低剂量CT视觉亚型联合定量参数预测慢性阻塞性肺疾病的价值

Value of low-dose spiral CT visual subtypes and quantitative parameters in predicting chronic obstructive pulmonary disease

目的探讨低剂量CT视觉亚型联合定量参数预测慢性阻塞性肺疾病(COPD)的价值。方法回顾性分析2021年1月至2022年3月在厦门大学附属第一医院因咳嗽、咳痰或呼吸短促/困难就诊的吸烟或既往吸烟172例患者的临床及影像资料。所有患者均接受肺功能检测及CT检查。根据全球慢性阻塞性肺疾病倡议诊断标准,172例患者分为非COPD组(79例)和COPD组(93例)。分析患者的影像视觉亚型分级并测量CT图像上肺密度低于-950 HU的百分比(LAA-950)、支气管壁厚度(WT)、支气管腔面积百分比(WA%)。2组间视觉亚型分级及CT参数的比较使用Mann-Whitney U检验,将差异有统计学意义的指标纳入多因素logistic回归分析,筛选出预测COPD的独立危险因素。最终建立临床特征联合CT影像特征的logistic预测模型,并采用受试者操作特征曲线的曲线下面积(AUC)分析模型预测COPD的效能。结果COPD组患者年龄及咳嗽、咳痰比例高于非COPD组,体重指数低于非COPD组(P<0.05)。在影像学特征中,COPD组与非COPD组间的影像视觉亚型、LAA-950、WT、WA%差异均具有统计学意义(P<0.001)。多因素logistic回归分析中,年龄(OR=1.06,95%CI 1.02~1.10,P=0.002)、咳痰(OR=2.86,95%CI 1.37~5.97,P=0.005)、影像视觉亚型(OR=1.73,95%CI 1.30~2.30,P<0.001)是预测COPD的独立危险因素。最终以年龄、体重指数、咳嗽/咳痰、影像视觉亚型分级、LAA-950、WA%联合构建logistic预测模型,其预测COPD的AUC为0.903,灵敏度为82.8%,特异度为81.0%。结论基于低剂量CT的LAA-950、视觉亚型分级和WA%联合临床特征诊断COPD具有良好的临床价值。

Objective To explore the value of low dose CT visual subtypes and quantitative parameters in predicting chronic obstructive pulmonary disease(COPD).Methods The clinical and imaging data of 172 patients with smoking or previous smoking who visited the First Affiliated Hospital of Xiamen University from January 2021 to March 2022 due to cough,expectoration or shortness of breath/difficulty were analyzed retrospectively.All patients underwent pulmonary function testing and CT examinations.According to the diagnostic criteria of the Global Initiative for COPD,172 patients were divided into a non COPD group(79 cases)and a COPD group(93 cases).The visual subtypes grading of the patient′s images were analyzed and the percentage of low attenuation area(LAA)-950,bronchial wall thickness(WT),and lumen area percentage(WA%)on CT images were measured.Visual subtype grading and CT parameters were compared between the 2 groups using the Mann-Whitney U test,and statistically significant differences were included in a multifactorial logistic regression analysis to screen for independent risk factors predicting COPD.Finally,a logistic prediction model of clinical features combined with CT imaging features was constructed,and the area under the curve(AUC)of the receiver operating characteristic curves was used to analyze the model to predict the efficacy of COPD.Results The age and proportion of cough and sputum in the COPD group were higher than those in the non COPD group,and the body mass index was lower than that in the non COPD group(P<0.05).In terms of imaging features,there were statistically significant differences in visual subtypes grading,LAA-950,WT,and WA%between the COPD group and the non COPD group(P<0.001).Logistic regression analysis showed age(OR=1.06,95%CI 1.02—1.10,P=0.002),expectoration(OR=2.86,95%CI 1.37—5.97,P=0.005),and visual subtype grading(OR=1.73,95%CI 1.30—2.30,P<0.001)were independent risk factors for predicting COPD.The final logistic prediction model was jointly constructed with age,body mass index,cough/sputum,imaging visual subtype grading,LAA-950,and WA%,and its AUC for predicting COPD was 0.903 with a sensitivity of 82.8%and a specificity of 81.0%.Conclusion The combined clinical features of LAA-950,visual subtype grading and WA%based on low-dose CT show good clinical value in diagnosing COPD.