Impaired lung deformation during expiration in chronic obstructive pulmonary disease using four-dimensional dynamic-ventilation CT

Introduction

Chronic obstructive pulmonary disease (COPD) is a heterogeneous disorder characterized by persistent respiratory symptoms and airflow limitation due to a combination of small airway disease and parenchymal destruction (emphysema) (1,2). While spirometry, specifically the ratio of forced expiratory volume in 1 s to forced vital capacity (FEV₁/FVC) measured after bronchodilator administration, remains the clinical cornerstone for diagnosis and staging (3), it provides only a global, integrated measure of lung function. This global metric cannot resolve the regional heterogeneity in mechanical impairment—a fundamental aspect of COPD pathophysiology that involves disparate contributions from loss of elastic recoil, airway remodeling, and gas trapping across different lung regions. Consequently, there is a recognized need for imaging biomarkers that can quantify this regional dysfunction to better understand disease progression and phenotypic variability (3).

Four-dimensional dynamic-ventilation computer tomography (4D-CT) (three spatial dimensions plus time) has emerged as a powerful tool for capturing lung motion in vivo throughout the respiratory cycle (4,5). By applying deformable image registration techniques, this modality allows for the voxel-wise calculation of biomechanical parameters, most notably principal strain. Principal strain, a tensor-derived measure of normal (tensile/compressive) deformation, quantifies the local magnitude of parenchymal expansion and contraction. In the context of lung mechanics, regional strain is directly influenced by parenchymal compliance and tethering forces. In healthy lungs, the coordinated deformation of elastic parenchyma ensures efficient ventilation. In COPD, however, the characteristic loss of elastic recoil due to emphysematous destruction, coupled with small airway obstruction and remodeling, disrupts this uniform mechanical behavior. Consequently, the lung’s ability to generate normal, homogeneous deformation during expiration is impaired. By capturing the diminution in both the magnitude and spatial uniformity of parenchymal deformation, these parameters may provide a direct correlation to the progressive decline in lung function that defines COPD severity. Therefore, strain parameters such as the maximum principal strain (PS_max) and mean principal strain (PS_mean) offer a promising, quantitative lens through which the integrity and homogeneity of regional lung tissue mechanics can be assessed (4,5).

Preliminary studies using 4D-CT strain analysis in smokers and patients with mild-to-moderate COPD have demonstrated reduced deformation in affected individuals compared to healthy controls (4,5). Intriguingly, one study suggested that parameters from the initial 2-s of expiration might be particularly sensitive (4). However, the natural history of COPD involves a progression towards severe and very severe airflow limitation [the Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages III–IV], where expiratory flow becomes profoundly prolonged and heterogeneous. The biomechanical behavior of the lung across this full severity spectrum, and particularly during the entirety of the prolonged expiration in advanced disease, remains poorly characterized. It is unclear whether strain analysis of the brief, effort-dependent initial exhalation or the complete, flow-limited expiratory phase provides a more robust correlation of overall mechanical impairment.

We hypothesize that 4D-CT-derived principal strain parameters can quantitatively reflect the progressive deterioration of lung tissue compliance and increasing mechanical heterogeneity in COPD. Furthermore, we hypothesize that in severe disease, strain metrics calculated over the entire expiratory phase demonstrate a stronger correlation with global disease severity than those from the initial 2-s phase, as the full expiratory trajectory better captures the dominant, abnormal resistive mechanics.

Therefore, the primary objectives of this study were: (I) to quantitatively evaluate and compare lung deformation patterns across the full spectrum of COPD severity, and (II) to investigate the association of expiration-phase strain parameters with severe airflow limitation, specifically comparing the utility of whole-expiration versus initial rapid-exhalation metrics. This work aims to advance the biomechanical understanding of COPD by providing a quantitative imaging framework for assessing regional parenchymal dysfunction. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1616/rc).

