Abstract
Background:According to the 2022 WHO classification, acute myeloid leukemia (AML) with CEBPA mutations is considered a unique subtype that often has a better prognosis. Although previous studies have shown that the prognosis of therapy-related AML is usually worse than that of de novo AML, there is currently no research to clarify the prognosis of therapy-related CEBPA mutated AML (t- CEBPA AML).
Aims:This study firstly investigated the clinical characteristics and prognosis of t-CEBPA AML by systematically comparing the differences between t-CEBPA AML, therapy-related AML with wild type CEBPA (t-AML), and de novo CEBPA mutated AML(dn-CEBPA AML).
Methods: This study compared the genetics, clinical baseline and outcomes of t-CEBPA AML, dn-CEBPA AML and t-AML. The diagnosis of CEBPA mutated AML was based on the 2022 WHO classification. All patients were newly diagnosed AML and received treatment at the Hematology Institute of Peking University from January 2017 to December 2024.
Results: In this study, we retrospectively compared and analyzed the clinical baseline and survival data of total 478 patients: 23 t-CEBPA AML patients, 153 t-AML patients and 302 dn-CEBPA AML patients. Our findings revealed that t-CEBPA AML and t-AML patients were older than dn-CEBPA AML (44 years vs 49 years vs 40 years, p<0.001) and exhibited a similar co-mutation spectrum, with lower incidences of GATA2 mutations (8.6% vs 1.9% vs 26.8%, p<0.001), and higher incidences of KIT (17.3% vs 14.4% vs 6.6%, p<0.001), TP53(12.9% vs 10.5% vs 0.6%) and AML-myelodysplasia-related (MR) gene mutations (21.7% vs 25.9% vs 12.5%, p<0.001). But for cytogenetics, t-CEBPA AML and dn-CEBPA AML showed the same high frequency of normal karyotype (73.9% of t-CEBPA; 73.9% of dn-CEBPA), unlike t-AML (28.8%; p <0 .001). Although t-CEBPA AML and dn-CEBPA AML demonstrated higher CRc rates after chemotherapy than t-AML (91.3% vs 96.0% vs 74.5%, p<0.001), t-CEBPA AML and t-AML had a higher relapse rate than dn-CEBPA (30.4% vs 28.7% vs 18.2%, p<0.001), which might result in poor 3-years-overall survival (OS) (50.3% vs 64.0% vs 76.1%, p<0.001) and 3-year-disease free survival (DFS) (46.1% vs 60.0% vs 70.0%, p=0.003). In addition, receiving allogeneic hematopoietic stem cell transplantation (allo-HSCT) significantly improved survival in t-CEBPA AML (3-year-OS:76.0% vs 38.1%, p=0.072; 3-year-DFS:66.8% vs 35.7%, p=0.041) and t-AML (3-year-OS:74.7% vs 36.5%, p<0.001; 3-year-DFS:69.4% vs 45.7%, p=0.018) patients, but this improvement was not significant in dn-CEBPA AML. Among t-CEBPA AML, multivariate analysis identified TET mutation (HR=1.78, p=0.032), AML-MR genes mutation (HR=1.36, p=0.047) and measurable residual disease (MRD) negative after induction therapy (HR=0.085, p=0.041) as independent risk factors for inferior OS. Moreover, by comparing the specific mutation information, we found that the proportion of basic leucine zipper domain (bZIP) in-frame mutations of dn-CEBPA is higher than that of t-CEBPA AML (91.1% vs 60%, p=0.031). Considering the difference in the proportion of bZIP-inf mutation, we compared the long-term survival of t-CEBPAbZIP-inf and dn-CEBPAbZIP-inf, and the results suggest that there seems to be little difference in survival between them (3-year-OS: 62.0 vs 76.3%, P=0.47; 3-year-DFS: 45.0% vs 72.1%, p=0.075).
Summary/Conclusion: This study analyzed that t-CEBPA AML and t-AML have similar clinical features and poor prognosis compared with dn-CEBPA AML. T-CEBPA AML and t-AML should receive allo-HSCT to improve long-term survival. But among CEBPAbZIP-inf mutations patients, the poor prognosis caused by “therapy-related” is offset.
Keywords acute myeloid leukemia, CEBPA, bZIP-inf, therapy-related