作者：王茹，傅宇飞，纵瑞龙，师毅冰，孙境熙

单位：

221000江苏省徐州市中心医院影像科

Units：

Department of Medical Imaging, Xuzhou Central Hospital, Xuzhou 221000, China

关键词：

孤立性肺结节；影响因素分析；预测模型

Keywords：

Solitary pulmonary nodule; Root cause analysis; Predictive model

CLC：

R 521.6

DOI：

10.12114/j.issn.1008-5971.2023.00.173

Funds：

2022年度徐州医科大学附属医院科技发展基金面上项目（XYFM202207）

摘要：

目的　探讨恶性孤立性肺小结节（SSPN）的影响因素，构建其预测模型并进行验证。方法　回顾性选取2015年6月至2020年5月就诊于徐州市中心医院的SSPN患者214例作为训练组，同时选取2020年6月至2022年6月就诊于该院的SSPN患者94例作为验证组。以肺活检结果诊断SSPN的良恶性。比较训练组良性SSPN患者与恶性SSPN患者的临床资料｛性别、年龄、吸烟史、肿瘤史、肿瘤标志物〔神经元特异性烯醇化酶（NSE）、癌胚抗原（CEA）、鳞癌相关抗原（SCC）及细胞角质蛋白19片段抗原21-1（CYFRA21-1）〕｝与影像学资料〔结节位置、结节侧别、CT征象（分叶征、毛刺征、钙化、胸膜凹陷征、空气支气管征）、有无纵隔淋巴结以及结节直径〕。采用多因素Logistic回归分析探讨恶性SSPN的影响因素，并构建其预测模型。采用ROC曲线评价本研究预测模型、王欣等预测模型、梅奥模型对验证组恶性SSPN的预测价值。结果　恶性SSPN患者年龄、结节直径大于良性SSPN患者，CEA、SCC及有分叶征、毛刺征、胸膜凹陷征、空气支气管征者占比高于良性SSPN患者，有钙化者占比低于良性SSPN患者（P＜0.05）。多因素Logistic回归分析结果显示，年龄增长、CEA升高、胸膜凹陷征、空气支气管征是恶性SSPN的危险因素（P＜0.05）。根据上述危险因素构建预测模型：P=ex/（1+ex），其中x=-8.309+0.118×年龄+0.232×CEA+1.434×胸膜凹陷征+0.882×空气支气管征。ROC曲线分析结果显示，本研究预测模型、王欣等预测模型、梅奥模型预测恶性SSPN的AUC分别为0.886〔95%CI（0.822，0.951）〕、0.773〔95%CI（0.676，0.869）〕、0.781〔95%CI（0.676，0.887）〕。结论　年龄增长、CEA升高、胸膜凹陷征、空气支气管征是恶性小SPN的危险因素，而根据上述危险因素构建的预测模型对恶性SSPN具有较好的预测价值。

Abstract：

　Objective To explore the influencing factors of malignant small solitary pulmonary nodules (SSPN) ,construct and validate its predictive model. Methods A total of 214 patients with SSPN admitted to Xuzhou Central Hospitalfrom June 2015 to May 2020 were retrospectively selected as the training group, and 94 patients with SSPN admitted to XuzhouCentral Hospital from June 2020 to June 2022 were selected as the validation group. The benign and malignant SSPN wereconfirmed by lung biopsy results. The clinical data {sex, age, smoking history, tumor history, tumor markers [neuron specificenolase (NSE) , carcinoembryonic antigen (CEA) , squamous carcinoma-associated antigen (SCC) , and cyto-keratin 19 fragmentantigen 21-1 (CYFRA21-1) ] } and imaging data [nodule location, nodule side, CT signs (lobe sign, burr sign, calcification,pleural indentation sign, air bronchial sign) , with or without mediastinal lymph node and nodule diameter] were compared betweenpatients with benign and malignant SSPN in the training group. Multivariate Logistic regression analysis was used to explore theinfluencing factors of malignant SSPN and to construct its prediction model. ROC curve was used to evaluate the predictive valueof this study's prediction model, Wang Xin et al. 's prediction model and Mayo model for malignant SSPN in the validation group.Results The age and nodule diameter of malignant SSPN patients were larger than those of benign SSPN patients, and CEA,SCC, and the proportion of lobulated sign, burr sign, pleural sag sign and air bronchial sign were higher than those of benignSSPN patients, and the proportion of calcification was lower than that of benign SSPN patients (P < 0.05) . Multivariate Logisticregression analysis showed that increased age, increased CEA, pleural depression sign and air bronchial sign were risk factorsfor malignant SSPN (P < 0.05) . A prediction model was constructed based on the above risk factors: P=ex/ (1+ex) , where x=-8.309+0.118×age+0.232×CEA+1.434×pleural depression sign+0.882×air bronchial sign. ROC curve analysis results showedthat the AUC of the prediction model in this study, the prediction model of Wang Xin et al., and the Mayo model for malignantSSPN were 0.886 [95%CI (0.822, 0.951) ] , 0.773 [95%CI (0.676, 0.869) ] , 0.781 [95%CI (0.676, 0.887) ] , respectively.Conclusion Increased age, elevated CEA, pleural indentation sign, and air bronchogram sign are independent risk factors formalignant SSPN. The model constructed based on the above risk factors has good predictive value for malignant SSPN.

ReferenceList：