Patients
A total of 213 patients matching the eligibility criteria were identified between 2013 and 2022. Thereof, 125 patients were treated with HMA-VEN and 88 with HMA monotherapy. Six patients had been excluded due to switch from HMA to HMA-VEN after having received more than 3 cycles of monotherapy, as described in the Methods section.
Patients in the HMA group were diagnosed earlier (interquartile range [IQR], January 2016 – November 2018) than patients who received HMA-VEN (December 2020 – April 2022). Only 3 patients (3.4%) still started HMA monotherapy after VEN received marketing authorization in April 2021.
Baseline characteristics were generally well balanced between groups (Table 1). Median age in the HMA-VEN and HMA cohorts were 76 (range, 37–91) and 74 years (54–94), respectively. There were no significant differences in the assignment to NCCN 2.2016 or ELN 2017 risk groups between cohorts. However, patients in the HMA-VEN cohort were in slightly better general condition with 56% having an ECOG PS ≤ 1 compared to 41% in the monotherapy group. Additionally, more patients in the combination group had de novo instead of secondary AML (64% vs. 47%).
Table 1
– Baseline Characteristics
Characteristic | HMA-VEN (N = 125) | HMA (N = 88) |
Sex – No. (%) | | |
Female | 46 (37) | 37 (42) |
Male | 79 (63) | 51 (58) |
Age – yr | | |
Median (Range) | 76 (37–91) | 74 (54–94) |
ECOG PS – No. (%) | | |
0–1 | 70 (56) | 36 (41) |
2–3 | 44 (35) | 48 (55) |
4 | 2 (2) | 0 (0) |
Missing | 9 (7) | 4 (5) |
Etiology – No. (%) | | |
De novo | 80 (64) | 41 (47) |
After MDS | 24 (19) | 28 (32) |
After MPN | 6 (5) | 8 (9) |
After MDS/MPN | 5 (4) | 2 (2) |
Therapy-related | 10 (8) | 9 (10) |
Missing | 0 (0) | 0 (0) |
ELN 2017 Risk – No. (%) | | |
Favorable | 14 (11) | 11 (12) |
Intermediate | 34 (27) | 27 (31) |
Adverse | 75 (60) | 48 (55) |
Missing | 2 (2) | 2 (2) |
NCCN Cytogenetic Risk – No. (%) | | |
Favorable | 2 (2) | 2 (2) |
Intermediate | 87 (70) | 54 (61) |
Poor | 33 (26) | 29 (33) |
Missing | 3 (2) | 3 (3) |
NPM1 Mutation – No. (%) | | |
Yes | 22 (18) | 10 (11) |
No | 94 (75) | 71 (81) |
Missing | 9 (7) | 7 (8) |
CEBPA Mutation* – No. (%) | | |
Yes | 1 (1) | 1 (1) |
No | 111 (89) | 77 (88) |
Missing | 13 (10) | 10 (11) |
RUNX1 Mutation – No. (%) | | |
Yes | 26 (21) | 4 (5) |
No | 91 (73) | 61 (69) |
Missing | 8 (6) | 23 (26) |
FLT3 TKD – No. (%) | | |
Yes | 6 (5) | 0 (0) |
No | 72 (58) | 41 (47) |
Missing | 47 (38) | 47 (53) |
FLT3 ITD – No. (%) | | |
Yes | 19 (15) | 4 (5) |
No | 67 (54) | 41 (47) |
Missing | 39 (31) | 43 (49) |
TP53 Mutation – No. (%) | | |
Yes | 22 (18) | 15 (17) |
No | 97 (78) | 58 (66) |
Missing | 6 (5) | 15 (17) |
ASXL1 Mutation – No. (%) | | |
Yes | 28 (22) | 11 (12) |
No | 92 (74) | 53 (60) |
Missing | 5 (4) | 24 (27) |
IDH1 Mutation – No. (%) | | |
Yes | 13 (10) | 4 (5) |
No | 68 (54) | 32 (36) |
Missing | 44 (35) | 52 (59) |
IDH2 Mutation – No. (%) | | |
Yes | 8 (6) | 6 (7) |
No | 73 (58) | 32 (36) |
Missing | 44 (35) | 50 (57) |
Bone marrow Blasts – % | | |
Median (Range) | 47 (11–95) | 34 (20–90) |
Missing – No. (%) | 6 (5) | 10 (11) |
Peripheral Blasts – % | | |
Median (Range) | 19.0 (0.0–91.0) | 6.5 (0.0–81.0) |
White Blood Count (/nl) | | |
Median (Range) | 7.0 (0.6–437.0) | 4.1 (0.4–239.0) |
Prior Treatments – No. (%) / Missing (%) | | |
HMA | 5 (4) / 8 (6) | 5 (6) / 1 (1) |
