CPX-351 (Liposomal Cytarabine and Daunorubicin) versus venetoclax plus hypomethylating agent therapy in newly diagnosed acute myeloid leukemia: a retrospective comparison involving 600 Mayo Clinic patients

doi:10.21203/rs.3.rs-8233179/v1

Download PDF

Article

CPX-351 (Liposomal Cytarabine and Daunorubicin) versus venetoclax plus hypomethylating agent therapy in newly diagnosed acute myeloid leukemia: a retrospective comparison involving 600 Mayo Clinic patients

https://doi.org/10.21203/rs.3.rs-8233179/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The comparative value of liposomal cytarabine/daunorubicin (CPX-351) versus venetoclax plus a hypomethylating agent (Ven-HMA) in the frontline treatment of older adults with primary (de novo) or secondary acute myeloid leukemia (AML) remains uncertain. In the current study, we retrospectively examined outcomes of 600 patients with newly diagnosed AML treated with CPX-351 (N = 112) or Ven-HMA (N = 488). AML subtypes included de novo (46%), post-myelodysplastic syndrome (MDS) (19%), post-myeloproliferative neoplasm (MPN) (12%), post-MDS/MPN (6%), therapy-related (17%) and AML with myelodysplasia-related genetic abnormalities (AML-MR; 46%). Patients receiving CPX-351 were younger (median 65 vs. 73 years; p < 0.01), predominantly female (50% vs. 38%; p = 0.02), more likely to have secondary AML (68% vs. 51%; p < 0.01), and less likely to harbor NPM1^MUT (5% vs. 12%; p = 0.02). Rates of complete response with or without count recovery (CR/CRi) were comparable between CPX-351 and Ven-HMA (55% vs. 60%; p = 0.30), including in AML-MR (60% vs. 63%; p = 0.70). Ven-HMA use was associated with fewer infectious complications (62% vs. 83%; p < 0.01) and yielded higher CR/CRi rates in males (60% vs. 45%; p = 0.04), de novo AML (68% vs. 50%; p = 0.03), and in the presence of STAG2^MUT (86% vs. 44%; p = 0.02), or CEBPA^MUT (88% vs. 50%; p = 0.03). Overall survival censored for transplant, was similar (median 10 vs. 13 months; p = 0.90), with Ven-HMA being superior in post-MDS AML (median 12 vs. 7 months; p = 0.02) and CPX-351 in the presence of SF3B1^MUT (not reached vs. 14 months; p < 0.01). Our findings suggest that Ven-HMA is as effective and less toxic than CPX-351 in newly diagnosed AML, including AML-MR, despite selection of younger, fitter patients for CPX-351.

Biological sciences/Cancer/Haematological cancer/Leukaemia/Acute myeloid leukaemia

Biological sciences/Cancer/Cancer epidemiology

AML

myelodysplasia

mutations

survival

response

CPX-351 (Vyxeos™) is a liposomal formulation of cytarabine and daunorubicin, that received Food and Drug Administration (FDA) approval in August 2017 for the treatment of newly-diagnosed secondary acute myeloid leukemia, including therapy-related AML (t-AML) and AML with myelodysplasia-related changes (MRC).^1–3 This formulation was designed to optimize synergistic cytotoxicity of cytarabine and daunorubicin by delivering the agents in a fixed 5:1 molar ratio ⁴. Preclinical studies further demonstrated enhanced intracellular drug delivery and prolonged exposure within leukemic blasts, resulting in superior antileukemic activity compared with the free-drug combination ^4,5. Based on these observations, a pivotal phase 3 trial compared CPX-351 with standard cytarabine and daunorubicin (7 + 3) induction in older adults (aged 60–75 years) with secondary AML including t-AML, post-MDS/chronic myelomonocytic leukemia (CMML) AML, or de novo AML with MDS-related cytogenetic abnormalities. In this trial, CPX-351 significantly improved overall survival (OS) (median 9.56 vs. 5.95 months; p < 0.01) and composite remission rate (complete response with or without count recovery (CR/CRi, 47.7% vs. 33.3%; p = 0.016).⁵

Since the approval of CPX-351, venetoclax (Ven) in combination with hypomethylating agent (HMA) has emerged as a widely used frontline therapy for older adults with AML and patients considered ineligible for intensive chemotherapy.^6,7 Despite the increasing use of Ven-HMA, comparative data on CPX-351 versus Ven-HMA in the frontline treatment of AML remains limited. ⁸ The primary objective of the current study was to provide a comprehensive comparison of response and survival outcomes between these two treatment strategies, in order to better inform clinical decisions in patients with newly diagnosed AML, particularly those with AML-MR.^9,10 In addition, we evaluated the comparative performance of each agent across distinct clinical and genomic subtypes of AML, identifying subtype-specific differences in drug sensitivity. Subsequently, to further refine treatment comparisons, we conducted a machine-learning–based analysis to develop treatment-specific risk models.

The current study was performed under an institutional review board approved minimum risk protocol that allowed retrospective collection and analysis of data from patients with newly diagnosed AML treated with CPX-351 or Ven-HMA at Mayo Clinic sites in Rochester, MN; Scottsdale, AZ; and Jacksonville, FL (IRB protocol number: 12-003574). All patients were treated outside clinical trials between August 2017 and September 2024 with follow-up updated in March 2025. Cytogenetic and molecular profiling were performed through conventional karyotyping and next-generation sequencing (NGS), respectively. AML-MR gene mutations or cytogenetic abnormalities were categorized according to the International Consensus Classification (ICC) 2022¹¹. Treatment response was assessed according to the 2022 European Leukemia Net (ELN) criteria ¹², typically after one or two cycles, based on treating physician discretion. Measurable residual disease (MRD) was quantified using multiparameter flow cytometry with a minimum sensitivity of 0.01%. OS analysis was censored at the time of allogeneic stem cell transplant (ASCT). Event free survival (EFS) was defined as time to relapse, disease progression (i.e. need to switch or discontinue therapy due to lack of response), or death. Conventional statistical methods were employed for group comparisons and survival analysis (JMP Pro 18.0.0 software SAS Institute, Cary, NC, USA). Cumulative incidence of relapses was estimated using competing risk methodology (Bluesky statistics).

