Follow-up of the Randomized ASAP Trial Shows No Survival Advantage for Patients with Poor Responsive or Relapsed AML Who Received Intensive Salvage Chemotherapy for Remission Induction Prior to Allogeneic Transplantation Compared to Immediate Allogeneic Transplantation

INTRODUCTION

To attempt inducing a complete remission (CR) prior to allogeneic hematopoietic cell transplantation (alloHCT) in patients with active AML after first induction or untreated first relapse is considered as standard of care. This standard, however, has never been tested in a randomized controlled trial (RCT). With the availability of highly potent sequential conditioning regimens the value of CR induction prior to alloHCT is unclear for patients with poor responsive or relapsed AML. Against this background, the ASAP trial was conducted (NCT 02461537, Stelljes et al Lancet Hematology 2024). Here, we report the long-term follow-up (FU) and outcome according to genetic risk groups.

METHODS

To test salvage chemotherapy prior to alloHCT, patients aged between 18 and 75 years with AML and poor response after first induction or untreated first relapse and available HLA-compatible donor were randomized 1:1 to remission induction (RIST) with high-dose cytarabine plus mitoxantrone or immediate alloHCT with sequential conditioning after disease control (DisC). Disease control consisted preferentially of watchful waiting only. Overall survival (OS) from randomization according to the intention-to-treat (ITT) was the major secondary endpoint and is reported here with a median FU of 5 years. To improve molecular risk assessment available baseline samples were re-analyzed and classified according to the ELN 2022 risk classification.

RESULTS

281 patients were enrolled between September 2015 and January 2022. Of 140 patients randomized to DisC 135 proceeded to alloHCT (96%) and of 141 patients randomized to RIST 128 patients (91%) were transplanted. Treatment success was documented for 82.9% (116/140) of patients in the DisC arm and 79.4% (112/141) of patients in the RIST arm, resulting in a difference for treatment success of +3.4% (95% CI, –5.8% to 12.6%) for DisC versus RIST. OS at 3 years from randomization analyzed according to ITT was 54.8% (95% CI, 46% to 63%) versus 54.5% (95% CI, 46% to 62%) (p=0.95) and at 5 years 46.1% (95% CI, 37% to 55%) versus 47.5% (95% CI, 39% to 56%) (p=0.8), for DisC versus RIST, respectively, indicating no benefit of remission induction prior to alloHCT. Among patients with treatment success, disease-free survival (DFS) was not statistically different by treatment arm (4-year DFS from treatment success, 48.3% (95% CI, 39% to 57%) for DisC versus 49.1% (39% to 58%) for RIST, p=0.9). In multivariable Cox regression models for OS from randomization, ELN risk and age significantly predicted survival, but not treatment arm. TP53-lesions were detected in 21 patients randomized to DisC and 9 patients randomized to RIST. Three-year OS since randomization in patients with adverse risk AML with TP53-lesions was 24% (95% CI, 9% to 34%) versus 33% (95% CI, 8% to 62%) (log-rank test, p=0.5) and with adverse risk AML without TP53-lesions 48% (95% CI, 33% to 62%) versus 45% (95% CI, 29% to 59%) (log-rank test, p=0.8) for DisC versus RIST, respectively. With intermediate risk AML 3-year OS was 60% (95% CI, 46% to 72%) compared to 63% (95% CI, 51% to 73%) for patients randomized to DisC or RIST (log-rank test, p=0.52), respectively. Patients with favorable risk AML (94% enrolled with untreated relapse) randomized to DisC had 3-year OS of 90% (95% CI, 64% to 97%) compared to 46% (95% CI, 18% to 70%) for patients randomized to RIST (log-rank test, p=0.01).

CONCLUSIONS

In summary, longer follow-up of the ASAP trial confirms our previous results and shows no survival advantage for remission induction prior to alloHCT as opposed to immediate alloHCT. This result questions the general concept of remission induction prior to alloHCT, since immediate alloHCT after sequential conditioning may reduce time in hospital and treatment exposure. More potent bridging concepts with targeted drugs prior to alloHCT are warranted especially for adverse risk AML and need to be tested in RCTs to demonstrate sustained survival advantage. Together with the profound impact of genetic risk on long-term survival the results may be interpreted as a sign that patients with poor responsive or relapsed AML benefit rather from relapse prevention after alloHCT than conventional CR induction prior to alloHCT.