Acute myeloid leukemia (AML) is a group of hematological malignancies with high heterogeneity in clinical characteristics and prognosis. Based on the leukemia-associated immunophenotypes (LAIPs) of AML, minimal residual disease (MRD) which is related to the outcome of the patients could be detected by multiparameter flow cytometry. Although 0.1% was commonly used to discriminate outcome in most studies so far, measurable MRD or MRD level below 0.1% has also been associated with prognostic significance, the precise threshold of MRD for prognosis prediction in AML still remains controversial.

In this study, a total of 292 adult patients diagnosed as AML (non-M3) were enrolled, and 36 kinds of LAIPs were detected by flow cytometry. Based on the expression level of these LAIPs in 47 normal or regenerating bone marrow samples, the individual baseline level of each kind of LAIP was established, which ranged from <2.00×10-5 to 5.71x10-4, much lower than 0.1%. MRD statement based on 0.1% or individual baseline expression level of each LAIP were termed as 0.1%-MRD and individual-MRD respectively.

The survival analysis showed that, comparing with the generally used MRD cut off value of 0.1%, individual-MRD threshold distinguished the patients with different outcome more precisely. A total of 273,162 and 163 samples were detected at the time point of achieved complete remission (post CR), after the first (post Con1) and the second (post Con2) consolidation course, respectively. At all three time points, the patients of individual MRDneg showed significantly better survival compared with those of 0.1%-MRDneg/individual-MRDpos or 0.1%-MRDpos (3y-OS: P<0.001, P<0.001; 3y-EFS: P<0.001, P<0.001). Multivariate analysis also showed that individual-MRD status presented independent prognostic value at each of three time point.

Notably, in patients of low or intermediate risk groups (LR or IR) according to the NCCN risk stratification, the Individual-MRD status at post CR could identify the patients with significantly different outcome. The higher 3-year estimated OS and EFS rate were observed in the patients with individual-MRDneg in both LR and IR groups (LR group: 3y-OS: 92.2% vs 65.6%, p=0.003; 3y-EFS: 72.9% vs 44.6%, p=0.001; IR group: 3y-OS: 60.6% vs 37.3%, p=0.023; 3y-EFS: 53.3% vs 23.3%, p=0.006), while 0.1%-MRD status could not distinguish these differences clearly. Furthermore, among the patients of LR or IR group which received chemotherapy only, those with individual-MRDneg status presented favorable survival which was even comparable with the patients accepted allogeneic hematopoietic stem cell transplantation (ASCT) (3y-OS: 77.8% vs 77.2%, p=0.941, 3y-EFS: 64.2% vs 66.5%, p=0.611).

In conclusion, our study established the individual MRD threshold according to LAIP baseline levels in normal or regenerating BM samples. The individual MRD status could predict prognosis more precisely, especially in NCCN low or intermediated risk cohorts. In addition, with combination of individual MRD status and cytogenetic-molecular risk classification, we distinguished the patients with superior survival, which might help to guide the choose of ASCT strategy in clinical practice.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract

Sign in via your Institution