Methods

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of China-Japan Friendship Hospital (approval No. 2023-KY-046). Written informed consent was obtained from all participants.

Subjects

The data for this study were sourced from China-Japan Friendship Hospital. Between August 2020 and May 2022, 74 COPD patients underwent spirometry and dynamic-ventilation computer tomography (CT) and were recruited in this study. COPD diagnosis was established based on post-bronchodilator FEV₁/FVC ratio <0.70, with disease severity subsequently classified according to GOLD criteria (6). Inclusion criteria were: (I) ≥18 years old; (II) Absence of acute respiratory infections or other significant pulmonary abnormalities (e.g., interstitial lung diseases) that might compromise quantitative CT measurements; (III) no history of thoracic surgery; and (IV) a time interval of ≤2 weeks between CT examinations and spirometry. Exclusion criteria were: (I) inability to comply with voice guidance during dynamic-ventilation CT scanning (n=8); (II) measurement items of spirometry were incomplete (n=2); and (III) poor image quality that cannot meet strain analysis (n=3). Totally, 61 COPD subjects (42 males and 19 females) with a mean age of 62.74 years (range, 33–85 years), including 29 GOLD I, 18 GOLD II, 6 GOLD III, and 8 GOLD IV, were enrolled in the final analysis. In addition, part of the subjects’ data utilized in this study were previously analyzed with a different objective (4). The related clinical characteristics of the subjects enrolled were illustrated in Table 1.

Patient preparation and CT scans protocol

Each subject received comprehensive CT imaging consisting of two protocols: (I) standard low-dose chest CT (static CT); and (II) dynamic-ventilation CT (dynamic CT), both performed in the supine position on a 320-detector row CT system (Aquilion ONE, Canon Medical Systems, Otawara, Japan). The complete acquisition parameters for both CT protocols are detailed in Appendix 1.

Subjects were trained in maintaining consistent deep breathing patterns before dynamic CT scanning and received a voice prompt during the examination to complete the whole scan. Sixteen of the subjects underwent dynamic CT, also with the assistance of a respiratory monitoring device (patent No. ZL 202320361612.8). The working mechanism of the respiratory monitoring device was described in Appendix 1.

Image analysis

Static CT: the percent low attenuation volume [threshold: <−950 Hounsfield units (HU)] of the entire lung, defined as the low-density (LD) index, along with total lung volume (TLV), was automatically quantified using dedicated commercial software (Lung Density Analysis, Canon Medical Systems).

Dynamic CT: the lung volume changes during the entire respiratory cycle automatically measured by the same commercial software made a lung volume-frame curve (Figure 1). The expiratory phase was operationally defined as the temporal interval extending from the peak inspiratory frame (corresponding to the first expiratory frame; maximum lung volume) to the peak expiratory frame (representing the minimum lung volume frame). Strain analysis was performed using expiratory phase data derived from dynamic CT imaging, processed with dedicated software (MicroVec V3.6.2, MicroVec Pte Ltd., Beijing, China), and detailed processing steps were summarized in Appendix 1.

Figure 1 A 75-year-old male smoker with COPD. (A) Coronal CT image with emphysema area (<−950 HU) marked in blue (LD index =45.1%). Emphysema area was more obvious in the upper-middle part of the whole lung (an inset, red dotted line level). (B) Lung volume-frame curve. (C) Serial changes of PS_max in pseudo-color fusion image. The color bar provides a quantitative scale for the non-normalized strain values. Red color indicates regions of high principal strain magnitude, while blue color indicates low strain magnitude. COPD, chronic obstructive pulmonary disease; HU, Hounsfield units; LD, low-density; PS_max, maximum principal strain.

