Clinical value of low attenuation area ratio and lung volume on CT in evaluating severity of spirometric abnormality / 上海交通大学学报（医学版）

ABSTRACT

Objective • To explore the clinical value of the ratio of low attenuation areas (LAA%) and lung volume calculated according to chest CT in evaluating the severity of spirometric abnormality. Methods • The patients who underwent chest CT scan and lung function test at the same time from January 2010 to July 2014 in Ruijin Hospital, Shanghai Jiao Tong University School of Medicine were retrospectively enrolled. LAA% and lung volume were calculated by Myrian software. The correlation analysis was performed between post-bronchodiation forced expiratory volume in 1 second (FEV1) or its percentage of predicted value (FEV1%pred) and LAA% or lung volume. Multi-variable models for post-bronchodiation FEV1 and FEV1%pred were developed based on LAA%, lung volume, gender, height, and weight. The ROC curves were depicted and the diagnostic values of LAA%, lung volume, and models were compared. The cut points were chosen according to Younden indexes and specificity. Results • A total of 1 150 patients were included in summarization group. LAA% was negatively correlated with post-bronchodiation FEV1 and FEV1%pred. Lung volume was positively correlated with post-bronchodiation FEV1. The regression models of post-bronchodiation FEV1 and FEV1%pred were FEV1=-2.700+0.111×lung volume-0.216×ln (LAA%+0.1)-0.025×age+0.154×gender+0.034×height and FEV1%pred=65.582+4.014×lung volume-7.508×ln (LAA%+0.1)-10.264×gender, respectively. The regression model performed better than LAA% and lung volume in estimating the degree of post-bronchodiation FEV1 decrease. LAA% and the regression model performed better than lung volume in estimating the degree of post-bronchodiation FEV1%pred decrease. The sensitivity was 75.6% and the specificity was 90.6% in estimating post-bronchodiation FEV1%pred1.61% as the criteria. The sensitivity was 58.9% and the specificity was 81.4% in estimating post-bronchodiation FEV1<1 L by using the value of regression model<1.76. Conclusion • The regression models of post-bronchodiation FEV1 and FEV1%pred were obtained based on lung volume and LAA% on chest CT. LAA% and lung volume can roughly estimate the severity of spirometric abnormality.