Additionally, a machine learning approach, the Risk-calibrated Supersparse Linear Integer Model (RISKSLIM) method was employed to develop a risk scoring model ¹³. RISKSLIM formulates risk score construction as a mixed-integer nonlinear optimization problem that balances predictive performance, measured by the Area Under the Curve (AUC), with model simplicity. It achieves this by limiting the number of predictors and constraining their coefficients to small integers. To solve this complex optimization problem, RISKSLIM employs a specialized cutting plane algorithm called the Lattice Cutting Plane Algorithm (LCPA), which is designed to efficiently find certifiably optimal solutions and scales linearly with the number of samples in the dataset. To partition the cohort into clusters, we fitted an Optimal Survival Tree (OST)¹⁴ using the risk scores generated by the RISKSLIM algorithm as the sole input variable. We then pruned the tree by traversing it bottom-up, iteratively merging nodes until all remaining leaf nodes demonstrated statistically significant differences in survival outcomes.

Patient characteristics

A total of 600 patients (median age 72 years, 60% males) with newly diagnosed AML treated with CPX-351 (N-112, 19%), or Ven-HMA (N = 488, 81%) were included. AML subtypes included de novo (N = 277, 46%), post-myelodysplastic syndrome (post-MDS, N = 114,19%), post-myeloproliferative neoplasm (post-MPN, N = 70, 12%), post-MDS/MPN (N = 36, 6%), and t-AML (N = 103, 17%, Table 1). Compared with those treated with Ven-HMA, patients receiving CPX-351 were younger (median age 65 vs. 73 years; p < 0.01), more likely to be female (50% vs. 38%; p = 0.02), more frequently diagnosed with secondary AML (68% vs. 51%; p < 0.01), and less likely to have NPM1^MUT (5% vs. 12%; p = 0.02) whereas ELN 2022 cytogenetic risk distribution was similar between the two treatment groups (adverse risk; 43% vs. 40%; p = 0.5, Table 1).

Table 1

Clinical characteristics, cytogenetics and co-mutation patterns at diagnosis of 600 patients with acute myeloid leukemia (AML) stratified by upfront treatment with liposomal daunorubicin and cytarabine (CPX-351) versus venetoclax plus a hypomethylating agent (Ven-HMA).
Variables	All patients N = 600	CPX-351 N = 112	Ven-HMA N = 488	Univariate P-value
Age in years; median (range)	72 (19–98)	65 (19–77)	73 (19–98)	< 0.01
Age ≥ 60 years; n (%)	529 (88)	80 (71)	449 (92)	< 0.01
Male; n (%)	360 (60)	56 (50)	304 (62)	0.02
De novo AML; n (%)	277 (46)	36 (32)	241 (49)	< 0.01
Secondary AML (n = 323) Post-Myelodysplastic syndrome (MDS); n (%) Post-Myeloproliferative neoplasm (MPN); n (%) Post-MDS/MPN; n (%) Therapy-related; n (%)	114 (35) 70 (22) 36 (11) 103 (32)	32 (42) 10 (13) 11 (14) 23 (30)	82 (33) 60 (24) 25 (10) 80 (32)	0.1
Prior HMA treatment; n (%)	69 (12)	24 (21)	45 (9)	< 0.01
Leukocytes, x10⁹/L; median (range) (Evaluable = 592)	4 (0.4–281)	3 (0.5–281)	4 (0.4–200)	0.3
Hemoglobin, g/dL; median (range) (Evaluable = 593)	8 (3–16)	8 (5–16)	8.5 (3–16)	0.9
Platelets, x10⁹/L; median (range) (Evaluable = 579)	58 (4-1002)	56 (4-1002)	59 (5-601)	0.3
Circulating blast %; median (range) (Evaluable = 554)	13 (0–93)	15 (0–80)	13 (0–93)	0.9
Bone marrow blast %; median (range) (Evaluable = 466)	38 (1–97)	31 (5–95)	40 (1–97)	< 0.01
ELN 2022 cytogenetic risk (Evaluable = 586) Adverse Non-adverse	237 (40) 349 (60)	48 (43) 63 (57)	189 (40) 286(60)	0.5
Presence of complex or monosomal karyotype; n (%) (Evaluable = 575)	184 (32)	35 (32)	149 (32)	0.9
ASXL1^MUT; n (%) (Evaluable = 477)	102 (21)	19 (22)	83 (21)	0.9
BCOR^MUT; n (%) (Evaluable = 477)	43 (9)	11 (13)	32 (8)	0.2
CBL^MUT; n (%) (Evaluable = 477)	19 (4)	6 (7)	13 (3)	0.2
CEBPA^MUT; n (%) (Evaluable = 553)	34 (6)	8 (8)	26 (6)	0.5
DDX41^MUT; n (%) (Evaluable = 469)	16 (3)	1 (1)	15 (4)	0.2
DNMT3A^MUT; n (%) (Evaluable = 549)	78 (14)	17 (17)	61 (14)	0.5
EZH2^MUT; n (%) (Evaluable = 477)	16 (3)	2 (2)	14 (4)	0.5
FLT3^MUT; n (%) (Evaluable = 572)	60 (11)	10 (9)	50 (11)	0.6
IDH1^MUT; n (%) (Evaluable = 567)	38 (7)	5 (5)	33 (7)	0.3
IDH2^MUT; n (%) (Evaluable = 567)	71 (13)	19 (18)	52 (11)	0.08
JAK2^MUT; n (%) (Evaluable = 478)	50 (11)	7 (8)	43 (11)	0.4
KRAS^MUT; n (%) (Evaluable = 544)	28 (5)	7 (7)	21 (5)	0.4
NRAS^MUT; n (%) (Evaluable = 544)	47 (9)	8 (8)	39 (9)	0.7
NPM1^MUT; n (%) (Evaluable = 561)	59 (11)	5 (5)	54 (12)	0.02
PHF6^MUT; n (%) (Evaluable = 477)	12 (3)	0 (0)	12 (3)	0.03
PTPN11^MUT; n (%) (Evaluable = 576)	16 (3)	4 (5)	12 (3)	0.5
RUNX1^MUT; n (%) (Evaluable = 543)	111 (20)	23 (23)	88 (20)	0.5
SETBP1^MUT; n (%) (Evaluable = 476)	11 (2)	0 (0)	11 (3)	0.03
SF3B1^MUT; n (%) (Evaluable = 476)	29 (6)	7 (8)	22 (6)	0.4
SRSF2^MUT; n (%) (Evaluable = 476)	100 (21)	13 (15)	87 (22)	0.1
STAG2^MUT; n (%) (Evaluable = 473)	37 (8)	9 (11)	28 (7)	0.3
TET2^MUT; n (%) (Evaluable = 477)	99 (21)	12 (14)	87 (22)	0.06
TP53^MUT; n (%) (Evaluable = 564)	136 (24)	19 (18)	117 (25)	0.1
U2AF1^MUT; n (%) (Evaluable = 476)	36 (8)	7 (8)	29 (7)	0.9
WT1^MUT; n (%) (Evaluable = 477)	16 (3)	6 (7)	10 (3)	0.07
ZRSR2^MUT; n (%) (Evaluable = 477)	9 (2)	3 (3)	6 (2)	0.3
Complete response with or without count recovery (CR/CRi); n (%)	357 (60)	62 (55)	295 (60)	0.3
Allogeneic stem cell transplant; n (%)	149 (25)	59 (53)	90 (18)	< 0.01

