Choice of hypomethylating agent for newly diagnosed TP53-mutant Acute Myeloid Leukemia: A command registry study

Abstract Introduction:Owing to inherent chemoresistance and poor outcomes in TP53 mutant (m) acute myeloid leukemia (AML), treatment strategies are aimed to optimize response while minimizing toxicity. Hypomethylating agents (HMA; azacitidine AZA or decitabine DEC) with or without venetoclax (VEN)/investigational agents has been a preferred approach. While AZA and DEC demonstrated comparable efficacy in treating patients (pts) with AML (Zeidan et al., Blood 2022), previous data suggest that DEC may induce superior TP53m clearance and response rates (Welch et al., NEJM 2016). However, there is no head-to-head comparison with AZA in TP53m AML. In this study, we sought to explore the difference in clinical outcomes between AZA- and DEC-based induction in TP53m AML. Methods: We conducted a multicenter observational study in collaboration with 15 U.S. academic centers under the COMMAND consortium. We reviewed the electronic medical records of 652 TP53m AML pts and identified 321 pts who received either DEC (n=183 57%) or AZA (n=138 43%) alone or in combination with VEN for induction and analyzed differences in clinical outcome. To minimize selection bias, we performed propensity score matching (PSM) analysis between two groups (gps) to further analyze difference in clinical outcome in matched cohort. Results: The median age (years) of the pts in AZA (71.5 range (R),44-90) and DEC (70 R,29-88) gps were comparable (p= 0.15). Proportion of pts with secondary (s) AML (41% vs 36%, p=0.41), bone marrow blast ≥ 30% (57% vs 56%, p= 0.99), complex cytogenetics (CG; 93% vs 88%, p= 0.25) and multi-hit TP53m (89% vs 91%, p=0.70) were also comparable between AZA and DEC gps, respectively. Similarly, the proportion of pts who received venetoclax combination during induction (75% vs 71%, p= 0.44) were comparable in AZA and DEC gps, respectively. 113 (82%) and 145 (79%) were evaluable for response post induction in AZA and DEC gps, respectively. The complete remission rates with or without count recovery (CR/CRi) were observed in 35% and 49.7% of pts in AZA and DEC gps (p= 0.02), respectively. Similarly, the CR/CRi rates were 36.5% and 63.5% with AZA+VEN and DEC+VEN, respectively (p= 0.01). 11 (10%) and 15 (11.5%) pts were bridged to allogeneic stem cell transplantation (allo-HCT) after CR1 in AZA and DEC gps (p= 0.75), respectively. The median follow-up on the entire cohort was 5.3 mo (R, 0.2 to 59). The median event free survival (EFS; 3.1 vs 4.5 mo 95% CI: 0.88-1.42, p= 0.49) and overall survival (OS; 5.8 vs 5.83 mo 95% CI: 0.84-1.35, p= 0.68) was not different between AZA and DEC gps, respectively. We conducted a multivariable analysis (MVA) for EFS and OS using parsimonious models to help identify the strongest predictors for survival. For EFS, achievement of CR/CRi (HR= 0.26, 95% CI: 0.18-0.37, p=0.001) retained favorable and complex CG (HR= 1.81, 95% CI: 1.03-3.23, p=0.03) retained unfavorable significance. Similarly, for OS, achievement of CR/CRi (HR= 0.32, 95% CI: 0.23-0.45, p=0.001) remained favorable and complex CG (HR= 2.42, 95% CI: 1.36-4.34, p=0.003) retained unfavorable significance. HMA plus venetoclax combination or allo-HCT did not retain significance for EFS or OS. We also analyzed response rate, survival, and conducted multivariable analysis in a propensity matched cohort of 216 pts, 108 each in AZA and DEC gps. The CR/CRi rates were not significantly different in AZA (45%) and DEC (55%) matched cohorts (p= 0.38), respectively. The median EFS (3.6 vs 5.16 95% CI: 0.80-1.42, p= 0.61) and OS (7.07 vs 6.6 95% CI: 0.84-1.35, p= 0.76) were also not significantly different. In MVA, achievement of CR/CRi (HR= 0.31, 95% CI: 0.22-0.44, p=0.001) and allo-HCT (HR= 0.19, 95% CI: 0.09-0.38, p=0.001) retained favorable and HMA+VEN combination (HR= 1.52, 95% CI: 1.06-2.18, p0.001) retained unfavorable significance for EFS. Similarly, for OS, achievement of CR/CRi (HR= 0.43, 95% CI: 0.31-0.60, p=0.001) and allo-HCT (HR= 0.15, 95% CI: 0.07-0.32, p=0.001) retained favorable significance. While complex CG (HR=2.42, 95% CI: 1.32-4.42, p=0.004) and HMA+VEN combination (HR= 1.15, 95% CI: 1.06-2.18, p0.001) retained unfavorable significance for OS. Conclusion: Our results from large multi-institutional collaboration suggest that AZA- and DEC-based induction has comparable outcome in TP53m AML. We also observed that the addition of venetoclax to HMA has a negative impact on survival in propensity matched cohort.