Although the pivotal phase-3 trial establishing CPX-351 as standard of care for therapy-related AML (tAML) or AML with myelodysplasia-related changes (AML-MRC) was performed only in patients aged 60–75 years [1, 2], it is approved for adults without age limit. Still, limited data resulted in restrictive recommendations regarding its use in younger patients (<60 years) by national and international guidelines [3]. To address this age issue (</≥60 years), to provide extended follow-up and outcome data after allogeneic hematopoietic cell transplantation (allo-HCT), we re-analyzed our real-world cohort, that had previously shown promising results of 188 patients receiving first-line induction with CPX-351 [4]. For this purpose, respective centers updated clinical data of 89 out of 95 patients (94%), who were alive at data cut of our previous study, resulting in 182 patients as cohort for this re-analyis. We compared patient-, disease- and transplant-related factors, response and transplant rates and OS from induction as well as post-transplant outcome regarding OS, relapse-free survival (RFS), relapse incidence (CIR) and non-relapse mortality (NRM) among both age groups. Details on statistics and definitions are given in the Supplemental Methods.

Patient-, disease- and transplant characteristics of the entire patient cohort have been previously described [4], but are here depicted separately for younger and elderly patients (<60 years, n = 46, 25%; ≥60 years, n = 136, 75%) in Supplemental Tables S1, S2. Median age was 53 and 68 years for the younger and the elderly subgroup. Both age groups had similar disease characteristics including WHO 2017 subtypes, blast counts,and adverse-risk features, e.g. ELN 2017 adverse risk and complex karyotype, suggesting a comparable disease biology. However, a lower proportion of younger patients had received hypomethylating agents prior CPX-351 (<60 yrs 2% vs. ≥60 yrs 14%, p = 0.04), while a higher proportion proceeded to allo-HCT compared to elderly (<60 yrs 85% vs. ≥60 yrs 61%, p = 0.003). Additionally, younger patients more frequently received allo-HCT from matched siblings, whereas elderly patients more often underwent haploidentical transplant.

Median follow-up from first-line CPX-351-based induction was 35.6 months (range, 0.5–57) with a median OS of 32.3 months (range, 0.6–60) and a 2-year OS rate of 52% for the entire cohort (Fig. S1). Among 168 patients (92%) with available response data, 118 (70%) achieved CR, CRi, or MLFS. Median OS in responders, who proceeded to allo-HCT (n = 94), was not reached, while it was 14.6 months for those without transplant (n = 24, Fig. S1). Of those 24 responders without subsequent allo-HCT, 13 did not receive further treatment due to poor performance status or comorbidities and of these 7 died due to treatment-related causes and 3 due to relapse. The other 11 patients received chemotherapy-based consolidation (CPX-351, n = 4; cytarabine, n = 7) and of these 2 died due to relapse, while 9 patients (ongoing CR, n = 8; active disease, n = 1). Of those 8 CR patients, 7 had favorable ELN 2017 risk (Fig. S2). Among responders, favorable-risk category was significantly more common in the non-transplant compared to the transplant subgroup (35% vs. 6%, p = 0.0007) and favorable-risk category was the main reason for not proceeding to transplant.

Given the promising outcome following CPX-351 induction and subsequent allo-HCT, we re-analyzed all patients undergoing allo-HCT with their disease and treatment course depicted in Fig. S3. Six additional patients underwent allo-HCT resulting in 122 allografted patients with a median follow-up of 33.7 months (range, 0.1 to 55) and a transplant rate of 67%. Median RFS from allo-HCT was 34.4 months (range, 0.1–45.4), while median OS was not reached (range, 0.1–54.9). The 2-year RFS and OS rates were 57 and 64% respectively, while CIR and NRM at 2 years were 20% and 24% (Fig. 1A).

Fig. 1: Outcome after allo-HCT.
figure 1

A Posttransplant RFS, OS, NRM and CIR for the entire cohort. B Posttransplant RFS, OS, NRM and CIR depending on disease status prior allo-HCT: patients in CR (n = 94) are depicted in orange (solid line), MRD-negative patients (n = 23) are depicted green (dotted line) and MRD-positive patients (n = 14) in red (dotted line), while R/R patients (n = 27) are depicted in black (solid line).

Full size image

Those 94 patients in remission at allo-SCT showed significantly better RFS (2-year RFS 60% vs. 41%, p = 0.01, HR 0.5, 95% CI 0.3–0.96) compared to those 27 patients with active disease, while numerical OS differences (2-year OS 67% vs 50%, p = 0.14, HR 0.6, 95% CI 0.3–1.3) were not statistically significant with a notable 2-year OS of 50% in patients with active disease (Fig. 1B). Among 94 patients proceeding to allo-SCT in remission, pretransplant flow-cytometry based measurable residual disease (MRD) was available for 37 patients (39%) with 14 patients (38%) being MRD-positive and 23 (62%) MRD-negative. With extended follow-up, post-transplant outcomes did not differ depending on MRD-status, despite an OS disadvantage in the MRD-positive subgroup by trend (Fig. 1B).

