Minimal Residual Diease Evaluation By Flow Cytometry Is a Complementary Tool To Cytogenetics For Treatment Decision In Acute Myeloid Leukemia

The two main criteria for risk-adapted treatment in AML are the presence of adverse cytogenetic or molecular features, and the response to induction treatment assessed by morphology, in which only failure to respond is truly informative. Therefore, more sensitive techniques are needed to evaluate the response. However, the potential value of MRD detection within different cytogenetic subgroups in AML has not yet been defined. This raises the question of whether a negative MRD result could counterbalance the adverse effect of poor-risk cytogenetics, or whether high MRD levels after induction modify the outcome of patients with otherwise favourable feature. In addition, it is not known whether the modality of intensification therapy modifies the influence of the level of MRD assessed after induction therapy.

We analyzed the prognostic impact of MRD level on the BM at CR after induction therapy using MFC in 306 non-APL AML patients. First, we have validated the prognostic value of MRD thresholds we had previously proposed (≥0.1%; ≥0.01-0.1%; and <0.01), with a 5-year RFS of 38%, 50% and 71%, respectively (p=0.002). The second aim was to investigate if MRD is independent of cytogenetics. Intermediate risk cytogenetics represent a grey zone for risk stratification, and our data show that MRD evaluation by MFC discriminate three prognostic subgroups, with a 5-year RFS of 38%, 50% and 70% for patients with high (n = 86), intermediate (n = 83), and low MRD levels (n = 18) (p = 0.03). In addition, MRD assessments yielded relevant information on favourable (poor prognosis for high MRD levels) and adverse (undetectable MRD overcomes adverse prognosis) cytogenetic groups. Regarding the role of different intensification approaches, we have observed that MRD levels (evaluated at the time of mCR) allowed the identification of three risk groups among the 118 patients who underwent autologous transplantation (p = 0.02), with 5-year RFS of 40%, 57% and 83% for patients with high, intermediate and low MRD, respectively. In patients intensified with allogeneic transplantation (n = 83), MRD levels also discriminated three subgroups with 5-year RFS of 53%, 62%, and 75% for cases with high, intermediate and low MRD levels (not statistically significant differences). In patients intensified with chemotherapy alone (n = 105), only cases that achieved low MRD had an acceptable outcome, with a 5-year RFS of 65%, whereas those with intermediate or high MRD levels had a very poor outcome (5-year RFS of 11% and 28%; p = 0.020). In addition, considering patients with intermediate or high risk MRD levels (n=169), intensification with transplant improved the outcome as compared with intensification with chemotherapy (p < 0.001), with a 5-year RFS of 75%, 43% and 20% for allogeneic transplanted patients, autologous transplantation, and chemotherapy treated patients, respectively. When only patients with high MRD levels (n = 86) were considered, those receiving autologous transplantation and chemotherapy had similarly very poor results (5-year RFS of 35% and 30%), and only allogeneic transplantation offered a favourable outcome for these patients, with a 5-year RFS of 66% (p = 0.06). By contrast, the type of intensification therapy did not influenced the outcome of patients with low MRD levels. Multivariate analysis (Cox regression) revealed an independent prognostic value for flow-MRD levels at the time of mCR, whether considered as a continuous (p=0.03) or categorical variable (p < 0.001), together with the cytogenetic classification (p = 0.001), and patient age (p = 0.002). A scoring system, easy to apply in clinical practice, was generated based on MRD level and cytogenetics, assigning 0 points if the variable fell in the low-risk category, 1 point for intermediate levels of MRD or medium-risk cytogenetics, and 2 points for high MRD levels or high-risk cytogenetics. Based on these scores, five significantly different AML risk groups could be defined for scores of 0 to 4. These groups had 5-year RFS of 100% (n = 4), 70% (n = 30), 54% (n = 102), 35% (n = 103) and 19% (n = 19), respectively (p < 0.001).

We conclude that immunophenotypic evaluation of MRD is a useful prognostic factor that could be used together with cytogenetics for the prognostic stratification of AML patients. Moreover, allogeneic transplant is the preferable option for patients with high MRD levels after induction therapy independently of the cytogenetic signature.