Classifying Future Healthcare Utilization in COPD Using Quantitative CT Lung Imaging and Two-Step Feature Selection via Sparse Subspace Learning with the CanCOLD Study.

RATIONALE: Although numerous candidate features exist for predicting risk of higher risk of healthcare utilization in patients with chronic obstructive pulmonary disease (COPD), the process for selecting the most discriminative features remains unclear. OBJECTIVE: The objective of this study was to develop a robust feature selection method to identify the most discriminative candidate features for predicting healthcare utilization in COPD, and compare the model performance with other common feature selection methods. MATERIALS AND METHODS: In this retrospective study, demographic, lung function measurements and CT images were collected from 454 COPD participants from the Canadian Cohort Obstructive Lung Disease study from 2010-2017. A follow-up visit was completed approximately 1.5 years later and participants reported healthcare utilization. CT analysis was performed for feature extraction. A two-step hybrid feature selection method was proposed that utilized: (1) sparse subspace learning with nonnegative matrix factorization, and, (2) genetic algorithm. Seven commonly used feature selection methods were also implemented that reported the top 10 or 20 features for comparison. Performance was evaluated using accuracy. RESULTS: Of the 454 COPD participants evaluated, 161 (35%) utilized healthcare services at follow-up. The accuracy for predicting subsequent healthcare utilization for the seven commonly used feature selection methods ranged from 72%-76% with the top 10 features, and 77%-80% with the top 20 features. Relative to these methods, hybrid feature selection obtained significantly higher accuracy for predicting subsequent healthcare utilization at 82% ± 3% (p < 0.05). Selected features with the proposed method included: DLCO, FEV1, RV, FVC, TAC, LAA950, Pi-10, LAA856, LAC total hole count, outer area RB1, wall area RB1, wall area and Jacobian. CONCLUSION: The hybrid feature selection method identified the most discriminative features for classifying individuals with and without future healthcare utilization, and increased the accuracy compared to other state-of-the-art approaches.

Abnormal Exertional Breathlessness on Cardiopulmonary Cycle Exercise Testing in Relation to Self-Reported and Physiological Responses in Chronic Airflow Limitation.

BACKGROUND: Exertional breathlessness is a cardinal symptom of cardiorespiratory disease. RESEARCH QUESTION: How does breathlessness abnormality, graded using normative reference equations during cardiopulmonary exercise testing (CPET), relate to self-reported and physiological responses in people with chronic airflow limitation (CAL)? STUDY DESIGN AND METHODS: An analysis was done of people aged ≥ 40 years with CAL undergoing CPET in the Canadian Cohort Obstructive Lung Disease study. Breathlessness intensity ratings (Borg CR10 scale [0-10 category-ratio scale for breathlessness intensity rating]) were evaluated in relation to power output (W), rate of oxygen uptake (V˙o2), and minute ventilation (V˙e) at peak exercise, using normative reference equations as follows: (1) probability of breathlessness normality (probability of having an equal or greater Borg CR10 rating among healthy; lower probability reflecting more severe breathlessness) and (2) presence of abnormal breathlessness (rating above the upper limit of normal). Associations with relevant participant-reported and physiologic outcomes were evaluated. RESULTS: We included 330 participants (44% women): mean ± SD age, 64 ± 10 years (range, 40-89 years); FEV1/FVC, 57.3% ± 8.2%; FEV1, 75.6% ± 17.9% predicted. Abnormally low exercise capacity (peak V˙o2 < lower limit of normal) was present in 26%. Relative to peak W, V˙o2, and V˙e, abnormally high breathlessness was present in 26%, 25%, and 18% of participants. For all equations, abnormally high exertional breathlessness was associated with worse lung function, exercise capacity, self-reported symptom burden, physical activity, and health-related quality of life; and greater physiological abnormalities during CPET. INTERPRETATION: Abnormal breathlessness graded using CPET normative reference equations was associated with worse clinical, physiological, and functional outcomes in people with CAL, supporting construct validity of abnormal exertional breathlessness.

Airway tree caliber heterogeneity and airflow obstruction among older adults.

