COPD profiles and treatable traits using minimal resources: identification, decision tree and stability over time

Acute exacerbation prediction of COPD based on Auto-metric graph neural network with inspiratory and expiratory chest CT images

Chronic obstructive pulmonary disease (COPD) is a widely prevalent disease with significant mortality and disability rates and has become the third leading cause of death globally. Patients with acute exacerbation of COPD (AECOPD) often substantially suffer deterioration and death. Therefore, COPD patients deserve special consideration regarding treatment in this fragile population for pre-clinical health management. Based on the above, this paper proposes an AECOPD prediction model based on the Auto-Metric Graph Neural Network (AMGNN) using inspiratory and expiratory chest low-dose CT images. This study was approved by the ethics committee in the First Affiliated Hospital of Guangzhou Medical University. Subsequently, 202 COPD patients with inspiratory and expiratory chest CT Images and their annual number of AECOPD were collected after the exclusion. First, the inspiratory and expiratory lung parenchyma images of the 202 COPD patients are extracted using a trained ResU-Net. Then, inspiratory and expiratory lung Radiomics and CNN features are extracted from the 202 inspiratory and expiratory lung parenchyma images by Pyradiomics and pre-trained Med3D (a heterogeneous 3D network), respectively. Last, Radiomics and CNN features are combined and then further selected by the Lasso algorithm and generalized linear model for determining node features and risk factors of AMGNN, and then the AECOPD prediction model is established. Compared to related models, the proposed model performs best, achieving an accuracy of 0.944, precision of 0.950, F1-score of 0.944, ad area under the curve of 0.965. Therefore, it is concluded that our model may become an effective tool for AECOPD prediction.

Demystification of artificial intelligence for respiratory clinicians managing patients with obstructive lung diseases

INTRODUCTION

Asthma and chronic obstructive pulmonary disease (COPD) are leading causes of morbidity and mortality worldwide. Despite all available diagnostics and treatments, these conditions pose a significant individual, economic and social burden. Artificial intelligence (AI) promises to support clinical decision-making processes by optimizing diagnosis and treatment strategies of these heterogeneous and complex chronic respiratory diseases. Its capabilities extend to predicting exacerbation risk, disease progression and mortality, providing healthcare professionals with valuable insights for more effective care. Nevertheless, the knowledge gap between respiratory clinicians and data scientists remains a major constraint for wide application of AI and may hinder future progress. This narrative review aims to bridge this gap and encourage AI deployment by explaining its methodology and added value in asthma and COPD diagnosis and treatment.

AREAS COVERED

This review offers an overview of the fundamental concepts of AI and machine learning, outlines the key steps in building a model, provides examples of their applicability in asthma and COPD care, and discusses barriers to their implementation.

EXPERT OPINION

Machine learning can advance our understanding of asthma and COPD, enabling personalized therapy and better outcomes. Further research and validation are needed to ensure the development of clinically meaningful and generalizable models.

Can Treatable Traits Be the Approach to Addressing the Complexity and Heterogeneity of COPD?

The complexity of COPD implies the need to identify groups of patients with similar clinical characteristics and prognosis or treatment requirements. This is why much attention has been paid to identifying the different clinical phenotypes by investigating the clinical expression of the disease, and endotypes by studying the biological networks that enable and limit reactions. However, this approach is complicated because one endotype gives rise to one or more clinical characteristics, and clinical phenotypes can be derived from several endotypes. To simplify the approach, a new taxonomic classification of COPD based on the different causes (or etiotypes) has been proposed, but these etiotypes have not yet been validated. A simpler method is the so-called tractable traits approach, which is free from any designation of the disorder to be treated and does not present the criticality of using etiotypes. A large randomised controlled trial on using the treatable traits approach in COPD is still lacking. Nevertheless, this approach is already applied by following the GOLD strategy. However, its application is complicated because several potentially treatable traits have been identified within the pulmonary domain, the extrapulmonary domain, and the behavioural/risk factor domain. In addition, the hierarchy of the dominant treatable traits has not yet been established, and they change over time both spontaneously and because of treatment. This means that the patients being treated according to the tractable traits approach must be constantly followed over time so that the therapy is focused on their temporal needs.

