Exploration Of A New Prognostic System Of 125 AML Patients In A Single Center

[Objective]Acute myeloid leukemia(AML)is a kind of malignant clonal diseases derived from hematopoietic stem and/or progenitor cells with strong heterogeneity.The traditional prognosis stratification based on cytogenetic and molecular defects is insufficient to provide a comprehensive and accurate perspective for the prognosis of AML in real world.This study aims to explore new prognostic models for prognosis stratification in AML by a mathematical operation mode which integrate a mass of clinical information.[Methods]This study retrospectively collected 125 patients with newly diagnosed AML(except APL)in the Department of Hematology of the First Affiliated Hospital of Xiamen University from January 1,2017 to October 1,2020.All the included patients have full clinical data,including targeted nex-generation sequencing(NGS)of mutation hotspots,cytogenetics and clinical baseline(age,sex,karyotype,number of mutations,blood routine results,LDH,bone marrow blast cell counts,ECOG score,etc.).The COX proportional hazard regression model and Lasso regression model were used to analyze the patients’overall survival(OS)and relapse-free survival(RFS),ultimately formatting a new prognostic scoring system.Finally,the new model was applied to the stratification of AML patient groups and the prediction of the risk of early recurrence and early death.[Results]The OS rate and and the RFS rate of 125 patients with AML were 60.8%and 48.8%,respectively.The median OS was 20 months,and the median RFS was 15 months.According to the Chinese guideline for the diagnosis and treatment of adult acute myeloid leukemia(non APL)(2017 version),23 patients(18.4%)were divided into low-risk group,43 patients(34.4%)were middle-risk group,and 59 patients(47.2%)were high-risk gruop.111 patients(88.8%)had hotspot gene mutations,among them,55 cases(44%)with 1-2 mutations,56 cases(44.8%)with 3 or more mutations.The commen mutations with frequencies≥10%were TET2,FLT3-ITD,DNMT3A,CEBPA,IDH2,GATA2,RUNX1,ASXL1,NPM1,NRAS.By combining patient’s genetics information and clinical indicators,Logstic regression analysis identified four adverse factors(IDH2 mutation,TP53 mutation,age≥60 years,ECOG score≥3 points)affecting the complete remission rate.Using COX regression and Lasso regression,we found that six indicators(age,white blood cell count,LDH,ECOG,FLT3-ITD,TP53)are associated with the patient’s OS,and seven factors(mutations number,age,white blood cell count,hemoglobin level,LDH,ECOG,FLT3-ITD)are related to the patient’s RFS.Finally,we established a new prognostic scoring system following Lasso regression analysis with the aforementioned factors.It could provide a more accurate model for the risk stratification of 125 patients as well as the prediction of the early death and recurrence.[Conclusion]This study explored new prognostic,predictive tools for estimating OS and RFS in AML.Compared with the conventional recommended stratification for AML prognosis,this new system is more accurate and comprehensive to predict the prognosis and to provide significant instruction for the treatment of AML.