AML: Negative Prognostic Impact of Early Blast Persistence Can be Overcome By Subsequent Remission Induction

Introduction: In acute myeloid leukemia [AML] intensive treatment usually includes an interim bone marrow assessment in the context of the first induction chemotherapy cycle. Early blast clearance has been shown to predict for achievement of a complete remission [CR], which makes the interim bone marrow result a widely used additive for further therapy decision making. However, there is an ongoing debate on the prognostic role of early blast clearance (day 14 - 21) and consistent clarification of this issue is still lacking. Here, we provide monocentric data from a large intensively treated AML cohort addressing the prognostic impact of early remission status on long-term survival.

Methods: Data of 1008 intensively treated AML patients who were diagnosed between 2000 and 2017 and aged between 18 and 86 years were retrieved from our local AML database. Evaluation of bone marrow aspirates and / or biopsies was conducted by cytologic, flow cytometric and histopathological methods. Remission status was determined in accordance with the European-Leukemia Net [ELN] criteria of 2017. Patients were grouped according to their remission status after induction 1 (day 14 - 21) and prior to further therapy (chemotherapy and / or allogeneic hematopoietic stem cell transplantation [HSCT]). Baseline characteristics of the two groups such as age, cytogenetic / molecular risk according to ELN 2010, performance score according to Eastern Cooperative Oncology Group [ECOG] and Charlson Comorbidity Index [CCI] were compared using chi-square test and Kruskal-Wallis H test followed by post-hoc testing. Overall- and relapse-free-survival [OS, RFS] were analyzed using Kaplan-Meier- and Cox regression models.

Results: In 57 % (n = 572) of the entire cohort, complete / incomplete remission [CR / CRi] or at least morphologic leukemia free state [MLFS] were observed after the first cycle of induction therapy. Partial response [PR] was found in 19 % (n = 196) and persistent disease [PD] in 24 % (n = 240) of all patients. The three different groups (CR / CRi / MLFS vs. PR vs. PD) did not significantly differ in age, ECOG-score or CCI, but - as expected - substantially with regard to ELN risk classification. Of the entire cohort, 62 % (n = 626) received a double induction, 61 % (n = 617) consolidation chemotherapy and 47 % (n = 477) allogeneic HSCT as a part of first-line therapy. Prior to HSCT, 85 % (n = 406) of the patients were in first CR / CRi / MLFS. Amongst these 406 patients, 51 % (n = 243) had only achieved PR / PD in early bone marrow puncture on day 14 - 21 of induction therapy. 71 % of these patients (n = 173) were converted into CR / CRi / MLFS prior to HSCT by additional chemotherapy. Within the non-transplanted group, 58 of 130 patients (44 % with initial PR / PD) achieved CR / CRi / MLFS after an additional cycle of induction chemotherapy. Early blast persistence and refractory disease were generally associated with inferior OS as compared with blast clearance (p < 0.001). Interestingly, early PR lost its negative prognostic impact on survival after conversion into CR / CRi / MLFS by additional therapy prior to post-induction therapy. This could particularly be observed in patients undergoing HSCT, but also in non-transplanted patients receiving a second cycle of induction therapy. In patients without allogeneic HSCT, early PR with blast clearance beyond induction 1 resulted in impaired RFS (p = 0.045) but not OS, possibly due to the influence of subsequent salvage therapy. Early PD remained prognostically unfavorable for OS and RFS in both transplanted and non-transplanted patients even after subsequent conversion into CR / CRi / MLFS (p = 0.002 and p = 0.021).

Conclusion: Our results show that the achievement of CR / CRi / MLFS prior to post-remission therapy overrides the negative impact of an early partial response on OS, irrespective of the favorable prognostic impact that early blast clearance generally has on OS. With regard to OS, our results suggest a predominant role of the final remission quality after induction therapy. However, the adverse impact of achieving no response at interim bone marrow assessment can apparently not be attenuated by a later remission induction. On the assumption that the final depth of remission after induction therapy is predictive for OS, MRD-guided decision making prior to consolidation treatment might further improve first-line AML therapy, particularly in the intermediate ELN risk group.