The strain-related parameters comprised: (I) PS_max, defined as the peak strain magnitude in the pixel displacement field; (II) PS_mean, defined as the spatial average of the principal strain magnitude across all pixels/voxels within the lung mask at a given time point, representing the global average deformation intensity; and (III) maximum displacement speed (Speed_max), corresponding to the maximum pixel displacement between two consecutive temporal images. Due to the 4D-CT data being acquired at 10 Hz, the time interval between successive frames is 0.1 s. The unit has not been converted into physical dimensions and is expressed as: pixel/0.1 s. The reported principal strain parameters (PS_max and PS_mean) represent normal strain components, quantifying tensile/compressive deformation along the primary axes. These parameters were derived from both the whole expiratory phase and the initial 2-s expiratory phase for comparative analysis.

The first expiratory frame served as the reference baseline, with its strain parameters and lung volume establishing reference values. For subsequent frames (2^nd to 10^th and 2^nd to peak expiration), all strain parameters were normalized by corresponding volume changes to account for expiratory effort variability. Cumulative strain metrics were then calculated for two intervals: (I) the first 2-s expiratory phase (frames 2–10, denoted as PS_max2s, PS_mean2s, and Speed_max2s); and (II) the whole expiratory phase (frames 2–peak, denoted as PS_max-all, PS_mean-all, and Speed_max-all).

Spirometry

All participants underwent standardized spirometry (MasterScreen PFT, Vyaire Medical GmbH, Hoechberg, Germany) in the seated position. Tests were administered by certified pulmonary function technicians according to the American Thoracic Society (ATS)/European Respiratory Society (ERS) guidelines (6). For each participant, a minimum of three acceptable and reproducible forced expiratory maneuvers were performed. The highest values of FVC and FEV₁ obtained from these maneuvers were used for analysis. Other measured parameters included peak expiratory flow (PEF), maximal mid-expiratory flows at 75%, 50%, and 25% of FVC (MMEF75%, MMEF50%, MMEF25%), and the mean mid-expiratory flow (MMEF25–75%). Complete spirometric data are presented in Table 1.

Statistical analysis

Statistical analyses were performed using SPSS (version 17.0 for Windows, SPSS Inc., Chicago, IL, USA). The normality of continuous variables was assessed using the Kolmogorov-Smirnov test, supplemented by visual inspection of graphical data distributions. Normally distributed data are presented as mean ± standard deviation (SD), while non-normally distributed data are summarized as median (25^th–75^th percentile). Categorical variables are expressed as frequencies and percentages.

Differences in clinical and spirometric parameters across GOLD stages I (mild), II (moderate), and III–IV (severe) were evaluated using the Kruskal-Wallis test or Pearson’s Chi-squared test, as appropriate. All strain-derived parameters (PS_max, PS_mean, and Speed_max) were adjusted for lung volume changes and analyzed separately for (I) the whole expiratory phase and (II) the first 2-s expiratory phase. Intergroup differences in strain parameters across GOLD stages were assessed using the Kruskal-Wallis test. In cases of statistical significance, post-hoc pairwise comparisons were conducted using Dunn’s test with Bonferroni correction, applying an adjusted significance level of α=0.0167.

Correlations between strain-related parameters and the degree of airflow limitation were examined using Spearman’s rank correlation analysis. A two-sided P value <0.05 was considered statistically significant for all tests, unless otherwise specified for post-hoc comparisons.

Results

Based on the severity of airflow limitation, COPD patients were stratified into three subgroups according to GOLD criteria (6): mild (GOLD I), moderate (GOLD II), and severe (GOLD III–IV). Clinical and Spirometric parameters are comprehensively summarized in Table 1. The average interval time between spirometry and CT scans was 5.4±2.1 days.

CT parameters among COPD subjects with different severity of airflow limitation

Both LD index and TLV demonstrated significant progressive increases with worsening airflow limitation (LD index: P<0.001; TLV: P<0.001). In contrast, all strain-derived parameters exhibited inverse correlations with disease severity (Table 2, Figure 2). To illustrate the magnitude of change, the median values of PS_max-all decreased by up to 74.69% in severe (GOLD III–IV) patients compared to the mild (GOLD I) group (see Table S1 for detailed percentage changes across all groups). The adjusted strain parameters, especially the ones derived from the whole expiration phase, showed significant correlations with spirometric indices, including FEV₁, FEV₁/FVC, and MMEF (ρ=0.271–0.520, P<0.05; see Table S2).

