Background: The study developed accurate explainable machine learning (ML) models for predicting first-time acute exacerbation of chronic obstructive pulmonary disease (COPD, AECOPD) at an individual level. Methods: We conducted a retrospective case–control study. A total of 606 patients with COPD were screened for eligibility using registry data from the COPD Pay-for-Performance Program (COPD P4P program) database at Changhua Christian Hospital between January 2017 and December 2019. Recursive feature elimination technology was used to select the optimal subset of features for predicting the occurrence of AECOPD. We developed four ML models to predict first-time AECOPD, and the highest-performing model was applied. Finally, an explainable approach based on ML and the SHapley Additive exPlanations (SHAP) and a local explanation method were used to evaluate the risk of AECOPD and to generate individual explanations of the model’s decisions. Results: The gradient boosting machine (GBM) and support vector machine (SVM) models exhibited superior discrimination ability (area under curve [AUC] = 0.833 [95% confidence interval (CI) 0.745–0.921] and AUC = 0.836 [95% CI 0.757–0.915], respectively). The decision curve analysis indicated that the GBM model exhibited a higher net benefit in distinguishing patients at high risk for AECOPD when the threshold probability was <0.55. The COPD Assessment Test (CAT) and the symptom of wheezing were the two most important features and exhibited the highest SHAP values, followed by monocyte count and white blood cell (WBC) count, coughing, red blood cell (RBC) count, breathing rate, oral long-acting bronchodilator use, chronic pulmonary disease (CPD), systolic blood pressure (SBP), and others. Higher CAT score; monocyte, WBC, and RBC counts; BMI; diastolic blood pressure (DBP); neutrophil-to-lymphocyte ratio; and eosinophil and lymphocyte counts were associated with AECOPD. The presence of symptoms (wheezing, dyspnea, coughing), chronic disease (CPD, congestive heart failure [CHF], sleep disorders, and pneumonia), and use of COPD medications (triple-therapy long-acting bronchodilators, short-acting bronchodilators, oral long-acting bronchodilators, and antibiotics) were also positively associated with AECOPD. A high breathing rate, heart rate, or systolic blood pressure and methylxanthine use were negatively correlated with AECOPD. Conclusions: The ML model was able to accurately assess the risk of AECOPD. The ML model combined with SHAP and the local explanation method were able to provide interpretable and visual explanations of individualized risk predictions, which may assist clinical physicians in understanding the effects of key features in the model and the model’s decision-making process.

Idiopathic pulmonary fibrosis (IPF) is characterized by fibrotic change in alveolar epithelial cells and leads to the irreversible deterioration of pulmonary function. Transforming growth factor-beta 1 (TGF-β1)-induced epithelial-mesenchymal transition (EMT) in type 2 lung epithelial cells contributes to excessive collagen deposition and plays an important role in IPF. Atractylodin (ATL) is a kind of herbal medicine that has been proven to protect intestinal inflammation and attenuate acute lung injury. Our study aimed to determine whether EMT played a crucial role in the pathogenesis of pulmonary fibrosis and whether EMT can be utilized as a therapeutic target by ATL treatment to mitigate IPF. To address this topic, we took two steps to investigate: 1. Utilization of anin vitro EMT model by treating alveolar epithelial cells (A549 cells) with TGF-β1 followed by ATL treatment for elucidating the underlying pathways, including Smad2/3 hyperphosphorylation, mitogen-activated protein kinase (MAPK) pathway overexpression, Snail and Slug upregulation, and loss of E-cadherin. Utilization of an in vivo lung injury model by treating bleomycin on mice followed by ATL treatment to demonstrate the therapeutic effectiveness, such as, less collagen deposition and lower E-cadherin expression. In conclusion, ATL attenuates TGF-β1-induced EMT in A549 cells and bleomycin-induced pulmonary fibrosis in mice.

Natural dietary components have become the subject of an increasing amount of interest due to the side effects of anticancer treatment. Pterostilbene, an analog of resveratrol, is primarily found in grapes, and has been suggested to exert antioxidant and anticancer effects in different tumor types. The present study aimed to investigate the antitumor effects and molecular mechanisms of pterostilbene in the human lung squamous cell carcinoma (SqCC) cell line, H520. The results of the present study indicate that pterostilbene significantly reduced cell viability and induced S phase arrest, and that treatment with pterostilbene was associated with the downregulation of cyclin A and cyclin E, as with the upregulation of p21 and p27 expression in H520 cells. In the apoptosis analysis, pterostilbene induced S phase accumulation and the activation of caspase‑3, ‑8 and ‑9 in H520 cells, potentially through the activation of extrinsic and intrinsic apoptotic pathways. Additionally, the in vivo study demonstrated that pterostilbene effectively inhibited lung SqCC growth in a H520 xenograft model. Given the in vitro and in vivo antitumor effects of pterostilbene demonstrated in the present study, pterostilbene may serve a novel and effective therapeutic agent to for patients with SqCC.

Background: Adult asthma is phenotypically heterogeneous with unclear aetiology. We aimed to evaluate the potential contribution of environmental exposure and its ensuing response to asthma and its heterogeneity.Methods: Environmental risk was evaluated by assessing the records of National Health Insurance Research Database (NHIRD) and residence-based air pollution (particulate matter with diameter less than 2.5 micrometers (PM_2.5) and PM_2.5-bound polycyclic aromatic hydrocarbons (PAHs)), integrating biomonitoring analysis of environmental pollutants, inflammatory markers and sphingolipid metabolites in case–control populations with mass spectrometry and ELISA. Phenotypic clustering was evaluated by t-distributed stochastic neighbor embedding (t-SNE) integrating 18 clinical and demographic variables.Findings: In the NHIRD dataset, modest increase in the relative risk with time-lag effect for emergency (N=209 837) and outpatient visits (N=638 538) was observed with increasing levels of PM_2.5and PAHs. Biomonitoring analysis revealed a panel of metals and organic pollutants, particularly metal Ni and PAH, posing a significant risk for current asthma (ORs=1.28–3.48) and its severity, correlating with the level of oxidative stress markers, notably Nε-(hexanoyl)-lysine (r=0.108–0.311, p<0.05), but not with the accumulated levels of PM_2.5exposure. Further, levels of circulating sphingosine-1-phosphate and ceramide-1-phosphate were found to discriminate asthma (p<0.001 and p<0.05, respectively), correlating with the levels of PAH (r=0.196, p<0.01) and metal exposure (r=0.202–0.323, p<0.05), respectively, and both correlating with circulating inflammatory markers (r=0.186–0.427, p<0.01). Analysis of six phenotypic clusters and those cases with comorbid type 2 diabetes mellitus (T2DM) revealed cluster-selective environmental risks and biosignatures.Interpretation: These results suggest the potential contribution of environmental factors from multiple sources, their ensuing oxidative stress and sphingolipid remodeling to adult asthma and its phenotypic heterogeneity.