Prognostic Impact of Combined Ex Vivo Drug Resistance to Fludarabine, Treosulphan and Mitoxantrone in Childhood Acute Myeloid Leukemia.

Combined ex vivo drug resistance profile to prednisolone, vincristine and L-asparaginase (PVA) has prognostic significance in childhood de novo ALL (Den Boer JCO 2003, Styczynski JCO 2004). So far, no data exist to support prognostic value of in vitro drug resistance profile in childhood acute myeloid leukemia (AML). The aim of this study was the analysis of prognostic value of ex vivo drug resistance profile in childhood AML. A total number of 90 children (46 male, 44 female) aged 0.1–17 yrs (median 10 yrs), with de novo AML were included into the study. All patients were recruited between 1999 and 2004 and were treated according to ANLL-98 protocol. Patients with early deaths were excluded from the study. Drug resistance profile was done by the MTT assay in one laboratory for 4–24 drugs. According to median cytotoxicity for each of tested drugs, all patients were scored as resistant or sensitive to this drug. Patients who received a HSC transplantation were censored at the time of transplantation. Drug resistance profile and other known prognostic parameters, were examined by Kaplan-Meier method and by multivariate analysis using the Cox regression proportional hazard model. Probability of overall survival was 0.57±0.05, p(LFS)=0.72±0.05, mean LFS 3.36 yrs (95%CI=2.95–3.77). Relapses occurred in 17/90 children; 40/90 children died during follow-up. Children who relapsed during follow-up showed median higher in vitro resistance of leukemic blasts to most of tested drugs, except for cytarabine, cladribine, vincristine, mercaptopurine and thioguanine. Better LFS was observed for patients with favourable cytogenetics (1.00 vs 0.74±0.09, p=0.055); early BM response by day 15 (0.76±0.07 vs 0.54±0.13, p=0.013); BM remission by day 28 of induction (0.74±0.07 vs 0.35±0.26, p=0.018); ex-vivo sensitive patients to cyclophosphamide (0.83±0.15 vs 0.65±0.09, p=0.12), doxorubicin (0.81±0.12 vs 0.64±0.09, p=0.18), epirubicin (0.72±0.13 vs 0.61±0.12, p=0.26), fludarabine (0.73±0.12 vs 0.62±0.11, p=0.22), mitoxantrone (0.77±0.12 vs 0.51±0.13, p=0.07), treosulphan (0.88±0.12 vs 0.62±0.11, p=0.29), etoposide (0.70±0.13 vs 0.63±0.09, p=0.4) and for patients with leukemic cells being sensitive to fludarabine, treosulphan and mitoxantrone ie. FTM score (0.73±0.12 vs 0.50±0.14, p=0.034). In multivariate analysis, two factors showed prognostic value: early BM response by day 15 (p=0.0021; HR=0.29, 95%CI=0.13–0.64) and combined ex vivo drug resistance profile to fludarabine, treosulphan and mitoxantrone (FTM score), p=0.0048; HR=0.38, 95%CI=0.19–0.77. Combined ex vivo drug resistance profile to fludarabine, treosulphan and mitoxantrone has prognostic significance in childhood acute myeloid leukaemia, however cytogenetics and early bone marrow response to therapy seems to have stronger prognostic value.