An Update on Outcomes for COPD Pharmacological Trials: A COPD Investigators Report - Reassessment of the 2008 American Thoracic Society/European Respiratory Society Statement on Outcomes for COPD Pharmacological Trials

Background: In 2008, a dedicated American Thoracic Society/European Respiratory Society task force published a paper on the possible use and limitations of clinical outcomes and biomarkers to evaluate the impact of pharmacological therapy in patients with chronic obstructive pulmonary disease. Since then, our scientific understanding of chronic obstructive pulmonary disease has increased considerably; there has been a progressive shift from a one-size-fits-all diagnostic and therapeutic approach to a personalized approach; and many new treatments currently in development will require new endpoints to evaluate their efficacy adequately. Objectives: The emergence of several new relevant outcome measures motivated the authors to review advances in the field and highlight the need to update the content of the original report. Methods: The authors separately created search strategies for the literature, primarily based on their opinions and assessments supported by carefully chosen references. No centralized examination of the literature or uniform criteria for including or excluding evidence were used. Measurements and Main Results: Endpoints, outcomes, and biomarkers have been revisited. The limitations of some of those reported in the American Thoracic Society/European Respiratory Society task force document have been highlighted. In addition, new tools that may be useful, especially in evaluating personalized therapy, have been described. Conclusions: Because the "label-free" treatable traits approach is becoming an important step toward precision medicine, future clinical trials should focus on highly prevalent treatable traits, and this will influence the choice of outcomes and markers to be considered. The use of the new tools, particularly combination endpoints, could help better identify the right patients to be treated with the new drugs.

Unleashing the Power of Very Small Data to Predict Acute Exacerbations of Chronic Obstructive Pulmonary Disease

Introduction

In this article, we explore to what extent it is possible to leverage on very small data to build machine learning (ML) models that predict acute exacerbations of chronic obstructive pulmonary disease (AECOPD).

Methods

We build ML models using the small data collected during the eHealth Diary telemonitoring study between 2013 and 2017 in Sweden. This data refers to a group of multimorbid patients, namely 18 patients with chronic obstructive pulmonary disease (COPD) as the major reason behind previous hospitalisations. The telemonitoring was supervised by a specialised hospital-based home care (HBHC) unit, which also was responsible for the medical actions needed.

Results

We implement two different ML approaches, one based on time-dependent covariates and the other one based on time-independent covariates. We compare the first approach with standard COX Proportional Hazards (CPH). For the second one, we use different proportions of synthetic data to build models and then evaluate the best model against authentic data.

Discussion

To the best of our knowledge, the present ML study shows for the first time that the most important variable for an increased risk of future AECOPDs is "maintenance medication changes by HBHC". This finding is clinically relevant since a sub-optimal maintenance treatment, requiring medication changes, puts the patient in risk for future AECOPDs.

Conclusion

The experiments return useful insights about the use of small data for ML.

Study protocol: pneumonia and inhaled corticosteroid treatment patterns in chronic obstructive pulmonary disease - a cohort study using sequence analysis (PICCS)

INTRODUCTION

Treatment with inhaled corticosteroids (ICS) is a widely used treatment in chronic obstructive pulmonary disease. The main effects include a reduction in the number of exacerbations and, for some patients, an increase in expected mortality. Unfortunately, the treatment is also linked to an increased risk of pneumonia, and very little is known about which patients experience this increased risk. There is a need for identification of patient characteristics associated with increased risk of pneumonia and treatment with ICS.

METHODS AND ANALYSIS

This is a register-based cohort study that uses the nationwide Danish registers. Data from several registers in the years 2008-2018 will be merged on an individual level using the personal identification numbers that are unique to every citizen in Denmark. Clusters based on pneumonia incidence and ICS treatment patterns will be explored with a sequence analysis in a 3-year follow-up period.

ETHICS AND DISSEMINATION

This is a register-based study and research ethics approval is not required according to Danish Law and National Ethics Committee Guidelines. The results will be submitted to peer-reviewed journals and reported at appropriate national and international meetings.

Treatable Traits in Airway Disease: From Theory to Practice

Chronic airway diseases such as asthma and chronic obstructive pulmonary disease are prevalent and complex conditions that often coexist in the same patient. To address this complexity in clinical practice, and to move forward toward personalized and precision medicine of airway diseases, a strategy based on the identification and treatment of so-called "treatable traits" (TTs) has been proposed. A TT is a recognizable phenotypic or endotypic characteristic that can be assessed and successfully targeted by therapy to improve a clinical outcome in a patient with airway disease. Importantly, TTs can coexist in the same patient, so they are not mutually exclusive. The TT strategy proposes to investigate in each individual patient with chronic airway disease the number and type of TTs present and to treat each of them according to guideline recommendations. This strategy is agnostic (ie, independent) to the traditional diagnostic labels (asthma, chronic obstructive pulmonary disease), so it can be applied to any patient with airway disease. Currently, there is firm evidence supporting the adequacy and validity of the TT strategy. Here, we review the current state of the art of this topic, first by presenting its theoretical background and then by discussing how to best implement it in clinical practice.

