Joseph K. Leader

Association of Radiographic Emphysema and Airflow Obstruction with Lung Cancer

RATIONALE To study the relationship between emphysema and/or airflow obstruction and lung cancer in a high-risk population. OBJECTIVE We studied lung cancer related to radiographic emphysema and spirometric airflow obstruction in tobacco-exposed persons who were screened for lung cancer using chest computed tomography (CT). METHODS Subjects completed questionnaires, spirometry, and low-dose helical chest CT. CT scans were scored for emphysema based on National Emphysema Treatment Trial criteria. Multiple logistic regressions estimated the independent associations between various factors, including radiographic emphysema and airflow obstruction, and subsequent lung cancer diagnosis. MEASUREMENTS AND MAIN RESULTS Among 3,638 subjects, 57.5, 18.8, 14.6, and 9.1% had no, trace, mild, and moderate-severe emphysema, and 57.3, 13.6, 22.8, and 6.4% had no, mild (Global Initiative for Chronic Obstructive Lung Disease [GOLD] I), moderate (GOLD II), and severe (GOLD III-IV) airflow obstruction. Of 3,638 subjects, 99 (2.7%) received a lung cancer diagnosis. Adjusting for sex, age, years of cigarette smoking, and number of cigarettes smoked daily, logistic regression showed the expected lung cancer association with the presence of airflow obstruction (GOLD I-IV, odds ratio [OR], 2.09; 95% confidence interval [CI], 1.33-3.27). A second logistic regression showed lung cancer related to emphysema (OR, 3.56; 95% CI, 2.21-5.73). After additional adjustments for GOLD class, emphysema remained a strong and statistically significant factor related to lung cancer (OR, 3.14; 95% CI, 1.91-5.15). CONCLUSIONS Emphysema on CT scan and airflow obstruction on spirometry are related to lung cancer in a high-risk population. Emphysema is independently related to lung cancer. Both radiographic emphysema and airflow obstruction should be considered when assessing lung cancer risk.

Automated lung segmentation in X-ray computed tomography

RATIONALE AND OBJECTIVES To develop and evaluate a reliable, fully-automated lung segmentation scheme for application in X-ray computed tomography. MATERIALS AND METHODS The automated scheme was heuristically developed using a slice-based, pixel-value threshold and two sets of classification rules. Features used in the rules include size, circularity, and location. The segmentation scheme operates slice-by-slice and performs three key operations: (1) image preprocessing to remove background pixels, (2) computation and application of a pixel-value threshold to identify lung tissue, and (3) refinement of the initial segmented regions to prune incorrectly detected airways and separate fused right and left lungs. RESULTS The performance of the automated segmentation scheme was evaluated using 101 computed tomography cases (91 thick slice, 10 thin slice scans). The 91 thick cases were pre- and post-surgery from 50 patients and were not independent. The automated scheme successfully segmented 94.0% of the 2,969 thick slice images and 97.6% of the 1,161 thin slice images. The mean difference of the total lung volumes calculated by the automated scheme and functional residual capacity plus 60% inspiratory capacity was -24.7 +/- 508.1 mL. The mean differences of the total lung volumes calculated by the automated scheme and an established, commonly used semi-automated scheme were 95.2 +/- 52.5 mL and -27.7 +/- 66.9 mL for the thick and thin slice cases, respectively. CONCLUSION This simple, fully-automated lung segmentation scheme provides an objective tool to facilitate lung segmentation from computed tomography scans.

Radiographic Emphysema Predicts Low Bone Mineral Density in a Tobacco-exposed Cohort