Treatment response

CR/CRi rates were comparable between CPX-351 and Ven-HMA (55% vs. 60%; p = 0.30), including among patients with adverse karyotype (44% vs. 44%; p = 0.98), NPM1 ^MUT (100% vs. 89%; p = 0.30), TP53 ^MUT (32% vs. 45%; p = 0.30), IDH1 ^MUT (60% vs. 76%; p = 0.50), IDH2 ^MUT (63% vs. 77%; p = 0.30), SF3B1 ^MUT (71% vs. 45%; p = 0.20), and CBL ^MUT (33% vs. 69%; p = 0.10) (Table 2). Compared with CPX-351, Ven-HMA yielded significantly higher CR/CRi rates in males (60% vs. 45%; p = 0.04), patients with de novo AML (68% vs. 50%; p = 0.03), STAG2^MUT (86% vs. 44%; p = 0.02), or CEBPA^MUT (88% vs. 50%; p = 0.03, Table 2). Total number of chemotherapy cycles in patients who achieved CR/CRi was 2 cycles (range 1–4) with CPX-351 and 5 cycles (range 1–65). The median time to CR/CRi was 1.3 months (range 0.5–3.6) with CPX-351 compared to 1.4 months (0.6–10) with Ven-HMA. Among the 188 patients (53%) with evaluable MRD following CR/CRi, MRD negativity by flow cytometry was more frequent with Ven-HMA compared with CPX-351 (75% vs. 48%; p = 0.01).

Table 2

Comparison of the rates of achievement of Complete response with or without count recovery (CR/CRi) in 600 patients with newly diagnosed acute myeloid leukemia (AML) treated with liposomal daunorubicin and cytarabine (CPX-351) versus venetoclax plus a hypomethylating agent (Ven-HMA).
Variables	Univariate analysis Rate of CR/CRi CPX-351 vs. Ven-HMA	P-value
Age ≥ 60 years N = 529	59% vs. 60%	0.8 (sex-adjusted p = 0.7)
Male sex N = 360	45% vs. 60%	0.035 (age-adjusted p = 0.04)
De novo AML N = 277	50% vs. 68%	0.03 (age-adjusted p = 0.1) (sex-adjusted p = 0.03)
Post-myeloproliferative neoplasm (post-MPN) AML N = 70	30% vs. 38%	0.6 (age-adjusted p = 0.2) (sex-adjusted p = 0.7)
Post-myelodysplastic syndrome (post-MDS) AML N = 114	47% vs. 49%	0.8 (age-adjusted p = 0.6) (sex-adjusted p = 0.8)
Post-MDS/MPN AML N = 36	55% vs. 44%	0.6 (age-adjusted p = 0.7) (sex-adjusted p = 0.5)
Therapy-related AML N = 128	79% vs. 64%	0.1 (age-adjusted p = 0.2) (sex-adjusted p = 0.1)
Presence of ELN adverse cytogenetic risk N = 237	44% vs. 44%	0.98 (age-adjusted p = 0.9) (sex-adjusted p = 0.9)
Presence of TP53 mutation N = 136	32% vs. 45%	0.3 (age-adjusted p = 0.3) (sex-adjusted p = 0.2)
Presence of RUNX1 mutation N = 111	57% vs. 53%	0.8 (age-adjusted p = 0.98) (sex-adjusted p = 0.9)
Presence of ASXL1 mutation N = 102	47% vs. 60%	0.3 (age-adjusted p = 0.6) (sex-adjusted p = 0.3)
Presence of SRSF2 mutation N = 100	54% vs. 74%	0.15 (age-adjusted p = 0.2) (sex-adjusted p = 0.09)
Presence of TET2 mutation N = 99	50% vs. 67%	0.3 (age-adjusted p = 0.17) (sex-adjusted p = 0.2)
Presence of DNMT3A mutation N = 78	71% vs. 77%	0.6 (age-adjusted p = 0.5) (sex-adjusted p = 0.6)
Presence of IDH2 mutation N = 71	63% vs. 77%	0.3 (age-adjusted p = 0.8) (sex-adjusted p = 0.1)
Presence of FLT3 mutation N = 60	70% vs. 58%	0.5 (age-adjusted p = 0.7) (sex-adjusted p = 0.4)
Presence of NPM1 mutation N = 59	100% vs. 89%	0.3 (age-adjusted p = 0.2) (sex-adjusted p = 0.2)
Presence of NRAS mutation N = 47	63% vs. 62%	0.95 (age-adjusted p = 0.6) (sex-adjusted p = 0.95)
Presence of BCOR mutation N = 43	66% vs. 64%	0.9 (age-adjusted p = 0.8) (sex-adjusted p = 0.9)
Presence of IDH1 mutation N = 38	60% vs. 76%	0.5 (age-adjusted p = 0.9) (sex-adjusted p = 0.5)
Presence of STAG2 mutation N = 37	44% vs. 86%	0.02 (age-adjusted p = 0.02) (sex-adjusted p = 0.03)
Presence of U2AF1 mutation N = 36	57% vs. 52%	0.8 (age-adjusted p = 0.8) (sex-adjusted p = 0.8)
Presence of CEBPA mutation N = 34	50% vs. 88%	0.03 (age-adjusted p = 0.02) (sex-adjusted p = 0.03)
Presence of SF3B1 mutation N = 29	71% vs. 45%	0.2 (age-adjusted p = 0.3) (sex-adjusted p = 0.3)
Presence of KRAS mutation N = 28	43% vs. 57%	0.5 (age-adjusted p = 0.4) (sex-adjusted p = 0.6)
Presence of CBL mutation N = 19	33% vs. 69%	0.1 (age-adjusted p = 0.5) (sex-adjusted p = 0.15)
Presence of EZH2 mutation N = 16	0% vs. 43%	0.15 (age-adjusted p = 0.2) (sex-adjusted p = 0.2)
Presence of PTPN11 mutation N = 16	50% vs. 50%	1.0 (age-adjusted p = 0.9) (sex-adjusted p = 0.9)
Presence of ZRSR2 mutation N = 9	33% vs. 33%	1.0 (age-adjusted p = 0.98) (sex-adjusted p = 1.0)