Impulse Oscillometry as a Diagnostic Test for Pulmonary Emphysema in a Clinical Setting

Body plethysmography (BP) is the standard pulmonary function test (PFT) in pulmonary emphysema diagnosis, but not all patients can cooperate to this procedure. An alternative PFT, impulse oscillometry (IOS), has not been investigated in emphysema diagnosis. We investigated the diagnostic accuracy of IOS in the diagnosis of emphysema. Eighty-eight patients from the pulmonary outpatient clinic at Lillebaelt Hospital, Vejle, Denmark, were included in this cross-sectional study. A BP and an IOS were performed in all patients. Computed tomography scan verified presence of emphysema in 20 patients. The diagnostic accuracy of BP and IOS for emphysema was evaluated with two multivariable logistic regression models: Model 1 (BP variables) and Model 2 (IOS variables). Model 1 had a cross-validated area under the ROC curve (CV-AUC) = 0.892 (95% CI: 0.654-0.943), a positive predictive value (PPV) = 59.3%, and a negative predictive value (NPV) = 95.0%. Model 2 had a CV-AUC = 0.839 (95% CI: 0.688-0.931), a PPV = 55.2%, and an NPV = 93.7%. We found no statistically significant difference between the AUC of the two models. IOS is quick and easy to perform, and it can be used as a reliable rule-out method for emphysema.

Lung Inflammation in alpha-1-antitrypsin deficient individuals with normal lung function

BACKGROUND

Alpha-1-antitrypsin deficient (AATD) individuals are prone to develop early age of onset chronic obstructive pulmonary disease (COPD) more severe than non-genetic COPD. Here, we investigated the characteristics of lower respiratory tract of AATD individuals prior to the onset of clinically significant COPD.

METHODS

Bronchoalveolar lavage was performed on 22 AATD with normal lung function and 14 healthy individuals. Cell counts and concentrations of proteases, alpha-1-antitrypsin and proinflammatory mediators were determined in the bronchoalveolar lavage fluid from study subjects. In order to determine the airway inflammation, we also analyzed immune cell components of the large airways from bronchial biopsies using immunohistochemistry in both study subjects. Finally, we made comparisons between airway inflammation and lung function rate of decline using four repeated lung function tests over one year in AATD individuals.

RESULTS

AATD individuals with normal lung function had 3 folds higher neutrophil counts, 2 folds increase in the proteases levels, and 2-4 folds higher levels of IL-8, IL-6, IL-1β, and leukotriene B4 in their epithelial lining fluid compared to controls. Neutrophil elastase levels showed a positive correlation with the levels of IL-8 and neutrophils in AATD epithelial lining fluid. AATD individuals also showed a negative correlation of baseline FEV1 with neutrophil count, neutrophil elastase, and cytokine levels in epithelial lining fluid (p < 0.05). In addition, we observed twofold increase in the number of lymphocytes, macrophages, neutrophils, and mast cells of AATD epithelial lining fluid as compared to controls.

CONCLUSION

Mild inflammation is present in the lower respiratory tract and airways of AATD individuals despite having normal lung function. A declining trend was also noticed in the lung function of AATD individuals which was correlated with pro-inflammatory phenotype of their lower respiratory tract. This results suggest the presence of proinflammatory phenotype in AATD lungs. Therefore, early anti-inflammatory therapies may be a potential strategy to prevent progression of lung disease in AATD individuals.

The presence of extra-pulmonary treatable traits increases the likelihood of responding to pulmonary rehabilitation

INTRODUCTION

Studies suggest that people with chronic obstructive pulmonary disease (COPD) who are worse at baseline respond better to pulmonary rehabilitation (PR). Identifying treatable traits (TTs) may help to distinguish responders from non-responders. We explored the impact of PR on extra-pulmonary traits of people with COPD and whether the presence of TT influences the type of response to PR.

METHODS

A comprehensive assessment of 9 TT including symptoms (dyspnoea, fatigue, anxiety and depression), functional capacity, deconditioning, balance, impact of the disease and health-related quality of life was conducted before and after a 12-week community-based PR programme. Pre-post differences between people with or without each TT at baseline were compared with independent samples t-tests or Mann-Whitney U tests. Proportion of responders between groups were explored with chi-square tests and odds ratio.

RESULTS

102 people with COPD were included (70 [65; 75] years old, 78% male, FEV₁ 47 [36; 60] %predicted). They had a median of 3 (out of 9) TTs per person and each patient responded on average to 5 (out of 9) outcomes of PR. People with TT were more responsive than those without them in all outcomes (p < 0.05) except for the 1-min sit-to-stand test. The presence of TT increased 4 to 20 times the likelihood of being a good responder.

CONCLUSIONS

Identification of baseline extra-pulmonary TT in people with COPD showed the potential to inform on PR responsiveness and might therefore be an important strategy for patient prioritization, treatment personalisation (i.e., activation of the most suitable components) and optimisation.