CCAAT/enhancer-binding protein alpha (CEBPA) mutations are a favorable prognostic factor for acute myeloid leukemia (AML) patients in adult and pediatric. However, several previous studies have examined relapse rate may be as high as 30%.

To explore some features that predict relapse, we found that 33/345 de novo AML patients (9.5%) harbored a mutation in CEBPA, of which 33 patients had 25 (75.8%) harbored CEBPA-dm and 8 (24.2%) harbored CEBPA-sm. Five (15.2%) relapsed patients were all classical CEBPA-dm children. There were no statistically significant differences in clinical characteristics and survival outcomes between CEBPA-sm and CEBPA-dm patients. The classical CEBPA-dm patients with N-terminal mutation sites among amino acid (AA)1-59 were less likely to relapse[n=0(0%) vs. n=5(45.5%), P=0.011] and had superior RFS and EFS than patients with N-terminal mutation sites concentrated in the AA range 60-120 (RFS: 100.0% vs. 41.6±19.7%, P=0.005; EFS: 92.3±7.4% vs. 41.6±19.7%, P=0.03). Gene expression profiling in 9 classical CEBPA-dm patients (including 3 patients in N- terminal AA mutation sites among 1-59 and 6 patients among 60-120) showed that over-expression of PDLIM1 may be an indicator of poor prognosis [RSF of PDLIM1 high vs. low PDLIM1 expression: 33.3±27.2% vs. 100.0%, P=0.053].

When we applied the "adjusted risk stratification", namely N-terminal AA mutation sites used as one of the stratification indicators based on NCCN guidelines, we observed a more pronounced difference in RFS among adjusted risk groups, especially for the 24 cases classical CEBPA-dm patients (P=0.358 vs. P=0.027 for classical CEBPA-dm patients; P=0.028 vs. P=0.022 for classical CEBPA-dm patients who chemotherapy-only). Multivariate analysis identified that higher bone marrow (BM) blast(≥75%) was an independent risk factors for EFS [hazard ratio (HR): 1.51, 95%CI: 1.25-1.84, P=0.032] and normal karyotype was an independent favorable prognostic factor for RFS (HR: 0.31, 95%CI: 0.12-0.59, P=0.049) in classical CEBPA-dm cohort. N-terminal AA sites range 1-59 appeared to be a protective factor for RFS (HR=0.82, 95%CI: 0.61-1.38, P=0.056). In classical CEBPA-dm cohort receiving chemotherapy-only, higher BM blast(≥75%) at diagnosis was independent prognostic factors for RFS and EFS (HR=0.82, 95%CI: 0.61-1.38, P=0.056).

In conclusion, our clinical data suggested that N- terminal AA mutation sites concentrated among 1-59 appear to be a protective factor for RFS in classical CEBPA-dm patients. The "adjusted risk stratification" may be useful in re-assessing the prognosis of classical CEBPA-dm AML particularly in patients belonging to the intermediate-risk subgroup. Over-expression of PDLIM1 may be an indicator of poor prognosis and need to be validated in our future large samples.

No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution