A novel predictive model of outcome in acute myeloid leukemia without favorable karyotype based on treatment strategy, karyotype and FLT3‐ITD mutational status

Prognosis in adult patients with acute myeloid leukemia (AML) without favorable karyotype is significantly worse than that of patients with favorable karyotype, especially in older patients; in 1 study of 1065 intensive chemotherapy (IC)-treated patients (median age of 66 years), the reported 5-year survival rates were 34% with favorable, 13% with intermediate-risk, and 2% with adverse karyotype.1 In younger AML patients without favorable cytogenetic abnormalities, especially in the presence of normal karyotype, the presence of internal tandem duplication of fms-like tyrosine kinase 3 (FLT3-ITD) has been associated with worse outcome2 and, in its absence, the presence of nucleophosmin 1 (NPM1)3 or biallelic mutation in CCAAT/enhancer binding protein gene (CEBPA) favored survival.4-6 In the current study of unselected and consecutive adult patients with AML, we appraised the prognostic performance of cytogenetic and FLT3-ITD/NPM1 mutation information and developed a practical risk model that accounted for not only genetic and clinical variables, but also the specific treatment strategy undertaken, including achievement or not of complete remission (CR) or CR with incomplete hematologic recovery (CRi) and the inclusion of allogeneic stem cell transplant (ASCT) for postremission consolidation. The current study was approved by the Mayo Clinic Institutional Review Board. Diagnosis of AML and its subclassification was according to World Health Organization criteria.7 Conventional response criteria were used for CR and CRi assignment.8 Conventional statistics was used for analysis of overall survival, calculated from the initial diagnosis of AML to date of death or last follow-up. Three hundred eighty adult patients (median age 63 years; range 21-89; 59% males) with AML met the study inclusion criteria: that is, karyotype was not favorable and all patients had FLT3-ITD and NPM1 mutation information. Presenting features, treatments received, and responses achieved are outlined in Supporting Information Table S1, which also depicts univariate and multivariate impact of these features on survival. AML subcategories included primary (n = 228; 60%), secondary (n = 111; 29%), and therapy-related (n = 41; 11%). Cytogenetic risk distribution, according to the European LeukemiaNet classification, was intermediate risk in 273 (72%) and adverse risk in 107 (28%) patients. FLT3-ITD and NPM1 mutational frequencies were 17% and 28%, respectively; in addition, CEBPA mutation information was available in 217 patients, including 24 (11%) mutated cases. Treatment included intensive chemotherapy (IC; “7 + 3” or equivalent) in 277 (73%) patients, less aggressive chemotherapy including the use of hypomethylating agents (HMAs) in 52 (14%) patients and supportive care alone in 51 (13%) patients. CR and CRi were achieved in 42% and 39% of IT-treated patients, respectively; neither CR nor CRi was seen with less aggressive chemotherapy or under supportive care management. ASCT was undertaken in 109 (29%) patients, including 105 patients with CR/CRi. Respective CR and CRi rates were 48% and 38% for primary, 30% and 39% for secondary, and 28% and 56% for therapy-related AML (P = .01); 45% and 41% for intermediate-risk karyotype and 33% and 35% for adverse risk karyotype (P < .01); 48% and 38% for FLT3-ITD mutated and 41% and 40% for its wild-type counterpart (P = .5); 63% and 32% for NPM1 mutated and 32% and 43% for its wild-type counterpart (P < .01); 36% and 41% for age >60 years and 48% and 38% for age 60 years or younger (P = .09); and 40% and 46% in women and 44% and 33% in men (P = .04). After a median follow-up of 13 months, 206 (54%) deaths were documented; median survival was 22 months with 1-, 3-, and 5-year survival rates of 63%, 41%, and 35%, respectively (Figure 1). Figure 1B-D illustrates survival data stratified by AML subcategories (Figure 1B), treatment received (Figure 1C), and response attained with IC (Figure 1D). In univariate analysis, significant risk factors for survival included advanced age, male sex, secondary and therapy-related AML subcategories, adverse karyotype, NPM1 wild type, FLT3-ITD/NPM1 profiles other than FLT3-ITD−/NPM1+, treatment other than IC, not achieving CR or CRi and not receiving ASCT (Supporting Information Table S1); in multivariable analysis that included all 380 patients revealed independent survival impact for adverse karyotype, FLT3-ITD, FLT3-ITD/NPM1 profiles other than FLT3-ITD−/NPM1+, not achieving CR or CRi and not receiving ASCT (Supporting Information Table S1). Age-adjusted multivariable analysis of patients receiving IC (n = 277) identified the following as independent predictors of favorable outcome: achievement of CR/CRi (HR 0.2, 95% CI 0.1-0.3; 2 risk points), undergoing ASCT (HR 0.4, 95% CI 0.3-0.6; 1 risk point) and absence of adverse karyotype (HR 0.3, 95% CI 0.2-0.5; 1 risk point) and absence of FLT3-ITD (HR 0.5, 95% CI 0.3-0.8; 1 risk point; Supporting Information Table S1). Subsequently, HR-weighted risk point allocation allowed development of a practical predictive model of survival outcome that took into consideration treatment strategy and response as well as genetic risk factors (Figure 2); 1-, 3-, and 5-year survival rates were 89%, 66%, and 58% for low risk, 61%, 30%, and 30% for intermediate risk and 21%, 0%, and 0% for high risk (Figure 2); the model was equally effective when applied to patients who did not receive ASCT (Figure 2B) or those with adverse karyotype (Figure 2C) or with FLT3-ITD (Figure 2D). Supporting Information Figure S1 provides a consort diagram that summarizes the above-elaborated events and underscores the relative need for ASCT in patients with adverse karyotype and, to a lesser extent, in those with FLT3-ITD. Finally, in patients not receiving induction chemotherapy (n = 104), only adverse karyotype (HR 2.0, 95% CI 1.3-3.1) predicted worse survival (median 3 vs 9 months) but 5-year survival remained <10%, regardless. The current study, which constitutes real-life experience in adult AML without favorable karyotype, confirms the primary importance of achieving CR or CRi, as an essential first step toward long-term remission; median survivals (5-year survival rates) were 69 months (53%) for IC-treated patients achieving CR/CRi; 8 months (0%) for IC-treated patients not achieving CR/CRi; 11 months (8%) for patients receiving HMA-like less aggressive chemotherapy; and 3 months (7%) for patients managed with supportive care only. In this regard, it was important to appreciate the fact that CR/CRi was attained only with IC and that survival outcome was not significantly different between CR and CRi (P = .3); both the former and latter observations differ from what has been reported by other investigators. It was also encouraging to witness adequate CR/CRi rates in patients older than 60 years (77% vs 85%; P = .09) and in those with secondary or therapy-related AML (70% vs 83% vs 86% in primary; P = .01), FLT3-ITD (86% vs 81%; P = .36) or adverse karyotype (68% vs 86%; P < .01). On the other hand, outcome was dismal in the absence of CR/CRi, regardless of treatment received. Once CR/CRi has been achieved, there is a reasonable chance of long-term remission, even with chemotherapy alone; median survivals (5-year survival rates) were “not reached” (60%) for CR/CRi followed by ASCT versus 43 months (46%) for CR/CRi that was not consolidated with ASCT (P < .01; Supporting Information Figure S1). The additional survival benefit of ASCT was most apparent in patients with adverse karyotype and, to a lesser extent, in those with FLT3-ITD. Put in other words, ASCT partially overcame the adverse impact of genetic risk factors, but was not essential for realizing long-term remission. Nothing to report. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.