Expiratory CT scan in patients with normal inspiratory CT scan: a finding of obliterative bronchiolitis and other causes of bronchiolar obstruction

Gaeta, Michele; Minutoli, Fabio; Girbino, Giuseppe; Murabito, Alessandra; Benedetto, Caterina; Contiguglia, Rosario; Ruggeri, Paolo; Privitera, Salvatore

doi:10.4081/mrm.2013.542

Cited by 4 publications

(4 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also repeated this analysis using PRM‐defined ROIs, which also supports this notion and furthermore indicates that model predictions are sensitive to structural indications of pathological alterations, for example, emphysema and fSAD. We do note however that these structural indications (i.e., gas trapping and emphysema) are less obvious in TLC scans, 53,55 which results in larger differences between DIR and predictions for single‐input TLC model. This is illustrated in Figure 9 where compared to the other models, the single‐input TLC model overestimated volume change in fSAD regions and underestimated volume change in normal regions.…”

Section: Discussionmentioning

confidence: 74%

“…Based on these structure–function relationships, we hypothesized that AI predictions of local tissue expansion using only a single lung CT image will be correlated with registration‐derived Jacobian estimations, albeit to a lesser degree compared with AI predictions using paired images. Secondly, we hypothesize that single‐image AI predictions will be more accurate using FRC images compared to total lung capacity (TLC) images, since FRC reveals pathophysiologic information not apparent at TLC, such as gas trapping 53–55 . While several deep learning‐based registration methods have been proposed which involve predicting the lung motion in three dimensions, 48 predicting the Jacobian image directly only involves estimating the tissue distensibility without having to estimate motion making it a less complex task.…”

Section: Introductionmentioning

confidence: 99%

“…Secondly, we hypothesize that singleimage AI predictions will be more accurate using FRC images compared to total lung capacity (TLC) images, since FRC reveals pathophysiologic information not apparent at TLC, such as gas trapping. [53][54][55] While several deep learning-based registration methods have been proposed which involve predicting the lung motion in three dimensions, 48 predicting the Jacobian image directly only involves estimating the tissue distensibility without having to estimate motion making it a less complex task.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Direct estimation of regional lung volume change from paired and single CT images using residual regression neural network

et al. 2023

View full text Add to dashboard Cite

BackgroundChest computed tomography (CT) enables characterization of pulmonary diseases by producing high‐resolution and high‐contrast images of the intricate lung structures. Deformable image registration is used to align chest CT scans at different lung volumes, yielding estimates of local tissue expansion and contraction.PurposeWe investigated the utility of deep generative models for directly predicting local tissue volume change from lung CT images, bypassing computationally expensive iterative image registration and providing a method that can be utilized in scenarios where either one or two CT scans are available.MethodsA residual regression convolutional neural network, called Reg3DNet+, is proposed for directly regressing high‐resolution images of local tissue volume change (i.e., Jacobian) from CT images. Image registration was performed between lung volumes at total lung capacity (TLC) and functional residual capacity (FRC) using a tissue mass‐ and structure‐preserving registration algorithm. The Jacobian image was calculated from the registration‐derived displacement field and used as the ground truth for local tissue volume change. Four separate Reg3DNet+ models were trained to predict Jacobian images using a multifactorial study design to compare the effects of network input (i.e., single image vs. paired images) and output space (i.e., FRC vs. TLC). The models were trained and evaluated on image datasets from the COPDGene study. Models were evaluated against the registration‐derived Jacobian images using local, regional, and global evaluation metrics.ResultsStatistical analysis revealed that both factors – network input and output space – were significant determinants for change in evaluation metrics. Paired‐input models performed better than single‐input models, and model performance was better in the output space of FRC rather than TLC. Mean structural similarity index for paired‐input models was 0.959 and 0.956 for FRC and TLC output spaces, respectively, and for single‐input models was 0.951 and 0.937. Global evaluation metrics demonstrated correlation between registration‐derived Jacobian mean and predicted Jacobian mean: coefficient of determination (r2) for paired‐input models was 0.974 and 0.938 for FRC and TLC output spaces, respectively, and for single‐input models was 0.598 and 0.346. After correcting for effort, registration‐derived lobar volume change was strongly correlated with the predicted lobar volume change: for paired‐input models r2 was 0.899 for both FRC and TLC output spaces, and for single‐input models r2 was 0.803 and 0.862, respectively.ConclusionsConvolutional neural networks can be used to directly predict local tissue mechanics, eliminating the need for computationally expensive image registration. Networks that use paired CT images acquired at TLC and FRC allow for more accurate prediction of local tissue expansion compared to networks that use a single image. Networks that only require a single input image still show promising results, particularly after correcting for effort, and allow for local tissue expansion estimation in cases where multiple CT scans are not available. For single‐input networks, the FRC image is more predictive of local tissue volume change compared to the TLC image.