Figure 2 Both PS_max-all (left) and PS_max2s (right) showed a decreased trend with progressive airflow limitation. Significance markers from post-hoc Dunn’s tests are shown, and the significance level was set at α=0.0167. The horizontal line in the boxes represents the median, the bottom and top of the boxes represent the 25^th and 75^th percentiles, respectively. The bars represent the upper adjacent value (75^th percentile plus 1.5 times the interquartile range) and the lower adjacent value (25^th percentile minus 1.5 times the interquartile range), and the stars/circles represent outliers. GOLD, Global Initiative for Chronic Obstructive Lung Disease; PS_max-all, maximum principal strain in the whole expiratory phase; PS_max2s, maximum principal strain in the first 2-s expiratory phase.

Correlations between CT parameters and GOLD classification

Both dynamic and static CT parameters showed moderate correlations with GOLD classification. In general, correlation coefficients with GOLD classification for the strain-related parameters derived from the whole expiration phase were higher than those derived from the first 2-s of expiration phase in dynamic CT, and the highest correlation coefficient in this study was provided by PS_max-all obtained from the whole expiration phase (ρ=−0.732, P<0.001) (Table 3).

Discussion

In this study, we demonstrated that lung deformation during expiration, quantified by 4D-CT-derived principal strain parameters, progressively declines with increasing severity of airflow limitation in COPD. The substantial reductions (e.g., exceeding 70% for PS_max-all in severe COPD) highlighted in Table S1 underscore the profound deterioration of regional lung mechanics with disease progression. Crucially, strain metrics calculated over the entire expiratory phase exhibited stronger correlations with disease severity than those derived from the initial 2-s period. This finding suggests that the whole expiratory effort provides a more comprehensive biomechanical signature of the disease, particularly in advanced stages where prolonged and heterogeneous flow limitation dominates.

In mild COPD, expiration is relatively rapid and complete, so the initial 2 s may capture most of the mechanical event. In severe COPD (GOLD III–IV), expiration is prolonged and flow-limited due to dynamic airway collapse and increased lung compliance. The initial effort-driven 2 s becomes less representative of the overall, obstructed expiratory mechanics, which are dominated by slow, heterogeneous emptying throughout the entire prolonged phase. Therefore, whole-expiration strain metrics, which integrate this entire abnormal process, correlate more strongly with the global measure of impairment (GOLD stage).

From a biomechanical perspective, the decrease in strain parameters’ value likely reflects the loss of effective parenchymal compliance and increased mechanical heterogeneity resulting from emphysematous destruction and small airway disease. The superior performance of whole-expiration parameters underscores a shift in dominant resistive mechanics in severe COPD, where the entire expiratory trajectory becomes a more integrated biomarker of global mechanical dysfunction.

An intriguing observation in this study was the presence of negative values for volume-adjusted strain in some patients, particularly those with severe (GOLD III–IV) airflow limitation. This finding, which suggests regional parenchymal expansion during the global expiratory phase, is physiologically plausible in the context of severe, heterogeneous COPD rather than being a mere artifact. Several non-mutually exclusive mechanisms are proposed to explain this phenomenon. First, it may reflect regional asynchrony (pendelluft). In severe obstructive disease, marked inequalities in regional time constants can lead to paradoxical motion, where severely obstructed lung units lag behind or even continue to expand while adjacent regions are contracting—a phenomenon previously documented using dynamic imaging techniques (5,7). Second, severe air trapping and dynamic hyperinflation in emphysematous regions could result in minimal local volume change during exhalation. When such a small regional volume change is used in the denominator of our global normalization formula, even minor displacement noise or minimal true motion in the numerator can yield disproportionately large or negative adjusted strain values. Third, from a methodological perspective, the normalization in the study uses global lung volume change. Consequently, any region exhibiting slight expansion against the predominant global deflation trend would mathematically result in a negative adjusted strain. These observations underscore the complex biomechanical dysfunction in advanced COPD, where heterogeneous mechanical properties lead to discoordinated and inefficient ventilation. The detection of such paradoxical motion patterns via strain analysis may itself be a biomarker of severe, heterogeneous disease.