Hydroxyurea | 4 (3) / 4 (3) | 2 (2) / 0 (0) |
HSCT§ | 2 (2) / 0 (0) | 2 (2) / 0 (0) |
Baseline refers to the status prior to initiation of the respective treatments. (*) CEBPA mutations include both in-frame bZIP mutations and any biallelic mutations, due to changes in classification. (§) HSCT for prior MDS or MPN. HMA denotes hypomethylating agent, VEN venetoclax, ECOG PS Eastern Cooperative Oncology Group Performance Status, MDS myelodysplastic syndrome, MPN myeloproliferative neoplasm, ELN European Leukemia Network, NCCN National Comprehensive Cancer Network, NPM1 nucleophosmin 1, CEBPA CCAAT/enhancer-binding protein-α, RUNX1 Runt-related transcription factor 1, FLT3 fms like tyrosine kinase 3, TKD tyrosine kinase domain, ITD internal tandem duplications, TP53 tumor protein p53, ASXL1 putative polycomb group protein ASXL1, IDH1/2 isocitrate dehydrogenase 1/2, and HSCT hematopoietic stem cell transplantation. |
In the HMA-VEN cohort, 122 (98%) and 3 (2%) patients received AZA and DEC, respectively, compared to 58 (66%) and 30 (34%) in the HMA cohort.
Survival outcomes
With a median time of follow-up of 12.8 months in the HMA-VEN cohort (range, 0.4–61.7) and 26.6 months in the HMA cohort (0.2–67.2), OS was significantly longer in patients receiving HMA-VEN (Fig. 1A). Median OS in the HMA-VEN and HMA cohorts was 7.9 months (range, 5.1–14.7) and 4.9 months (3.1–7.1), respectively, with 42% (95% CI, 33–54) and 19% (12–30) of patients being alive after 1 year of treatment. The HR for death was 0.64 (95% CI, 0.46–0.88; p = 0.006).
Favorable OS in the HMA-VEN cohort compared to the HMA cohort was consistently found across most subgroups defined by baseline patient or tumor characteristics, including age, NPM1 and TP53 status, and NCCN and ELN 2017 risk groups (Suppl. Figure 1). Median OS in the HMA-VEN and HMA cohorts was 20.2 (95% CI, 20.2-not calculable [NC]) and 7.6 months (1.4-NC), respectively, in patients with ELN 2017 favorable risk disease compared to 9.5 (6.2-NC) and 4.9 months (2.8–11.0) with intermediate and 7.0 (4.2–12.0) and 4.6 months (2.6–7.1) with adverse risk AML (Suppl. Figure 2).
Surprisingly, in patients with ASXL1-mutant AML, OS was numerically shorter with HMA-VEN compared to HMA monotherapy (Suppl. Figure 3A). Multiple regression including an interaction analysis suggests a differential impact of ASXL1 status depending on the treatment received (P value for interaction, 0.04; Suppl. Figure 3B). However, numbers are too low to draw conclusions.
TTNT was significantly longer in the HMA-VEN cohort, with a median of 5.3 months (95% CI, 4.2–7.4) compared to 3.4 months (2.5–4.9) in the HMA cohort (Fig. 1B). HR for treatment switch or death was 0.63 (95% CI, 0.47–0.85; p = 0.002) favoring the combination cohort.
Treatment response
Of the 125 patients in the HMA-VEN cohort and 88 patients in the HMA cohort, 49 (39%) and 29 (33%) patients, respectively, were evaluable for response. More patients in the combination group achieved composite complete remission (cCR) as best response, which was defined as either CR or CR with incomplete (CRi) or partial (CRh) hematological recovery (Table 2). In the HMA-VEN group, 28 of 49 (57%; 95% CI, 42–71) evaluable patients had cCR compared to 4 of 29 (14%, 4–32) in the HMA group. In contrast, 9 of 49 (31%, 9–32) and 13 of 29 (45%, 26–64) evaluable patients had no response, respectively.
