The primary goal of induction therapy in acute myeloid leukemia (AML) is to induce complete remission with hematologic recovery (CR) or incomplete hematologic recovery (CRi) [1]. In patients receiving intensive chemotherapy, guidelines suggest considering bone marrow evaluation at day 14 and re-induction if residual disease ( > 5% blasts) is present [2]. However, despite prior studies reporting a prognostic role of residual disease at D-14, its utility has not been widely accepted as an independent prognostic factor, and other studies have reported no consistent association with long-term outcomes [3,4,5,6]. Additionally, depending on underlying genetic risk and performance status, patients achieving CR/CRi may proceed with either allogeneic stem cell transplant (ASCT) or multiple cycles of consolidation chemotherapy, which may include anthracycline-based or cytarabine-only regimens [2, 7].
The objective of the current retrospective study was to examine how these pre- and post-CR/CRi factors impact ASCT-censored relapse-free survival (RFS) and overall survival (OS), as well as post-transplant survival, following intensive induction chemotherapy.
The current study was conducted under IRB-approved minimum risk protocols that allowed retrospective collection and analysis of patient data. Formal criteria were used for diagnosis, European LeukemiaNet (ELN) 2022 cytogenetic risk stratification, and response [1]. We included patients who received standard “7 + 3” regimens (3 days of daunorubicin or idarubicin, combined with 7 days of cytarabine) as the first induction therapy and achieved CR/CRi. RFS and OS were calculated from the date of CR/CRi. Conventional statistical methods were utilized (JMP Pro version 17.1.0, SAS Institute, Cary, NC, USA).
A total of 279 AML patients diagnosed between 2004 and 2020 were included (Supplementary Table 1). The median age was 58 years; 167 patients (60%) were male. Disease type included de novo (n = 210; 75%), post-chronic myeloid neoplasm (CMN; n = 45; 16%), and therapy-related (n = 24; 9%). Induction regimens included idarubicin 12 mg/m² plus cytarabine (n = 183; 66%), daunorubicin 60 mg/m² plus cytarabine (n = 76; 27%) and daunorubicin 90 mg/m² plus cytarabine (n = 19; 7%), one patient had daunorubicin 45 mg/m2 and cytarabine. Re-induction was utilized to achieve CR/CRi in 72 patients (27%), at a median of 18 days from the start of first induction therapy. Indications for re-induction included residual bone marrow blasts (median 30%; range, 4–90%) on day 14 (n = 53) or days 21–28 (n = 15), persistent extramedullary disease (n = 2), or circulating blasts (n = 2). At a median transplant-censored follow-up of 26 months, 93 deaths (33%), 117 relapses (42%), and 134 ASCTs (48%) were documented.
Median RFS was 18 months (95% CI: 14–23), with respective 1-, 2-, 3- year RFS of 63% (95% CI 56–69), 42% (35–49) and 35% (28–42). The cumulative incidence of relapse accounting for death as a competing risk was 32% (95%CI 27-40), 51% (95%CI 45–60) and 58% (95% CI 51–68) at 1-, 2- and 3- year, respectively. In multivariable analysis (MVA), age ≥65 years (HR 1.5, p = 0.047), post-CMN (HR 1.7, p = 0.03), ELN adverse karyotype (vs intermediate HR 1.7, p < 0.01; vs. favorable HR 2.6, p < 0.01), and the need for re-induction to achieve CR/CRi (HR 1.9, p = 0.01) were independently predictive of relapse. The median RFS was 14 months (95% CI 8–16) in those who need re-induction vs 22 months (95% CI 17–29; p < 0.01) in those not requiring re-induction. RFS was not significantly impacted by NPM1 mutation (HR 0.4, p = 0.08) or FLT3-ITD (HR 1.4, p = 0.2).
The median OS was 76 months (95% CI: 37–124) with a 5-year survival probability of 52%. In MVA (Table 1), age ≥65 years (HR 2, p = 0.01), ELN adverse karyotype (vs. intermediate HR 2.7, p < 0.01; vs. favorable HR 5, p < 0.01), FLT3-ITD (HR 1.9, p = 0.046), and the need for re-induction (HR 2, p = 0.02) independently predicted OS; NPM1 mutation did not impact OS, HR 0.3, p = 0.2. The median OS was 37 months (95% CI 13–85) in those who need re-induction vs 106 months (95% CI 56–not reached[NR]; p < 0.01), in those not requiring re-induction (Fig. 1).