INTRODUCTION: Smaller mean airway tree caliber is associated with airflow obstruction and chronic obstructive pulmonary disease (COPD). We investigated whether airway tree caliber heterogeneity was associated with airflow obstruction and COPD. METHODS: Two community-based cohorts (MESA Lung, CanCOLD) and a longitudinal case-control study of COPD (SPIROMICS) performed spirometry and computed tomography measurements of airway lumen diameters at standard anatomic locations and total lung volume. Percent-predicted airway lumen diameters were calculated using sex-specific reference equations accounting for age, height and lung volume. The association of airway tree caliber heterogeneity, quantified as the standard deviation (SD) of percent-predicted airway lumen diameters, with baseline forced expired volume in 1-second (FEV1), FEV1/forced vital capacity (FEV1/FVC) and COPD, as well as longitudinal spirometry, were assessed using regression models adjusted for age, sex, height, race-ethnicity, and mean airway tree caliber. RESULTS: Among 2,505 MESA Lung participants (mean±SD age: 69±9 years; 53% female, mean airway tree caliber: 99±10% predicted, airway tree caliber heterogeneity: 14±5%; median follow-up: 6.1 years), participants in the highest quartile of airway tree caliber heterogeneity exhibited lower FEV1 (adjusted mean difference: -125 ml, 95%CI:-171,-79), lower FEV1/FVC (adjusted mean difference: -0.01, 95%CI:-0.02,-0.01), and higher odds of COPD (adjusted OR 1.42, 95%CI:1.01-2.02) when compared with the lowest quartile, whereas longitudinal changes in FEV1 and FEV1/FVC did not differ significantly. Observations in CanCOLD and SPIROMICS were consistent. CONCLUSION: Among older adults, airway tree caliber heterogeneity was associated with airflow obstruction and COPD at baseline but was not associated with longitudinal changes in spirometry.

Physiological Characterization of Preserved Ratio Impaired Spirometry in the CanCOLD Study: Implications for Exertional Dyspnea and Exercise Intolerance.

RATIONALE: It is increasingly recognized that adults with preserved ratio impaired spirometry (PRISm) are prone to increased morbidity. However, the underlying pathophysiological mechanisms are unknown. OBJECTIVES: Evaluate the mechanisms of increased dyspnea and reduced exercise capacity in PRISm. METHODS: We completed a cross-sectional analysis of the CanCOLD population-based study. We compared physiological responses in 59 participants meeting PRISm spirometric criteria (post-bronchodilator FEV1<80% predicted and FEV1/FVC≥0.7), 264 controls, and 170 ever-smokers with chronic obstructive pulmonary disease (COPD), at rest and during cardiopulmonary exercise testing (CPET). MEASUREMENTS AND MAIN RESULTS: PRISm had lower total lung, vital and inspiratory capacities than controls (all p<0.05), and minimal small airway, pulmonary gas-exchange, and radiographic parenchymal lung abnormalities. Compared with control, PRISm had higher dyspnea/oxygen uptake [V̇O2] ratio at peak exercise (4.0±2.2vs2.9±1.9, Borg units/L/min, p<0.001) and lower V̇O2peak (74±22vs96±25% predicted, p<0.001). At standardized submaximal work rates, PRISm had greater tidal volume/inspiratory capacity (VT%IC, p<0.001), reflecting inspiratory mechanical constraint. In contrast to PRISm, COPD had characteristic small airways dysfunction, dynamic hyperinflation, and pulmonary gas-exchange abnormalities. Despite these physiological differences between the 3 groups, the relationship between increasing dyspnea and VT%IC during CPET was similar. Resting IC significantly correlated with V̇O2peak (r=0.65, p<0.001) in the entire sample, even after adjusting for airflow limitation, gas-trapping and diffusing capacity. CONCLUSION: In PRISm, lower exercise capacity and higher exertional dyspnea than healthy controls were mainly explained by lower resting lung volumes and earlier onset of dynamic inspiratory mechanical constraints at relatively low work rates.

A fully automated pipeline for the extraction of pectoralis muscle area from chest computed tomography scans.

