A Comparative Evaluation of Computed Tomography Images for the Classification of Spirometric Severity of the Chronic Obstructive Pulmonary Disease with Deep Learning

Abstract

Recently, deep learning applications in medical imaging have been widely applied. However, whether it is sufficient to simply input the entire image or whether it is necessary to preprocess the setting of the supervised image has not been sufficiently studied. This study aimed to create a classifier trained with and without preprocessing for the Global Initiative for Chronic Obstructive Lung Disease (GOLD) classification using CT images and to evaluate the classification accuracy of the GOLD classification by confusion matrix. According to former GOLD 0, GOLD 1, GOLD 2, and GOLD 3 or 4, eighty patients were divided into four groups (n = 20). The classification models were created by the transfer learning of the ResNet50 network architecture. The created models were evaluated by confusion matrix and AUC. Moreover, the rearranged confusion matrix for former stages 0 and ≥1 was evaluated by the same procedure. The AUCs of original and threshold images for the four-class analysis were 0.61 ± 0.13 and 0.64 ± 0.10, respectively, and the AUCs for the two classifications of former GOLD 0 and GOLD ≥1 were 0.64 ± 0.06 and 0.68 ± 0.12, respectively. In the two-class classification by threshold image, recall and precision were over 0.8 in GOLD ≥1, and in the McNemar–Bowker test, there was some symmetry. The results suggest that the preprocessed threshold image can be possibly used as a screening tool for GOLD classification without pulmonary function tests, rather than inputting the normal image into the convolutional neural network (CNN) for CT image learning.

Diagnostic efficacy of visual subtypes and low attenuation area based on HRCT in the diagnosis of COPD

“…Previous studies have shown that quantitative CT analysis is correlated with PFT results and pathology-based quantification of emphysema [ 21 , 27 , 28 ]. Moreover, there are numerous reports using quantitative CT imaging in COPD patients, demonstrating its wide use as a simple imaging method to obtain information on the whole lung [ 29 , 30 ]. Similarly, based on our research results, we found that LAA and visual subtypes were crucial in identifying COPD patients and assessing the severity of COPD patients.…”

A Novel Melspectrogram Snippet Representation Learning Framework for Severity Detection of Chronic Obstructive Pulmonary Diseases

“…In recent years, deep learning techniques have also been employed in COPD severity classification [26], [27], [17], [18], [19]. Sugimori et al [28] had used CT scan images to classify five class COPD severities using convolutional neural network (CNN). They have used a ResNet50 [29] architecture as CNN backbone follwed by dense classifier and achieved a classification accuracy of 44% for five class COPD severity classification.…”

“…Table X, represents the overall comparative study by considering different datasets, methods, classifiers, and performance measures. It can be observed from the table that existing works exploited the use of subjective measurements including symptoms, medical assessments, physical examinations, questionnaire responses [5] and different biomedical input modalities including CT scans [55], [28], [27], PRM method [26], spirometric measures [5], electrocardiogram (ECG) [20] and tracheal sounds [25] for COPD severity classification. However, the proposed framework exploits the potential of time-frequency melspectrogram representation of lung sound signals and a pretrained audio classification network i.e., YAMNet, which help to achieve the highest classification performance both in binary classification and multi-class classification for COPD severity as compared with existing works.…”

Predicting early-onset COPD risk in adults aged 20–50 using electronic health records and machine learning

“…Extensive research has demonstrated the accuracy of ML models in predicting the development of COPD based on genetic information and electronic medical records data ( Peng et al, 2020 ; Zhang et al, 2022a ). In particular, studies by Peng et al (2020) and Cosentino et al (2023) showcased the effectiveness of ML algorithms in detecting early signs of COPD from chest CT scans and in predicting the case-control status of COPD from raw high-dimensional spirograms, respectively ( Sugimori et al, 2021 ). Makimoto’s study, focusing on CT-based quantitative models to predict severe exacerbations of COPD ( Makimoto & Kirby, 2023 ).…”