Measurable residual disease (MRD) assessment as an early end point oftreatment efficacy in acute myeloid leukemia (AML): A pethema studyof 2,623 patients (Pts) treated with intensive therapy

Abstract

Abstract Background: Despite advances in treatment of AML pts considered fit for intensive therapy, significant unmet medical needs remain. Early end points for survival can accelerate the approval of novel therapies. MRD has been recently accepted as an early end point in other hematological malignancies, but important questions remain unanswered to support its acceptance in AML. Is multiparameter flow cytometry (MFC) more sensitive than the morphological definition of complete remission (CR)? Are MRD kinetics more prognostic than a single assessment? Are there genetic risk groups in whom MRD is not informative? Is the prognostic value of MRD independent of induction and consolidation therapy? Are MRD negative rates achieved with each therapy associated with different survival? Aim: Investigate the potential role of MRD assessed by MFC as an early endpoint for survival in AML pts treated with intensive therapy. Methods: The PETHEMA registry had 2,623 newly-diagnosed AML pts treated in routine practice and clinical trials, who had at least one MRD assessment at first CR using local MFC methods. Pts were mainly treated with 3+7 (n=2,160), followed by 3+7 plus a FLT3 inhibitor (n=137), FLAG-based regimens (n=97). Additional treatments included CPX-351 (n=59), 3+7+Etoposide (n=44), 3+7+GO (n=37), and others (n=89). Pts were stratified into low (CBF, n=348), intermediate-low (CEBPA/NPM1+ and FLT3-, n=392), intermediate-high (MRC 2010 intermediate or normal karyotype with FLT3- and NPM1-, n=738) and high (MRC 2010 high or FLT3+, n=784) genetic risk. Relapse-free (RFS) and overall survival (OS) were landmarked at the time of MRD assessment. Results: Of the 2623 pts in CR1, 1256 (48%) showed MRD ≥0.1% and 532 (20%) had MRD levels ≥0.01% - <0.1%. When compared to pts with <0.01% MRD (n=835, 32%), those with MRD levels ≥0.01% - <0.1% and ≥0.1% showed inferior RFS (medians of 35, 23 and 17, p<.001) and OS (medians of 94, 55 and 33 months, p<.001). Hence, the 0.01% cutoff was selected to define negative vs positive MRD status. Positive MRD was associated with inferior RFS (HR= 1.3, p<.001) and OS (HR= 1.4, p<.001). The respective global odds ratio estimated by the bivariate Plackett Copula model were 8.0 (95% CI, 5-16.7) and 9.8 (95% CI, 5.7-27.8), which indicates a strong correlation between MRD status at CR1 and survival outcomes. Of note, a sub-analysis in pts with ≥2 MRD assessments (n=322) showed that MRD kinetics are more prognostic than a single MRD test. When compared to pts with sustained MRD negativity and those having positive and negative MRD result, pts who were systematically MRD positive showed inferior RFS (median of 61, 48 and 16 months, p<.001) and OS (median 96 vs 105 vs 35 months, p<.001). In multivariate analysis adjusted for genetic risk and the different treatments, MRD status retained independent prognostic value for RFS (HR= 1.3, p<.001) and OS (HR= 1.3, p<.001). MRD status predicted different outcome in pts with genetic high-risk and was borderline significant in those with low-risk. By contrast, there were no differences in intermediate-risk groups, which could be partially associated with MRD-guided intensification. Accordingly, pts who were not transplanted showed different OS according to MRD status. By contrast, MRD status at CR1 was not prognostic in pts undergoing autologous or allogeneic transplant. The association between the treatment effect in MRD and survival was investigated in the three most frequent induction regimens. MRD negative rates were higher after 3+7 plus a FLT3 inhibitor vs 3+7 vs FLAG-based regimens (42% vs 31% vs 29%, p=.02). The higher MRD negative rates preceded longer RFS (median 32 vs 21 vs 14 months, p=.01) and OS (median 58 vs 47 vs 22 months, p=.01) with 3+7 plus a FLT3 inhibitor vs 3+7 vs FLAG-based regimens. Conclusion: MRD assessment by MFC proved to be more sensitive than CR and showed a strong patient-level association with survival. MRD negative rates achieved with each treatment were also associated with differences in survival. While these results qualify MRD as a potential early end point of treatment efficacy in AML pts considered fit for intensive therapy, attention should be paid to genetically-defined intermediate risk groups in whom MRD status may guide treatment decisions that could ameliorate the poor prognosis of positive MRD. MRD kinetics may be required to improve prognostication of such pts, particularly in those who are transplanted.