Comparison of the variability of the annual rates of change in FEV₁ determined from serial measurements of the pre- versus post-bronchodilator FEV₁ over 5 years in mild to moderate COPD: results of the lung health study

Spirometry test values can be estimated from a single chest radiograph

Introduction

Physical measurements of expiratory flow volume and speed can be obtained using spirometry. These measurements have been used for the diagnosis and risk assessment of chronic obstructive pulmonary disease and play a crucial role in delivering early care. However, spirometry is not performed frequently in routine clinical practice, thereby hindering the early detection of pulmonary function impairment. Chest radiographs (CXRs), though acquired frequently, are not used to measure pulmonary functional information. This study aimed to evaluate whether spirometry parameters can be estimated accurately from single frontal CXR without image findings using deep learning.

Methods

Forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), and FEV1/FVC as spirometry measurements as well as the corresponding chest radiographs of 11,837 participants were used in this study. The data were randomly allocated to the training, validation, and evaluation datasets at an 8:1:1 ratio. A deep learning network was pretrained using ImageNet. The input and output information were CXRs and spirometry test values, respectively. The training and evaluation of the deep learning network were performed separately for each parameter. The mean absolute error rate (MAPE) and Pearson's correlation coefficient (r) were used as the evaluation indices.

Results

The MAPEs between the spirometry measurements and AI estimates for FVC, FEV1 and FEV1/FVC were 7.59% (r = 0.910), 9.06% (r = 0.879) and 5.21% (r = 0.522), respectively. A strong positive correlation was observed between the measured and predicted indices of FVC and FEV1. The average accuracy of >90% was obtained in each estimation of spirometry indices. Bland-Altman analysis revealed good agreement between the estimated and measured values for FVC and FEV1.

Discussion

Frontal CXRs contain information related to pulmonary function, and AI estimation performed using frontal CXRs without image findings could accurately estimate spirometry values. The network proposed for estimating pulmonary function in this study could serve as a recommendation for performing spirometry or as an alternative method, suggesting its utility.

Lifetime Risk Factors for Pre- and Post-Bronchodilator Lung Function Decline. A Population-based Study

Rationale: Interactions between early life and adult insults on lung function decline are not well understood, with most studies investigating prebronchodilator (pre-BD) FEV₁ decline.Objectives: To investigate relationships between adult risk factors and pre- and post-BD lung function decline and their potential effect modification by early life and genetic factors.Methods: Multiple regression was used to examine associations between adult exposures (asthma, smoking, occupational exposures, traffic pollution, and obesity) and decline in both pre- and post-BD spirometry (forced expiratory volume in 1 s [FEV₁], forced vital capacity [FVC], and FEV₁/FVC) between ages 45 and 53 years in the Tasmanian Longitudinal Health Study (n = 857). Effect modification of these relationships by childhood respiratory risk factors, including low childhood lung function and GST (glutathione S-transferase) gene polymorphisms, was investigated.Results: Baseline asthma, smoking, occupational exposure to vapors/gases/dusts/fumes, and living close to traffic were associated with accelerated decline in both pre- and post-BD FEV₁. These factors were also associated with FEV₁/FVC decline. Occupational exposure to aromatic solvents was associated with pre-BD but not post-BD FEV₁ decline. Maternal smoking accentuated the effect of personal smoking on pre- and post-BD FEV₁ decline. Lower childhood lung function and having the GSTM1 null allele accentuated the effect of occupational exposure to vapors/gases/dusts/fumes and personal smoking on post-BD FEV₁ decline. Incident obesity was associated with accelerated decline in FEV₁ and more pronounced in FVC.Conclusions: This study provides new evidence for accentuation of individual susceptibility to adult risk factors by low childhood lung function, GSTM1 genotype, and maternal smoking.

Prediction of Obstructive Lung Disease from Chest Radiographs via Deep Learning Trained on Pulmonary Function Data

Background

Chronic obstructive pulmonary disease (COPD), the third leading cause of death worldwide, is often underdiagnosed.

Purpose

To develop machine learning methods to predict COPD using chest radiographs and a convolutional neural network (CNN) trained with near-concurrent pulmonary function test (PFT) data. Comparison is made to natural language processing (NLP) of the associated radiologist text reports.

Materials and Methods

This IRB-approved single-institution retrospective study uses 6749 two-view chest radiograph exams (2012–2017, 4436 unique subjects, 54% female, 46% male), same-day associated radiologist text reports, and PFT exams acquired within 180 days. The Image Model (Resnet18 pre-trained with ImageNet CNN) is trained using frontal and lateral radiographs and PFTs with 10% of the subjects for validation and 19% for testing. The NLP Model is trained using radiologist text reports and PFTs. The primary metric of model comparison is the area under the receiver operating characteristic curve (AUC).

