Prognostic Factors For Overall Survival In Elderly Patients With Relapsed Acute Myeloid Leukemia – Retrospective Study On Behalf Of The East German Study Group For Hematology and Oncology (OSHO)

In patients with relapsed acute myeloid leukemia (AML) > 60 years of age we analyzed age at relapse, interval from first complete remission (CR1) to relapse, cytogenetic risk at initial diagnosis, prior allogeneic stem cell transplantation (alloSCT) and FLT3/NPM1 mutational status as possible prognostic factors for overall survival (OS).

After achieving CR1 more than 50% of elderly AML patients eventually relapse. Prognostic factors for OS are poorly defined in this patient population. For younger patients with relapsed AML a risk score has been described including age at relapse, interval from CR1 to relapse, cytogenetic risk at initial diagnosis and prior stem cell transplantation (SCT) as prognostic factors. We sought to investigate whether these are also prognostic factors in elderly patients with relapsed AML. In addition, we assessed the prognostic impact of FLT3- and NPM1 mutational status (wild-type (wt) or mutated (mut)) at diagnosis.

In the ongoing multicenter OSHO trial #69 for AML patients > 60 years we evaluated data of all relapsed patients. Overall survival was calculated from the day of first relapse until the day of death using the Kaplan Meier method. Univariate analysis was performed to test for the influence of age at relapse, interval from CR1 to relapse, cytogenetic risk at initial diagnosis, prior alloSCT and FLT3/NPM1 mutational status. Subsequently, independent prognostic factors were defined in a multivariate analysis with age at relapse, time from CR1 to relapse, cytogenetic risk at initial diagnosis and prior alloSCT as covariates.

From April 2005 until April 2013 904 patients were registered. 733 of these received intensive induction chemotherapy which resulted in CR1 in 447 (61%) pts. In this patient group 260 relapses were observed after a median interval, calculated from the day of CR1, for living patients of 2.7 years (range 0.1 to 7.5). Median age at relapse was 69 years (range 60 – 85) with 129 (49.6%) pts. being 60 to 68 years old, 102 (39.2%) pts. being 69 to 74 years old and 29 (11.1%) pts. being 75 to 85 years old. Median interval from CR1 to relapse was 0.58 years (0.07 – 6.28). 114 (43.8%) relapses occurred up to 6 months after CR1, 119 (45.8%) between 7 and 18 months after CR1 and 27 (10.4%) later than 18 months after CR1. Only five (1.9%) relapsed pts. showed good risk cytogenetics at diagnosis, whereas it was of intermediate risk in 159 (61.1%) pts., of poor risk in 68 (26.2%) pts. and unknown in 28 (10.8%) pts. Forty-one (15.8%) pts. had received prior alloSCT in CR1. Information on FLT3- and NPM1 mutational status at diagnosis was available in 194 (74.6%) pts. 110 (42.3%) pts. had FLT3/NPM1 wt/wt, 48 (18.5%) pts. had FLT3/NPM1 wt/mut, 23 (8.8%) pts. had FLT3/NPM1 mut/wt and 13 (5.0%) pts. had FLT3/NPM1 mut/mut.

OS rate at 2 years of all relapsed pts. was 13 ± 2%. For patients younger than 69 years and for those 69 years of age or older OS rate at 2 years was 17 ± 4% and 9 ± 3%, respectively (p=0.03). The interval between CR1 and first relapse also affected 2 year-OS with 7 ± 3%, 15 ± 4% and 36 ± 12% for pts. with relapse up to 6 months, 7 to 18 months and later than 18 months after CR1, respectively ( 18 months: p=0.009). OS rate at 2 years was also influenced by cytogenetic risk at initial diagnosis with 17 ± 3% for pts. having good or intermediate risk cytogenetics and 3 ± 2% for those with poor risk cytogenetics (p< 0.0005). Prior alloSCT had a negative influence on OS. Two-year OS rate was 10 ± 5 and 13 ± 3% (p= .015) for patients with prior alloSCT vs. those without prior alloSCT, respectively. FLT3/NPM1 mutational status at diagnosis had no impact on OS.

In univariate analysis age at relapse (p<0.04), interval from CR1 to relapse (p< 0.0005), cytogenetic risk at initial diagnosis (p<0.02) and prior alloSCT (p<0.02) were shown to be prognostic factors for OS, whereas FLT3/NPM1 mutational status was not significant (p=0.82). In multivariate analysis the same factors remained significant but only interval from CR1 to relapse (p<0.0005) and prior alloSCT (p=0.003) were independent.

In AML patients >60 years in first relapse OS is poor. Longer interval from CR1 to relapse and no prior alloSCT are independent beneficial prognostic factors for OS. FLT3/NPM1 mutational status at diagnosis has no prognostic impact on OS.

Wedding:Roche: Speakers Bureau; Amgen: Speakers Bureau; Chugai: Speakers Bureau; Janssen-Cilag: Speakers Bureau; Novartis: Speakers Bureau; Cephalon: Speakers Bureau; Prostarkan: Speakers Bureau; Pfizer: Speakers Bureau. Niederwieser:Novartis: Consultancy.