预测肺腺癌浸润性:基于临床特征和普通CT平扫影像学特征的联合模型

Predicting the invasiveness of lung adenocarcinoma:based on combined model utilizing clinical characteristics and plain CT imaging features

目的探讨结合临床特征和普通CT平扫影像学特征的联合模型在预测直径不超过30 mm原发型肺腺癌浸润性方面的有效性。方法回顾性分析2020年1月~2023年12月在我院接受术前胸部CT平扫的107例肺结节患者。根据病理诊断,将患者分为非典型腺瘤样增生、原位腺癌、微浸润性腺癌及浸润性腺癌4组。将非典型腺瘤样增生、原位腺癌和微浸润性腺癌患者归类为非侵袭组(n=73,男性22例,年龄49.73±13.92岁),浸润性腺癌患者归类为侵袭组(n=34,男性14例,年龄57.53±12.00岁),收集患者年龄、性别以及影像资料等,比较组间差异。通过单因素和多因素分析确定预估肺腺癌浸润性的独立危险因素,并据此构建预测模型。采用ROC曲线评估模型的诊断效能。结果两组性别和年龄的差异有统计学意义(P<0.05)。两组的肺结节特征,包括结节类型、病变部位、平均直径、形状、边缘、边界、毛刺征、胸膜牵拉征、肺气肿、空气支气管征、空泡征及血管类型的差异均有统计学意义(P<0.05)。结节平均直径和血管类型被确定为肺腺癌侵袭性的独立危险因素(P<0.05)。ROC曲线分析显示,联合诊断肺腺癌侵袭性的AUC为0.891(95%CI:0.816~0.956)。在最佳约登指数0.514的条件下,结节平均直径截断值为9.75 mm,模型的敏感度为0.706,特异度为0.808。结论肺结节的普通CT平扫影像学特征对于术前评估肺腺癌的侵袭性具有重要意义,结节的平均直径和血管类型是预测肺腺癌侵袭性的独立危险因素。

Objective To evaluate the utility of a combined model integrating clinical characteristics and plain CT imaging features in predicting the invasiveness of primary lung adenocarcinoma with diameters less than 30 mm.Methods A retrospective analysis was conducted on 107 patients with pulmonary nodules who underwent preoperative chest CT scans at our hospital from January 2020 to December 2023.Patients were categorized into four groups based on pathological diagnosis:atypical adenomatous hyperplasia,adenocarcinoma in situ,minimally invasive adenocarcinoma,and invasive adenocarcinoma.Among them,73 patients with atypical adenomatous hyperplasia,adenocarcinoma in situ,and minimally invasive adenocarcinoma formed the non-invasive group(including 22 males,mean age 49.73±13.92),while 34 patients with invasive adenocarcinoma constituted the invasive group(including 14 males,mean age 57.53±12.00).Clinical data including age,gender,and imaging details were collected and compared between the two groups.Univariate and multivariate analyses were performed to identify independent predictors of lung adenocarcinoma invasiveness,leading to the development of a predictive model.Model performance was assessed using ROC curve analysis.Results Significant differences were observed between the two groups in terms of gender and age(P<0.05).Various CT imaging features such as nodule type,location,average diameter,shape,margin,spiculation,pleural indentation,presence of emphysema,air bronchogram,vacuole sign,and vascular type exhibited statistically significant differences between the groups(P<0.05).Average nodule diameter and vascular type were identified as independent predictors of lung adenocarcinoma invasiveness(P<0.05).The ROC curve analysis showed an AUC of 0.891(95%CI:0.816-0.956)for the combined model in diagnosing lung adenocarcinoma invasiveness.Using the optimal Youden index of 0.514,the cutoff value for average nodule diameter was determined to be 9.75 mm,with a sensitivity of 0.706 and specificity of 0.808.Conclusion Plain CT imaging features of pulmonary nodules play a crucial role in preoperatively assessing the invasiveness of lung adenocarcinoma.Average nodule diameter and vascular type independently predict lung adenocarcinoma invasiveness.