皮肌炎/多发性肌炎并发间质性肺病的CT定量分析与肺功能的相关性研究

Correlation between CT quantitative analysis and pulmonary function of interstitial lung disease in dermatomyositis/polymyositis

目的:探讨CT定量分析对皮肌炎/多发性肌炎并发间质性肺病(DM/PM-ILD)的评估价值。方法:回顾性分析2017年1月至2020年10月我院确诊的61例DM/PM-ILD患者同一时期的胸部HRCT及肺功能检查资料,同时选取60例性别、年龄及BMI相匹配且胸部HRCT无异常的健康人作为对照。使用3D Slicer软件根据阈值分割法获得CT定量指标,包括正常肺区域百分比(NL%)、磨玻璃影区域百分比(GGO%)、纤维化区域百分比(F%)及总的异常病变区域百分比(AA%)。肺功能参数包括肺总量占预计值的百分比(TLC%)、用力肺活量占预计值的百分比(FVC%)、第一秒用力呼气量占预计值的百分比(FEV1%)、一秒率(FEV1/FVC)、肺一氧化碳弥散量占预计值的百分比(DLCO%)。采用Spearman或Pearson相关分析及多元逐步线性回归分析各组数据之间的关系,使用ROC曲线确定鉴别DM/PM-ILD组与对照组的最佳CT定量指标。结果:ROC曲线分析结果显示,GGO%鉴别DM/PM-ILD患者组与对照组具有最大的AUC为0.91(95%CI:0.85~0.96),以GGO%=7.1%为阈值,敏感度为85.2%,特异度为90.0%。CT定量指标与肺功能参数除FEV1/FVC之外均具有相关性,其中F%与TLC%、FEV1%、FVC%、DLCO%相关系数均最高(r=-0.512、-0.609、-0.622、-0.519,P值均<0.001)。多元逐步线性回归分析结果显示,F%及NL%是预测TLC%的最佳指标(R2=0.303,P<0.001);F%是预测FEV1%、FVC%的最佳指标(R2=0.372、0.381,P均<0.001);F%及GGO%是预测DLCO%的最佳指标(R2=0.305,P<0.001)。结论:CT定量分析可用于DM/PM-ILD的诊断,且可评估肺功能损害情况。

Objective:To explore the value of CT quantitative analysis in the evaluation of DM/PM-ILD.Methods:The chest HRCT and pulmonary function data of 61 DM/PM-ILD cases diagnosed in our hospital from January 2017 to October 2020 were retrospectively analyzed.At the same time,60 healthy people matched by gender,age and BMI with normal chest HRCT were selected as controls.The 3D Slicer software was used to obtain quantitative indices using threshold segmentation method,including the percentage of normal lung(NL%),the percentage of ground glass opacity(GGO%),the percentage of fibrosis(F%)and the percentage of abnormal area(AA%).Pulmonary function parameters included the percentage of total lung volume to predicted value(TLC%),the percentage of forced vital capacity to predicted value(FVC%),the percentage of forced expiratory volume 1 second to predicted value(FEV1%),one second rate(FEV1/FVC)and the percentage of lung diffusion capacity for carbon monoxide to predicted value(DLCO%).The relationships between the groups were analyzed using Spearman or Pearson correlation analysis and multiple stepwise linear regression,and the ROC curve was used to determine the best quantitative CT index to distinguish the DM/PM-ILD group from the control group.Results:The ROC curve showed that the GGO%best differentiated the two groups,with a maximum AUC of 0.91(95%CI:0.85~0.96),a sensitivity of 85.2%and specificity of 90.0%,using GGO%=7.1%as the threshold.There was a correlation between CT quantitative index and pulmonary function parameters except FEV1/FVC.F%had the highest correlation coefficient with TLC%,FEV1%,FVC%,and DLCO%(r=-0.512,-0.609,-0.622,-0.519,P<0.001).Multiple stepwise linear regression showed that F%and NL%were the best indexes to predict TLC%(R 2=0.303,P<0.001).F%was the best index to predict FEV1%and FVC%(R 2=0.372,0.381,P<0.001).F%and GGO%were the best indices to predict DLCO%(R 2=0.305,P<0.001).Conclusion:CT quantitative analysis can help in the diagnosis of DM/PM-ILD and the assessment of lung function impairment.