Background: Computed tomography (CT)-derived pectoralis muscle area (PMA) measurements are prognostic in people with or at-risk of COPD, but fully automated PMA extraction has yet to be developed. Our objective was to develop and validate a PMA extraction pipeline that can automatically: 1) identify the aortic arch slice; and 2) perform pectoralis segmentation at that slice. Methods: CT images from the Canadian Cohort of Obstructive Lung Disease (CanCOLD) study were used for pipeline development. Aorta atlases were used to automatically identify the slice containing the aortic arch by group-based registration. A deep learning model was trained to segment the PMA. The pipeline was evaluated in comparison to manual segmentation. An external dataset was used to evaluate generalisability. Model performance was assessed using the Dice-Sorensen coefficient (DSC) and PMA error. Results: In total 90 participants were used for training (age 67.0±9.9 years; forced expiratory volume in 1 s (FEV1) 93±21% predicted; FEV1/forced vital capacity (FVC) 0.69±0.10; 47 men), and 32 for external testing (age 68.6±7.4 years; FEV1 65±17% predicted; FEV1/FVC 0.50±0.09; 16 men). Compared with manual segmentation, the deep learning model achieved a DSC of 0.94±0.02, 0.94±0.01 and 0.90±0.04 on the true aortic arch slice in the train, validation and external test sets, respectively. Automated aortic arch slice detection obtained distance errors of 1.2±1.3 mm and 1.6±1.5 mm on the train and test data, respectively. Fully automated PMA measurements were not different from manual segmentation (p>0.05). PMA measurements were different between people with and without COPD (p=0.01) and correlated with FEV1 % predicted (p<0.05). Conclusion: A fully automated CT PMA extraction pipeline was developed and validated for use in research and clinical practice.

Prevalence of chronic cough, its risk factors and population attributable risk in the Burden of Obstructive Lung Disease (BOLD) study: a multinational cross-sectional study.

Background: Chronic cough is a common respiratory symptom with an impact on daily activities and quality of life. Global prevalence data are scarce and derive mainly from European and Asian countries and studies with outcomes other than chronic cough. In this study, we aimed to estimate the prevalence of chronic cough across a large number of study sites as well as to identify its main risk factors using a standardised protocol and definition. Methods: We analysed cross-sectional data from 33,983 adults (≥40 years), recruited between Jan 2, 2003 and Dec 26, 2016, in 41 sites (34 countries) from the Burden of Obstructive Lung Disease (BOLD) study. We estimated the prevalence of chronic cough for each site accounting for sampling design. To identify risk factors, we conducted multivariable logistic regression analysis within each site and then pooled estimates using random-effects meta-analysis. We also calculated the population attributable risk (PAR) associated with each of the identifed risk factors. Findings: The prevalence of chronic cough varied from 3% in India (rural Pune) to 24% in the United States of America (Lexington,KY). Chronic cough was more common among females, both current and passive smokers, those working in a dusty job, those with a history of tuberculosis, those who were obese, those with a low level of education and those with hypertension or airflow limitation. The most influential risk factors were current smoking and working in a dusty job. Interpretation: Our findings suggested that the prevalence of chronic cough varies widely across sites in different world regions. Cigarette smoking and exposure to dust in the workplace are its major risk factors. Funding: Wellcome Trust.

Dysanapsis is not associated with exertional dyspnoea in healthy male and female never-smokers aged 40 years and older.

In healthy adults, airway-to-lung (i.e., dysanapsis) ratio is lower and dyspnoea during exercise at a given minute ventilation (V̇E) is higher in females than in males. We investigated the relationship between dysanapsis and sex on exertional dyspnoea in healthy adults. We hypothesized that females would have a smaller airway-to-lung ratio than males and that exertional dyspnoea would be associated with airway-to-lung ratio in males and females. We analyzed data from n = 100 healthy never-smokers aged ≥40 years enrolled in the Canadian Cohort Obstructive Lung Disease (CanCOLD) study who underwent pulmonary function testing, a chest computed tomography scan, and cardiopulmonary exercise testing. The luminal area of the trachea, right main bronchus, left main bronchus, right upper lobe, bronchus intermedius, left upper lobe, and left lower lobe were 22%-37% smaller (all p < 0.001) and the airway-to-lung ratio (i.e., average large conducting airway diameter relative to total lung capacity) was lower in females than in males (0.609 ± 0.070 vs. 0.674 ± 0.082; p < 0.001). During exercise, there was a significant effect of V̇E, sex, and their interaction on dyspnoea (all p < 0.05), indicating that dyspnoea increased as a function of V̇E to a greater extent in females than in males. However, after adjusting for age and total lung capacity, there were no significant associations between airway-to-lung ratio and measures of exertional dyspnoea, regardless of sex (all r < 0.34; all p > 0.05). Our findings suggest that sex differences in airway size do not contribute to sex differences in exertional dyspnoea.

Normative Reference Equations for Breathlessness Intensity during Incremental Cardiopulmonary Cycle Exercise Testing.

