Clinical Outcomes Associated with <i>NPM1</i> Mutations in Newly Diagnosed AML

Background: Mutations of the Nucleophosmin 1 gene (NPM1mt) form the most common alteration in acute myeloid leukemia (AML) detected in 20-30% of cases. Despite the relatively favorable prognosis associated with NPM1mt, long-term outcomes remain suboptimal. Predictors of outcomes have not been fully elucidated. With the advent of menin inhibitors as a targeted therapy for NPM1mt AML, upcoming combination studies with these agents are intended; however expected outcomes with current therapies are largely unknown. Therefore, we conducted a retrospective analysis to delineate predictors of outcomes in patients (pts) with NPM1mt AML. Methods: We screened adult pts treated at our center for newly diagnosed AML between January 2012 and December 2023. We identified 396 pts (18%) with NPM1mt and 1819 pts (82%) with NPM1wt. The composite complete remission or complete remission with incomplete hematologic recovery (CRc) was identified by subgroup, and MRD was assessed by flow cytometry (sensitivity: 10-4). Event-free survival (EFS) was defined as the duration from the initiation of treatment until the earliest occurrence of disease relapse, disease progression, death from any cause, or failure of achieving a response (FDA Definition, i.e. event at day 0), with overall survival (OS) defined as the duration from therapy start to death or last follow-up. Predictors of survival were identified using Cox regression where factors with P&amp;lt;0.1 on univariate analysis were selected for multivariate analysis (MVA). Results: Compared with NPM1wt AML, pts with NPM1mt were more commonly female (56% vs 43%; P&amp;lt;0.001) and had significantly higher bone marrow blasts at diagnosis (P&amp;lt;0.001). Those with NPM1mt more commonly had a monocytic phenotype (27% vs 9%, P&amp;lt;0.001), a diploid karyotype (75% vs 24%, P&amp;lt;0.001), and myelofibrosis (MF-1 or higher) (30% vs 23%, P=0.008). Extramedullary disease (EMD) had similar incidence in both groups (3.5% vs 2.4%; P=0.2). Most common co-occurring mutations included DNMT3A (47% vs 18%, P&amp;lt;0.001), FLT3-ITD (42% vs 13%, P&amp;lt;0.001), KRAS/NRAS (29% vs 25%, P=0.2), SRSF2 (23% vs 28%, P=0.4) and TET2 (22% vs 14%, P&amp;lt;0.001). Therapy-related AML was 9% of NPM1mt vs 19% of NPM1wt AML (P&amp;lt;0.001). There were 16 pts (4%) with NPM1mt and an adverse karyotype; 44 pts (11%) had NPM1mt and co-occurring adverse risk mutations. High intensity chemotherapy (HIC) was given for 190 NPM1mt pts (48%), of which 46 pts (12%) had the addition of venetoclax (HIC+Ven), whereas 66 pts (17%) received a hypomethylating agent + Ven (HMA+Ven), 32 pts (8%) received cladribine with low-dose cytarabine + Ven (Clad LDAC Ven) and 106 pts (27%) received other lower-intensity treatments. The CRc rate for those with NPM1mt who received HIC without ven was 94% vs 96% for HIC+Ven, with MRD-neg rates of 86% and 82% respectively. Those with NPM1mt who received HMA+Ven had a CRc rate of 86%, whereas Clad LDAC Ven led to a 94% rate (MRD neg of 96% and 97% respectively). With a median follow-up of 52.6 months, NPM1mt pts treated with HIC had a median EFS and OS of 87.8 and 87.8 months with 2-year rates of 63% and 65%, respectively. NPM1mt pts treated with Clad LDA Ven had a median EFS and OS of 59.6 and 59.6 months with 2-year rates of 73% and 76% respectively. HMA+ven in NPM1mt (all of whom were &amp;gt;60 years) led to a median EFS and OS of 15.4 and 23.6 months, with 2-year rates of 48% and 51% respectively. In those with NPM1mt and age&amp;gt;75 years, the median EFS and OS were 10.3 months and 10.9 months, with 2-year rates of 32% and 32% respectively. Among pts with NPM1mt who received HIC, predictors of worse EFS and OS in MVA included age [Hazard ratio (HR) for OS: 1.03; P=0.009], FLT3-ITD (HR for OS: 1.92; P=0.014), EMD (HR for OS: 6.03; P=0.001), and a performance status &amp;gt; 2 (HR for OS: 8.01; P=0.001). The addition of venetoclax to HIC did not independently predict for outcomes. Among pts with NPM1mt who received HMA+Ven, the only predictor of worse EFS and OS was age (HR for OS: 1.10; P=0.016). Notably, therapy-related status or adverse risk features by ELN did not predict outcomes. Conclusion: In conclusion, older age, worse performance status, FLT3-ITD co-mutation, and EMD predicted for worse outcomes in NPM1mt AML treated with HIC whereas older age was the only predictor of outcomes in those receiving HMA + venetoclax.