构建并验证nomogram模型预测ⅠA期肺腺癌肺泡间转移

Development and validation of a nomogram for predicting spread through air spaces in stageⅠA lung adenocarcinoma

背景和目的:肺泡间转移(spread through air spaces,STAS)是早期肺腺癌的不良预后因素,尤其是楔形切除的患者,术前预测STAS有助于选择更佳的手术方式。本研究旨在建立并验证基于术前临床和影像学特征的列线图(nomogram)来预测ⅠA期肺腺癌的STAS。方法:回顾性分析2017年1月—2018年12月在复旦大学附属中山医院接受手术治疗的595例ⅠA期肺腺癌患者,通过4%的甲醛溶液固定石蜡包埋组织切片评估STAS结果。基于术前临床资料和胸部计算机体层成像(computed tomography,CT),4种临床特征和11种影像学特征纳入分析。通过logistic回归筛选临床和影像学特征中预测STAS的独立预测因素并构建nomogram模型。通过一致性指数(concordance index,C-index)、受试者工作特征(receiver operating characteristic,ROC)曲线的曲线下面积(area under curve,AUC)、校准图来评估模型的效能。结果:在595例ⅠA期肺腺癌患者中,STAS阳性为87例(14.6%)。单因素及多因素logistic回归结果显示分叶征(OR=8.156,95%CI:1.021~65.099)、毛刺征(OR=5.258,95%CI:2.506~11.032)和实性成分占比(consolidation tumor ratio,CTR)(0.50<CTR≤0.75,OR=16.955,95%CI:3.579~80.309;0.75<CTR≤1.00,OR=20.793,95%CI:4.383~98.636)是STAS阳性的独立预测因素。基于这些预测因素建立的nomogram模型,对于STAS有较好的预测效能;C-index为0.901,AUC为0.897,且校准曲线拟合良好。结论:本研究建立并验证了预测ⅠA期肺腺癌STAS的nomogram模型。本模型操作简单,有较好的预测效能,有利于术前选择合理的手术方式。

Background and purpose:Spread through air spaces(STAS)is a poor prognostic factor for early lung adenocarcinoma,especially in patients with wedge resection.Preoperative prediction of STAS is helpful to select a better surgical treatment.This study aimed to develop and validate a nomogram based on preoperative clinical and computed tomography(CT)characteristics to predict STAS in stageⅠA lung adenocarcinoma.Methods:A total of 595 patients with stageⅠA lung adenocarcinoma who underwent surgical treatment in Zhongshan Hospital Fudan University from January 2017 to December 2018 were retrospectively analyzed.The results of STAS were evaluated by paraffin embedded tissues fixed with 4%formaldehyde solution.Based on preoperative clinical data and chest CT,4 clinical characteristics and 11 CT characteristics were analyzed.The independent predictors of STAS in clinical and CT characteristics were identified by logistic regression analysis and then used to build a nomogram.Concordance index(C-index),area under the curve(AUC)of receiver operating characteristic(ROC)and calibration plots were used to evaluate the performance of the model.Results:Among the 595 stageⅠA lung adenocarcinoma patients,87 patients(14.6%)were STAS positive.Univariate and multivariate logistic regression analyses showed that lobulation(OR=8.156,95%CI:1.021-65.099),spiculation(OR=5.258,95%CI:2.506-11.032)and consolidation tumor ratio(CTR)(0.50<CTR≤0.75;OR=16.955,95%CI:3.579-80.309;0.75<CTR≤1.00;OR=20.793,95%CI:4.383-98.636)were independent predictors of STAS positivity.The nomogram based on these predictors achieved good predictive performance for STAS with a C-index of 0.901,an AUC of 0.897 and a well-fitted calibration curve.Conclusion:This study developed and validated a nomogram for predicting STAS in stageⅠA lung adenocarcinoma.This model is simple to operate and has good predictive performance,which is conducive to the selection of reasonable surgical methods before operation.