Genetic Abnormalities Associated with Outcomes in Unfit Patients with Acute Myeloid Leukemia Receiving Less Intensive Therapy in the Real World

Background

Less intensive regimens including azacitidine plus venetoclax or low-dose cytarabine based therapy are recommended as standard chemotherapy in elderly or unfit patients with acute myeloid leukemia (AML) except APL. ELN 2022 risk classification by genetics at initial diagnosis of AML was based on data from patients younger than 60 years receiving intensive chemotherapy in clinical trials. There is still limited data on whether this classification is applicable to unfit patients receiving less intensive therapy in the real word.

Objectives

This study aims to assess whether the ELN 2022 risk classification is applicable to the unfit patients receiving less intensive therapy and explore the variables which significantly affect their outcomes.

Methods

Data of consecutive newly-diagnosed elderly patients or > 55 years unfit patients with AML receiving less intensive chemotherapy at Peking University People's Hospital from January 2016 to June 2024 were reviewed. Data of cytogenetics and gene mutations were obtained. Cox regression model was utilized to identify variables associated with survival.

Results

A total of 488 patients was included in the study. Median age was 64 (IQR, 61-68) years. 227 (47%) was female. 148 (30%) had ≥ 1 comorbidity(ies). According to ELN risk classification, 113 (23%) patients was in favorable-risk; 119 (24%), intermediate-risk; 256 (53%), adverse-risk. 252 (52%) patients received azacitidine plus venetoclax regimen; 236 (48%), low-dose cytarabine based chemotherapy. With a median follow-up of 9 (IQR, 3-24) months, 284 (58%) patients achieved CR/CRi after first induction therapy; 375 (77%), ultimately. 2-year survival rate was 50% (95%CI: 44-60%). There was no difference in survival between those receiving azacitidine plus venetoclax regimen and low-dose cytarabine based chemotherapy (p = 0.229). By ELN risk classification, favorable-risk group (67%, 55-79%) had higher 2-year survival rate than intermediate-risk group (40%, 29-52%; p < 0.001); while, no significant difference was observed between intermediate-risk and high-risk groups (42%, 34-50%; p = 0.425).

In multivariate analysis ASXL1 (p = 0.036), FLT3-ITD+/NPM1- (p = 0.038), FLT3-TKD mutation (p = 0.017), GATA2 (p < 0.001), -5/del(5q)/t(5q)/add(5q) (p = 0.001), and -7/del(7q) (p = 0.003) were genetic abnormalities significantly-associated with poor survival; while, CEBPA-bZIP in frame mutation (p = 0.002) and IDH2 (p = 0.008), favorable survival. In addition, female, higher blasts and lower PLT count were adverse clinical variables. Next, multivariate analyses were performed to investigate the prognostic variables in the azacitidine plus venetoclax regimen group and low-dose cytarabine based chemotherapy group, respectively. Besides the identified genetic abnormalities with prognostic significance, monosomal karyotype emerged as an adverse indicator for survival in the azacitidine plus venetoclax group; FLT3-ITD-/NPM1+, favorable indicator in low-dose cytarabine based chemotherapy group.

Based on number of genetic abnormalities with prognostic significance, patients were classified into the low- (1-2; n = 107; 22%), intermediate- (3; n = 198; 40%), high-risk (4; n = 121; 25%) and very high-risk (5-7; n = 62; 13%) cohorts with significant difference in survival with 2-year survival rates of 80% (70-90%), 48% (39-57%), 33% (21-45%) and 17% (3-31%) (p < 0.001). HRs (95% CI) with the low-risk cohort as reference were 2.9 (1.7, 5.0; p < 0.001), 5.6 (3.2, 9.7; p < 0.001) and 10.3 (5.7, 18.7; p < 0.001) (p-value for trend < 0.001). Sensitivity analyses showed that an integral system based on these identified genetic abnormalities can significantly distinguish the prognostic risk in the patients receiving azacitidine plus venetoclax regimen or low-dose cytarabine based chemotherapy. The results remain consistent regardless of whether censored at hematopoietic stem cell transplantation or not.

Conclusions

The ELN 2022 risk classification may be inadequate for unfit patients with AML receiving less intensive therapy. We identified genetic abnormalities as prognostic variables which might facilitate a more precise stratification of disease risk for these patients in the real world.