Table 2
– Best Response Best response according to European Leukemia Network 2022 criteria stratified by treatment. Patients without response assessment during the first three months of treatment were considered non-evaluable. (*) percentages in the response categories are calculated as the number of patients in the respective category divided by the number of evaluable patients in the treatment cohort. HMA denotes hypomethylating agent, VEN venetoclax, CR complete remission, CRi complete remission with incomplete recovery, CRh complete remission with partial recovery, MLFS morphologic leukemia-free state, and PR partial remission.
Response | HMA-VEN (N = 125) | HMA (N = 88) |
Non-Evaluable – No. (%) | 76 (61) | 59 (67) |
Evaluable* – No. (%) | 49 (39) | 29 (33) |
cCR | 28 (57) | 4 (14) |
CR | 17 (35) | 3 (10) |
CRi | 10 (20) | 1 (3) |
CRh | 1 (2) | 0 () |
MLFS | 8 (16) | 4 (14) |
PR | 4 (8) | 8 (28) |
No Response | 9 (18) | 13 (45) |
Supplementary Data |
In patients treated with HMA-VEN, achievement of cCR was associated with favorable OS with a HR for death of 0.25 (95% CI, 0.10–0.59; Fig. 2A). Median OS in HMA-VEN patients with cCR was 24.3 months (95% CI, 11.7-NC) compared to 5.0 (3.5-NC) in patients without cCR.
Since a large proportion of patients in a palliative treatment setting do not undergo response assessments outside of clinical trials, we evaluated whether the recovery of peripheral blood counts was also prognostic for OS. In the HMA-VEN cohort, 58 of 125 patients (46%) achieved blood count recovery, defined as neutrophils ≥ 1/nl and platelets ≥ 100/nl, and 67 (54%) did not. Patients with recovery exhibited a longer OS with a HR for death of 0.52 (95% CI 0.33–0.84; Fig. 2B). Median OS was 16.4 months (95% CI, 8.1–22.5) and 3.8 months (3.2–9.5), respectively.
Subsequent treatments
In the HMA-VEN and HMA cohorts, 27 of 125 (22%) and 31 of 88 (35%) of patients received at least one subsequent treatment for AML with the majority being low intensity regimens (Suppl. Table 2). Only 3 patients (3%) in the HMA cohort received venetoclax as part of any subsequent therapy. Intensive treatment was given in 1 (1%) and 10 (11%) patients in the HMA-VEN and HMA cohorts, respectively. HSCT was performed in 14 (11%) and 8 (9%) patients. Without censoring for HSCT, OS continued to favor the HMA-VEN cohort with a HR for death of 0.71 (95% CI, 0.52–0.98; Suppl. Figure 4A). Post-transplantation survival is shown in Suppl. Figure 4B.
Multiple regression
In the univariable regression models, treatment group, NCCN poor cytogenetic risk (HR 1.99; 95% CI, 1.40–2.83), ECOG PS ≥ 2 (2.67; 1.89–3.77) and presence of a TP53 mutation (1.94; 1.27–2.97) were associated with shorter OS (Fig. 3).
Based on the criteria defined above, treatment, age, NCCN risk group, ECOG PS, leukocytes as well as NPM1, RUNX1, and TP53 status were included into the multiple regression model. In addition, we decided to also include the choice of HMA because of large imbalances between the cohorts. In the multivariable analysis, treatment with HMA-VEN remained significantly associated with a prolonged OS with a HR for death of 0.48 (95% CI, 0.29–0.77; p = 0.003; Fig. 3). While patients with higher age or assignment to the NCCN poor risk group showed a tendency towards shorter OS, only TP53 mutations (HR, 1.82; 95% CI, 1.01–3.30) and ECOG PS ≥ 2 (2.78; 1.77–4.37) were significantly coupled with poorer survival in the multivariable regression model. The use of DEC was associated with numerically improved OS compared to AZA.
A second analysis using the ELN 2017 risk stratification instead of the NCCN 2.2016 classification of cytogenetic risk, showed similar results. The HR for death was 0.51 (95% CI, 0.34–0.78) for treatment with HMA-VEN compared to HMA alone. Moreover, favorable and intermediate ELN 2017 risk were associated with improved OS compared to adverse risk with HRs of 0.46 (95% CI, 0.24–0.85) and 0.59 (0.39–0.88), respectively (Suppl. Figure 5).