Rationale: Cardiopulmonary exercise testing (CPET) is the gold standard to evaluate exertional breathlessness, a common and disabling symptom. However, the interpretation of breathlessness responses to CPET is limited by a scarcity of normative data. Objectives: We aimed to develop normative reference equations for breathlessness intensity (Borg 0-10 category ratio) response in men and women aged ⩾40 years during CPET, in relation to power output (watts), oxygen uptake, and minute ventilation. Methods: Analysis of ostensibly healthy people aged ⩾40 years undergoing symptom-limited incremental cycle CPET (10 W/min) in the CanCOLD (Canadian Cohort Obstructive Lung Disease) study. Participants had smoking histories <5 pack-years and normal lung function and exercise capacity. The probability of each Borg 0-10 category ratio breathlessness intensity rating by power output, oxygen uptake, and minute ventilation (as an absolute or a relative value [percentage of predicted maximum]) was predicted using ordinal multinomial logistic regression. Model performance was evaluated by fit, calibration, and discrimination (C statistic) and externally validated in an independent sample (n = 86) of healthy Canadian adults. Results: We included 156 participants (43% women) from CanCOLD; the mean age was 65 (range, 42-91) years, and the mean body mass index was 26.3 (standard deviation, 3.8) kg/m2. Reference equations were developed for women and men separately, accounting for age and/or body mass. Model performance was high across all equations, including in the validation sample (C statistic for men = 0.81-0.92, C statistic for women = 0.81-0.96). Conclusions: Normative reference equations are provided to compare exertional breathlessness intensity ratings among individuals or groups and to identify and quantify abnormal breathlessness responses (scores greater than the upper limit of normal) during CPET.

The modified Borg/6-min walk distance ratio: a method to assess exertional breathlessness and leg discomfort using the 6-min walk test.

Background: The 6-min walk test (6MWT) is widely used to assess exercise capacity across chronic health conditions, but is currently not useful to assess symptoms, as the scores do not account for the 6-min walk distance (6MWD). We aimed to 1) develop normative reference equations for breathlessness and leg discomfort intensity expressed as modified Borg (mBorg)/6MWD ratios; and 2) validate the equations in people with COPD. Methods: Analysis of people aged ≥40 years who performed two 6MWTs (on a 20-m course) in the Canadian Cohort Obstructive Lung Disease (CanCOLD) study: a healthy cohort (n=291; mean±sd age 67.5±9.4 years; 54% male) with normal 6MWD and lung function, and a COPD cohort (n=156; age 66.2±9.0 years; 56% male; forced expiratory volume in 1 s (FEV1)/forced vital capacity 56.6±8.2%; FEV1 74.4±18.6% pred). The mBorg score was calculated as the Borg 0-10 category ratio intensity rating of breathlessness or leg discomfort recorded at the end of the 6MWT +1 (range 1-11), to avoid zeros and yield ratios proportional to the symptom score and 6MWD-1. Results: Using data from the healthy cohort, sex-specific normative reference equations for breathlessness and leg discomfort mBorg/6MWD ratios were developed using multivariable linear regression, accounting for age, and body mass or body mass index. In the COPD cohort, abnormal breathlessness and leg discomfort (mBorg/6MWD>upper limit of normal) showed strong concurrent validity with worse airflow limitation, Medical Research Council breathlessness and COPD Assessment Test scores. Conclusion: Normative references for the mBorg/6MWD ratio are presented to assess breathlessness and leg discomfort responses to the 6MWT in COPD.

CT Imaging With Machine Learning for Predicting Progression to COPD in Individuals at Risk.

BACKGROUND: Identifying individuals at risk of progressing to COPD may allow for initiation of treatment to potentially slow the progression of the disease or the selection of subgroups for discovery of novel interventions. RESEARCH QUESTION: Does the addition of CT imaging features, texture-based radiomic features, and established quantitative CT scan to conventional risk factors improve the performance for predicting progression to COPD in individuals who smoke with machine learning? STUDY DESIGN AND METHODS: Participants at risk (individuals who currently or formerly smoked, without COPD) from the Canadian Cohort Obstructive Lung Disease (CanCOLD) population-based study underwent CT imaging at baseline and spirometry at baseline and follow-up. Various combinations of CT scan features, texture-based CT scan radiomics (n = 95), and established quantitative CT scan (n = 8), as well as demographic (n = 5) and spirometry (n = 3) measurements, with machine learning algorithms were evaluated to predict progression to COPD. Performance metrics included the area under the receiver operating characteristic curve (AUC) to evaluate the models. DeLong test was used to compare the performance of the models. RESULTS: Among the 294 at-risk participants who were evaluated (mean age, 65.6 ± 9.2 years; 42% female; mean pack-years, 17.9 ± 18.7), 52 participants (23.7%) in the training data set and 17 participants (23.0%) in the testing data set progressed to spirometric COPD at follow-up (2.5 ± 0.9 years from baseline). Compared with machine learning models with demographics alone (AUC, 0.649), the addition of CT imaging features to demographics (AUC, 0.730; P < .05) or CT imaging features and spirometry to demographics (AUC, 0.877; P < .05) significantly improved the performance for predicting progression to COPD. INTERPRETATION: Heterogeneous structural changes occur in the lungs of individuals at risk that can be quantified using CT imaging features, and evaluation of these features together with conventional risk factors improves performance for predicting progression to COPD.

