%, and catheter-related 16%. Response rates were CR 39%, partial remission 7%, persistent leukemia 36%, early death 18%. Median event-free survival (EFS) was 2.4 months, meidan relapse-free survival was 5.9 months, and median overall survival was 6.2 months. In 38% of patients a change of karyotype between diagnosis and relapse occurred. In general, cytogenetics (CG) at relapse had a higher prognostic impact as compared to CG at diagnosis and therefore was included in the following analyses of prognostic parameters. CR rate was significantly related to duration of CR1 (CR1 0 months 29%; CR1 <6 months 14%; CR1 >6<18 months 52%; CR1 >18 months 56%; p<0.0001) and CG at relapse (favorable CG 85%; intermediate CG 44%; unfavorable CG 21%; p<0.0001) but not to age < vs. >60 years (40% vs. 39%). EFS and OS were significantly related to duration of CR1 (p=0.0001 and p=0.0004) and CG at relapse (p=0.0002 and p=0.0009). Logistic regression analysis revealed CG at relapse (p=0.006) as well as duration of CR1 (p=0.004) being independently related to CR rate. Cox regression analysis revealed CG at relapse (p=0.001) and duration of CR1 (p=0.014) being independently related to EFS. CG at relapse was the only parameter independently related to OS (p=0.001). The inclusion of the therapy-dependent parameter, residual bone marrow blasts at day 18 (day 18 blasts), revealed day 18 blasts being independently related to CR rate, EFS, and OS. These data indicate that 1) the S-HAI regimen confers a significant anti-leukemic efficacy in patients with relapsed and refractory AML unless unfavorable CG are present; and 2) CG at relapse is the most important prognostic parameter in these patients and day 18 blasts may be used to early identify treatment failure and guide the decision about alternative treatment approaches.