Genetic alteration patterns and clinical outcomes of elderly and secondary acute myeloid leukemia

Abstract To illustrate the clinical and genetic features of elderly and secondary acute myeloid leukemia (AML) patients, we compared 145 elderly AML (e‐AML) and 55 secondary AML (s‐AML) patients with 451 young de novo AML patients. Both e‐AML and s‐AML patients showed lower white blood cell (WBC) and bone marrow (BM) blasts at diagnosis. NPM1, DNMT3A, and IDH2 mutations were more common while biallelic CEBPA and IDH1 mutations were less seen in e‐AML patients. s‐AML patients carried a higher frequency of KMT2A‐AF9. In treatment response and survival, e/s‐AML conferred a lower complete remission (CR) rate and shorter duration of event‐free survival (EFS) and overall survival (OS) compared with young patients. In multivariate analysis, s‐AML was an independent risk factor for OS but not EFS in the whole cohort. Importantly, intensive therapy tended to improve the survival of e/s‐AML patients without increasing the risk of early death, and hematopoietic stem cell transplantation (HSCT) could rescue the prognosis of s‐AML, which should be recommended for the treatment of fit patients.

past decades, however, most of previous studies focused on de novo AML especially those patients with young age, 7,8 while reports regarding genetic alterations and their prognostic significance in e/s-AML are still rare.
More importantly, the treatment of e/s-AML remains controversial. Various modalities, such as hypomethylation agents as exemplified as decitabine and azacitidine, and low doses chemotherapy were tried in this group of patients; however, no therapeutic regimen was proved to be significantly superior to traditional chemotherapy. To some extent, the treatment decision was strongly dependent on the fitness of AML patients.
In this study, we examined genetic alterations and post-treatment minimal residual diseases (MRD) in order to illustrate their distribution and prognostic impact in e/s-AML and to provide treatment recommendations for those patients.

Cytogenetic risk stratification was based on 2017 European
LeukmiaNet (ELN) recommendations. 9 This study was approved by the ethic committee of Ruijin Hospital. All patients had given informed consent for both treatment and cryopreservation of bone marrow (BM) and peripheral blood according to the Declaration of Helsinki.

| Treatment protocols
For young de novo patients (younger than 60 years old), standard intensive "3 + 7" induction regimens (idarubicin 10-12 mg/m 2 or daunorubicin 45-60 mg/m 2 , D1-3; cytarabine 100 mg/m 2 D1-7) were given as the initial induction therapy. If CR was achieved, four cycles of high-dose cytarabine (2 g/m 2 ) was given as consolidation. For e-AML (60 years and older) and s-AML patients, the treatment was mainly decided by the physician in consideration of the fitness of patients and risk of disease. Fit patients received treatment similar to young patients but reduced cycles of consolidation to 2 cycles of high-dose cytarabine; unfit patients received "3 + 7" regimens with reduced dose, hypomethylation treatment or palliative treatment according to the physician's decision. Genetic alterations including FLT3-ITD/TKD, KMT2A-PTD, NPM1,   NRAS, CKIT, CEBPA, DNMT3A, IDH1, IDH2, RUNX1-RUNXT1T1, CBFβ-MYH11, KMT2A rearrangements were detected as previously reported. 10 Bone marrow aspirate samples were processed according to the standard procedure of our institution as previously reported. 11 Detection of MRD after induction therapy was based on leukemiaassociated immunophenotype (LAIP) at diagnosis and performed by ten-color multiparametric flow cytometry. MRD was considered positive when leukemia cells were greater than or equal to 0.01%.

| Statistical analyses
Complete remission (CR) was defined by the criteria of the International Working Group. 12

| Characteristics of patients
The baseline characteristics of patients were shown in Table 1.
Patients with s-AML presented female predominance (P = .041), most of whom had a previous history of breast carcinoma (43%). Older age (P < .001), lower white blood cell (WBC) count (P = .031), hemoglobin (HB, P = .004), and BM blasts (P < .001) were observed in s-AML as compared with young patients. Similarly, elderly patients showed lower WBC (P = .036) and BM blasts (P = .009) at diagnosis. In WHO subtype distribution, higher frequency of pure erythroid leukemia was seen in s-AML (P = .013). Both e-AML and s-AML patients received less intensive induction, but more hypomethylation treatment and palliative treatment than younger patients (all P < .001).

| Cytogenetic and genetic alterations
In cytogenetic classification, elderly patients had a significantly higher proportion of intermediate risk cytogenetics (P = .011). Favorable cytogenetic alterations were less frequent in both elderly and secondary patients (P = .008 and.014, respectively) as compared with young patients.
As for the association between genetic abnormalities and clinical features, NPM1 mutations were associated with higher WBC in elderly patients (P = .037). Moreover, s-AML patients with KMT2A-AF9 were prone to having higher BM blasts (P = .068) (Table S1).

| Treatment responses
In total cohort, CR rate and ED rate were 76.1% and 10.5%, respectively. Both s-AML and e-AML patients conferred reduced CR rate as compared with young patients (s-AML vs young: 58% vs 83%, P < .001; e-AML vs young, 60.7% vs 83%, P < .001). Additionally, a higher frequency of ED (e-AML vs young: 16.6% vs 8%, P = .003) was observed in e-AML (Table 3). In order to find significant factors that can independently predict ED and CR, we conducted univariate and multivariate analyses (Tables S1 and 4).
Among patients achieving CR, 258 young, 56 elderly, and 23 secondary AML patients had a definite LAIP feature before treatment, and the MRD of whom could be monitored. The frequency of positive MRD was higher in s-AML than in young patients (P = .039, Table 3).
When e-AML and s-AML patients were put together, those who were treated with intensive therapy had a higher CR rate (74.2% vs 44.3%, P < .001) and tended to have a lower incidence of positive MRD (60.8% vs 82.1%, P = .051) than those treated with other therapy categories. In addition, e/s-AML patients receiving intensive therapy tended to have a lower ED rate than those who were treated with palliative treatment (12.7% vs 22.5%, P = .090).

| Impact of prognostic factors on survival
The median follow-up in all patients was 27 months (range, 0-66 months). Generally, e/s-AML patients had inferior EFS and OS com- Univariate analysis for EFS and OS was shown in Table S3. In order to explore the prognostic significance of increased age and s-AML after accounting for other recognized prognostic factors, we conducted multivariate analysis (Table 4). In whole cohort, s-AML relative to de novo AML was an independent risk factor for OS (P = .009), while it was not associated with EFS. Notably, the independent prognostic impact of s-AML on OS was lost when HSCT was not regarded as a censored event, suggesting that HSCT may abrogate the adverse impact of s-AML on survival to a certain extent (Table S4).

| DISCUSSION
Acute myeloid leukemia is a hematologic malignancy with a relative high incidence rate especially in high Human Development Index (HDI) countries. 13 The incidence of AML increases with age, which makes AML a tumor of the elderly population. 1    Recently, a study reported that CPX-351 could improve the response rates and survival of patients aged 60 to 75 with s-AML compared with standard 3 + 7 treatment. 34 The Food and Drug Administration (FDA) approved glasdegib and venetoclax for the treatment of patients over 75 years old, or young patients who have comorbidities that are not suitable for intensive induction chemotherapy. 35,36 We expect that the frontline use of these new drugs may improve the outcomes of e/s-AML individuals, which need to be compared with traditional intensive therapy in prospective research.
In summary, the incidence of e/s-AML is increasing and will be more common in the future, which merits our attention. Both elderly and secondary AML presented with distinct clinical, cytogenetic, and molecular features, whose prognosis remains dismal compared with