The association of spirometric small airways obstruction with respiratory symptoms, cardiometabolic diseases, and quality of life: results from the Burden of Obstructive Lung Disease (BOLD) study.

BACKGROUND: Spirometric small airways obstruction (SAO) is common in the general population. Whether spirometric SAO is associated with respiratory symptoms, cardiometabolic diseases, and quality of life (QoL) is unknown. METHODS: Using data from the Burden of Obstructive Lung Disease study (N = 21,594), we defined spirometric SAO as the mean forced expiratory flow rate between 25 and 75% of the FVC (FEF25-75) less than the lower limit of normal (LLN) or the forced expiratory volume in 3 s to FVC ratio (FEV3/FVC) less than the LLN. We analysed data on respiratory symptoms, cardiometabolic diseases, and QoL collected using standardised questionnaires. We assessed the associations with spirometric SAO using multivariable regression models, and pooled site estimates using random effects meta-analysis. We conducted identical analyses for isolated spirometric SAO (i.e. with FEV1/FVC ≥ LLN). RESULTS: Almost a fifth of the participants had spirometric SAO (19% for FEF25-75; 17% for FEV3/FVC). Using FEF25-75, spirometric SAO was associated with dyspnoea (OR = 2.16, 95% CI 1.77-2.70), chronic cough (OR = 2.56, 95% CI 2.08-3.15), chronic phlegm (OR = 2.29, 95% CI 1.77-4.05), wheeze (OR = 2.87, 95% CI 2.50-3.40) and cardiovascular disease (OR = 1.30, 95% CI 1.11-1.52), but not hypertension or diabetes. Spirometric SAO was associated with worse physical and mental QoL. These associations were similar for FEV3/FVC. Isolated spirometric SAO (10% for FEF25-75; 6% for FEV3/FVC), was also associated with respiratory symptoms and cardiovascular disease. CONCLUSION: Spirometric SAO is associated with respiratory symptoms, cardiovascular disease, and QoL. Consideration should be given to the measurement of FEF25-75 and FEV3/FVC, in addition to traditional spirometry parameters.

Factors Associated with Attrition in a Longitudinal Cohort of Older Adults in the Community.

Introduction: Retaining participants in longitudinal studies increases their power. We undertook this study in a population-based longitudinal cohort of adults with COPD to determine the factors associated with increased cohort attrition. Methods: In the longitudinal population-based Canadian Cohort of Obstructive Lung Disease (CanCOLD) study, 1561 adults > 40 years old were randomly recruited from 9 urban sites. Participants completed in-person visits at 18-month intervals and also were followed up every 3 months over the phone or by email. The cohort retention for the study and the reasons for attrition were analyzed. Hazard ratios and robust standard errors were calculated using Cox regression methods to explore the associations between participants who remained in the study and those who did not. Results: The median follow-up (years) of the study is 9.0 years. The overall mean retention was 77%. Reasons for attrition (23%) were: dropout by participant (39%), loss of contact (27%), investigator-initiated withdrawal (15%), deaths (9%), serious disease (9%), and relocation (2%). Factors independently associated with attrition were lower educational attainment, higher pack-year tobacco consumption, diagnosed cardiovascular disease, and a higher Hospital Anxiety and Depression Scale score: adjusted hazard ratios (95% confidence interval) were 1.43(1.11, 1.85); 1.01(1.00, 1.01); 1.44(1.13, 1.83); 1.06(1.02, 1.10) respectively. Conclusions: Identification and awareness of risk factors for attrition could direct targeted retention strategies in longitudinal studies. Moreover, the identification of patient characteristics associated with study dropout could address any potential bias introduced by differential dropouts.

