Minimal Residual Disease Assessment by Flow Cytometry In AML Patients at Time of Aplasia Post-Induction Is Highly Predictive of Outcome and Strongly Correlates with Molecular MRD In a Subgroup of NPM1 Positive Patients

Quantification of minimal residual disease (MRD) by multiparameter flowcytometry (MFC) in patients with acute myeloid leukemia (AML) provides significant prognostic information. Previous analyses have shown that positivity of leukemia-associated aberrant immunophenotype (LAIP) after induction and consolidation therapy is a risk factor concerning relapse and reflects over-all survival and relapse-free survival. Albeit molecular MRD monitoring has translated into therapeutic strategies in certain AML subgroups, e.g. APL, flowcytometric MRD assessment is still not integrated into therapeutic strategies.

We performed a retrospective analysis of flowcytometric MRD data of 201 adult patients with AML. Patients were treated within the clinical study of the German AML cooperation group (AMLCG99 or sHAM); patients with APL were excluded. 165 patients were in the intermediate risk group according to karyotype; NPM1 and Flt3 positivity was present in 80 and 59 patients, respectively. 51% of patients relapsed in the course of disease, the median overall survival (OS) and relapse-free survival (RFS) were 518 and 292 days, respectively. MRD assessment by 3-color-flowcytometry was performed at initial diagnosis, during bone marrow aplasia (day 16 – 18 of induction therapy), after induction, after consolidation and at time of relapse. MRD flow was compared to molecular MRD assessment in patients with mutated NPM1.

Data on flowcytometric MRD assessment were available in 99% of patients at initial diagnosis, 74% on day16 of induction therapy, 80% post-induction, 42% post-consolidation and 62% at time of relapse. MRD positivity was defined as percentage of LAIP positive cells > 0,15%. In addition, the post-therapeutic degree of LAIP reduction compared to initial diagnosis was assessed: an indequate degree of LAIP reduction was set at a log difference (initial diagnosis/ day16) < 2. We could confirm that flowcytometric MRD predicts OS and RFS. The absolute level of MRD as well as the relative degree of MRD reduction on day16 of induction therapy defined significantly distinct risk groups concerning OS and RFS (OS: p = 0,03, RFS: p = 0,002). Discriminating the different types of LAIPs, we could demonstrate that cross-lineage phenotypes have the best sensitivity and specificity for predicting relapse (sensitivity 64%, specificity 88% on day16). The course of LAIP and NPM1 as assessed by rtPCR, highly correlates on day16 (correlation coefficient 0,84).

Flowcytometric MRD assessment provides significant prognostic information at a very early point in induction therapy. In a subpopulation of NPM1+ AML patients we found a strong correleation of flowcytometric MRD and molecular MRD assessment. LAIP reduction on day 16 post-induction could therefore be integrated into prognostic models and can improve risk stratification especially in the absence of molecular MRD markers.

No relevant conflicts of interest to declare.