Construction and Validation of the New Prognostic Model Based on Genetic Risk Stratification for Chinese AML Patients

Acute myeloid leukemia (AML) is a heterogeneous malignancy characterized by the clonal expansion of aberrant hematopoietic stem and progenitor cells. The heterogeneity of AML is underscored by recurrent cytogenetic and molecular abnormalities that significantly impact patient prognosis. Risk stratification is critical for guiding patient treatment and assessing prognosis, necessitating ongoing comparison and improvement of the existing stratification systems. In China, where the incidence of AML predominates, it represents a staggering 70% of all acute leukemia cases. Given that China's population accounts for approximately 18% of the global population, this prevalence underscores the nation's substantial contribution to the worldwide burden of AML, highlighting the critical importance of AML research and treatment advancements in China to the international medical community. This retrospective study enrolled a total of 378 patients who received intensive induction chemotherapy from three centers in China (training cohort: n = 298; external validation cohort: n = 80) to validate the 2023 China AML genetic risk stratification (2023 CN) and compared it with the 2017 and 2022 European Leukemia Network (ELN) genetic risk stratification (2017 ELN and 2022 ELN). Our results suggest that the 2023 CN has a significant advantage in predicting relapse and relapse-free survival (RFS) compared to the 2017 ELN and 2022 ELN. By comparing the differences in risk stratification of specific genetic variants, we proposed a new risk stratification system (n-2023 CN) more suitable for Chinese AML patients. The n-2023 CN recommended that patients with t(8;21)(q22;q22) or inv(16)(p13q22)/t(16;16)(p13;q22) combined with mutated KIT at the D816 locus (KITD816) in the intermediate-risk group. Mutated FLT3-ITD with a high allele ratio (≥0.5; FLT3-ITDhigh) and mutated FLT3-ITD with a low allele ratio (<0.5; FLT3-ITDlow) were categorized as adverse-risk and intermediate-risk groups, respectively, regardless of concurrent NPM1 mutation. Significant differences were observed between the groups in the n-2023 CN. It performs better in predicting complete remission (CR), 3-year OS, and 3-year RFS, with the area under the receiver operating characteristic curve (AUC) values of 0.697, 0.68, and 0.616, respectively. We further found that age ≥60 years, positive CD15 expression of leukemia cells, and del(7q) were independent adverse factors for OS. Combining these factors with the n-2023 CN, we constructed a new prognostic model, which showed better efficacy in predicting CR, 1-year OS, and 3-year OS, with AUCs of 0.791, 0.737, and 0.714, respectively. The nomogram of the new prognostic model demonstrated good forecasting ability in internal validation through bootstrap with 1,000 repetitions, with the concordance index being 0.717 (0.679-0.755). We also applied the new prognostic model to the external validation cohort. We found that it could stratify patients into three risk categories with significantly different survival rates， and the model showed favorable efficacy in predicting 3-year OS, with an AUC of 0.736. In conclusion, the n-2023 CN and the new prognostic model may provide essential references for prognostic assessment and personalized treatment of AML patients in China. Our study, encompassing patient cohorts from diverse regions and healthcare facilities, is designed to ensure a robust sample size and comprehensive data set. By tapping into the distinctive features of the Chinese patient population, our research contributes to the global diversity of AML studies, delivering invaluable insights and data to the international medical community.