Results

The Image Model achieves an AUC of 0.814 for prediction of obstructive lung disease (FEV1/FVC <0.7) from chest radiographs and performs better than the NLP Model (AUC 0.704, p<0.001) from radiologist text reports where FEV1 = forced expiratory volume in 1 second and FVC = forced vital capacity. The Image Model performs better for prediction of severe or very severe COPD (FEV1 <0.5) with an AUC of 0.837 versus the NLP model AUC of 0.770 (p<0.001).

Conclusion

A CNN Image Model trained on physiologic lung function data (PFTs) can be applied to chest radiographs for quantitative prediction of obstructive lung disease with good accuracy.

An Individualized Prediction Model for Long-term Lung Function Trajectory and Risk of COPD in the General Population

BACKGROUND

Prediction of future lung function will enable the identification of individuals at high risk of developing COPD, but the trajectory of lung function decline varies greatly among individuals. This study involved the development and validation of an individualized prediction model of lung function trajectory and risk of airflow limitation in the general population.

METHODS

Data were obtained from the Framingham Offspring Cohort, which included 4,167 participants ≥ 20 years of age and who had ≥ 2 valid spirometry assessments. The primary outcome was prebronchodilator FEV₁; the secondary outcome was the risk of airflow limitation (defined as FEV₁/FVC less than the lower limit of normal). Mixed effects regression models were developed for individualized prediction, and a machine learning algorithm was used to determine essential predictors. The model was validated in two large, independent multicenter cohorts (N = 2,075 and 12,913, respectively).

RESULTS

With 20 common predictors, the model explained 79% of the variation in FEV₁ decline in the derivation cohort. In two validation datasets, the model had low error in predicting FEV₁ decline (root mean square error range, 0.18-0.22 L) and high discriminative power in predicting risk of airflow limitation (C-statistic range, 0.86-0.87). This model was implemented in a freely accessible website-based application, which allows prediction based on flexible sets of predictors (http://resp.core.ubc.ca/ipress/FraminghamFEV1).

CONCLUSIONS

The individualized predictor is an accurate tool to predict long-term lung function trajectories and risk of airflow limitation in the general population. This model enables identifying individuals at higher risk of COPD, who can then be targeted for preventive therapies.

Individual FEV1 Trajectories Can Be Identified from a COPD Cohort

OBJECTIVE

We aim to make use of clinical spirometry data in order to identify individual COPD-patients with divergent trajectories of lung function over time.

STUDY DESIGN AND SETTING

Hospital-based COPD cohort (N = 607) was followed on average 4.6 years. Each patient had a mean of 8.4 spirometries available. We used a Hierarchical Bayesian Model (HBM) to identify the individuals presenting constant trends in lung function.

RESULTS

At a probability level of 95%, one third of the patients (180/607) presented rapidly declining FEV1 (mean -78 ml/year, 95% CI -73 to -83 ml) compared to that in the rest of the patients (mean -26 ml/year, 95% CI -23 to -29 ml, p ≤ 2.2 × 10(-16)). Constant improvement of FEV1 was very rare. The rapid decliners more frequently suffered from exacerbations measured by various outcome markers.

CONCLUSION

Clinical data of unique patients can be utilized to identify diverging trajectories of FEV1 with a high probability. Frequent exacerbations were more prevalent in FEV1-decliners than in the rest of the patients. The result confirmed previously reported association between FEV1 decline and exacerbation rate and further suggested that in clinical practice HBM could improve the identification of high-risk individuals at early stages of the disease.

Annual rates of change in pre- vs. post-bronchodilator FEV1 and FVC over 4 years in moderate to very severe COPD

While the slope of decline in FEV1 has traditionally been calculated from the post- rather than the pre-bronchodilator measurement in COPD interventional trials, it is not clear whether and to what extent these two slopes differ in symptomatic patients with COPD. Therefore, we used data from the 4-year UPLIFT trial of tiotropium 18 mcg QD vs. placebo to compare annual rates of change in pre- vs. post-bronchodilator FEV1 in 5041 patients with moderate to very severe COPD (mean FEV1 48% pred) in whom the post-bronchodilator FEV1 was measured after 4 inhalations of two different classes of short-acting inhaled bronchodilators at baseline and 1 month and every 6 months post-randomization over 4 years. Linear mixed effects models were used to estimate annual rates of decline in FEV1 and FVC pre- and post-bronchodilator in each treatment group separately, after adjusting for height, gender, smoking status, baseline % predicted FEV1 or FVC, and baseline acute % improvement in lung function. The slopes of the post-bronchodilator FEV1 and FVC were significantly steeper than the pre-bronchodilator slopes regardless of treatment arm (p < 0.001), while the estimated variances of the slopes were similar. Post-bronchodilator increases in FEV1 and FVC diminished progressively and significantly (p < 0.0001) over the 4-year trial, suggesting a possible explanation for the significant differences between the pre- and post-bronchodilator slopes. While the reasons for these differences are not completely clear, they are important to consider when assessing treatment effects on rates of decline in FEV1 and FVC.