Quantitative computed tomography and visual emphysema scores: association with lung function decline.

Background: Computed tomography (CT) visual emphysema score is a better predictor of mortality than single quantitative CT emphysema measurements in COPD, but there are numerous CT measurements that reflect COPD-related disease features. The purpose of this study was to determine if linear combinations of quantitative CT measurements by principal component analysis (PCA) have a greater association with forced expiratory volume in 1 s (FEV1) lower limit of normal (LLN) annualised change (ΔFEV1) than visual emphysema score in COPD. Methods: In this retrospective, longitudinal study, demographic, spirometry and CT images were acquired. CT visual emphysema score and quantitative analysis were performed; low attenuation area <950 HU (LAA950) and 12 other quantitative CT measurements were investigated. PCA was used for CT feature extraction. Multiple linear regression models for baseline FEV1 LLN and 6-year ΔFEV1 were used to determine associations with visual emphysema score and CT measurements. A total of 725 participants were analysed (n=299 never-smokers, n=242 at-risk and n=184 COPD). Results: Quantitative CT measures (LAA950 and PCA components) were independently statistically significant (p<0.05) in predicting baseline FEV1 LLN, whereas visual emphysema score was not statistically significant in any baseline model. When predicting 6-year ΔFEV1, only visual emphysema score was significant (p<0.05) in models with LAA950 and PCA combination of emphysema measurements. In the model with PCA using all CT measurements predicting 6-year ΔFEV1, visual emphysema score (p=0.021) along with one PCA component (p=0.004) were statistically significant. Conclusions: PCA with a combination of CT measurements reflecting several different COPD-related disease features independently predicted baseline lung function and increased the relative importance of quantitative CT compared with visual emphysema score for predicting lung function decline.

Impaired Spirometry and COPD Increase the Risk of Cardiovascular Disease: A Canadian Cohort Study.

BACKGROUND: Individuals with COPD and preserved ratio impaired spirometry (PRISm) findings in clinical settings have an increased risk of cardiovascular disease (CVD). RESEARCH QUESTION: Do individuals with mild to moderate or worse COPD and PRISm findings in community settings have a higher prevalence and incidence of CVD compared with individuals with normal spirometry findings? Can CVD risk scores be improved when impaired spirometry is added? STUDY DESIGN AND METHODS: The analysis was embedded in the Canadian Cohort Obstructive Lung Disease (CanCOLD). Prevalence of CVD (ischemic heart disease [IHD] and heart failure [HF]) and their incidence over 6.3 years were compared between groups with impaired and normal spirometry findings using logistic regression and Cox models, respectively, adjusting for covariables. Discrimination of the pooled cohort equations (PCE) and Framingham risk score (FRS) in predicting CVD were assessed with and without impaired spirometry. RESULTS: Participants (n = 1,561) included 726 people with normal spirometry findings and 835 people with impaired spirometry findings (COPD Global Initiative for Chronic Obstructive Lung Disease [GOLD] stage 1 disease, n = 408; GOLD stage ≥ 2, n = 331; PRISm findings, n = 96). Rates of undiagnosed COPD were 84% in GOLD stage 1 and 58% in GOLD stage ≥ 2 groups. Prevalence of CVD (IHD or HF) was significantly higher among individuals with impaired spirometry findings and COPD compared with those with normal spirometry findings, with ORs of 1.66 (95% CI, 1.13-2.43; P = .01∗) (∗ indicates statistical significane with P < .05) and 1.55 (95% CI, 1.04-2.31; P = .033∗), respectively. Prevalence of CVD was significantly higher in participants having PRISm findings and COPD GOLD stage ≥ 2, but not GOLD stage 1. CVD incidence was significantly higher, with hazard ratios of 2.07 (95% CI, 1.10-3.91; P = .024∗) for the impaired spirometry group and 2.09 (95% CI, 1.10-3.98; P = .024∗) for the COPD group compared to individuals with normal spirometry findings. The difference was significantly higher among individuals with COPD GOLD stage ≥ 2, but not GOLD stage 1. The discrimination for predicting CVD was low and limited when impaired spirometry findings were added to either risk score. INTERPRETATION: Individuals with impaired spirometry findings, especially those with moderate or worse COPD and PRISm findings, have increased comorbid CVD compared with their peers with normal spirometry findings, and having COPD increases the risk of CVD developing.

Quantitative CT Lung Imaging and Machine Learning Improves Prediction of Emergency Room Visits and Hospitalizations in COPD.

