外周ⅠA期肺小腺癌(≤2 cm)气腔播散的相关因素分析及诺莫图模型的建立

Analysis of factors associated with spread through air spaces(STAS) of small adenocarcinomas(≤2 cm) in peripheral stage ⅠA lungs and modeling of nomograms

目的探究外周ⅠA期肺小腺癌(≤2 cm)气腔播散(spread through air spaces, STAS)与临床及CT形态学特征等相关因素之间的关系, 并构建诺莫图模型。方法收集2017年至2022年在南通大学附属医院行肺部外科手术且术后病理诊断为外周ⅠA期肺小腺癌患者的相关临床、病理及影像学资料, 其中, 2017年至2021年符合纳入标准的病例作为训练组, 2022年符合纳入标准的病例作为验证组。采用单因素分析及多因素logistic回归分析研究外周ⅠA期肺小腺癌发生STAS的独立危险因素, 在此基础上构建诺莫图预测模型, 并用受试者工作特征曲线(ROC)、校正模型等对该模型进行性能评价。结果共纳入430例符合标准的患者, 其中, 训练组351例(STAS阳性109例, STAS阴性242例), 验证组79例(STAS阳性23例, STAS阴性56例)。单因素分析显示:两组患者在年龄(>58岁)、性别、吸烟史、肿瘤位置(胸膜下、非胸膜下)、胸膜牵拉、结节类型、结节最大径、实性成分最大径、肿瘤实性占比(consolidation tumor ratio, CTR)、分叶征、毛刺征、支气管截断征、血管征(包括结节内/周围血管增粗、扭曲)、卫星灶、磨玻璃带征方面差异有统计学意义(P<0.05)。多因素logistic回归分析结果显示:CTR(OR=4.98, P<0.001)、分叶征(OR=4.07, P=0.013)、毛刺征(OR=3.66, P<0.001)、卫星灶(OR=3.56, P=0.009)是STAS发生的独立危险因素。应用上述因素构建诺莫图模型并对模型进行验证, 结果显示:经诺莫图预测模型绘制ROC曲线, 训练组的ROC曲线下面积(AUC)0.840(灵敏度0.835, 特异度0.734), 验证组的AUC 0.852(灵敏度0.786, 特异度0.783), 训练组和验证组校正曲线与理想曲线有较好的重合度。结论 CTR、分叶征、毛刺征、卫星灶是STAS的独立危险因素, 本研究构建的诺莫图模型具有较好的预测价值。

Objective To investigate the relationship between spread through air spaces(STAS)of peripheral stageⅠA small adenocarcinoma of the lung(≤2 cm)and related factors such as clinical and CT morphological features,and to construct a nomogram model.Methods Relevant clinical,pathological and imaging data of patients who underwent lung surgery and were diagnosed as peripheral stageⅠA small lung adenocarcinoma by postoperative pathology in the Affiliated Hospital of Nantong University from 2017 to 2022 were collected,of which cases that met the inclusion criteria from 2017 to 2021 served as the training group,and those that met the inclusion criteria in 2022 served as the validation group.The independent risk factors for the occurrence of STAS in peripheral stageⅠA lung small adenocarcinoma were investigated by using univariate analysis and multifactorial logistic regression analysis,based on which a nomogram prediction model was constructed,and the subjects were analyzed by using the receiver operating characteristic curve(ROC),correction model,etc.were used to evaluate the model.Results A total of 430 patients who met the criteria were included,including 351 patients in the training group(109 STAS-positive and 242 STAS-negative)and 79 patients in the validation group(23 STAS-positive and 56 STAS-negative).Univariate analysis showed that the patients in the two groups showed a significant difference in age(>58 years old),gender,smoking history,tumor location(subpleural,non-subpleural),pleural pull,nodule type,nodule maximal diameter,solid component maximal diameter,consolidation tumor ratio(CTR),lobulation sign,burr sign,bronchial truncation sign,vascular sign(includes thickening and distortion of blood vessels in/around the nodes),satellite lesions,and ground-glass band sign were statistically significant(P<0.05).The results of multifactorial logistic regression analysis showed that CTR(OR=4.98,P<0.001),lobulation sign(OR=4.07,P=0.013),burr sign(OR=3.66,P<0.001),and satellite lesions(OR=3.56,P=0.009)were the independent risk factors for the occurrence of STAS.Applying the above factors to construct the nomogram model and validate the model,the results showed that the ROC curve was plotted by the nomogram prediction model,and the area under the ROC curve(AUC)of the training set was 0.840(sensitivity 0.835,specificity 0.734),and the validation set had an AUC value of 0.852(sensitivity 0.786,specificity 0.783),and the training set and validation set calibration curves have good overlap with the ideal curve.Conclusion CTR,lobular sign,burr sign,and satellite lesions are independent risk factors for STAS,and the nomogram model constructed in this study has good predictive value.