Abstract
Introduction: Response to hypomethylating agents (HMA) plus BCL-2 inhibitor is a key determinant of clinical outcomes in elderly Acute Myeloid Leukemia (AML). Recently, Donher et al., developed a 4-mutation risk model including FLT3, N/K RAS, and TP53 that accurately allocated HMA plus BCL-2 inhibitor responses to superior, intermediate and low benefit, a model subsequently validated by investigator at the MD Anderson Cancer Center. However, generalization and adoption of the molecular Prognostic Risk Signature (mPRS) requires real-world applicability. The primary study aim was to evaluate Complete Remission (CR) plus Complete Remission incomplete (CRi) rate in elderly Acute Myelogenous Leukemia (AML) patients when accounted for mPRS. Secondly, given FLT3 ITD and N/KRAS preferential distribution in monocytic/granulocytic immunophenotypic subtypes, we explore mPRS effect on treatment response by AML immunophenotypic differentiation arrest, as defined by multiparameter flow cytometry (e.g., CD34, CD117, HLADR, CD33/CD13, MP0, CD14, CD64 etc. variable expression). Methods: After IRB approval, 123 AML cases were selected. All patients received HMA plus BCL-2 inhibitor. Response was evaluated after 1 cycle of therapy. Descriptive statistics summarized data. Chi-square and t-test were used to evaluate associations between molecular risk categories and treatment outcome. Results: Median age was 71 years (y), range 60-93. 73/123 (59.3%) were male. ELN22 risk showed that 13/123 (11%), 22/123 (18%) and 88/123 (72%) of cases were favorable (fav), intermediate (int) and adverse (adv). CR+CRi rate was observed in 9/56 (9%) of fav, 14/56 (25%) of int, 37/56 (66%) of adv ELN22 risk groups, p=0.09. However, using the 4-gene mPRS, the CR+CRi rate was 43 % in the superior benefit group, 37.5 % in the intermediate group, and 19 % in the low-benefit group, p=0.03. The low-benefit subgroup had a higher proportion of complex karyotype (91%) compared to the intermediate (10%) and superior (7.6%) groups (p <.0001). Normal karyotype was observed in 41%, 55% and 5% of superior, intermediate and low risk benefit group, p= 0.008. To validate the effect of the 4 -gene classifier by AML immunophenotypic differentiation arrest subtype, we examined HMA plus BCL-2 inhibitor response by HPC [Hemopoietic stem cell+ Common myeloid progenitor+ Multipotent progenitor] and maturing [Monocytic + granulocytic] like AML cases. Among HPC-like AML cases, superior, intermediate, and low response rate were observed in 56%, 22% and 22%, p=0.04, respectively, while the corresponding rates in maturing-like cases were 31%, 50%, and 19%. Interestingly, P53 mutations (43.1%) and myelodysplasia related mutations [e.g., ASXL1, BCOR, RUNX1, splicing factors] were more frequently observed in HPC-like AML. However, FLT3 ITD mutations were observed in 33/43 (79%) of maturing-like and 8/43 (21%) of HPC-like AML subtypes (p <0.001), while N/KRAS mutations were observed in 62% and 38%, respectively (p=0.1). Conclusions: The mPRS appears to improve stratification of treatment benefit among older AML patients receiving HMA plus BCL-2 inhibitor therapy. By clustering AML cases in differentiation arrest subtypes, HPC-like AML aggregated most of the superior risk benefit. Our data suggests that “immunophenotypic differentiation arrest” could assist mPRS in predicting HMA+ BCL2 inhibitor response in AML.