Fig. 1: Impact of need for re-induction therapy on relapse-free, post-remission, and post-transplant survival.
A Relapse-free survival (RFS): Patients requiring re-induction had significantly shorter RFS compared with those not requiring re-induction (median RFS 14 months [95% CI, 8–16] vs. 22 months [95% CI, 17–29]; p = 0.001). B Post-remission overall survival: Requirement for re-induction was associated with inferior post-remission survival (median 37 months [95% CI, 13–85] vs. 106 months [95% CI, 56–not reached]; p = 0.004). C Post-transplant survival: Among patients who underwent allogeneic hematopoietic stem cell transplantation, post-transplant survival did not differ significantly by re-induction status (median 61 months [95% CI, 10–131] vs. 76 months [95% CI, 26–216]; p = 0.6).
Table 1 Cox regression analyses of factors associated with relapse-free survival, overall survival, and post-transplant survival.
Figure 2 shows the impact of three different consolidation strategies: anthracycline-based regimen (idarubicin 12 mg/m² or daunorubicin 60 mg/m² on days 1 to 2/3 plus cytarabine 100 mg/m² on days 1 to 5/7; N = 84), high-dose cytarabine (HiDAC; 3 g/m² every 12 h on days 1, 3, and 5; N = 99), and intermediate-dose cytarabine (IDAC; 1–2 g/m² every 12 h on days 1, 3, and 5; N = 58) on RFS and OS.
Fig. 2: Impact of first consolidation regimen on relapse-free survival and overall survival.
A Relapse-free survival (RFS) in the overall cohort. RFS did not differ significantly across consolidation strategies (p = 0.9), with median RFS of 16 months for high-dose cytarabine (HiDAC), 24 months for intermediate-dose cytarabine (IDAC), and 18 months for anthracycline plus cytarabine. Comparisons between HiDAC and IDAC (p = 0.7), anthracycline plus cytarabine and IDAC (p = 0.7), and anthracycline plus cytarabine and HiDAC (p = 0.9) were not significant. B RFS among patients aged <65 years. No significant differences in RFS were observed across consolidation strategies (p = 0.3). Median RFS was 16 months for HiDAC, 27 months for IDAC, and 41 months for anthracycline plus cytarabine, with no significant differences among individual comparisons (HiDAC vs IDAC, p = 0.6; anthracycline plus cytarabine vs IDAC, p = 0.2; anthracycline plus cytarabine vs HiDAC, p = 0.2). C RFS among patients aged ≥65 years. Anthracycline plus cytarabine was associated with inferior RFS compared with IDAC (p = 0.001), with median RFS of 8 and 23 months, respectively. D Overall survival (OS) in the overall cohort. OS did not differ significantly across consolidation strategies (p = 0.4), with median OS of 85 months for HiDAC, 40 months for IDAC, and 125 months for anthracycline plus cytarabine. No significant differences were observed between individual regimens (HiDAC vs IDAC, p = 0.3; anthracycline plus cytarabine vs IDAC, p = 0.2; anthracycline plus cytarabine vs HiDAC, p = 0.95). E OS among patients aged <65 years. A trend toward differential OS was observed across consolidation strategies but did not reach statistical significance (p = 0.07). Median OS was 85 months for HiDAC, 40 months for IDAC, and not reached for anthracycline plus cytarabine. OS was superior with anthracycline plus cytarabine compared with HiDAC (p = 0.02), while comparisons with IDAC were not significant (HiDAC vs IDAC, p = 0.7; anthracycline plus cytarabine vs IDAC, p = 0.2). F OS among patients aged ≥65 years. Anthracycline plus cytarabine was associated with inferior OS compared with IDAC (p = 0.003), with median OS of 12 versus 40 months, respectively. *Consistent with institutional protocol, HiDAC was not administered to patients within this age group.
On age-stratified analysis, in older patients (age ≥65 years; Fig. 2), RFS was significantly better with IDAC (N = 36; median 23 months) compared with anthracycline-based regimen (N = 25; 8 months; p < 0.01) as the first consolidation treatment cycle. Similarly, median OS was longer with IDAC (40 months) vs. anthracycline-based regimen (12 months; p < 0.01). In patients younger than 65 years, RFS did not differ significantly across consolidation strategies (p = 0.3), with median RFS of 16 months for HiDAC, 27 months for IDAC, and 41 months for anthracycline-based regimen. For OS, although the overall comparison across consolidation strategies did not reach statistical significance (p = 0.07), median OS was 85 months for HiDAC, 40 months for IDAC, and NR for anthracycline-based regimen. Notably, anthracycline-based regimen was associated with superior OS compared with HiDAC (p = 0.02), while comparisons involving IDAC were not significant.