Survival outcomes

At a median follow-up of 30 months (range; 3–88) for surviving patients, 416 (69%) deaths, 179 relapses (37% in CPX-351 and 49% in Ven-HMA responders, p = 0.09) and 149 ASCTs (53% in CPX-351 and 18% in Ven-HMA, p < 0.01) were documented. OS censored for transplant was similar between CPX-351 and Ven-HMA, (median 10 vs. 13 months, p = 0.9), including among patients ≥ 60 years (median 10 vs. 11 months, p = 0.9), and < 60 years (13 vs. 10 months, p = 0.9, Fig. 1). Likewise, no significant differences in OS were observed in patients with adverse karyotype (10 vs. 8 months; p = 0.2), TP53^MUT (9 vs. 7 months; p = 0.60), IDH1^MUT (not reached vs. 20 months; p = 0.60), or IDH2^MUT (21 vs. 40 months; p = 0.60 (Table 3). By contrast, OS favored Ven-HMA in post-MDS AML (median 12 vs. 7 months; age-adjusted p = 0.02, Table 3, Fig. 2, Supplemental Fig. 1), and CPX-351-in patients with SF3B1 ^MUT (median not reached vs. 14 months; age-adjusted p < 0.01, Supplemental Fig. 2). Notably, EFS was inferior with CPX-351 compared to Ven-HMA (5 vs. 9 months, p < 0.01) and the cumulative incidence of relapse/progression was higher with CPX-351 at 1-/2-/5- years (55%/59%/62% vs. 30%/42%/48%) with Ven-HMA (p < 0.01, Supplemental Fig. 3).

Table 3

Comparison of Overall survival (OS) censored at allogeneic stem cell transplant (ASCT) in 600 patients with newly diagnosed acute myeloid leukemia (AML) treated with liposomal daunorubicin and cytarabine (CPX-351) versus venetoclax plus a hypomethylating agent (Ven-HMA).
Variables	Univariate analysis Overall survival CPX-351 vs. Ven-HMA	P-value
Age ≥ 60 years N = 529	10 vs. 13 months	0.7 (sex-adjusted p = 0.5)
Male sex N = 360	9 vs. 11 months	0.8 (age-adjusted p = 0.7)
De novo AML N = 277	13 vs. 15 months	0.7 (age-adjusted p = 0.8) (sex-adjusted p = 0.8)
Post-myeloproliferative neoplasm (post-MPN) AML N = 70	Not reached vs. 9 months	0.2 (age-adjusted p = 0.4) (sex-adjusted p = 0.4)
Post-myelodysplastic syndrome (post-MDS) AML N = 114	7 vs. 12 months	0.01 (age-adjusted p = 0.02) (sex-adjusted p = 0.02)
Post-MDS/MPN AML N = 36	10 vs. 6 months	0.8 (age-adjusted p = 0.6) (sex-adjusted p = 0.9)
Therapy-related AML N = 128	13 vs. 11 months	0.4 (age-adjusted p = 0.4) (sex-adjusted p = 0.7)
Presence of ELN adverse cytogenetic risk N = 237	10 vs. 8 months	0.2 (age-adjusted p = 0.2) (sex-adjusted p = 0.3)
Presence of TP53 mutation N = 136	9 vs. 7 months	0.6 (age-adjusted p = 0.8) (sex-adjusted p = 0.6)
Presence of RUNX1 mutation N = 111	13 vs. 14 months	0.8 (age-adjusted p = 0.8) (sex-adjusted p = 0.8)
Presence of ASXL1 mutation N = 102	10 vs. 13 months	0.3 (age-adjusted p = 0.3) (sex-adjusted p = 0.3)
Presence of SRSF2 mutation N = 100	9 vs. 16 months	0.3 (age-adjusted p = 0.6) (sex-adjusted p = 0.2)
Presence of TET2 mutation N = 99	8 vs. 13 months	0.2 (age-adjusted p = 0.2) (sex-adjusted p = 0.2)
Presence of DNMT3A mutation N = 78	Not reached vs. 20 months	0.4 (age-adjusted p = 0.6) (sex-adjusted p = 0.4)
Presence of IDH2 mutation N = 71	21 vs. 40 months	0.6 (age-adjusted p = 0.9) (sex-adjusted p = 0.8)
Presence of FLT3 mutation N = 60	10 vs. 13 months	0.9 (age-adjusted p = 0.5) (sex-adjusted p = 0.9)
Presence of NPM1 mutation N = 59	21 vs. 20 months	0.7 (age-adjusted p = 0.9) (sex-adjusted p = 0.9)
Presence of NRAS mutation N = 47	14 vs. 12 months	0.5 (age-adjusted p = 0.8) (sex-adjusted p = 0.4)
Presence of BCOR mutation N = 43	39 vs. 19 months	0.5 (age-adjusted p = 0.4) (sex-adjusted p = 0.4)
Presence of IDH1 mutation N = 38	Not reached vs. 20 months	0.6 (age-adjusted p = 0.5) (sex-adjusted p = 0.5)
Presence of STAG2 mutation N = 37	14 vs. 19 months	0.5 (age-adjusted p = 0.8) (sex-adjusted p = 0.6)
Presence of U2AF1 mutation N = 36	7 vs. 9 months	0.8 (age-adjusted p = 0.8) (sex-adjusted p = 0.6)
Presence of CEBPA mutation N = 34	8 vs. 13 months	0.7 (age-adjusted p = 0.6) (sex-adjusted p = 0.7)
Presence of SF3B1 mutation N = 29	Not reached vs. 14 months	0.04 (age-adjusted p < 0.01) (sex-adjusted p < 0.01)
Presence of KRAS mutation N = 28	8 vs. 13 months	0.99 (age-adjusted p = 0.4) (sex-adjusted p = 0.9)
Presence of CBL mutation N = 19	4 vs. 15 months	0.1 (age-adjusted p = 0.3) (sex-adjusted p = 0.2)
Presence of EZH2 mutation N = 16	4.5 vs. 9 months	0.1 (age-adjusted p = 0.2) (sex-adjusted p = 0.1)
Presence of PTPN11 mutation N = 16	13.5 vs. 8 months	0.02 (age-adjusted p = 0.08) (sex-adjusted p = 0.06)
Presence of ZRSR2 mutation N = 9	Not reached vs. 6 months	0.7 (age-adjusted p = 0.7) (sex-adjusted p = 0.7)