RATIONALE: Predicting increased risk of future healthcare utilization in chronic obstructive pulmonary disease (COPD) patients is an important goal for improving patient management. OBJECTIVE: Our objective was to determine the importance of computed tomography (CT) lung imaging measurements relative to other demographic and clinical measurements for predicting future health services use with machine learning in COPD. MATERIALS AND METHODS: In this retrospective study, lung function measurements and chest CT images were acquired from Canadian Cohort of Obstructive Lung Disease study participants from 2010 to 2017 (https://clinicaltrials.gov, NCT00920348). Up to two follow-up visits (1.5- and 3-year follow-up) were performed and participants were asked for details related to healthcare utilization. Healthcare utilization was defined as any COPD hospitalization or emergency room visit due to respiratory problems in the 12 months prior to the follow-up visits. CT analysis was performed (VIDA Diagnostics Inc.); a total of 108 CT quantitative emphysema, airway and vascular measurements were investigated. A hybrid feature selection method with support vector machine classifier was used to predict healthcare utilization. Performance was determined using accuracy, F1-measure and area under the receiver operating characteristic curve (AUC) and Matthews's correlation coefficient (MC). RESULTS: Of the 527 COPD participants evaluated, 179 (35%) used healthcare services at follow-up. There were no significant differences between the participants with or without healthcare utilization at follow-up for age (p = 0.50), sex (p = 0.44), BMI (p = 0.05) or pack-years (p = 0.76). The accuracy for predicting subsequent healthcare utilization was 80% ± 3% (F1-measure = 74%, AUC = 0.80, MC = 0.6) when all measurements were considered, 76% ± 6% (F1-measure = 72%, AUC = 0.77, MC = 0.55) for CT measurements alone and 65% ± 5% (F1-measure = 60%, AUC = 0.67, MC = 0.34) for demographic and lung function measurements alone. CONCLUSION: The combination of CT lung imaging and conventional measurements leads to greater prediction accuracy of subsequent health services use than conventional measurements alone, and may provide needed prognostic information for patients suffering from COPD.

Comparison of Feature Selection Methods and Machine Learning Classifiers for Predicting Chronic Obstructive Pulmonary Disease Using Texture-Based CT Lung Radiomic Features.

RATIONALE: Texture-based radiomics analysis of lung computed tomography (CT) images has been shown to predict chronic obstructive pulmonary disease (COPD) status using machine learning models. However, various approaches are used and it is unclear which provides the best performance. OBJECTIVES: To compare the most commonly used feature selection and classification methods and determine the optimal models for classifying COPD status in a mild, population-based COPD cohort. MATERIALS AND METHODS: CT images from the multi-center Canadian Cohort Obstructive Lung Disease (CanCOLD) study were pre-processed by resampling the image to a 1mm isotropic voxel volume, segmenting the lung and removing the airways (VIDA Diagnostics Inc.), and applying a threshold of -1000HU-to-0HU. A total of 95 texture features were then extracted from each CT image. Combinations of 17 feature selection methods and 9 classifiers were tested and evaluated. In addition, the role of data cleaning (outlier removal and highly correlated feature removal) was evaluated. The area under the curve (AUC) from the receiver operating characteristic curve was used to evaluate model performance. RESULTS: A total of 1204 participants were evaluated (n = 602 no COPD, n = 602 COPD). There were no significant differences between the groups for female sex (no COPD = 46.3%; COPD = 38.5%; p = 0.77), or body mass index (no COPD = 27.7 kg/m2; COPD = 27.4 kg/m2; p = 0.21). The highest AUC value for predicting COPD status (AUC = 0.78 [0.73, 0.84]) was obtained following data cleaning and feature selection using Elastic Net with the Linear-SVM classifier. CONCLUSION: In a population-based cohort, the optimal combination for radiomics-based prediction of COPD status was Elastic Net as the feature selection method and Linear-SVM as the classifier.

Physical Activity and Symptom Burden in COPD: The Canadian Obstructive Lung Disease Study.