RATIONALE Studies demonstrating an association between chronic obstructive pulmonary disease and low bone mineral density (BMD) implicate factors distinct from treatments and severity of lung disease in the pathogenesis of osteoporosis. Whereas emphysema has been independently associated with vascular disease and other comorbidities, its association with BMD has not been well studied. OBJECTIVES We explored the associations of BMD with computed tomography (CT) measures of emphysema and other risk factors in current and former smokers. METHODS One hundred ninety subjects completed a CT scan, pulmonary function testing, questionnaires, and dual x-ray absorptiometry measurements of hip and lumbar spine BMD. Subjects were classified as having normal BMD, osteopenia, or osteoporosis. Demographic, physiologic, and radiographic characteristics were compared and the association of BMD with radiographic emphysema, airflow obstruction, and osteoporosis risk factors was assessed. MEASUREMENTS AND MAIN RESULTS No difference existed in age, tobacco exposure, oral steroid use, or physical activity across BMD categories. Both osteopenia and osteoporosis were associated with the presence of airflow obstruction, inhaled corticosteroid use, and female sex, and demonstrated a significant relationship with the presence of visual emphysema (P = 0.0003). Quantitative emphysema, but not CT-measured indices of airway wall thickness, was inversely associated with BMD. Visual emphysema alone was a significant predictor of osteopenia/osteoporosis (odds ratio = 2.55; 95% confidence interval, 1.24-5.25) in a model including obstruction severity, age, sex, and inhaled and oral steroid use. CONCLUSIONS Radiographic emphysema is a strong, independent predictor of low BMD in current and former smokers. This relationship suggests a common mechanistic link between emphysema and osteopenia/osteoporosis.

A Computational Geometry Approach to Automated Pulmonary Fissure Segmentation in CT Examinations

Identification of pulmonary fissures, which form the boundaries between the lobes in the lungs, may be useful during clinical interpretation of computed tomography (CT) examinations to assess the early presence and characterization of manifestation of several lung diseases. Motivated by the unique nature of the surface shape of pulmonary fissures in 3-D space, we developed a new automated scheme using computational geometry methods to detect and segment fissures depicted on CT images. After a geometric modeling of the lung volume using the marching cubes algorithm, Laplacian smoothing is applied iteratively to enhance pulmonary fissures by depressing nonfissure structures while smoothing the surfaces of lung fissures. Next, an extended Gaussian image based procedure is used to locate the fissures in a statistical manner that approximates the fissures using a set of plane ldquopatchesrdquo. This approach has several advantages such as independence of anatomic knowledge of the lung structure except the surface shape of fissures, limited sensitivity to other lung structures, and ease of implementation. The scheme performance was evaluated by two experienced thoracic radiologists using a set of 100 images (slices) randomly selected from 10 screening CT examinations. In this preliminary evaluation 98.7% and 94.9% of scheme segmented fissure voxels are within 2 mm of the fissures marked independently by two radiologists in the testing image dataset. Using the scheme detected fissures as reference, 89.4% and 90.1% of manually marked fissure points have distance les2 mm to the reference suggesting a possible under-segmentation of the scheme. The case-based root mean square (rms) distances (ldquoerrorsrdquo) between our scheme and the radiologist ranged from 1.48plusmn0.92 to 2.04plusmn3.88 mm. The discrepancy of fissure detection results between the automated scheme and either radiologist is smaller in this dataset than the interreader variability.

Pulmonary Nodule Detection with Low-Dose CT of the Lung: Agreement Among Radiologists

OBJECTIVE The purpose of our study was to assess relative intra- and interobserver agreement in detecting pulmonary nodules when interpreting low-dose chest CT screening examinations. MATERIALS AND METHODS Two hundred ninety-three selected low-dose CT examinations of the lung were independently interpreted by three radiologists to detect and classify pulmonary nodules. The data set selected was enriched with examinations depicting pulmonary nodules. A subset of 30 examinations was interpreted twice. All pulmonary nodules greater than 1.0 mm were marked. All nodules greater than 3.0 mm were marked, measured, and scored as to their probability of being benign or malignant. Nodule-based and examination-based relative reviewer agreements were evaluated using percentage of agreement and kappa statistics. Similar assessments were performed on the subset of examinations interpreted twice. RESULTS The three radiologists identified a total of 470, 729, and 876 pulmonary nodules of which 395, 641, and 778 were rated as noncalcified with some level of suspicion for being malignant. Nodule-based interobserver agreement among the radiologists was poor (highest kappa value in a paired comparison, 0.120). Examination-based agreement was higher (highest kappa value in a paired comparison, 0.458). Intraobserver agreement was higher than interobserver agreement for examination-based agreement (highest kappa = 0.889) but lower for nodule-based agreement (highest kappa = -0.035). Agreement improved as the suspicion of malignancy increased. CONCLUSION Unaided intra- and interobserver agreement in detecting pulmonary nodules in low-dose CT of the lung is relatively low. Computer-assisted detection may provide the consistency that is needed for this purpose.