show abstract

Section: Discussionmentioning

confidence: 74%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Direct estimation of regional lung volume change from paired and single CT images using residual regression neural network

et al. 2023

View full text Add to dashboard Cite

show abstract

“…However, for CT, it has been reported that various degrees of air trapping can be observed in subjects with normal pulmonary function, irrespective of current smoking status and cigarette consumption. 34 In particular in asymptomatic healthy subjects with normal pulmonary function, air trapping is usually limited to fewer than three adjacent secondary pulmonary lobules (“lobular air-trapping”). 35 Since it has been shown that the sensitivity of MRI is lower compared to CT, it may be speculated, that such limited, clinically insignificant air trapping will not be seen with MRI, while any air trapping large enough to detect with MRI is probably clinically significant, which would make MRI the less sensitive, but potentially more specific test.…”

Section: Dynamic Airway Imagingmentioning

confidence: 99%

MR imaging of the airways

Biederer

2023

BJR

View full text Add to dashboard Cite

The need for airway imaging is defined by the limited sensitivity of common clinical tests like spirometry, lung diffusion (DLCO) and blood gas analysis to early changes of peripheral airways and to inhomogeneous regional distribution of lung function deficits. Therefore, X-ray and computed tomography (CT) are frequently used to complement the standard tests. As an alternative, magnetic resonance imaging (MRI) offers radiation free lung imaging, but at lower spatial resolution. Non-contrast enhanced MRI shows healthy airways down to the first sub segmental level/4th order (CT: eighth). Bronchiectasis can be identified by wall thickening and fluid accumulation. Smaller airways become visible, when altered by peribronchiolar inflammation or mucus retention (tree-in-bud sign). The strength of MRI is functional imaging. Dynamic, time-resolvedMRI directly visualizesexpiratory airway collapse down to the lobar level (CT: segmental level). Obstruction of even smaller airways becomes visible as air trapping on the expiratory scans. MRI with hyperpolarized noble gases (3He, 129Xe) directly shows the large airways and peripheral lung ventilation. Dynamic contrast enhanced MRI (DCE MRI) indirectly shows airway dysfunction as perfusion deficits resulting from hypoxic vasoconstriction of the dependent lung volumes. Further promising scientific approaches such as non-contrast enhanced, ventilation-/perfusion-weighted MRI from periodic signal changes of respiration and blood flow are in development. In summary, MRI of the lungs and airways excels with its unique combination of morphologic and functional imaging capacities for research (e.g., in chronic obstructive lung disease or asthma) as well as for clinical imaging (e.g., in cystic fibrosis).

show abstract

BreathVisionNet: A pulmonary-function-guided CNN-transformer hybrid model for expiratory CT image synthesis

Zhang,

Pang,

et al. 2025

Computer Methods and Programs in Biomedicine

View full text Add to dashboard Cite

Expiratory CT scan in patients with normal inspiratory CT scan: a finding of obliterative bronchiolitis and other causes of bronchiolar obstruction

Cited by 4 publications

References 18 publications

Direct estimation of regional lung volume change from paired and single CT images using residual regression neural network

Direct estimation of regional lung volume change from paired and single CT images using residual regression neural network

MR imaging of the airways

BreathVisionNet: A pulmonary-function-guided CNN-transformer hybrid model for expiratory CT image synthesis

Contact Info

Product

Resources

About