基于数据挖掘的鼻咽癌T_(3~4)分期风险预测模型构建  

Construction of risk prediction model for nasopharyngeal carcinoma with T_(3~4) stage based on data dig

作　　者：荀丽雪 王凡 陆进 邵倩倩 蒋成义 马海军 张浩轩 XUN Lixue;WANG Fan;LU Jin;SHAO Qianqian;JIANG Chengyi;MA Haijun;ZHANG Haoxuan(Department of Human Anatomy,Bengbu Medical University,Bengbu Anhui 233030;Anhui Key Laboratory of Computational Medicine and Intelligent Health,Bengbu Medical University,Bengbu Anhui 233030;Department of Otolaryngology,The First Affiliated Hospital of Bengbu Medical University,Bengbu Anhui 233004,China;College of Clinical Medical,Bengbu Medical University,Bengbu Anhui 233030)

机构地区：[1]蚌埠医科大学人体解剖学教研室,安徽蚌埠233030 [2]蚌埠医科大学数字医学与智慧健康安徽省重点实验室,安徽蚌埠233030 [3]蚌埠医科大学第一附属医院耳鼻喉科,安徽蚌埠233004 [4]蚌埠医科大学临床医学院,安徽蚌埠233030

出　　处：《蚌埠医科大学学报》2025年第1期74-80,共7页Journal of Bengbu Medical University

基　　金：安徽省高等学校自然科学研究重点项目(KJ2020A0553);安徽省重点实验室开放项目(AHCM2023Z003);安徽省研究生创新创业实践项目(2022cxcysj171);蚌埠医学院研究生科研创新计划项目(Byycxz22004)。

摘　　要：目的:探讨基于数据挖掘的鼻咽癌(NPC)T_(3~4)分期风险预测模型的构建及其在识别NPC高风险人群中的价值。方法:回顾性采集NPC病人326例,以8∶2的比例将其随机分为训练组和验证组。采用LASSO算法筛选NPC病人T_(3~4)分期的相关因素,对筛选出的因素进行logistic回归分析并构建风险预测模型。采用C指数(CI)、受试者工作特征曲线下面积(AUC)、校准曲线(CC)、决策曲线(DCA)对模型进行评价。结果:NPC风险预测模型训练组的CI=0.770、AUC=0.714、DCA净获益率为18%~80%、90%~98%,验证组的CI=0.835、AUC=0.781、DCA净获益率为28%~98%。结论:基于数据挖掘构建的NPCT_(3~4)分期风险预测模型在识别NPC高风险人群较为准确,可作为一种无创方法对NPC高危人群进行预测,并指导临床决策。Objective:To construct the risk prediction model for nasopharyngeal carcinoma(NPC)with T_(3-4) stage based on data dig,and explore its value in identifying the high risk NPC groups.Methods:A total of 326 NPC patients were retrospectively collected,and randomly divided into the training group and verification group with a ratio of 8∶2.LASSO algorithm was used to screen the factors related to the T_(3-4) stage of NPC patients,the selected factors were analyzed by logistic regression,and the risk prediction model was built.The C index(CI),area under the receiver operating characteristic curve(AUC),calibration curve(CC)and decision curve analysis(DCA)were used to evaluate the model.Results:In the training group of NPC risk prediction model,the CI was 0.770,the AUC was 0.714,and the DCA net benefit rate were 18%–80%and 90%–98%.In the verification group,the CI was 0.835,the AUC was 0.781,and the DCA net benefit rate was 28%–98%.Conclusions:The NPC T_(3-4) stage risk prediction model built based on data dig is more accurate in identifying high-risk NPC groups,and can be used as a non-invasive method to predict high-risk NPC groups,and guide clinical decision-making.

关 键 词：鼻咽肿瘤 风险预测模型 数据挖掘 

分 类 号：R739.8[医药卫生—肿瘤]