Impact Of The Pretreatment Characteristics As Well As Cyto- and Molecular-Genetic Profile On Outcome After Relapse In Acute Myeloid Leukemia

Cyto- and molecular-genetic abnormalities evaluated at initial diagnosis are the most powerful prognostic and in part also predictive markers in acute myeloid leukemia (AML) with regard to achievement of complete remission (CR) and survival. Nonetheless, after relapse the prognostic impact of clinical characteristics and genetic abnormalities assessed at initial diagnosis with respect to achievement of subsequent CR and survival are less clear.

To evaluate the probability of CR achievement and survival in relapsed AML patients in correlation to clinical characteristics and genetic abnormalities assessed at initial diagnosis as well as treatment strategy.

The study includes intensively treated adults with newly diagnosed AML enrolled in 5 prospective AMLSG treatment trials between 1993 and 2009. Patients with acute promyelocytic leukemia were excluded. All patients received intensive therapy, including allogeneic (allo) and autologous (auto) hematopoietic stem cell transplantation (HSCT) during first line therapy.

A total of 3218 patients (median age, 54 years; range, 16-85 years) were enrolled in 5 AMLSG treatment trials. Of these, 1307 (41%) patients (16-60 years, n=958; ≥61 years, n=349) experienced relapse, n=194 after alloHSCT, n=75 after autoHSCT and 1038 after chemotherapy. Salvage strategies were as follows: (i) n=907, intensive chemotherapy (INT) followed in n=450 by HSCT (matched related donor [MRD], n=114; matched unrelated donor [MUD], n=303; cord blood graft [CB], n=3; haplo-identical family donor [HID], n=18; autoHSCT, n=12); (ii) n=100, direct alloHSCT (MRD, n=31; MUD, n=63; HID, n=4) or n=2 autoHSCT (TPL); (iii) n=29, donor lymphocyte infusions (DLI) in patients after alloHSCT in CR1; (iv) n=60, demethylating agents/low-dose cytarabine (NON-INT); (v) n=24, experimental treatment within phase I/II studies (EXP); (vi) all other patients (n=187) received best supportive care (BSC). After salvage therapy CR rate was 38% and after the different treatment approaches as follows: INT, 37%; TPL, 73%; DLI, 38%; NON-INT, 8%; EXP, 29%. After failure to respond to INT, n=159 additional patients achieved a CR2 after HSCT resulting in an overall CR2 rate of 50%. A logistic regression model revealed CEBPA double-mutant (dm) (OR, 6.42; p=0.0001), core-binding factor (CBF) AML (OR, 2.87; p=0.0002), a direct HSCT strategy (OR, 3.32; p=0.0002), and mutated NPM1 (OR, 1.59; p=0.02) as favorable (only if response after HSCT was included) and FLT3-ITD (OR, 0.66; p=0.04), age (difference of 10 years; OR, 0.82; p=0.003), NON-INT (OR, 0.08; p=0.0001) and in trend a previous alloHSCT in CR1 (OR, 0.65; p=0.08) as unfavorable independent parameters for achievement of CR2. Median follow-up for survival after relapse was 4.3 years and survival after 4 years was 22% (95%-CI, 19-25%). Patients proceeding to alloHSCT after first relapse (n=536; MRD, n=145; MUD, n=366; HID, n=22; CB, n=3) had a 4-year survival of 36% (95%-CI, 32-41%) and those not proceeding to alloHSCT of 8% (95%-CI, 6-11%). In univariable analyses the combined genotype mutated NPM1 in the absence of FLT3-ITD (p=0.66) was not associated with a favorable outcome. A multivariable regression model including alloHSCT as a time-dependent co-variable revealed alloHSCT performed after relapse (HR, 0.34; p<0.0001), CEBPAdm (HR, 0.48; p=0.002), CBF- AML (HR, 0.50; p<0.0003) and DLI in relapsed patients with a previous alloHSCT performed in CR1 (HR, 0.40; p=0.002) as significant favorable factors, whereas FLT3-ITD (HR, 1.35; p=0.005) and in trend NON-INT (OR, 1.40; p=0.06) were unfavorable factors. Due to collinearity of FLT3-ITD with duration of first remission (cut point at 1 yr), the latter was not included into the multivariable models.

Of 561 patients achieving CR2, 252 experienced 2nd relapse (REL2) and 114 died in CR2. Most REL2 patients (n=117) received INT whereas n=54 received BSC only. Allo- and autoHSCT were performed in 55 and 3 REL2 patients, respectively. CR3 rate in patients who received treatment was overall 40% including response to HSCT of 58%.

Patients with relapsed AML have an overall probability of less than 50% to achieve a CR2 and CR3 after intensive salvage chemotherapy; the only exceptions are AML with CEBPAdm and CBF-AML. AlloHSCT either as direct treatment of relapse or as salvage therapy after failure of intensive chemotherapy may overcome chemo-resistance.

Schlenk:Celgene: Honoraria, Research Funding; Pfizer: Honoraria, Research Funding; Chugai: Research Funding; Amgen: Research Funding; Novartis: Research Funding; Ambit: Honoraria. Off Label Use: Pomalidomide in Myelofibrosis. Kindler:Novartis: Membership on an entity’s Board of Directors or advisory committees.