While a significant association between 4D-CT-derived strain parameters and spirometric severity is established in this study, a pragmatic assessment of its clinical pathway is needed. The current protocol—with dual-volume acquisition and specialized processing—is best suited for targeted evaluation in complex cases, such as pre-procedural planning for lung volume reduction or phenotyping in therapeutic trials, rather than population screening. Compared to two-dimensional (2D) techniques like dynamic chest radiography (DCR) (8), which offer low-dose screening of global ventilation, 4D-CT strain analysis provides volumetric, voxel-level insight into regional parenchymal mechanics, capturing the biomechanical failure underlying structural heterogeneity. This functional information complements conventional CT metrics (e.g., LD index), moving beyond identifying where emphysema is to explain how poorly affected regions deform.

However, the above observations are representative of only a subset of patients, and many questions about COPD still remain to be answered. Yamashiro et al. have reported that asynchrony of respiratory movements between the pulmonary lobes was observed in patients with severe airflow limitation (7). In the present study, some patients with heterogeneously emphysematous destruction showed asynchronous motion between the pulmonary lobes, and were hard to carry out spirometry. Since completion of spirometry is one of the inclusion criteria, those part of patients have been ruled out at the first step. Even though the lung function was significantly impaired can be speculated from clinical symptoms, it is hard for us to establish a relationship between strain-related parameters and lung function in those patients. On the other side, spirometric measurements represent integrated functional outcomes influenced by complex biomechanical interactions and heterogeneous pathophysiological processes. Consequently, the degree of structural deterioration does not exhibit a linear relationship with pulmonary functional impairment (9-12). We are currently designing prospective studies to evaluate strain measurements in COPD patients ineligible for spirometry, particularly those being considered for interventional therapies or surgical procedures. These investigations may demonstrate the clinical utility of strain analysis in guiding therapeutic decision-making for advanced COPD cases.

Our study employs 4D-CT-based strain analysis to directly quantify parenchymal deformation, offering a complementary approach to computational methods like computational fluid dynamics (CFD). While CFD simulations model airflow dynamics from airway geometry and provide valuable insights into flow resistance, they depend on geometric reconstruction and assumed boundary conditions (13). In contrast, our technique derives tissue-level strain directly from in vivo image sequences, capturing the integrated mechanical response of lung parenchyma to breathing forces. Thus, it assesses the end-organ mechanical behavior rather than airway flow dynamics. Whereas structural CT identifies emphysematous regions and spirometry measures global function, 4D-CT strain analysis adds a dynamic, regional biomechanical perspective. This direct measurement of impaired deformation may aid in phenotyping severe, heterogeneous COPD and could inform interventions where understanding regional mechanics is critical.

