Outcomes following lower-intensity therapy with venetoclax in patients with NPM1-mutated Acute Myeloid Leukemia are highly dependent upon co-occurring mutations

Abstract

Abstract Background Mutated nucleophosmin-1 (NPM1m) is present in ~ 30% of patients (pts) with newly-diagnosed acute myeloid leukemia (ND-AML). A subset of pts with NPM1mhave improved overall survival (OS) following venetoclax (VEN) combined with hypomethylating agents (HMA; azacitidine or decitabine) or low-dose cytarabine (LDAC) treatment. Yet outcomes of pts with NPM1m AML treated with VEN-based therapy are highly dependent on co-mutations whereby the presence of signaling pathway mutations associate with worse OS (Döhner et. al. Blood 2024; Othman et. al. Blood 2024). Current mutational subgroup analyses are limited by sample size. Which co-occurring mutations best stratify outcomes in NPM1m AML following VEN-based lower intensity therapy is unclear. We hypothesized that studying pts with NPM1m AML who experienced poor clinical outcomes would better identify those who benefit most from investigational therapies. Methods We assembled a cohort of 2,089 pts withND-AML receiving frontline VEN+HMA (n=2,057) or LDAC (n=32) at academic medical centers in the US, UK, France, or Germany with available clinical, cytogenetic, and molecular data. OS was measured from therapy start, estimated with Kaplan-Meier curves, and compared using the log-rank test. Categorical features were compared using the Fisher Exact test. Multivariable analysis utilized AIC-based backwards selection of variables with allogeneic hematopoietic cell transplantation (HCT) included as a time-dependent covariate. Results 301 (14%) pts had NPM1m AML. Median age was 74 years (range: 20-93). Most (91%) received HMA-based therapy and had de novo AML (79%), while 21% had secondary AML including 14% arising from an antecedent hematologic disorder. 63/172 pts (37%) with available data had monocytic AML. Hematopoietic cell transplantation (HCT) was performed in 24 (8%) pts. Frequent co-mutations included DNMT3A (33%), TET2 (27%), FLT3-ITD (26%), IDH2 (22%), SRSF2 (16%), NRAS (14%), IDH1 (12%), FLT3-TKD (12%), and PTPN11 (8%). 4-gene classifier/molecular prognostic risk (mPRS) was higher in 60%, intermediate in 39%, and lower-benefit in 2%. A diploid karyotype was most common (70%), followed by other ELN 2022 intermediate-risk cytogenetic abnormalities (19%). ELN 2022 adverse-risk cytogenetics (1%) and complex karyotype (1%) were uncommon. After median follow-up of 22 months, median OS (mOS) was 22 months (95% CI: 14.4-29.6). No OS difference was observed in pts receiving HMA vs. LDAC+VEN (p=0.89). OS varied across mPRS higher, intermediate, and lower-benefit groups with mOS of 26.6, 13.8, and NR (24-month OS: 53%)(p=0.40). Co-occurring mutations in IDH1/2 associated with a reduced risk (HR: 0.62, 95% CI: 0.42-0.89, p=0.011) while mutations in FLT3-ITD (HR: 1.42, 95% CI: 0.99-2.04, p=0.055) and TET2 (HR: 1.76, 95% CI: 1.24-2.50, p=0.002) associated with an increased risk of death. Median OS in pts with vs. without triple-mutant NPM1/DNMT3A/FLT3-ITD (n=33) was 10.9 vs. 26.6 months (p=0.010). Other signaling pathway variants demonstrated no significant impact on OS including NRAS (p=0.73), PTPN11 (p=0.218), KRAS (p=0.69), or FLT3-TKD (p=0.91). When stratifying pts with mutated IDH1/2 and wild-type signaling pathway genes (FLT3-ITD, N/KRAS, PTPN11) (n=66) vs. mutated signaling pathway genes (n=133) vs. other mutations (n=102), mOS was 45 vs. 16.1 vs. 14.7 months (p=0.20). Backward AIC-based selection of gene mutations identified in ≥10% of the study population identified FLT3-ITD (p=0.04) and TET2 (p=0.004) as the top candidate predictors of OS in NPM1m AML. When dichotomized into NPM1-low (n=164; FLT3-ITD and TET2 wild-type) vs. NPM1-high (n=137; FLT3-ITD or TET2 mutated) risk groups, mOS was 43 vs. 11.5 months (HR: 0.53, p < 0.001), with consistent effects observed after inclusion of HCT (HR: 0.54, p < 0.001), and IDH1/2 mutation status (HR: 0.59, p=0.003). Pts with NPM1-high AML had a mOS similar to NPM1 wild-type AML (11.5 vs. 11.7 months, p=0.46). NPM1-low vs. high was associated with differences in complete response/complete response with incomplete count recovery rates (CR/CRi; 80% vs. 69%, p=0.028), and MRD-negative remissions measured via NPM1 qPCR (n=49; 63% vs. 38%, p=0.13). Conclusion Survival outcomes following lower-intensity venetoclax-based therapy in NPM1m AML are highly dependent upon co-occurring mutations. Development of risk stratification tools may identify patients who experience sub-optimal outcomes and benefit most from alternative therapeutic approaches.