Similar results were obtained when survival analyses were not censored for transplant; OS remained similar between CPX-351 and Ven-HMA (14 vs. 13 months, p = 0.1), while EFS remained inferior with CPX-351 (4 vs. 8 months, p < 0.01). Among non-transplanted patients (N = 451), OS was noted to be superior among patients who received Ven-HMA (10 vs. 8 months, p = 0.02, Fig. 3). A total of 149 patients (25%) underwent ASCT with a higher proportion in the CPX-351 group (53% vs.18%, p < 0.01); post-transplant survival was comparable between CPX-351 and Ven-HMA (46 months vs. not reached, p = 0.9, Fig. 3).

In multivariable analysis, several factors were independently associated with OS. Favorable factors for OS included ASCT (HR 0.3 95% CI 0.2–0.4, p < 0.01), and IDH2^MUT (HR 0.6 95% CI 0.4–0.9, p = 0.03). In contrast, male gender (HR 1.5 95% CI 1.2–1.9, p < 0.01), TP53 ^MUT(HR 1.6 95% CI 1.1–2.2, p < 0.01), ELN adverse cytogenetics (HR 2.0 95% CI 1.5–2.7, p < 0.01), EZH2^MUT(HR 2.1 95% CI 1.2–3.6, p < 0.01), PTPN11^MUT(HR 2.3 95% CI 1.3–4.1, p < 0.01) and post-MDS/MPN AML (HR 2.1 95% CI 1.3–3.4, p < 0.01) were associated with inferior OS. Notably, frontline treatment with CPX-351 showed a trend toward inferior OS compared with Ven-HMA (HR 1.3 95% CI 0.9–1.8, p = 0.09, Supplemental Fig. 4).

Given that patients in the CPX-351 cohort were more likely to have received prior HMA therapy (21% vs. 9%, p < 0.01), we performed an analysis restricted to patients with prior HMA exposure (N = 69). In this subgroup, CR/CRi rates (44% vs. 56%, p = 0.3) and OS (median 6 months with each treatment, p = 0.7) were comparable between CPX-351 and Ven-HMA, indicating that prior HMA exposure did not account for observed differences in overall outcomes.

Outcomes in AML-MR (N = 274)

We performed a separate analysis of 274 patients (median age 73 years, 62% males) with ICC defined AML-MR gene mutations or cytogenetic abnormalities, treated with CPX-351 (N = 57) or Ven-HMA (N = 217) (Supplemental Table 1). CPX-351 treated patients were younger (median age 66 vs. 74 years; p < 0.01), more likely to be female (56% vs. 33%; p < 0.01), more likely to harbor IDH2^MUT (22% vs. 12%; p = 0.06), CBL^MUT (12% vs. 5%; p = 0.07). ZRSR2^MUT (6% vs. 1%; p = 0.02), and less likely to have PHF6^MUT (0% vs. 5%; p = 0.04), SETBP1^MUT (0% vs. 5%; p = 0.04) or SRSF2^MUT (22% vs. 35%; p = 0.08) (Supplemental Table 1).

Despite these baseline differences, CR/CRi and OS rates were comparable between patients treated with CPX-351 and Ven-HMA. CR/CRi rates were 60% with CPX-351 versus 63% with Ven-HMA (p = 0.68), and median OS was 10 versus 14 months (p = 0.61), respectively (Supplemental Table 1; Fig. 4). However, EFS was notably inferior with CPX-351 (median 5 vs. 9 months, p = 0.01; Fig. 4). We then examined predictors of outcomes within this cohort. CR/CRi rates were higher with Ven-HMA among patients with CEBPA^MUT (91% vs. 40%, age-adjusted p < 0.01) and trended higher in those with SRSF2^MUT (72% vs. 45%; age-adjusted p = 0.09, Supplemental Table 2). Similarly, OS favored Ven-HMA in patients with post-MDS AML (15 vs. 8 months; age-adjusted p = 0.02). By contrast, OS favored CPX-351 in patients with SF3B1^MUT (NR vs. 14 months; age-adjusted p = 0.02) (Supplemental Table 3). These findings highlight that, while overall outcomes were similar between CPX-351 and Ven-HMA in AML-MR patients, specific molecular subgroups may derive differential benefit from one therapy over the other.

Outcomes in secondary AML (N = 253)

Next, we restricted our analysis to patients with secondary AML (t-AML, post-MDS and post-MDS/MPN, Supplemental Table 4). Overall, CR/CRi and OS rates were comparable between patients treated with CPX-351 (N = 66) and Ven-HMA (N = 187), with response rates of 38% vs. 43% (p = 0.5, Supplemental Table 5) and median OS of 9 vs. 11 months (p = 0.52, Fig. 5), respectively. EFS was numerically longer with Ven-HMA (median 8 vs. 7 months, p = 0.05, Fig. 5), suggesting a potential benefit in achieving more durable remission.