Background: The relationship between symptom burden and physical activity (PA) in chronic obstructive pulmonary disease (COPD) remains poorly understood with limited data on undiagnosed individuals and those with mild to moderate disease. Objective: The primary objective was to evaluate the relationship between symptom burden and moderate-to-vigorous intensity PA (MVPA) in individuals from a random population-based sampling mirroring the population at large. Methods: Baseline participants of the Canadian Cohort Obstructive Lung Disease (n=1558) were selected for this cross-sectional sub-study. Participants with mild COPD (n=406) and moderate COPD (n=331), healthy individuals (n=347), and those at risk of developing COPD (n=474) were included. The Community Healthy Activities Model Program for Seniors (CHAMPS) questionnaire was used to estimate MVPA in terms of energy expenditure. High symptom burden was classified using the COPD Assessment Test ([CAT] ≥10). Results: Significant associations were demonstrated between high symptom burden and lower MVPA levels in the overall COPD sample (ß=-717.09; 95% confidence interval [CI]=-1079.78, -354.40; p<0.001) and in the moderate COPD subgroup (ß=-694.1; 95% CI=-1206.54, -181.66; p=0.006). A total of 72% of the participants with COPD were previously undiagnosed. The undiagnosed participants had significantly higher MVPA than those with physician diagnosed COPD (ß=-592.41 95% CI=-953.11, -231.71; p=0.001). Conclusion: MVPA was found to be inversely related to symptom burden in a large general population sample that included newly diagnosed individuals, most with mild to moderate COPD. Assessment of symptom burden may help identify patients with lower MVPA, especially for moderate COPD and for relatively inactive individuals with mild COPD.

Small airways obstruction and its risk factors in the Burden of Obstructive Lung Disease (BOLD) study: a multinational cross-sectional study.

BACKGROUND: Small airways obstruction is a common feature of obstructive lung diseases. Research is scarce on small airways obstruction, its global prevalence, and risk factors. We aimed to estimate the prevalence of small airways obstruction, examine the associated risk factors, and compare the findings for two different spirometry parameters. METHODS: The Burden of Obstructive Lung Disease study is a multinational cross-sectional study of 41 municipalities in 34 countries across all WHO regions. Adults aged 40 years or older who were not living in an institution were eligible to participate. To ensure a representative sample, participants were selected from a random sample of the population according to a predefined site-specific sampling strategy. We included participants' data in this study if they completed the core study questionnaire and had acceptable spirometry according to predefined quality criteria. We excluded participants with a contraindication for lung function testing. We defined small airways obstruction as either mean forced expiratory flow rate between 25% and 75% of the forced vital capacity (FEF25-75) less than the lower limit of normal or forced expiratory volume in 3 s to forced vital capacity ratio (FEV3/FVC ratio) less than the lower limit of normal. We estimated the prevalence of pre-bronchodilator (ie, before administration of 200 µg salbutamol) and post-bronchodilator (ie, after administration of 200 µg salbutamol) small airways obstruction for each site. To identify risk factors for small airways obstruction, we performed multivariable regression analyses within each site and pooled estimates using random-effects meta-analysis. FINDINGS: 36 618 participants were recruited between Jan 2, 2003, and Dec 26, 2016. Data were collected from participants at recruitment. Of the recruited participants, 28 604 participants had acceptable spirometry and completed the core study questionnaire. Data were available for 26 443 participants for FEV3/FVC ratio and 25 961 participants for FEF25-75. Of the 26 443 participants included, 12 490 were men and 13 953 were women. Prevalence of pre-bronchodilator small airways obstruction ranged from 5% (34 of 624 participants) in Tartu, Estonia, to 34% (189 of 555 participants) in Mysore, India, for FEF25-75, and for FEV3/FVC ratio it ranged from 5% (31 of 684) in Riyadh, Saudi Arabia, to 31% (287 of 924) in Salzburg, Austria. Prevalence of post-bronchodilator small airways obstruction was universally lower. Risk factors significantly associated with FEV3/FVC ratio less than the lower limit of normal included increasing age, low BMI, active and passive smoking, low level of education, working in a dusty job for more than 10 years, previous tuberculosis, and family history of chronic obstructive pulmonary disease. Results were similar for FEF25-75, except for increasing age, which was associated with reduced odds of small airways obstruction. INTERPRETATION: Despite the wide geographical variation, small airways obstruction is common and more prevalent than chronic airflow obstruction worldwide. Small airways obstruction shows the same risk factors as chronic airflow obstruction. However, further research is required to investigate whether small airways obstruction is also associated with respiratory symptoms and lung function decline. FUNDING: National Heart and Lung Institute and Wellcome Trust. TRANSLATIONS: For the Dutch, Estonian, French, Icelandic, Malay, Marathi, Norwegian, Portuguese, Swedish and Urdu translations of the abstract see Supplementary Materials section.