Pulmonary Lobe Segmentation in CT Examinations Using Implicit Surface Fitting

Lobe identification in computed tomography (CT) examinations is often an important consideration during the diagnostic process as well as during treatment planning because of their relative independence of each other in terms of anatomy and function. In this paper, we present a new automated scheme for segmenting lung lobes depicted on 3-D CT examinations. The unique characteristic of this scheme is the representation of fissures in the form of implicit functions using radial basis functions (RBFs), capable of seamlessly interpolating ldquoholesrdquo in the detected fissures and smoothly extrapolating the fissure surfaces to the lung boundaries resulting in a ldquonaturalrdquo segmentation of lung lobes. A previously developed statistically based approach is used to detect pulmonary fissures and the constraint points for implicit surface fitting are selected from detected fissure surfaces in a greedy manner to improve fitting efficiency. In a preliminary assessment study, lobe segmentation results of 65 chest CT examinations, five of which were reconstructed with three section thicknesses of 0.625 mm, 1.25 mm, and 2.5 mm, were subjectively and independently evaluated by two experienced chest radiologists using a five category rating scale (i.e., excellent, good, fair, poor, and unacceptable). Thirty-three of 65 examinations (50.8%) with a section thickness of 0.625 mm were rated as either ldquoexcellentrdquo or ldquogoodrdquo by both radiologists and only one case (1.5%) was rated by both radiologists as ldquopoorrdquo or ldquounacceptable.rdquo Comparable performance was obtained with a slice thickness of 1.25 mm, but substantial performance deterioration occurred in examinations with a section thickness of 2.5 mm. The advantages of this scheme are its full automation, relative insensitivity to fissure completeness, and ease of implementation.

C-X-C Motif Chemokine 13 (CXCL13) Is a Prognostic Biomarker of Idiopathic Pulmonary Fibrosis

RATIONALE C-X-C motif chemokine 13 (CXCL13) mediates B-cell trafficking and is increased, proportionately to disease activity, in many antibody-mediated syndromes. Dysregulated B cells have recently been implicated in idiopathic pulmonary fibrosis (IPF) pathogenesis. OBJECTIVES To determine if CXCL13 is associated with IPF progression. METHODS CXCL13 was measured in lungs by DNA microarray and immunohistochemistry, and in plasma by ELISA. MEASUREMENTS AND MAIN RESULTS CXCL13 mRNA was threefold and eightfold greater in IPF lungs (n = 92) compared with chronic obstructive pulmonary disease (COPD) (n = 191) and normal (n = 108) specimens, respectively (P < 0.0001). IPF lungs also showed increased CXCL13 staining. Plasma CXCL13 concentrations (pg/ml) were greater in 95 patients with IPF (94 ± 8) than in 128 subjects with COPD (53 ± 9) and 57 normal subjects (35 ± 3) (P < 0.0001). Circulating CXCL13 levels were highest in patients with IPF with pulmonary artery hypertension (P = 0.01) or acute exacerbations (P = 0.002). Six-month survival of patients with IPF in the highest quartile of plasma CXCL13 was 65 ± 10% versus 93 ± 10% in the others (hazard ratio, 5.5; 95% confidence interval, 1.8-16.9; P = 0.0008). CXCL13 increases by more than 50% in IPF serial assays, irrespective of initial values, also presaged respiratory failure (hazard ratio, 7.2; 95% confidence interval, 1.3-40.0; P = 0.008). In contrast, CXCL13 clinical associations in subjects with COPD were limited to modest correlations with FEV1 (P = 0.05) and progression of radiographic emphysema (P = 0.05). CONCLUSIONS CXCL13 is increased and is a prognostic biomarker in patients with IPF, and more so than in patients with COPD. This contrast indicates CXCL13 overexpressions are intrinsic to IPF, rather than an epiphenomenon of lung injury. The present data implicate CXCL13 and B cells in IPF pathogenesis, and support considerations for trials of specific B-cell-targeted therapies in patients with this intractable disease.

The influence of radiographic phenotype and smoking status on peripheral blood biomarker patterns in chronic obstructive pulmonary disease.