We then investigated outcomes in molecularly defined subgroups within this cohort. CR/CRi was notably higher with CPX-351 in patients with PTPN11^MUT (67% vs. 0%, age-adjusted p < 0.01, Supplemental Table 5) and OS favored CPX-351 in patients with SF3B1^MUT (NR vs. 13 months, age-adjusted p = 0.05, Supplemental Table 6, Supplemental Fig. 2). By contrast, OS was superior with Ven-HMA in patients with post-MDS AML (12 vs. 7 months; age-adjusted p = 0.02).Taken together, these results indicate that, while overall response and survival outcomes were similar between CPX-351 and Ven-HMA in secondary AML, specific molecular subgroups (SF3B1) may derive differential benefit from one therapy over the other.

Survival risk models

In order to further refine treatment comparisons, machine-learning algorithms were utilized to construct treatment specific three-tiered prognostic models. Among CPX-351 treated patients, key risk factors included: (i) ELN adverse karyotype (1 point), (ii) presence of TP53^MUT (2 points), (iii) presence of TET2^MUT (1 point), (iv) post-MDS or post-MDS/MPN AML (3 points), and absence of SF3B1^MUT (4 points). Using these variables, patients were stratified into low (score 0–4, N = 29, median OS 39 months), intermediate (score 5–7, N = 39, median OS 10 months) and high-risk (score ≥ 8, N = 20, median OS 5 months) groups, (p < 0.01), with AUC values at 1 and 2 years of 84.74 and 92.05, respectively (Supplemental Fig. 5). Patient with SF3B1^MUT more likely to undergo ASCT (86% vs.14%, p = 0.05).

Similarly, among Ven-HMA treated patients, the following risk factors were identified: (i) ELN adverse karyotype (3 points), (ii) post-MPN AML (2 points), (iii) male gender (1 point), (iv) presence of TP53^MUT (1 point), (v) presence of TET2^MUT (1 point), (vi) absence of STAG2^MUT (2 points), (vii) absence of IDH1^MUT (2 points) and (viii) absence of IDH2^MUT (1 point). The above risk factors stratified patients into: low (score 0–5, N = 224, median OS 29 months), intermediate (score 6–9, N = 93, median OS 13 months) and high-risk (score ≥ 10, N = 66, median OS 7 months) groups, (p < 0.001), with AUC values at 1 and 2 years of 75.32 and 73.87, respectively (Supplemental Fig. 6). Of note, STAG2^MUT was less likely to have an accompanying ELN adverse cytogenetics (11% vs. 89%, p < 0.01), and TP53^MUT (4% vs. 96%, p < 0.01). These treatment-specific models underline how distinct clinical and molecular features differentially influence prognosis depending on the therapy received.

Approximately 15 months following CPX-351 approval, Ven-HMA was approved for use in elderly (> 75 years) and/or unfit patients with AML who are ineligible for intensive therapy. Notably, Ven-HMA similar to CPX-351 has shown favorable responses in patients with AML-MR^15–17. Since then, Ven based regimens have become ubiquitous for older/unfit adults with AML, and the role of CPX-351 in the current treatment landscape is unclear.¹⁸ This shift raises a fundamental clinical question regarding the comparative efficacy and safety of CPX-351 vs. Ven-HMA. Although randomized comparisons are lacking, several retrospective studies have sought to address this question. ^8,19,20 For instance, in an observational study including 656 patients with newly diagnosed AML treated with either CPX-351 (N = 217) or Ven and azacitidine (Ven/Aza, N = 439), ⁸ CR/CRi rates (61% vs. 65%) and OS (median 11 vs. 13 months) were similar, regardless of molecular profile⁸. Consistent with our findings, patients who received CPX-351 were younger and more likely to have secondary AML, while those treated with Ven/Aza were more likely to have de novo AML (52% vs. 29%). ASCT was performed more frequently in the CPX-351 group (28% vs.10%, p < 0.01) and was associated with improved OS (HR 0.3, p < 0.01); however frontline therapy did not significantly impact OS after adjusting for ASCT (HR 0.9, p = 0.78)⁸. In terms of safety, Ven-Aza was associated with lower incidences of febrile neutropenia (54% vs. 90%), fewer culture-positive infections (36% vs. 67%), and shorter inpatient hospitalization (mean: 15 vs. 41 days).⁸ Similarly, in our study, we observed no significant differences in CR/CRi rates (55% vs. 60%) including in AML-MR (60% vs. 63%, p = 0.7) and secondary AML (38% vs. 43%, p = 0.5). Transplant-censored OS (median: 10 vs. 13 months), or post-transplant survival (52 months vs. not reached) were also similar between the two groups. However, treatment with CPX-351 was associated with a notably higher incidence of infections (83% vs. 62%). Incidentally, a randomized study in younger (< 60 years), fit patients compared 7 + 3 with Ven–decitabine and showed similar overall CR/CRi rates (79% vs. 89%), but Ven–decitabine yielded superior responses in ELN 2022 adverse-risk disease (91% vs. 42%) and in patients with epigenetic mutations (91% vs. 67%).²¹ Notably, Ven–decitabine was also associated with fewer grade ≥ 3 infections (32% vs. 67%) and a shorter duration of severe thrombocytopenia (13 vs. 19 days).²¹