Regarding the impact of age, response (CR/CRi/MLFS) to CPX-351 (<60 years 67% vs ≥60 years 71%, n.s.) and treatment-related toxicities reflected by day 30 (<60 years 7% vs ≥60 years 8%, n.s.) and day60 mortality (<60 years 9% vs ≥60 years 15%, n.s., Fig. S4) were comparable between younger and elderly patients. However, in univariable analysis younger patients had significantly longer OS from CPX-351 induction (HR 0.48, 95% CI 0.31–0.75, p = 0.008) (Fig. 1C) and exhibited a significantly higher transplant rate (<60 years, n = 39, 85% and ≥60 years, n = 83, 61%, p = 0.003; Fig. S4). With a similar median post-transplant follow-up (<60 years, 34 months vs. ≥60 years, 33.3 months, p = 0.98) comparable day+100 mortality rates (<60 years 8% vs. ≥60 years 11%, p = 0.75) indicated similar early posttransplant toxicities among both groups. Regarding long-term posttransplant outcome, 2-year RFS and CIR also did not differ between the groups (Fig. 2). However, the younger group had a significantly lower 2-year NRM (<60 years 13% vs. ≥60 years 28%, p = 0.04), while 2 year OS differences were not statistically significant ( < 60 years 74% vs. ≥60 years 59%, p = 0.18, Fig. 2). Since NRM in elderly patients still increased beyond month +6 compared to younger patients, we investigated GvHD and causes of death: while younger patients had a higher overall aGvHD incidence (p = 0.02), rates of grade 3/4 aGvHD and of cGvHD were similar (Table S2) among the groups (Table S3). In the elderly cohort, infections were main cause of death (20/27, 74%) with a majority of 70% having a history of GvHD. In contrast, only 4 young patients died from treatment-associated complications.

Fig. 2: Outcome after allo-HCT depending on age (<60 vs. ≥ 60 yrs).
figure 2

RFS, OS, NRM and CIR from transplant for patients <60 years depicted in purple (n = 39) compared to patients ≥60 years depicted in blue (n = 83).

Full size image

Our updated real-world analysis of 182 patients receiving CPX-351 as first-line AML therapy represents the largest cohort comparing younger and elderly patients with the longest follow-up outside the phase-3 trial. The key finding are: (1) durable promising outcome for the entire cohort; (2) CPX-351 induction followed by allo-HCT facilitates long-term survival, even when allografted with active disease; (3) some (especially favorable risk) patients can achieve long-term survival with conventional consolidation; (4) CPX-351 is also effective in younger patients <60 years with a 2-year OS of 71%.

Our median OS of 32.3 months and 2-year OS of 52% appear favorable compared to the pivotal phase-3 trial [1, 2] and other retrospective studies [5,6,7], and is probably related to the higher transplant rate (67% vs. 35–45%) in our cohort. However, our promising posttransplant 2-year OS of 64% is in line with the other real-world analyses and the posttransplant 3- and 5-year OS (56% and 52%) in the post-hoc analysis of the prospective trial [4,5,6,7,8], further supporting the impact of allo-HCT to improve survival of high-risk AML patients. Converserly, long-term survival without allo-HCT was mainly seen in favorable risk patients - a subgroup without intention for first-line transplantation in general.

Disease burden at allo-HCT definitively impacted outcome, since MRD-negative patients had the best prognosis. Still, the 2-year OS rate of 50% in refractory patients is remarkable and in accordance to results from the ETAL3-ASAP trial, which demonstrated non-inferiority of direct allo-SCT without remission induction [9]. As long as new salvage therapies have not demonstrated a survival benefit [10, 11], this advocates for a timely scheduled allo-HCT after CPX-351 induction in high-risk AML regardless of remission. Likewise in other real-world analyses MRD-negativity at transplant facilitated improved survival, but reached no statistical significance [12]. Still the lack of biosamples not only hampered stringent analyses of MRD in a sufficient amount of patients, but also to investigate the impact of myelodysplasia-related (MR) gene mutations, which were associated with improved OS after CPX-351 induction in a randomized trial [13]. Given the ubiquitous use of the combination of hypomethylating agents/venetoclax, a comparison between both approaches is of particular interest. Although a retrospective comparison suggested similar outcomes [14], several differences between both patients groups (age, frequency of secondary AML and allo-HCT) impede interpretation and do not enable definitive conclusions. Indeed, for a unbiased direct comparison a well-designed, prospective randomized trial is required. Finally, we demonstrated a promising 2-year OS of 71% after CPX-351 induction in patients below 60 years, which was mainly related to a higher transplant rate in younger patients, since other clinical factors, response and toxicities were similar to the elderly patients. While posttransplant relapse rates were comparable, we observed a significantly higher NRM due to GvHD and infections in elderly patients beyond month 6 after transplant. These findings emphasize the need for stringent GvHD management and anti-infective prophylaxis especially in elderly patients. In summary, our re-analysis with expanded follow-up shows that CPX-351 is an effective treatment for high-risk AML independent of age facilitating allo-HCT for the majority of patients. However, the impact of age, the role of MR mutations and pretransplant disease status needs further clarification in ongoing trials (AMLSG 30/18 and ALFA2101).