There are several limitations that should be noted in the present study. First, the sample sizes in the GOLD III and IV subgroups were relatively small. While combining them allowed for comparative analysis, this limits the statistical power for conclusions specific to the severe cohort and may mask differences between GOLD III and IV stages. Therefore, findings related to severe COPD should be considered exploratory and require validation in larger, severity-stratified cohorts. Frankly speaking, with the deterioration of lung function, there was an increased risk of disorders (e.g., respiratory infection, atelectasis) in pulmonary that did not meet the inclusion criteria. Second, the respiratory monitoring device was not used uniformly in all participants, which could theoretically introduce heterogeneity in respiratory patterns during scanning. Although our pre-analysis quality control—mandating a high correlation coefficient (≥0.8) between the upper and lower lung volume curves for inclusion—ensured consistent respiratory motion in all analyzed scans, and a post-hoc analysis showed no significant difference in correlation coefficients between the device and non-device groups, the lack of standardization presents a potential source of heterogeneity. Future studies would benefit from the standardized application of such monitoring aids to ensure uniform respiratory patterns across all subjects. Third, we focused on the expiration phase only in the current study. The strain measurements for inspiration phase may provide additional information for understanding the pathophysiology of COPD and other modalities, and should be performed in the further study. Fourth, the dynamic-ventilation CT with two volume scans to cover the whole lung (4) was used in the current study, increased radiation exposure was unavoidable. However, a new CT scanner equipped with AiCE (advanced intelligent clear-IQ engine), a deep learning-based reconstruction technique, has been installed in our department. While its primary function is to enhance image quality at a comparable radiation dose, the improved noise performance may allow for future optimization of scan protocols with the potential for significant radiation dose reduction (14,15). Fifth, our statistical analysis did not adjust for potential confounders such as age, sex, or smoking history due to sample size constraints. Future studies with larger cohorts should incorporate multivariate models to isolate the independent contribution of strain parameters to functional impairment.

Conclusions

This study demonstrates that 4D-CT-derived strain parameters, which reflect lung tissue deformation during expiration, progressively decrease with worsening COPD severity as classified by GOLD stage. Notably, strain metrics calculated over the entire expiration phase correlate more strongly with disease severity than those derived from the initial 2-s phase. This suggests that capturing the complete, often prolonged expiratory effort—characteristic of advanced, flow-limited COPD—provides a more comprehensive biomechanical signature of the disease. These findings support the potential of 4D-CT strain analysis as a quantitative imaging tool to assess regional lung mechanics, complementing traditional spirometry and structural CT in the evaluation of COPD severity and heterogeneity.

Acknowledgments

We thank all the staff in our hospital who participated in the study for their assistance in recruiting patients for our study.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1616/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1616/dss

