Construction and Validation of an Impulse Oscillometry (IOS)-Based Platform for Screening, Diagnosis, and Management of Chronic Airway Diseases

Background Chronic obstructive pulmonary disease (COPD) and bronchial asthma are the most common chronic airway diseases, affecting approximately 600 million people worldwide. In China, the prevalence of COPD among adults aged 20 years or older is 8.6%, and the prevalence of asthma is 4.2%, representing a substantial disease burden. Despite increasing healthcare expenditures, key clinical outcomes have not improved accordingly. The prevention and management of these diseases face three major challenges: low rates of early diagnosis, non-standardized management, and weak diagnostic capacity at the primary care level.

Conventional spirometry, as the primary method for assessing respiratory function, relies on forced expiratory maneuvers and is therefore difficult to perform in elderly, pediatric, and severely ill patients. Moreover, it is relatively insensitive for detecting small airway dysfunction, an early pathological change in chronic airway diseases. Impulse oscillometry (IOS), the most widely used forced oscillation technique, measures respiratory system impedance-including resistance and reactance parameters-non-invasively based on the spectral relationship between pressure and airflow during tidal breathing. IOS is highly sensitive to small airway changes and requires minimal patient cooperation, making it a valuable complementary tool to spirometry.

Previous studies have demonstrated the clinical value of IOS parameters in the diagnosis and assessment of COPD and asthma. IOS parameters such as R5-R20, Fres, and AX correlate significantly with FEF25-75 measured by spirometry and show higher sensitivity. In subjects with chronic respiratory symptoms but preserved pulmonary function, IOS is more sensitive than spirometry for detecting small airway dysfunction. Recent domestic studies have further explored new IOS-based severity classification systems for COPD. However, the clinical application of IOS in China still faces key bottlenecks: the lack of diagnostic criteria and standardized pathways based on the Chinese population; unclear value in dynamic disease management and prognosis prediction; and the absence of high-level evidence on its effectiveness in real-world settings, especially in primary care. Therefore, systematic research on IOS technology and platform construction is urgently needed to promote its precise and standardized application.

Our previous work has explored the value of IOS in COPD screening, auxiliary diagnosis, and severity assessment. We first constructed an IOS-based screening pathway for COPD using a prediction model for the FEV1/FVC ratio, published in The Lancet Regional Health - Western Pacific, and applied for a corresponding invention patent (202411415494.X). In asthma patients, we were the first to discover the effectiveness of combining IOS with FeNO in identifying patients with inadequately controlled asthma. These preliminary studies provide a key methodological and evidentiary basis for the systematic construction of a "screening-diagnosis-management" platform in this project.

Study Objectives To construct and validate IOS-based diagnostic models for chronic airway diseases: Participants will be enrolled from respiratory and critical care medicine outpatient clinics and physical examination centers across multiple centers. Using respiratory specialists' clinical diagnosis as the gold standard, diagnostic models for COPD and asthma based on IOS parameters will be established, optimal diagnostic thresholds will be determined, and diagnostic performance indicators including sensitivity, specificity, and area under the curve (AUC) will be systematically evaluated. In addition, external validation of the screening models for COPD and asthma will be conducted in multicenter community populations.

To construct IOS-based prognostic warning models for chronic airway diseases: In patients with confirmed stable COPD or asthma, a 48-week prospective cohort follow-up will be conducted to analyze the association between IOS parameters and future acute exacerbation risk, and to construct a Cox proportional hazards regression prediction model for dynamic management and prognostic warning of chronic airway diseases.

Study Design This is a multicenter, observational study consisting of two phases. Phase 1: Construction and Validation of IOS-Based Screening and Diagnostic Models for Chronic Airway Diseases This phase is a multicenter, cross-sectional study. Participants aged 18-80 years will be enrolled from respiratory and critical care medicine outpatient clinics, physical examination centers, and community health centers in multiple regions across China (Northeast, Central South, East China, Northwest, and Southwest). Using respiratory specialists' final clinical diagnosis (COPD/asthma/other) as the gold standard, which is based on symptoms, spirometry results (including bronchodilation/bronchial challenge tests), and other clinical indicators, IOS-based diagnostic models for COPD and asthma will be constructed. Data collected include demographic information, clinical information, screening questionnaires (COPD-PS, COPD-SQ, ECRHS), IOS parameters (R5, R20, R5-R20, AX, Fres), spirometry parameters (FEV1, FVC, FEV1/FVC, MMEF), and inflammatory biomarkers (blood eosinophil count, total IgE, FeNO). Feature selection will be performed using LASSO regression or stepwise logistic regression. Multivariable logistic regression will be used to construct diagnostic models, and a visual nomogram will be developed to facilitate clinical decision-making. Model performance will be evaluated in a multicenter external validation set using ROC curves, calibration curves, and decision curve analysis (DCA). The optimal diagnostic thresholds will be determined, and the models will be compared with traditional screening questionnaires.

Phase 2: Construction of an IOS-Based Dynamic Management and Prognostic Warning Platform for Chronic Airway Diseases This phase is a multicenter, prospective cohort study. Patients aged 18-80 years with confirmed stable COPD (meeting GOLD criteria) or asthma (meeting GINA criteria) will be enrolled. Stable disease is defined as no acute exacerbation within the past 4 weeks, without the need for systemic glucocorticoids or increased routine medication. Participants will be followed up every 12 weeks for 48 weeks. Data collected at each visit include respiratory symptoms, acute exacerbation events (frequency, severity, emergency department visits, and hospitalizations), new comorbidities, mortality, and treatment adjustments. Symptom and quality of life assessments will be performed using mMRC, CAT, ACQ, AQLQ, and SGRQ. The primary outcome is acute exacerbation events, defined as symptom deterioration requiring systemic glucocorticoids, antibiotics, emergency department visits, or hospitalization. Secondary outcomes include annual decline in lung function, changes in quality of life scores, and all-cause and respiratory-related mortality. Cox proportional hazards regression models will be used to analyze the association between IOS parameters and future acute exacerbation risk and to construct a prognostic prediction model. Model performance will be evaluated using the C-index, calibration curves, and time-dependent ROC curves for predicting exacerbation risk at 3, 6, and 12 months.

