PF260 PROGNOSTIC AND PREDICTIVE IMPACT OF NPM1/FLT3‐ITD GENOTYPES AS DEFINED BY 2017 EUROPEAN LEUKEMIANET RISK CATEGORIZATION FROM AML PATIENTS TREATED WITHIN THE INTERNATIONAL RATIFY STUDY

Background: Patients (pts) with AML harboring FLT3 internal tandem duplications (ITD) have poor outcomes, in particular pts with a high ITD mutant to wild‐type (wt) allelic ratio (AR; ≥0.5). The 2017 update of the European LeukemiaNet (ELN) recommendations addressed the importance of the ITD AR in the genetic risk stratification and defined distinct NPM1 / FLT3 ‐ITD genotypes. Aims: In this exploratory post‐hoc analysis from randomized pts treated within the RATIFY trial evaluating the addition of midostaurin to standard chemotherapy, we investigated the prognostic impact of the NPM1/FLT3 ‐ITD genotypes according to the 2017 ELN risk categories. Methods: In total, 319 of 717 pts with FLT3 ‐ITD positive AML were included who gave informed consent for biomarker analyses and who could be categorized according to 2017 ELN genetic risk stratification. Median age was 47.9 years (range, 18–60); median follow‐up time was 4.8 years. Results: Pts were categorized as follows: (i) favorable‐risk (n = 85), NPM1 mut / FLT3 ‐ITD low , irrespective of secondary chromosome abnormalities and concurrent gene mutations; (ii) intermediate‐risk (n = 111), NPM1 mut / FLT3 ‐ITD high and NPM1 wt / FLT3 ‐ITD low , both subgroups without adverse cytogenetics and without concurrent RUNX1 , ASXL1 , TP53 mutations; (iii) adverse‐risk (n = 123), NPM1 wt / FLT3 ‐ITD high ; as well as NPM1 mut / FLT3 ‐ITD high and NPM1 wt / FLT3 ‐ITD low with adverse‐risk cytogenetic and/or molecular markers; frequencies of RUNX1 , ASXL1 , and TP53 mutations in adverse‐risk AML were 29.2%, 19.8%, and 1.9%, respectively. Rates of allogeneic transplantation (alloSCT) in first complete remission (CR) were not different among the groups, with 29.4%, 35.1%, and 35.0% in pts. with favorable‐, intermediate‐, and adverse‐risk AML, respectively. Rates of CR (including responses after induction 1 and 2) were negatively associated with adverse‐risk genetics (51.2% vs 69.4% and 64.0% in favorable‐ and intermediate‐risk, respectively; p = 0.02); multivariable logistic regression analysis revealed adverse‐risk as significantly different compared to favorable‐risk (odds ratio [OR] 0.534, 95% confidence interval [CI] 0.287–0.979; p = 0.044), whereas no difference was seen between the two other risk groups; treatment with midostaurin had no impact (p = 0.32). Overall survival (OS) differed among the 3 ELN‐risk groups; 5‐year survival rates were 63%, 43%, 33% for favorable, intermediate‐ and adverse‐risk groups, respectively. The OS curves did not show differential effects of midostaurin among the 3 risk groups (Figure 1), with a benefical effect seen for midostaurin vs placebo in all 3 groups with hazard ratios (HR) estimated as HR = 0.52, HR = 0.51, and HR = 0.55 for favorable‐, intermediate‐, and adverse‐risk groups, respectively. A multivariable Cox model adjusted for age, WBC, sex, and bone marrow blasts was used to test the effects of treatment and risk classification. The resulting model revealed an increasing HR with increasing risk category ( p  < 0.001; intermediate vs. favorable, HR = 1.70, 95% CI 1.08–2.68; adverse vs. favorable risk, HR = 2.48, 95% CI 1.59–3.86), while treatment with midostaurin significantly reduced risk compared to placebo ( p  < 0.001; HR = 0.53, 95% CI 0.38–0.72). Summary/Conclusion: Data from this post‐hoc exploratory analysis confirm the prognostic value of the 2017 ELN risk categories also among AML pts with the distinct NPM1 / FLT3 ‐ITD genotypes. A Cox model revealed a significant effect for midostaurin compared to placebo independent of the ELN risk groups. Analysis with regard to potential confounding effects of allogenic transplantation are ongoing.