【摘要】 目的探讨手术切除的肺腺癌脉管癌栓及神经侵犯与CT特征之间的关系,为判断预后及调整临床治疗提供指导。方法回顾性分析143例手术切除的肺腺癌患者术前胸部增强CT图像,共评估16个CT分类变量及7个连续变量。采用ROC曲线分析CT参数对肺腺癌脉管癌栓及神经侵犯的诊断效能。通过拟合建立最终多参数模型提高诊断效能。结果脉管癌栓多发生于实性成分,呈分叶状及有钙化的肿瘤中(P均<0.05);神经侵犯多见于伴支气管充气征的肿瘤中(P<0.05)。平扫CT值对脉管癌栓及神经侵犯的诊断效能最高。多参数模型预测脉管癌栓和神经侵犯的AUC值(曲线下面积)分别为0.79、0.88,准确度分别76.22%、81.82%,较单因素分析诊断效能更高。结论增强CT可术前预测肺腺癌脉管癌栓及神经侵犯,多参数模型预测脉管癌栓和神经侵犯的诊断效能较单因素分析更高。平扫CT值是肺腺癌神经侵犯的独立危险因素,可为临床治疗提供相应参考价值。更多还原

【Abstract】 Objective To investigate the value of computed tomography(CT) features in surgically resected lung adenocarcinomas in predicting lymphovascular and neural invasion, which provides guidance for judging prognosis and adjusting clinical treatment. Methods Preoperative chest contrast-enhanced(CE) CT images were retrospectively evaluated in 143 surgically resected lung adenocarcinomas, including a total of 16 CT categorical variables and 7 continuous variables. The receiver operating characteristic(ROC) curve analysis was employed to compare the diagnostic efficiency of CT parameters for lymphovascular and neural invasion in lung adenocarcinoma. The final multi-parameter model was established by fitting to improve the diagnostic efficiency. Results Lymphovascular invasion was found more frequently in tumors with solid composition, those were lobulated and with calcification(all P<0.05); neural invasion was more common in tumors with air bronchogram(P<0.05). The diagnostic efficiency of non-CE CT attenuation values was highest in predicting lymphovascular and neural invasion. And the multi-parameter models were employed to predict lymphovascular and neural invasion respectively, the areas under the curve(AUC) were 0.79 and 0.88, and the accuracies were 76.22% and 81.82%, which were higher than the single factor analysis. Conclusion CE CT can be used to predict lymphovascular and neural invasion in lung adenocarcinoma. The diagnostic efficiency of multi-parameter model in predicting lymphovascular and neural invasion is higher than that of the single factor analysis. Non-CE CT attenuation value is an independent risk factor of neural invasion of lung adenocarcinoma, which can provide corresponding reference for clinical treatment.更多还原