Search

Background: We seek to evaluate an automated method developed to assess central airway shape (cross-sectional and longitudinal) variations in the context of a multi-center study, SPIROMICS, seeking to evaluate phenotypes of COPD. Methods: 717 subjects (GOLD 1: 28, GOLD 2: 275, GOLD 3: 293, GOLD 4: 121) randomly selected from SPIROMICS baseline scans. Matching TLC and RV datasets were analyzed (VIDA Diagnostics) with a focus on the trachea and the left and right mainstem bronchi (LMB, RMB). Measurements included: extent of concave sections and deviation from circularity in lumen cross-sections, and straightness along the airway centerline. Results: Concavity and circularity: there is a significant increase in concavity and decrease in circularity between TLC and RV for all GOLD groups. Kruskal-Wallis test indicates significant concavity and circularity differences at RV in the trachea and LMB between GOLD groups 2 and 4, as well as 3 and 4. Straightness: statistically significant but physically small change in straightness between TLC and RV for all three major airways. Significant differences at RV in straightness across GOLD groups for trachea and LMB, but not for RMB. Conclusions: Central airway shapes change significantly across the vital capacity with distinctly different shape changes detectable across GOLD status.

Background COPD is associated with cardiovascular disease (CVD), and coronary artery calcification (CAC) provides incremental prognostic information. CAC is measured using Agatston scores on electrocardiographically gated CT scans. With increasing use of non-gated CT scans in clinical practice, we hypothesized that a simple visual Weston score would perform as well in predicting prevalent coronary artery disease (CAD) and incident CVD in COPD. Methods Current and former smokers (45-80 years) enrolled in the COPDGene study and followed for 5 years. At baseline, significant CAC was defined as Agatston >400 and Weston >7; incident CVD recorded as one/more of incident acute coronary event, stroke, peripheral arterial disease and coronary intervention. Results 1913 subjects were included; 813 with and 1100 without airflow obstruction. CAD was reported by 140(7.3%) subjects at baseline. 439(22.9%) and 748(39.1%) had significant CAC by Weston and Agatston respectively; the two methods correlated significantly (r=0.51;p 7 identified more patients with undiagnosed CAD who developed incident CVD (6.3% vs. 10.2%;p=0.01) than Agatston >400 (6.9% vs. 7.2%;p = 0.77). Conclusions A simple visual score for CAC performs well in predicting prevalent CAD and incident CVD in smokers with and without COPD.

Background: We seek to establish baseline computed tomography (CT)-based airway measures for 5 standardized paths into the lung that include segmental bronchi RB1, RB4, RB10, LB1, LB10, incl. differences between normal male (M) and female (F). Methods : 56 normal subjects (34 female, 22 male) of 100 imaged using a standardized CT protocol: std kernel, 0.5mm slice thick., 100mAs, 120kV, breath-hold at 90% vital cap. have been analyzed to date. Subjs age range: 20 - 70. Data analysis included airway segmentation, labeling (incl. sub-segmental), and measurements (Apollo® 2.0, VIDA Diagn.). All steps were automated but visually verified by human experts and edited as needed to assure data quality. Results : Common lumen diameter (D) and wall thickness (WT) were observed across the population, with segment specific differences. M and F subjects differ significantly in most but not all segments, with D in M being bigger where significant. E.g., 1.12 mm D diff. in RB1, 0.45 mm in LB1, but no significant diff. in RB4. WT are generally greater in M than in F. Unique branching patterns (lumen dimension & branching angles) common across the population were observed for sub-segmental branches in LB/RB 3, RB9, LB/RB10 (monopodal rather than bipodal, with the latter being dominant in the rest of the human airway tree). Path-based Pi10 measurements to date have not differed significantly, between individual sub-trees or between F and M. Conclusions: A common monopodial branching pattern has been identified for 3 branches. Normative measures for a standardized set of paths into the normal human lung have been established and provide a basis for evaluating changes associated with pathologic processes.

Background: Pulmonary blood volume(PBV) changes with lung inflation and this can be measured as a change in non-air volume via quantitative CT, assuming that tissue volume remains constant. Methods: 40 subjects(20 female, 20 male) were selected across various COPD classes from SPIROMICS multicenter research study. TLC and RV pair was acquired for each subject. Lobar segmentation for TLC was obtained via Apollo lung analytics platform(VIDA Diagnostics, US) with occasional edits from expert analysts. RV lobar mask was obtained via VIDA9s automatic image registration whereby TLC lobar mask was warped into the RV domain. Lung tissue mass (parenchyma plus blood) was derived from Hounsfield units. Results: PBV is greater at TLC compared to RV: mean relative tissue mass difference of 3.5% (95% confidence interval:[2.1%, 5.0%]) for the whole lung. Relative tissue mass (thus PBV) changes between TLC and RV are similar across all lobes with a mean value of 6.0%(significantly different from whole lung, p=0.001). In addition, as seen in, there is considerable difference in distribution of PBV changes across subjects. Conclusion: Decreases in PBV with lung inflation demonstrate subject variability when a mixed COPD population is evaluated suggesting lung volume effects must be considered when reporting regional PBV, especially when using blood volume as a surrogate for regional perfusion as is done with Dual Energy CT.

Background: Objective characterization of the magnitude and extent of fissure completeness and emphysematous disease at various stages of COPD is critical to select best therapeutic options. Objective: Determine distribution of fissure integrity (FI) and quantitative CT (QCT) measurements of emphysema by GOLD stages. Methods: 249 patients at various stages of COPD disease (Gold 0: 20, Gold 1: 36, Gold 2: 35, Gold 3: 89, Gold 4: 69) were collected from two institutions (Iowa Comprehensive Lung Imaging Center, Heidelberg Thoraxklinik). 92 QCT measurements of lung volume, emphysema, heterogeneity score, and FI were extracted across the lungs and analyzed using Apollo® software (VIDA Diagnostics, IA). One-way ANOVA analysis was conducted to determine QCT measurements with significant changes across GOLD stages. The distribution of patients in terms of gender and height was homogenous across GOLD stages (p>0.30). GOLD-4 population was on average younger than the other sub-populations (64 yrs, p Results: Lung volumes increased significantly at GOLD3-4 stages. QCT measurements of emphysema were significantly different between all GOLD stages, except between Gold-1 and Gold-2. Completeness of fissures in the right lung was significantly different between GOLD-4 and early COPD stages (RUL: Gold-4≠Gold1-3, p=0.009; RLL: Gold-4≠Gold-1, p=0.02; RH: Gold-4≠Gold-0, p=0.006). No significant changes were observed for FI in the left lung and emphysema heterogeneity across GOLD stages. Conclusion: Greater prevalence of incomplete fissure was associated with late stage COPD, and was only observed in the right lung. These findings have implications on the selection of therapeutic options for emphysema patients.