对孤立性肺结节恶性概率预测模型的验证、比较和改良  

Verification,comparison and melioration of different prediction models for solitary pulmonary nodule

作　　者：鲍彤[1] 肖飞[1] 郭永庆[1] 石彬[1] 宋之乙[1] 梁朝阳[1] 孙宏亮[2] 刘德若[1] BAO Tong;XlAO Fei;GUO Yongqing;SHI Bin;SONG Zhiyi;LIANG Chaoyang;LIU Deruo(Department of Thoracic Surgery, China-Japan Friendship Hospital, Beifing, 100029, P.R.China;Department of Radiology, China-Japan Friendship Hospital, Belting, 100029, P.R.China)

机构地区：[1]中日友好医院胸外科,北京100029 [2]中日友好医院放射诊断科,北京100029

出　　处：《中国胸心血管外科临床杂志》2018年第6期471-476,共6页Chinese Journal of Clinical Thoracic and Cardiovascular Surgery

基　　金：中日友好医院青年科技英才计划(2014-QNYC-B-09)

摘　　要：目的分析影响孤立性肺结节(solitary pulmonary nodule,SPN)恶性概率的危险因素,对已有的SPN恶性概率预测模型进行验证、比较和改良。方法纳入2017年3～9月在中日友好医院接受诊治、术后获得明确病理诊断的117例SPN患者,其中男59例、女58例,年龄24～83(59.10±11.31)岁。分析影像学上结节最大径、所处部位、毛刺状、分叶征、钙化灶等特点,记录癌胚抗原(CEA)和细胞角蛋白19片段抗原(Cyfra21-1)等血清学检验结果,应用单因素分析模型分析临床资料与术后病理良恶性诊断的关系,应用不同SPN恶性概率预测模型绘制受试者工作特征曲线(receiver operating characteristic curve,ROC曲线),得到各模型的曲线下面积(area under curve,AUC)、灵敏度、特异度、阳性预测值、阴性预测值等并进行比较,探索模型改进方法。结果 117例患者中经病理诊断为恶性肿瘤93例(79.5%),良性肿物24例(20.5%),单因素分析发现患者年龄、结节最大径、血清CEA和Cyfra21-1水平、毛刺状、分叶征及钙化表现在良、恶性SPN之间的差异有统计学意义。应用现有的SPN恶性概率预测模型绘制ROC曲线,得出的AUC值分别为0.813±0.051(Mayo模型)、0.697±0.066(VA模型)和0.854±0.045(北京大学人民医院模型)。结论患者年龄、结节最大径、血清CEA和Cyfra21-1水平、毛刺状、分叶征及钙化表现是判断SPN良恶性的可能危险因素。北大人民医院模型对SPN良恶性鉴别的准确性较高,更适合国人。将血清学检查指标如CEA和Cyfra21-1整合入模型并调整年龄所占权重可能会提高预测模型的准确性。Objective To identify risk factors that affect the verification of malignancy in patients with solitary pulmonary nodule （SPN） and verify different prediction models for malignant probability of SPN. Methods We retrospectively analyzed the clinical data of 117 SPN patients with definite postoperative pathological diagnosis who underwent surgical procedure in China-Japan Friendship Hospital from March to September 2017. There were 59 males and 58 females aged 59.10± 11.31 years ranging from 24 to 83 years. Imaging features of the nodule including maximum diameter, location, spiculation, lobulation, calcification and serum level of CEA and Cyfra21-1 were assessed as potential risk factors. Univariate analysis was used to establish statistical correlation between risk factors and postoperative pathological diagnosis. Receiver operating characteristic （ROC） curve was drawn by different predictive models for the malignant probability of SPN to get areas under the curves （AUC）, sensitivity, specificity, positive predictive values, negative predictive values for each model. The predictive effectiveness of each model was statistically assessed subsequently. Results Among 117 patients, 93 （79.5%） were malignant and 24 （20.5%） were benign. Statistical differencewas found between the benign and malignant group in age, maximum diameter, serum level of CEA and Cyfra21-1, spiculation, lobulation and calcification of the nodules. The AUC value was 0.813±0.051 （Mayo model）, 0.697±0.066 （VA model） and 0.854±0.045 （Peking University People＇s Hospital model）, respectively. Conclusion Age, maximum diameter of the nodule, serum level of CEA and Cyfra21-1, spiculation, lobulation and calcification are potential independent risk factors associated with the malignant probability of SPN. Peking University People＇s Hospital model is of high accuracy and clinical value for patients with SPN. Adding serum index into the prediction model as a new risk factor and adjusting the weight of age in the model

关 键 词：孤立性肺结节 肺癌 预测模型 

分 类 号：R563[医药卫生—呼吸系统]