In the anthracycline-based consolidation cohort, patients who received 2-3 additional HiDAC/IDAC had better median OS (NR) compared to those who received none or only one cycle (20 months; p < 0.01). The respective median RFS was 34 months and 9 months, p = 0.06. In both the HiDAC and IDAC cohorts, RFS and OS did not differ significantly between patients who received 3–4 versus 1–2 consolidation cycles. In the HiDAC cohort, median RFS was 16 months versus 22 months (p = 0.5), and median OS was NR versus 45 months for patients receiving 3–4 versus 1–2 cycles, respectively (p = 0.2). Similarly, in the IDAC cohort, median RFS was 25 months versus 23 months (p = 0.4), and median OS was 177 months versus 40 months for patients receiving 3–4 versus 1–2 cycles, respectively (p = 0.1).
The need for re-induction did not influence post-ASCT survival (p = 0.6; Fig. 1), which was, however, significantly longer in patients who received Ida + Cytarabine induction (vs. Da 60 mg/m2 + Cytarabine, HR 0.6, p = 0.04), were transplanted in CR1 (HR 0.4, p < 0.01), or whose AML was de novo (vs. post-CMN, HR 0.5, p = 0.02), (Table 1).
Beyond established prognostic variables such as advanced age and adverse ELN cytogenetic risk, our study identified need for re-induction to achieve CR/CRi based on residual day-14 disease as an additional independent marker of inferior outcomes. Previous studies have reported worse OS in patients with early persistent disease, with some demonstrating association with inferior survival while others report limited predictive value [5, 8,9,10]. In our cohort, this observation remained prognostic even after stratification by baseline cytogenetic risk, and in de novo AML (Supplementary fig. 1), suggesting that early treatment response captures biologic features that may not be fully accounted for by current cytogenetic risk stratification. While assessment of depth of response such as measurable residual disease (MRD) is widely accepted for its prognostic value and sensitivity [11], early morphologic assessment of depth of response, despite historical debate [12, 13], may still provide actionable prognostic insight, with the added advantages of earlier availability, broad accessibility and relatively low cost. Notably, the adverse prognostic impact of residual day-14 disease was mitigated by ASCT, indicating that early consideration of ASCT may offset its negative impact.
In patients younger than 65 years, anthracycline-based consolidation was associated with improved OS with a numerically longer RFS compared with IDAC or HiDAC. The efficacy of anthracycline combined with standard-dose cytarabine as a first consolidation regimen has not been well characterized, as studies of anthracycline-containing consolidation regimens in this age group vary in regimen composition and have yielded mixed results [14]. Notably, most incorporated anthracyclines in combination with high- or intermediate-dose cytarabine, which may have contributed to greater hematologic toxicity [14]. In contrast, a study by Bradstock et al. reported improved leukemia-free survival without a significant increase in non-hematologic toxicity with increasing cumulative dose of idarubicin [15]. Our findings suggest a potential survival advantage with a standard-dose, induction-like anthracycline-based regimen as first consolidation cycle in younger patients and warrant further investigation. Among those aged ≥65 years, outcomes with anthracycline-based consolidation were less favorable, whereas IDAC was associated with improved survival in this group. These findings underscore the importance of age-informed consolidation strategies, particularly in older adults, where treatment tolerance, comorbidity, and cumulative toxicity may have a greater impact on long-term outcomes.
Summary of our findings: First, the need for re-induction to achieve CR/CRi based on day-14 residual disease independently predicts inferior RFS/OS but not post-ASCT survival, supporting early transplant consideration in these patients. Secondly, in patients younger than 65 years, inclusion of anthracyclines at the first consolidation cycle was associated with improved OS without compromising RFS, whereas in patients aged ≥65 years, IDAC (compared with anthracycline-based regimen) was associated with superior RFS and OS, supporting age-informed consolidation strategies.
The study is limited by its retrospective design and the lack of comprehensive mutational data beyond FLT3 and NPM1, precluding full ELN 2022 stratification; therefore, these findings should be interpreted within this context, and validation in more detailed molecularly annotated cohorts is warranted.
