Convolutional Neural Network (CNN)-based models using Computed Tomography (CT) images classify Chronic Obstructive Pulmonary Disease (COPD) patients with high accuracy, but studies have used various different input images and it is unclear what input images are optimum, particularly in a milder COPD cohort. We propose a novel approach using 2D airway-optimized topological multi-planar reformat (airway-optimized tMPR) images as well as novel 3D fusion methods and compared the performance of these models with various established 2D/3D CNN-based methods in a population-based mild COPD cohort. Participants from the CanCOLD study were evaluated. We implemented several 2D/3D models adapted from the literature. Existing CNN-based models were trained using 2D collages of axial/coronal/sagittal slices, and colored and binary airway images. 3D models consisting of 15 axial inspiratory/expiratory slices were selected, and input and output combination methods were investigated. For the proposed models, 2D airway-optimized tMPR images were constructed using cut-surface renderings to convey shape and interior/contextual information. 3D output fusion of axial/coronal/sagittal images, as well as output fusion of the axial and 3D airway tree, were also investigated. Finally, the output fusion of 2D airway-optimized tMPR methods and 3D lungs combined method was investigated. 742 participants were used for training/validation and 309 for testing. The 2D and 3D methods adapted from the literature had accuracy ranging from 61%-72% in the mild COPD cohort. The 2D airway-optimized tMPR model achieved 73% accuracy. The proposed 3D model of combining axial/coronal/sagittal images had an accuracy of 75%. The proposed model output combining 2D colored airways and inspiratory combined 3D images, and the 3D collage of axial/coronal/sagittal images, resulted in 74% and 73% accuracy, respectively. However, the output fusion of the airway-optimized tMPR and 3D lung model of combining axial/coronal/sagittal images reached the highest accuracy of 78%. While the CNN model with 2D airway/lung-optimized images had improved performance with reduced computational resources as compared to the 3D models proposed, as well as the other published CNN-based models, the combination of this 2D method with the 3D CNN model of combining axial/coronal/sagittal images achieved the highest performance in this mild cohort.
We propose two-dimensional (2D) dual-energy (DE) x-ray imaging of lung structure and function for the assessment of COPD, and investigate the resulting image quality theoretically using the human observer detectability index (d') as a figure of merit. We modeled the ability of human observers to detect ventilation defects in xenon enhanced DE (XeDE) images and emphysema in unenhanced DE images. Our model of d' accounted for the extent of emphysematous destruction and functional impairment as a function of defect/lesion contrast, spatial resolution, x-ray scatter, quantum and background anatomical noise power spectrum (NPS), and the efficiency of human observers. The effect of x-ray spectrum and exposure allocation factor on d' was also explored. Our results suggest that, the detectability is maximized for exposure allocation factors that minimize quantum NPS. The optimal combination of tube voltage was found to be ~50/140 kV or 60/140 kV depending on the task and patient at an x-ray exposure equal to that of a standard chest x-ray. In 2D DE x-ray imaging of COPD, the detectability is primarily limited by low contrast, x-ray scatter, and anatomic noise, the latter two of which reduce the detectability of individual defects by 30% and ~>90%, respectively.
We propose a two-dimensional (2D) contrast-enhanced dual-energy (DE) approach for functional x-ray imaging of respiratory disease. With this approach, non-radioactive xenon is used to provide contrast between ventilated regions of the lung and unventilated regions of the lung (i.e. ventilation defects); DE subtraction is used to suppress rib structures from 2D thoracic images. We modeled theoretically the signal-to-noise ratio (SNR) and area under the receiver operating characteristic curve (AUC) of a human observer for a defect present vs. defect absent binary classification task under signal-known-exactly/background-known-exactly conditions. Our model accounted for the size of ventilation defects, contrast of ventilation defects, quantum noise, finite spatial resolution, x-ray attenuation and observer efficiency. We modeled spherical defects with diameters up to 2.5 cm, and contrast and noise levels relevant for imaging of children, adolescents, adult males and adult females. Quantum noise and spatial resolution properties were calculated assuming an ideal energy-integrating x-ray detector. All calculations were performed assuming low-energy and high-energy applied tube voltages of 70 kV and 140 kV, respectively, with 2 mm of added copper filtration on the high-energy spectrum, and a total entrance exposure of 18 mR, which is typical for anterior-posterior thoracic imaging procedures. Our analysis shows that an AUC of 0.85 can be achieved for defect diameters as small as 1.1 cm, 1.2 cm, 1.3 cm and 1.4 cm for children ages 2 to 8, adolescents ages 9 to 14, adult males and adult females, respectively. Our results suggest that the DE approach proposed here warrants further investigation as a low-dose, low-cost alternative to existing approaches for functional imaging of respiratory disease.