Similar findings were also reported in a smaller study of 50 patients with secondary AML treated with CPX-351 (N = 20) or Ven-HMA (N = 30)¹⁹. Composite CR rates (70% vs. 50%, p = 0.15), MRD negativity (66% vs. 50%, p = 0.12), and OS (11.3 vs. 10.1 months, p = 0.76) were not significantly different between the two regimens, although ASCT was performed more frequently following CPX-351 induction (65% vs. 35%, p = 0.049)¹⁹. In another multicenter retrospective analysis, Shimony et al²⁰, evaluated 395 patients with molecularly defined secondary AML treated with 7 + 3 (N = 167,42%), CPX-351 (N = 66, 17%) or Ven-HMA (N = 162, 40%). Composite CR rates were similar across treatment groups (56%, 44%, and 56%, respectively; p = 0.22). ASCT was more frequent following intensive therapy (59% with 7 + 3 and 50% with CPX-351 vs. 26%) compared to Ven-HMA (26%), p < 0.01). Importantly, age-adjusted analysis did not show a significant survival advantage of CPX-351 over Ven-HMA.²⁰. In the particular study, the presence of NRAS/KRAS mutations was associated with worse OS in patients treated with both CPX-351 and Ven-HMA, whereas OS favored Ven-HMA over CPX-351 in patients with splicing factor mutations.²⁰ By contrast, our subgroup analyses demonstrated different patterns: Ven-HMA was associated with superior outcomes in patients with post-MDS AML, while CPX-351 was associated with improved OS in those with SF3B1^MUT. Favorable outcomes with CPX-351 in spliceosome-mutated AML have been previously reported^22,23, and it is notable that in our study, 6 of 7 patients with SF3B1^MUT treated with CPX-351 proceeded to ASCT, which may have contributed to the observed survival advantage.

To further delineate predictors of treatment benefit, machine-learning algorithms were applied and enabled identification of both shared and treatment-specific prognostic factors, supporting the development of two treatment-specific risk models. Across both treatment groups, ELN adverse cytogenetics, TP53^MUT, TET2^MUT and post-MPN or post-MDS/post MDS/MPN AML were associated with inferior OS. However, certain genotypes demonstrated treatment-specific favorability; SF3B1^MUT was associated with improved outcomes with CPX-351, whereas STAG2^MUT and IDH1/2^MUT were associated with favorable outcomes with Ven-HMA.

Unsurprisingly, ASCT emerged as the most important determinant of OS, independent of frontline treatment and post-transplant survival was comparable across both treatment groups. However, it is noteworthy that EFS was longer, and MRD negativity rates were higher among patients treated with Ven-HMA, findings that are consistent with prior reports.¹⁹

Taken together, our results suggest that Ven-HMA was as effective and potentially less toxic than CPX-351 for the treatment of newly diagnosed AML, including AML-MR, despite the bias toward younger and fitter patients treated with the latter regimen. These observations support the use of Ven-HMA as a preferred frontline therapy in elderly or unfit patients. Nevertheless, the differential outcomes observed in patients with SF3B1^MUT or post-MDS AML, warrant further validation in prospective studies.

Conflict of interest:

SF: Nothing to disclose; LR: Nothing to disclose; NoG: Nothing to disclose; MR: Nothing to disclose; RI: Nothing to disclose; AA: Nothing to disclose; KM: Nothing to disclose; MA: Nothing to disclose; AAK: Nothing to disclose; HA: Nothing to disclose; KHB: Nothing to disclose; AAM: Research funding from Novartis, BMS, Solu Therapeutics and Sanofi; AM: Nothing to disclose; ANS: Nothing to disclose; MHT: Nothing to disclose; MRL: BeiGene Shanghai, Amgen,Astellas Pharma, Actinium Pharmaceuticals, Syndax, Consultancy; WH: Nothing to disclose; MS: Nothing to disclose; MMP: Research funding from Kura Oncology, Stemline therapeutics, Epigenetix, Solutherapeutics, Polaris and has served on the advisory board for AstraZeneca and SOBI pharmaceuticals, AP: Nothing to disclose; TB: Nothing to disclose; HM: Nothing to disclose; JF: Nothing to disclose; JP: Nothing to disclose; NP: Nothing to disclose; LS: Nothing to disclose; NK: Nothing to disclose; CAY: Curio Science, Consultancy; AT: Nothing to disclose; NG: Advisory board to DISC Medicine and Agios; AT and NG are members of the editorial board for BCJ.

Ethics approval:

This study was approved by the Institutional Review Board of the Mayo Clinic (IRB protocol number: 12-003574). All methods were performed in accordance with the Declaration of Helsinki, the relevant guidelines, and regulations.

Author contributions:

SF, NoG, MR, RI, AA, MA, AT and NG were involved in gathering of data; NG, AT and ST designed the study, LR, SF and NG conducted the data analysis; KM, AAK, HA, KHB, AAM, AM, ANS, MHT, MRL, WH, MS, MMP, AP, TB, HM, JF, JP, NP, LS, NK, CAY and NG participated in patient care; SF, AT and NG wrote the paper. All authors approved the final script.

Acknowledgments:

None

Data availability statement:

By email request to the corresponding author.

