Feasibility of Predicting Lung Volume Using Lung Field Area on Chest CT Localization Images

LIANG Xiaoyun; XIA Ziyang; ZHAO Ruomeng; CHEN Qiongzhu; FAN Liyuan; SONG Yuquan; LI Qianyi

doi:10.16458/j.cnki.1007-0893.2024.20.020

Abstract：Objective To explore the feasibility of using chest computed tomography(CT)localization images to predict lung volume.Methods From March 2022 to August 2022,chest CT images of 150 patients were restrospectively selected from the Picture Archiving and Communication System(PACS)of the First Affiliated Hospital of Guangzhou Medical University as the experimental group,and another 55 patients were selected as the external validation group.The lung field range on the chest CT localization images was delineated and the lung field area was measured.Subsequently,lung volume segmentation was performed on the chest CT scan images to calculate lung volume.The correlation between chest CT localization images and CT lung volume in the experimental group was analyzed,and a regression equation for predicting lung volume based on lung field area was established using demographic indicators of the experimental group.The results were then validated with the external validation group,and the correlation between the predicted lung volume and CT lung volume was analyzed.Results In the experimental group,the lung field area was positively correlated with CT lung volume for the total lung,right lung,and left lung(r=0.893,0.855,0.867,respectively;P<0.05).The multiple linear regression results showed that the prediction equation for lung volume(V)based on lung field area(S)was:V=-586.209+13.177×S+11.898×body mass+5.106×age-189.279×gender(male:1;female:2),with R2=0.837,adjusted R2=0.832,P<0.001,and an error rate of 8%.The lung volumes obtained from the prediction equation in both the experimental and external validation groups were positively correlated with CT lung volumes(r=0.915,0.804,respectively;P<0.001).Conclusion The lung field area on chest CT localization images has a high correlation with the lung volume measured by CT and can be used to predict lung volume.