We designed and generated pulmonary imaging biomarker pipelines to facilitate high-throughput research and point-of-care use in patients with chronic lung disease. Image processing modules and algorithm pipelines were embedded within a graphical user interface (based on the .NET framework) for pulmonary magnetic resonance imaging (MRI) and x-ray computed-tomography (CT) datasets. The software pipelines were generated using C++ and included: (1) inhaled He3 / Xe129 MRI ventilation and apparent diffusion coefficients, (2) CT-MRI coregistration for lobar and segmental ventilation and perfusion measurements, (3) ultrashort echo-time H1 MRI proton density measurements, (4) free-breathing Fourier-decomposition H1 MRI ventilation/perfusion and free-breathing H1 MRI specific ventilation, (5) multivolume CT and MRI parametric response maps, and (6) MRI and CT texture analysis and radiomics. The image analysis framework was implemented on a desktop workstation/tablet to generate biomarkers of regional lung structure and function related to ventilation, perfusion, lung tissue texture, and integrity as well as multiparametric measures of gas trapping and airspace enlargement. All biomarkers were generated within 10 min with measurement reproducibility consistent with clinical and research requirements. The resultant pulmonary imaging biomarker pipeline provides real-time and automated lung imaging measurements for point-of-care and high-throughput research.
The objective was to develop an automated optical coherence tomography (OCT) segmentation method. We evaluated three ex-vivo porcine airway specimens; six non-sequential OCT images were selected from each airway specimen. Histology was also performed for each airway and histology images were co-registered to OCT images for comparison. Manual segmentation of the airway luminal area, mucosa area, submucosa area and the outer airway wall area were performed for histology and OCT images. Automated segmentation of OCT images employed a despecking filter for pre-processing, a hessian-based filter for lumen and outer airway wall area segmentation, and K-means clustering for mucosa and submucosa area segmentation. Bland-Altman analysis indicated that there was very little bias between automated OCT segmentation and histology measurements for the airway lumen area (bias=-6%, 95% CI=-21%-8%), mucosa area, (bias=-4%, 95% CI=-14%-5%), submucosa area (bias=7%, 95% CI=-7%-20%) and outer airway wall area segmentation results (bias=-5%, 95% CI=-14%-5%). We also compared automated and manual OCT segmentation and Bland-Altman analysis indicated that there was negligible bias between luminal area (bias=4%, 95% CI=1%-8%), mucosa area (bias=-3%, 95% CI=-6%-1%), submucosa area (bias=-2%, 95% CI=-10%-6%) and the outer airway wall (bias=-3%, 95% CI=-13%-6%). The automated segmentation method for OCT airway imaging developed here allows for accurate and precise segmentation of the airway wall components, suggesting that translation of this method to in vivo human airway analysis would allow for longitudinal and serial studies.
Although there are more women than men dying of chronic obstructive pulmonary disease (COPD) in the United States and elsewhere, we still do not have a clear understanding of the differences in the pathophysiology of airflow obstruction between the sexes. Optical coherence tomography (OCT) is an emerging imaging technology that has the capability of imaging small bronchioles with resolution approaching histology. Therefore, our objective was to compare OCT-derived airway wall measurements between males and females matched for lung size and in anatomically matched small airways. Subjects 50-80 yrs were enrolled in the British Columbia Lung Health Study and underwent OCT and spirometry. OCT was performed using a 1.5mm diameter probe/sheath in anatomically matched airways for males and females; the right lower lobe (RB8 or RB9) or left lower lobe (LB8 or LB9) during end-expiration. OCT airway wall area (Aaw) was obtained by manual segmentation. For males and females there was no significant difference in OCT Aaw (p=0.12). Spearman correlation coefficients indicated that the forced expiratory volume in 1 second (FEV1) and Aaw were significantly correlated for males (r=-0.78, p=0.004) but not for females (r=-0.20, p=0.49) matched for lung size. These novel OCT findings demonstrate that while there were no overall sex differences in airway wall thickness, the relationship between lung function and airway wall thickness was correlated only in men. Therefore, factors other than airway remodeling may be driving COPD pathogenesis in women and OCT may provide important information for investigating airway remodeling and its relationship with COPD progression.