Background Chronic obstructive pulmonary disease (COPD) is characterized by both airway remodeling and parenchymal destruction. The identification of unique biomarker patterns associated with airway dominant versus parenchymal dominant patterns would support the existence of unique phenotypes representing independent biologic processes. A cross-sectional study was performed to examine the association of serum biomarkers with radiographic airway and parenchymal phenotypes of COPD. Methodology/Principal Findings Serum from 234 subjects enrolled in a CT screening cohort was analyzed for 33 cytokines and growth factors using a multiplex protein array. The association of serum markers with forced expiratory volume in one second percent predicted (FEV1%) and quantitative CT measurements of airway thickening and emphysema was assessed with and without stratification for current smoking status. Significant associations were found with several serum inflammatory proteins and measurements of FEV1%, airway thickening, and parenchymal emphysema independent of smoking status. The association of select analytes with airway thickening and emphysema was independent of FEV1%. Furthermore, the relationship between other inflammatory markers and measurements of physiologic obstruction or airway thickening was dependent on current smoking status. Conclusions/Significance Airway and parenchymal phenotypes of COPD are associated with unique systemic serum biomarker profiles. Serum biomarker patterns may provide a more precise classification of the COPD syndrome, provide insights into disease pathogenesis and identify targets for novel patient-specific biological therapies.

The Relationship Between Pulmonary Emphysema and Kidney Function in Smokers

BACKGROUND It has been reported that the prevalence of kidney dysfunction may be increased in patients exposed to tobacco with airflow obstruction. We hypothesized that kidney dysfunction would associate with emphysema rather than with airflow obstruction measured by the FEV₁. METHODS Five hundred eight current and former smokers completed a chest CT scan, pulmonary function tests, medical questionnaires, and measurement of serum creatinine. Glomerular filtration rates (eGFRs) were estimated using the method of the Chronic Kidney Disease Epidemiology Collaboration. Quantitative determinants of emphysema and airway dimension were measured from multidetector chest CT scans. RESULTS The mean age was 66 ± 7 years, and mean eGFR was 101 ± 22 mL/min/1.73 m². Univariate and multivariate analysis showed a significant association between radiographically measured emphysema and eGFR: Participants with 10% more emphysema had an eGFR that was lower by 4.4 mL/min/1.73 m² (P = .01), independent of airflow obstruction (FEV₁), age, sex, race, height, BMI, diabetes mellitus, hypertension, coronary artery disease, patient-reported dyspnea, pack-years of smoking, and current smoking. There was no association between eGFR and either FEV₁ or quantitative CT scan measures of airway dimension. CONCLUSIONS More severe emphysema, rather than airflow obstruction, is associated with kidney dysfunction in tobacco smokers, independent of common risk factors for kidney disease. This finding adds to recent observations of associations between emphysema and comorbidities of COPD, including osteoporosis and lung cancer, which are independent of the traditional measure of reduced FEV₁. The mechanisms and clinical implications of kidney dysfunction in patients with emphysema need further investigation.

Computer-aided detection of early interstitial lung diseases using low-dose CT images

This study aims to develop a new computer-aided detection (CAD) scheme to detect early interstitial lung disease (ILD) using low-dose computed tomography (CT) examinations. The CAD scheme classifies each pixel depicted on the segmented lung areas into positive or negative groups for ILD using a mesh-grid-based region growth method and a multi-feature-based artificial neural network (ANN). A genetic algorithm was applied to select optimal image features and the ANN structure. In testing each CT examination, only pixels selected by the mesh-grid region growth method were analyzed and classified by the ANN to improve computational efficiency. All unselected pixels were classified as negative for ILD. After classifying all pixels into the positive and negative groups, CAD computed a detection score based on the ratio of the number of positive pixels to all pixels in the segmented lung areas, which indicates the likelihood of the test case being positive for ILD. When applying to an independent testing dataset of 15 positive and 15 negative cases, the CAD scheme yielded the area under receiver operating characteristic curve (AUC = 0.884 ± 0.064) and 80.0% sensitivity at 85.7% specificity. The results demonstrated the feasibility of applying the CAD scheme to automatically detect early ILD using low-dose CT examinations.