Funding: This work was supported by the National Natural Science Foundation of China (No. 82202288), the Beijing Physician Scientist Training Project (No. BJPSTP-2024-18), and the CAMS Innovation Fund for Medical Sciences (Nos. 2021-I2M-1-049 and 2022-I2M-C&T-B-107).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1616/coif). Yanyan Xu reports receiving grants from the National Natural Science Foundation of China and the Beijing Physician Scientist Training Project. T.Y. reports receiving grants from the CAMS Innovation Fund for Medical Sciences. M.J. reports that she is a current employee of Beijing MicroVec. Inc. Yinghao Xu reports that she is a current employee of Canon Medical Systems. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of China-Japan Friendship Hospital (approval No. 2023-KY-046). Written informed consent was obtained from all participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Celli BR, Wedzicha JA. Update on Clinical Aspects of Chronic Obstructive Pulmonary Disease. N Engl J Med 2019;381:1257-66. [Crossref] [PubMed]
2. Sandelowsky H, Weinreich UM, Aarli BB, Sundh J, Høines K, Stratelis G, Løkke A, Janson C, Jensen C, Larsson K. COPD - do the right thing. BMC Fam Pract 2021;22:244. [Crossref] [PubMed]
3. Lange P, Halpin DM, O'Donnell DE, MacNee W. Diagnosis, assessment, and phenotyping of COPD: beyond FEV1. Int J Chron Obstruct Pulmon Dis 2016;11:3-12. [Crossref] [PubMed]
4. Xu Y, Liang T, Ma Y, Xie S, Sun H, Wang L, Xu Y. Strain Analysis in Patients at High-Risk for COPD Using Four-Dimensional Dynamic-Ventilation CT. Int J Chron Obstruct Pulmon Dis 2022;17:1121-30. [Crossref] [PubMed]
5. Xu Y, Yamashiro T, Moriya H, Tsubakimoto M, Nagatani Y, Matsuoka S, Murayama SACTIve Study Group. Strain measurement on four-dimensional dynamic-ventilation CT: quantitative analysis of abnormal respiratory deformation of the lung in COPD. Int J Chron Obstruct Pulmon Dis 2019;14:65-72. [Crossref] [PubMed]
6. Qaseem A, Wilt TJ, Weinberger SE, Hanania NA, Criner G, van der Molen T, Marciniuk DD, Denberg T, Schünemann H, Wedzicha W, MacDonald R, Shekelle PAmerican College of Physicians. American College of Chest Physicians; American Thoracic Society; European Respiratory Society. Diagnosis and management of stable chronic obstructive pulmonary disease: a clinical practice guideline update from the American College of Physicians, American College of Chest Physicians, American Thoracic Society, and European Respiratory Society. Ann Intern Med 2011;155:179-91. [Crossref] [PubMed]
7. Yamashiro T, Moriya H, Matsuoka S, Nagatani Y, Tsubakimoto M, Tsuchiya N, Murayama S. Asynchrony in respiratory movements between the pulmonary lobes in patients with COPD: continuous measurement of lung density by 4-dimensional dynamic-ventilation CT. Int J Chron Obstruct Pulmon Dis 2017;12:2101-9. [Crossref] [PubMed]
8. Yu D, Xie W, Guo Y, Zhu X, Liang W, Li M, et al. Dynamic Chest Radiography as an Alternative to Pulmonary Function Tests for Chronic Obstructive Pulmonary Disease. Radiology 2025;317:e251296. [Crossref] [PubMed]
9. Kang HS, Bak SH, Oh HY, Lim MN, Cha YK, Yoon HJ, Kim WJ. Computed tomography-based visual assessment of chronic obstructive pulmonary disease: comparison with pulmonary function test and quantitative computed tomography. J Thorac Dis 2021;13:1495-506. [Crossref] [PubMed]
10. Feldhaus FW, Theilig DC, Hubner RH, Kuhnigk JM, Neumann K, Doellinger F. Quantitative CT analysis in patients with pulmonary emphysema: is lung function influenced by concomitant unspecific pulmonary fibrosis? Int J Chron Obstruct Pulmon Dis 2019;14:1583-93. [Crossref] [PubMed]
11. Jacob J, Bartholmai BJ, Rajagopalan S, Kokosi M, Egashira R, Brun AL, Nair A, Walsh SLF, Karwoski R, Wells AU. Serial automated quantitative CT analysis in idiopathic pulmonary fibrosis: functional correlations and comparison with changes in visual CT scores. Eur Radiol 2018;28:1318-27. [Crossref] [PubMed]
12. Bates JH. Systems physiology of the airways in health and obstructive pulmonary disease. Wiley Interdiscip Rev Syst Biol Med 2016;8:423-37. [Crossref] [PubMed]
13. Ning J, Ding M, Li Z, Cai M, Liu X, Cai Y. Quantitative correlation between small airway morphology with respiratory function during disease progression in COPD: CFD analysis of human airways based on CT and OCT imaging. Comput Methods Programs Biomed 2025;272:109066. [Crossref] [PubMed]
14. Nakamura Y, Higaki T, Tatsugami F, Honda Y, Narita K, Akagi M, Awai K. Possibility of Deep Learning in Medical Imaging Focusing Improvement of Computed Tomography Image Quality. J Comput Assist Tomogr 2020;44:161-7. [Crossref] [PubMed]
15. Singh R, Digumarthy SR, Muse VV, Kambadakone AR, Blake MA, Tabari A, Hoi Y, Akino N, Angel E, Madan R, Kalra MK. Image Quality and Lesion Detection on Deep Learning Reconstruction and Iterative Reconstruction of Submillisievert Chest and Abdominal CT. AJR Am J Roentgenol 2020;214:566-73. [Crossref] [PubMed]