Lancet JE, Uy GL, Cortes JE, et al: CPX-351 (cytarabine and daunorubicin) Liposome for Injection Versus Conventional Cytarabine Plus Daunorubicin in Older Patients With Newly Diagnosed Secondary Acute Myeloid Leukemia. J Clin Oncol 36:2684–2692, 2018
Kantarjian H, Borthakur G, Daver N, et al: Current status and research directions in acute myeloid leukemia. Blood Cancer Journal 14:163, 2024
Krauss AC, Gao X, Li L, et al: FDA Approval Summary: (Daunorubicin and Cytarabine) Liposome for Injection for the Treatment of Adults with High-Risk Acute Myeloid Leukemia. Clin Cancer Res 25:2685–2690, 2019
Lim W-S, Tardi PG, Dos Santos N, et al: Leukemia-selective uptake and cytotoxicity of CPX-351, a synergistic fixed-ratio cytarabine:daunorubicin formulation, in bone marrow xenografts. Leukemia Research 34:1214–1223, 2010
Lancet JE, Uy GL, Cortes JE, et al: CPX-351 (cytarabine and daunorubicin) Liposome for Injection Versus Conventional Cytarabine Plus Daunorubicin in Older Patients With Newly Diagnosed Secondary Acute Myeloid Leukemia. Journal of Clinical Oncology 36:2684–2692, 2018
DiNardo CD, Jonas BA, Pullarkat V, et al: Azacitidine and Venetoclax in Previously Untreated Acute Myeloid Leukemia. N Engl J Med 383:617–629, 2020
Gangat N, Karrar O, Iftikhar M, et al: Venetoclax and hypomethylating agent combination therapy in newly diagnosed acute myeloid leukemia: Genotype signatures for response and survival among 301 consecutive patients. Am J Hematol 99:193–202, 2024
Matthews AH, Perl AE, Luger SM, et al: Real-world effectiveness of CPX-351 vs venetoclax and azacitidine in acute myeloid leukemia. Blood Adv 6:3997–4005, 2022
Gao Y, Jia M, Mao Y, et al: Distinct Mutation Landscapes Between Acute Myeloid Leukemia With Myelodysplasia-Related Changes and De Novo Acute Myeloid Leukemia. Am J Clin Pathol 157:691–700, 2022
Fang H, He R, Chiu A, et al: Genetic Factors in Acute Myeloid Leukemia With Myelodysplasia-Related Changes. Am J Clin Pathol 153:656–663, 2020
Arber DA, Orazi A, Hasserjian RP, et al: International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood 140:1200–1228, 2022
Döhner H, Wei AH, Appelbaum FR, et al: Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood 140:1345–1377, 2022
Rudin BUaC: Learning Optimized Risk Scores. Journal of Machine Learning Research 20:150 (2019): 1–75, 2019
Bertsimas D DJ, Gibson E, Orfanoudaki A.: Optimal survival trees. Machine learning, 2022
DiNardo CD, Pratz KW, Letai A, et al: Safety and preliminary efficacy of venetoclax with decitabine or azacitidine in elderly patients with previously untreated acute myeloid leukaemia: a non-randomised, open-label, phase 1b study. Lancet Oncol 19:216–228, 2018
Shimony SO, Murdock H, Keating J, et al: AML-MR Mutations Drive the Benefit of CPX-351 over 7 + 3 in the Pivotal Phase 3 AML Trial. Blood 144:60, 2024
Uy GL, Newell LF, Lin TL, et al: Transplant outcomes after CPX-351 vs 7 + 3 in older adults with newly diagnosed high-risk and/or secondary AML. Blood Adv 6:4989–4993, 2022
Gangat N, Dinardo CD: Newly diagnosed acute myeloid leukemia in unfit patients: 2026 treatment algorithms. Blood Cancer Journal 15:139, 2025
Salhotra A, Aribi A, Ngo D, et al: Outcome of secondary acute myeloid leukemia treated with hypomethylating agent plus venetoclax (HMA-Ven) or liposomal daunorubicin-cytarabine (CPX-351). American Journal of Hematology 96:E196-E200, 2021
Shimony S, Bewersdorf JP, Shallis RM, et al: Hypomethylating agents plus venetoclax compared with intensive induction chemotherapy regimens in molecularly defined secondary AML. Leukemia 38:762–768, 2024
Lu J, Xue S-l, Wang Y, et al: Venetoclax and decitabine vs intensive chemotherapy as induction for young patients with newly diagnosed AML. Blood 145:2645–2655, 2025
Chiche E, Rahmé R, Bertoli S, et al: Real-life experience with CPX-351 and impact on the outcome of high-risk AML patients: a multicentric French cohort. Blood Adv 5:176–184, 2021
Gangat N, Tefferi A: CPX-351 (Vyxeos) in Acute Myeloid Leukemia: Time to Move on? American Journal of Hematology 100:1271–1276, 2025

Yes there is potential conflict of interest. AAM: Research funding from Novartis, BMS, Solu Therapeutics and Sanofi; MRL: BeiGene Shanghai, Amgen,Astellas Pharma, Actinium Pharmaceuticals, Syndax, Consultancy; MMP: Research funding from Kura Oncology, Stemline therapeutics, Epigenetix, Solutherapeutics, Polaris and has served on the advisory board for AstraZeneca and SOBI pharmaceuticals, CAY: Curio Science, Consultancy; NG: Advisory board to DISC Medicine and Agios; AT and NG are members of the editorial board for BCJ.

Supplementaltable1patientcharectersiticsbasedonupfrontCPX351andHMAandVenAMLMR.docx
Supplemental Table 1
Supplementaltable2ComparisonoftheratesofachievementofcompleteresponseCPXvsHMAAMLMR.docx
Supplemental table 2
Supplementaltable3ComparisonofOSinCPXvsHMA20thjuly.docx
Supplemental Table 3
CPX351vsVenHMAsupplementalfigure5.tif
Supplemental figure 5
Supplementaltable5ComparisonoftheratesofachievementofcompleteresponseCPXvsHMAJuly20thsecondaryaml.docx
Supplemental Table 5
Supplementaltable4patientcharectersiticsbasedonupfrontCPX351andHMAandVennondenovo.docx
Supplementary Table 4
CPX351vsVenHMAsupplementalfigure1.tif
Supplemental Figure 1
Supplementaltable6ComparisonofOSinCPXvsHMAsecondaryaml.docx
Supplemental Table 6
CPX351vsVenHMAsupplementalfigure2.tif
Supplemental Figure 2
CPX351vsVenHMAsupplementalfigure4.tif
Supplemental Figure 4
CPX351vsVenHMAsupplementalfigure3.tif
Supplemental Figure 3
CPX351vsVenHMAsupplementalfigure6.tif
Supplemental figure 6

Download PDF

Review #2 received at journal
29 Dec, 2025
Reviewer #2 agreed at journal
15 Dec, 2025
Review #1 received at journal
12 Dec, 2025
Reviewer #1 agreed at journal
12 Dec, 2025
Reviewers invited by journal
11 Dec, 2025
Editor assigned by journal
05 Dec, 2025
Submission checks completed at journal
05 Dec, 2025
First submitted to journal
28 Nov, 2025

You are reading this latest preprint version

CPX-351 (Liposomal Cytarabine and Daunorubicin) versus venetoclax plus hypomethylating agent therapy in newly diagnosed acute myeloid leukemia: a retrospective comparison involving 600 Mayo Clinic patients

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Patient characteristics

Treatment response

Survival outcomes

Outcomes in AML-MR (N = 274)

Outcomes in secondary AML (N = 253)

Survival risk models

Discussion

Declarations

Conflict of interest:

Author contributions:

Acknowledgments:

Data availability statement:

References

Additional Declarations

Supplementary Files

Status:

Version 1