急性髓系白血病流式细胞术全程自动化诊断技术研究  

Automated flow cytometry analysis for diagnosis of acute myeloid leukemia

作　　者：郭玉娟 李智伟[2] 芮东升[1] 雷伟 摆文丽 陈鉴 李敏 农卫霞[3] 张向辉[1] 王奎 GUO Yujuan;LI Zhiwei;RUI Dongsheng;LEI Wei;BAI Wenli;CHEN Jian;LI Min;NONG Weixia;ZHANG Xianghui;WANG Kui(Department of Preventive Medicine,School of Medicine,Shihezi University,Shihezi,Xinjiang 832000,China;Clinical Laboratory Center,People’s Hospital of Xinjiang Uygur Autonomous Region,Urumqi,Xinjiang 830001,China;The Hematology Department of the Frist Affiliated Hospital,School of Medicine,Shihezi University,Shihezi,Xinjiang 832000,China)

机构地区：[1]石河子大学医学院预防医学系,新疆石河子832000 [2]新疆维吾尔自治区人民医院临床检测中心,新疆乌鲁木齐830001 [3]石河子大学医学院第一附属医院血液风湿科,新疆石河子832000

出　　处：《石河子大学学报（自然科学版）》2022年第4期431-437,共7页Journal of Shihezi University(Natural Science)

基　　金：国家自然科学基金项目(81860374)。

摘　　要：目的为解决急性髓系白血病(acute myeloid leukemia,AML)应用流式细胞术进行诊断时遇到的技术门槛和分析效率问题,研究实用的自动化AML诊断分析技术,推动流式实验室AML全程自动化分析技术的发展。方法本文开发ArcDia(Automated Registration of Cell populations and Diagnosis and Immunophenotyping of AML)方法,对来源于公共数据库的AML基准数据(数据1),和来源于流式实验室的AML检测数据(数据2)进行细胞聚类分析,采用有监督分类模型对亚群中心即宏细胞进行二级分类并标注,最后提取亚群特征并进行AML诊断。结果聚类方法为正态混合模型时,数据1诊断结果的灵敏度、特异度、准确率、F-measure为1、0.99、0.99、0.99,数据2为0.91、0.95、0.94、0.94;使用FlowSOM时,相应诊断结果的4个指标,数据1为1、0.98、0.98、0.98,数据2为0.88、0.95、0.93、0.93。结论诊断结果准确率较高,自动化过程完整且贴近临床分析,这展示了该自动化诊断技术的准确性和可用性,可以在临床流式检验室中推广应用。Objective Aiming at the technical obstacles and analysis inefficiency in the diagnosis of acute myeloid leukemia(AML)by flow cytometry,to develop practical automated AML diagnostic analysis toolset ArcDia(Automated Registration of Cell populations and Diagnosis and Immunophenotyping of AML)and to promote the application of full-process automated AML analysis technology in flow cytometry laboratory.Methods Cluster analysis were performed for AML benchmark data from public database(Data 1)and AML text experiment data from a flow laboratory(Data 2),A supervised procedure was proposed for meta cell classification and registration,cell subgroup characteristics were extracted and AML diagnosis was performed.Results When the clustering method was normal mixed model,the sensitivity,specificity,accuracy and F-measure of AML diagnosis were 1,0.99,0.99 and 0.99 for Data 1,and were 0.91,0.95,0.94 and 0.94 for Data 2;when using FlowSOM to cluster cells,four corresponding indicators were 1,0.98,0.98 and 0.98 for Data 1,and were 0.88,0.95,0.93 and 0.93 for Data 2.Conclusion The accuracy of AML diagnosis results is high,the automated process is complete and similar to that of clinical lab analysis,and ArcDia can be promoted widely in clinical flow laboratories.

关 键 词：急性髓系白血病 流式细胞术 聚类分析 亚群标注 自动化诊断 

分 类 号：R733[医药卫生—肿瘤]