Allogeneic Stem Cell Transplantation in Patients with Therapy-Related Acute Myeloid Leukemia and Myelodysplastic Syndromes: Effect of Patient Characteristics and Treatment History.

Abstract 4577

Patients with therapy-related acute myeloid leukemia (tAML) and myelodysplastic syndromes (tMDS) have a bad prognosis. Allogeneic stem cell transplantation (alloSCT) is potentially curative, but the majority of these patients die because of the high incidence of relapse (RI) and non relapse mortality (NRM). Since it is critical to understand which patients should receive alloSCT, we designed this retrospective study to assess which patients’ characteristics may predict the alloSCT outcome in tAML/tMDS. All the patients affected by tAML or tMDS and allografted in 4 Italian hematology units between 1998 and 2009 were included. Total patients were 28: 24 (86%) had tAML, 4 (14%) had tMDS. Patients had a median age of 49 years (range 21–65) at transplant; 15 patients were female (54%). Previous neoplasia was lymphoma in 23 patients (82%, 15 Hodgkin and 8 non-Hodgkin), or non-hematologic cancer (5, 18%) like breast carcinoma (2), seminoma (1), testicular embryonal carcinoma (1), and osteosarcoma (1). Previous cancer had been treated with chemotherapy (CT, 7 patients, 25%), radiotherapy (RT, 4 patients, 14%), or both (17, 61%). Fourteen patients (50%) had received >=2 therapy lines, 6 (21%) had received autologous stem cell transplant. tAML/tMDS occurred after a median time of 86 months (range 13–253) after the previous cancer treatment. Cytogenetic analysis was performed in 24 patients (86%): 8 patients (33%) had intermediate risk and 16 (57%) high risk cytogenetics according to Medical Research Council AML10 Trial definitions. Twenty-three patients (82%) received induction CT for tAML/tMDS, 19 (68%) received consolidation; 5 patients (18%) received upfront alloSCT. Induction consisted of idarubicin+cytarabine+/−etoposide (12 patients, 43%) or fludarabine+cytarabine+/−idarubicin (11 patients, 39%); 15 patients (54%) had an infection after induction. Median time from tAML/tMDS diagnosis to alloSCT was 5.7 months (range 0–25). At transplant, 10 patients (36%) had a Karnofsky Performance Status (KPS)<=80% and 13 (46%) a Sorror comorbidity score >1. Donor was identical sibling for 9 patients (32%), and alternative for 19 patients (68%): mismatched related (1, 3%), matched unrelated (15, 54%) or haploidentical (3, 11%). Disease status of patients at transplant was as following: 12 patients (43%) were in CR (11 in CR1), 4 (14%) in PR and 7 (25%) in PD; 5 patients (18%) received alloSCT at diagnosis. Patients underwent reduced intensity (11 RIC, 39%) or myeloablative (17, 61%) alloSCT. Myeloablative conditioning was mainly busulfan-based (15 patients, 54%); the majority of RIC patients received thiotepa+cyclophosphamide-based conditioning (9, 32%). Twelve patients (43%) are alive at last follow-up, 7 patients (25%) died of disease, 9 (32%) died of NRM. The main reason of death by NRM was infection (8 patients, 29%). Median follow-up of surviving patients was 528 days (55-1704). One- and 2-years overall survival (OS) was 50% and 36%, progression free survival (PFS) was 42% and 38%. RI was 28% at both 1 and 2 years, NRM was 18% at 100 days, 30% at 1 year and 35% at 2 years. Nine (32%) patients had acute GVHD of grade >=2, 8 patients had chronic GVHD (cGVHD, 29%), 3 patients (11%) had extensive cGVHD. OS and PFS were reduced in patients with high-risk cytogenetics (p=0.03 and p=0.01, respectively) and KPS<=80% (p=0.008 and p<0.001). RI was higher in patients with KPS<=80% (p=0.02); high-risk cytogenetics increased RI (trend, p=0.08). Patients treated for previous neoplasia with >2 CT lines had a worse NRM (p=0.005) and a reduced OS (p=0.05); treatment type (CT vs RT vs CT+RT) did not have a significant impact on OS, PFS nor NRM. NRM was higher in patients who had infections after induction (p=0.009), or received consolidation (p=0.03). AlloSCT after >180 days from tAML/tMDS diagnosis increased NRM (p=0.05). Conditioning (RIC vs myeloablative), donor type (identical sibling vs alternative) and type of previous neoplasia (hematologic vs non-hematologic) did not affect survival (OS and PFS), NRM and RI. In conclusion, alloSCT is effective in a minority of patients with tAML/tMDS. Survival is reduced by high risk cytogenetics and poor KPS. Number of previous treatments, infections after induction, and consolidation CT increase NRM. Thus, patients’ selection for alloSCT in tAML/tMDS setting should be based on cytogenetics, KPS, and treatment history. Prospective trials are awaited in order to confirm these results.