KEYWORDS: Image segmentation, Magnetic resonance imaging, Lung, 3D metrology, Chronic obstructive pulmonary disease, 3D image processing, 3D acquisition, Algorithm development, Image registration, Image processing algorithms and systems
A semi-automated method for generating hyperpolarized helium-3 (3He) measurements of individual slice (2D) or whole
lung (3D) gas distribution was developed. 3He MRI functional images were segmented using two-dimensional (2D) and
three-dimensional (3D) hierarchical K-means clustering of the 3He MRI signal and in addition a seeded region-growing
algorithm was employed for segmentation of the 1H MRI thoracic cavity volume. 3He MRI pulmonary function
measurements were generated following two-dimensional landmark-based non-rigid registration of the 3He and 1H
pulmonary images. We applied this method to MRI of healthy subjects and subjects with chronic obstructive lung
disease (COPD). The results of hierarchical K-means 2D and 3D segmentation were compared to an expert observer's
manual segmentation results using linear regression, Pearson correlations and the Dice similarity coefficient. 2D
hierarchical K-means segmentation of ventilation volume (VV) and ventilation defect volume (VDV) was strongly and
significantly correlated with manual measurements (VV: r=0.98, p<.0001; VDV: r=0.97, p<.0001) and mean Dice
coefficients were greater than 92% for all subjects. 3D hierarchical K-means segmentation of VV and VDV was also
strongly and significantly correlated with manual measurements (VV: r=0.98, p<.0001; VDV: r=0.64, p<.0001) and the
mean Dice coefficients were greater than 91% for all subjects. Both 2D and 3D semi-automated segmentation of 3He
MRI gas distribution provides a way to generate novel pulmonary function measurements.
Hyperpolarized helium-3 (3He) magnetic resonance imaging (MRI) has emerged as a non-invasive research method for
quantifying lung structural and functional changes, enabling direct visualization in vivo at high spatial and temporal
resolution. Here we described the development of methods for quantifying ventilation dynamics in response to
salbutamol in Chronic Obstructive Pulmonary Disease (COPD). Whole body 3.0 Tesla Excite 12.0 MRI system was
used to obtain multi-slice coronal images acquired immediately after subjects inhaled hyperpolarized 3He gas.
Ventilated volume (VV), ventilation defect volume (VDV) and thoracic cavity volume (TCV) were recorded following
segmentation of 3He and 1H images respectively, and used to calculate percent ventilated volume (PVV) and ventilation
defect percent (VDP). Manual segmentation and Otsu thresholding were significantly correlated for VV (r=.82, p=.001),
VDV (r=.87 p=.0002), PVV (r=.85, p=.0005), and VDP (r=.85, p=.0005). The level of agreement between these
segmentation methods was also evaluated using Bland-Altman analysis and this showed that manual segmentation was
consistently higher for VV (Mean=.22 L, SD=.05) and consistently lower for VDV (Mean=-.13, SD=.05) measurements
than Otsu thresholding. To automate the quantification of newly ventilated pixels (NVp) post-bronchodilator, we used
translation, rotation, and scaling transformations to register pre-and post-salbutamol images. There was a significant
correlation between NVp and VDV (r=-.94 p=.005) and between percent newly ventilated pixels (PNVp) and VDP (r=-
.89, p=.02), but not for VV or PVV. Evaluation of 3He MRI ventilation dynamics using Otsu thresholding and
landmark-based image registration provides a way to regionally quantify functional changes in COPD subjects after
treatment with beta-agonist bronchodilators, a common COPD and asthma therapy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.