Multicentre survey to explore current survival of patients with acute myeloid leukaemia who failed induction chemotherapy.

Background: Acute myeloid leukaemia not responsive to first induction chemotherapy (PIF-AML) still remains a challenge, and there are only few recent epidemiological data regarding the outcome of these patients. In this multicentre survey, we evaluate the prognosis and outcome of patients with PIF-AML, who were diagnosed and treated in the last 5 yrs in four Italian institutions.

Results: One hundred PIF-AML were recorded, 57 males and 43 females, with a median age of 63 yrs (19-79), 42% were younger than 60 yrs; 42% had a secondary AML and 40% had an adverse karyotype. According to cytogenetic/molecular risk stratification at diagnosis, 33% of patients were classified as favourable/intermediate-1 risk and 56% as intermediate-2/adverse risk. After a median follow-up of 11 months (1-49), 77% of patients died, while 23% were alive (with 12/23 in cCR). Thirty-six patients underwent allogeneic SCT, and of these, 11 of 36 (31%) were alive at last follow-up. The 12- and 24-month OS probability of the whole population was 45% and 21%, respectively. In multivariate analysis, the probability of OS of the whole population was significantly improved by Allo-SCT procedure (12-month OS probability 60% vs. 35%; P < 0.0001) and was better in patients with favourable/intermediate-1 risk at diagnosis (12-month OS probability 58% vs. 40%; P = 0.028). In transplanted cases, a pretransplant responsive disease was the only significant factor to predict a favourable outcome after Allo-SCT (P = 0.006).

Conclusion: Treatment options of PIF-AML still are limited and the prognosis, even recently, remains extremely poor. This survey shows that PIF-AML is still rarely cured without Allo-SCT and confirms the importance of initiating an urgent unrelated donor search in cases without a matched sibling donor. Moreover, the outcome of Allo-SCT is better in patients who achieve a good AML debulking before transplant. To reach this goal, new predictive scores and new protocols of salvage therapy (with target drugs or combinations) need to be explored urgently in PIF-AML.