Study Population Phase 1: Participants aged 18-80 years from respiratory clinics, physical examination centers, and community health centers across China. Inclusion criteria: clear consciousness, basic communication and comprehension ability, able to cooperate with questionnaires and IOS/spirometry examinations, and willing to sign written informed consent. Exclusion criteria: respiratory infection within the past 4 weeks; other severe pulmonary diseases affecting lung function (e.g., active tuberculosis, extensive bronchiectasis, pulmonary fibrosis, chest wall deformity, lung cancer); severe dysfunction of heart, brain, liver, or kidney (e.g., NYHA class III-IV, severe arrhythmia, uremia) or uncontrolled malignancy; cognitive impairment or psychiatric disorders; pregnancy or lactation; contraindications to pulmonary function tests.

Phase 2: Patients aged 18-80 years with confirmed stable COPD (GOLD criteria) or asthma (GINA criteria). Exclusion criteria: comorbid severe respiratory diseases (e.g., bronchiectasis, pulmonary fibrosis, lung cancer, active tuberculosis); severe heart, liver, or kidney dysfunction (e.g., NYHA class III-IV, end-stage liver disease, dialysis-dependent renal disease); conditions affecting pulmonary function test accuracy (e.g., chest wall deformity, pneumothorax, recent thoracic or abdominal surgery); severe cognitive impairment, psychiatric disorders, or communication barriers; planned relocation within 1 year or inability to complete regular follow-up; pregnancy or lactation.

Sample Size A total of 2,800 participants will be enrolled. Phase 1: Approximately 1,800 participants (600 from hospitals and 1,200 from communities). For the diagnostic model, 14 predictor variables will be included. Based on the principle of subjects per variable (SPV) > 10, the training set requires at least 140 participants. Assuming a COPD and asthma prevalence of approximately 25% in the hospital population, a total of 600 participants will be needed from hospital outpatient and physical examination centers (at least 150 COPD and 150 asthma patients). For external validation in the community, based on the confidence interval width of the AUC, with a 95% confidence level (two-sided) and a required 95% CI width not exceeding 0.15, and assuming an AUC of 0.80 for both COPD and asthma, at least 55 COPD patients (corresponding to approximately 400 participants) and 50 asthma patients (corresponding to approximately 1,200 participants) will be needed. Therefore, a total of approximately 1,200 community participants will be required for external validation.

Phase 2: Approximately 1,000 participants. Based on the requirements for constructing a Cox proportional hazards regression model with 14 predictor variables and SPV > 10, the training set requires at least 140 participants. Assuming an annual exacerbation rate of 40% for COPD and 30% for asthma, with a 15% dropout rate, the required sample size for COPD patients is (140/0.40) × 1.15 ≈ 400 (at least 160 with exacerbations), and for asthma patients is (140/0.30) × 1.15 ≈ 550 (at least 165 with exacerbations). Therefore, a total of 1,000 patients with confirmed COPD or asthma will be enrolled.

Measurements IOS parameters: R5, R20, R5-R20, AX, Fres. Spirometry parameters: FEV1, FVC, FEV1/FVC, MMEF. Inflammatory biomarkers: blood eosinophil count, total IgE, FeNO. Questionnaires: COPD-PS, COPD-SQ, ECRHS, mMRC, CAT, ACQ, AQLQ, SGRQ. Imaging: chest HRCT (inspiratory and expiratory phases) in Phase 2. Outcomes: acute exacerbation events, lung function decline, quality of life changes, and mortality.

Statistical Analysis Phase 1: Descriptive statistics, correlation analysis (Pearson/Spearman), multiple imputation for missing data, LASSO or stepwise logistic regression for feature selection, multivariable logistic regression for model construction, nomogram development, ROC curve analysis (AUC, 95% CI, Youden index, sensitivity, specificity, PPV, NPV), calibration curve with Hosmer-Lemeshow test, decision curve analysis (DCA), and DeLong test for model comparison.

Phase 2: Descriptive statistics, Spearman correlation or Kruskal-Wallis test for clinical association analysis, LASSO regression for feature selection, Cox proportional hazards regression for prognostic model construction, C-index for discrimination, calibration curve for calibration, and time-dependent ROC curves for predicting exacerbation risk at 3, 6, and 12 months. R software (version 4.3.0 or above) will be used. All tests will be two-sided, and P \< 0.05 will be considered statistically significant.

Expected Outcomes Establishment of IOS-based diagnostic models for chronic airway diseases, with optimal diagnostic thresholds and performance indicators for COPD and asthma in the Chinese population.

Construction of an IOS-based dynamic management and prognostic warning platform, including a Cox regression risk prediction model for acute exacerbations, forming a new strategy covering the full cycle of "screening-diagnosis-warning" for chronic airway diseases.

Generation of high-quality, localized evidence to support the clinical promotion of domestic portable IOS devices, technological iteration, and the formulation of related health policies (e.g., primary care respiratory disease screening standards and device configuration criteria).