Early Diagnosis of High-Risk Chronic Obstructive Pulmonary Disease Based on Quantitative High-Resolution Computed Tomography Measurements

Abstract

Wenxiu Zhang,1,* Yu Zhao,2,* Yuchi Tian,1 Xiaoyun Liang,1 Chenghao Piao2 1Institute of Research and Clinical Innovations, Neusoft Medical Systems Co, Ltd, Shanghai, People’s Republic of China; 2Radiology Department, Second Affiliated Hospital of Shenyang Medical College, Shenyang, Liaoning, People’s Republic of China*These authors contributed equally to this workCorrespondence: Chenghao Piao, First Floor, Radiology Department, Second Affiliated Hospital of Shenyang Medical College, No. 64, Qishan West Road, Huanggu District, Shenyang, Liaoning, People’s Republic of China, Tel [phone], Email [email protected]: Quantitative computed tomography (QCT) techniques, focusing on airway anatomy and emphysema, may help to detect early structural changes of COPD disease. This retrospective study aims to identify high-risk COPD participants by using QCT measurements.Patients and Methods: We enrolled 140 participants from the Second Affiliated Hospital of Shenyang Medical College who completed inspiratory high-resolution CT scans, pulmonary function tests (PFTs), and clinical characteristics recorded. They were diagnosed Non-COPD by PFT value of FEV1/FVC > 70% and divided into two groups according percentage predicted FEV1 (FEV1%), low-risk COPD group: FEV1% ≥ 95%, high-risk group: 80% < FEV1% < 95%. The QCT measurements were analyzed by the Student’s t-test (or Mann–Whitney U-test) method. Then, feature candidates were identified using the LASSO method. Meanwhile, the correlation between QCT measurements and PFTs was assessed by the Spearman rank correlation test. Furthermore, support vector machine (SVM) was performed to identify high-risk COPD participants. The performance of the models was evaluated in terms of accuracy (ACC), sensitivity (SEN), specificity (SPE), F1-score, and area under the ROC curve (AUC), with p < 0.05 considered statistically significant.Results: The SVM based on QCT measurements achieved good performance in identifying high-risk COPD patients with 85.71% of ACC, 88.34% of SEN, 84.00% of SPE, 83.33% of F1-score, and 0.93 of AUC. Further, QCT measurements integration of clinical data improved the performance with an ACC of 90.48%. The emphysema index (%LAA− 950) of left lower lung was negatively correlated with PFTs (P < 0.001). The airway anatomy indexes of lumen diameter (LD) were correlated with PFTs.Conclusion: QCT measurements combined with clinical information could provide an effective tool for an early diagnosis of high-risk COPD. The QCT indexes can be used to assess the pulmonary function status of high-risk COPD.Keywords: early diagnosis, QCT measurements, COPD, SVM

Keywords

• early diagnosis
• qct measurements
• copd
• svm