Construction and Validation of A Nomogram to Identify Mucus Obstruction In Patients With Chronic Obstructive Pulmonary Disease

Summary

This study aimed to identify independent clinical predictors of computed tomography (CT)–detected small airway mucus plugs in patients with chronic obstructive pulmonary disease (COPD) and to construct and validate a nomogram for individualized risk prediction.

Abstract

Small airway mucus impaction in chest computed tomography (CT) is a clinically significant finding in chronic obstructive pulmonary disease (COPD), associated with accelerated pulmonary function decline, increased frequency of acute exacerbations, and higher susceptibility to respiratory infections. However, a validated predictive tool for identifying patients at risk of CT-detected mucus plugs is currently lacking. This study aimed to develop and validate a nomogram to predict small airway mucus obstruction in patients with COPD. We retrospectively enrolled 212 COPD patients from Shenzhen Second People’s Hospital (January 2021 to June 2022), of whom 47 had CT-confirmed mucus plugs (mucus plug group, MP) and 165 did not (non-mucus plug group, NMP). Univariate and receiver operating characteristic (ROC) analyses were used to identify candidate predictors. Multivariate logistic regression was conducted to construct the final predictive model, which was then transformed into a nomogram. Internal validation was performed using bootstrap sampling (1000 iterations). Bronchiectasis, chronic rhinosinusitis (CRS), body mass index (BMI), forced expiratory flow at 25–75% of predicted (FEF25–75%pred), residual volume-to-total lung capacity ratio (RV/TLC), and serum 25-hydroxyvitamin D [25(OH)D] were identified as independent risk factors for CT mucus plugs. The nomogram demonstrated excellent predictive value with an AUC of 0.9611. Calibration curves and decision curve analyses demonstrated good clinical utility. Bootstrap internal validation further supported the model’s predictive stability. This nomogram provides a practical, individualized tool to facilitate early identification and personalized management of COPD patients at risk of small-airway mucus obstruction.

Introduction

Chronic obstructive pulmonary disease (COPD) is characterized by persistent and largely irreversible airflow limitation. The World Health Organization states that it is projected to become the third leading cause of death globally by 2030¹. The disease primarily initiates in the small airways (airways with an internal diameter of less than 2 mm), which represent a fundamental site of COPD pathology. Structural and inflammatory changes in these regions often precede the emergence of clinical symptoms by several years, yet contribute substantially to the airflow obstruction. Pathological hallmarks of small airway disease in COPD include infiltration by inflammatory cells,^2,3,4 impairment of epithelial defense mechanisms^5,6 airway remodeling and fibrosis^7,8,9and the formation of mucus plugs (MP)^10,11.

Airway mucus plugs in COPD represent a pathological accumulation of mucus within the airway lumen, resulting in airflow limitation¹². Mucus plug formation is associated with a pro-inflammatory milieu, characterized by elevated eosinophil counts and upregulation of type 2 cytokine gene expression¹³. Excessive intraluminal mucus impairs oxygen diffusion and causes hypoxia in airway epithelial cells, creating conditions favorable to persistent bacterial colonization and recurrent lower respiratory tract infections¹⁴. These infections exacerbate disease severity and increase mortality risk¹⁵. Elevated airway mucus secretion has further been identified as a precursor to acute exacerbation events in COPD¹⁶. This highlights the critical need for early detection and a mechanistic understanding of the factors contributing to mucus plugs in patients with COPD.

A range of risk factors have been associated with airway mucus plug formation in chronic airway diseases, including viral infections^17,18, colonization by Pseudomonas aeruginosa^19,20 recurrent acute exacerbation episodes, impaired pulmonary function as measured by forced expiratory volume in 1 second (FEV1)²¹, smoking history²², elevated eosinophil peroxidase levels²³, intrabronchial mucin 5B (MUC5B) protein concentrations, and 25-hydroxyvitamin D (25(OH)D) levels, as well as infections attributable to mycoplasma and Aspergillus. species^24,25,26. Nevertheless, the specific risk profile for mucus plug development in COPD patients remains incompletely characterized, and the prognostic utility of individual risk factors in isolation is limited.

A multifactorial approach integrating several predictors may yield more clinically meaningful risk stratification. Nomograms have been widely applied across medical specialties, including oncology, cardiology, and pulmonology, to facilitate survival predictions, risk stratification, and therapeutic decision-making²⁷. They provide a nuanced, interpretable way to capture complex interactions among diverse clinical variables. Despite their broad utility, no validated nomogram exists to predict CT-detected mucus plugs in COPD patients. This study addresses this gap by identifying independent risk factors for mucus plug formation in COPD and developing a validated predictive nomogram to enable individualized risk assessment. Such a tool could be readily integrated into routine COPD management workflows, particularly in centers with access to HRCT imaging and spirometry, to support early targeted interventions and reduce the burden of exacerbations in at-risk patients.

Protocol

The present study was approved by the Ethics Committee of Shenzhen Second People’s Hospital (Protocol No. 20193357024). Informed consent was obtained from all participants or their legal representatives prior to enrollment.

Study population and methodology

This study was designed as a single-center, retrospective cohort study. Medical records of patients with a primary diagnosis of COPD at the Department of Respiratory Medicine, Shenzhen Second People’s Hospital, from January 2021 to June 2022 were reviewed. All adult patients (≥18 years) with a primary diagnosis of COPD were initially screened using International Classification of Diseases (ICD-10) coding and chart review from the hospital’s electronic medical record (EMR) system.

Inclusion criteria

(1) Confirmed diagnosis of COPD in accordance with the Global Initiative for chronic obstructive lung disease (GOLD) guidelines; (2) Availability of high-resolution computed tomography (HRCT) of the chest performed within one week of hospitalization; (3) Availability of complete spirometry and laboratory data; and (4) At least one year of follow-up data for acute exacerbation monitoring.

Exclusion criteria

(1) Active pulmonary infections (e.g., pneumonia or tuberculosis) at the time of HRCT imaging; (2) Coexisting lung malignancy; (3) Prior thoracic surgery with potential impact on airway anatomy; and (4) Missing critical clinical data or non-evaluable imaging due to motion artifacts. After applying these criteria, a final cohort of 212 patients was enrolled, comprising 47 patients in the mucus plug-positive (MP) group and 165 patients in the non-mucus plug (NMP) group. Representative HRCT images are illustrated in Figure 1. Patients in the NMP group (n = 165) served as internal controls, enabling statistical comparison of clinical characteristics, pulmonary function indices, and laboratory biomarkers between groups. All analyses were conducted on this internally controlled cohort to support hypothesis-driven model development.

Data collection

Data extraction followed a structured, sequential protocol. Demographic variables collected included age, sex, body mass index (BMI), and smoking status. Clinical history variables comprised COPD duration, acute exacerbation frequency, and comorbidities. Spirometry parameters retrieved including FEV1%, FEV1 to forced vital capacity(FVC), Vital capacity(VC), forced expiratory flow(FEF25–75%pred), Residual volume(RV), Total lung capacity(TLC), and the RV/TLC ratio. Laboratory indices included serum total immunoglobulin E (IgE), 25-hydroxyvitamin D(25(OH)D), serum calcium (Ca²⁺), phosphorus, Carbohydrate antigen (CA199), and fractional exhaled nitric oxide (FeNO), and conducting airway nitric oxide (CaNO). Comorbidity screening included sinusitis, asthma, bronchiectasis, fungal and bacterial colonization, and cardiovascular and metabolic diseases. All data was retrieved from the hospital’s electronic medical record (EMR) system. HRCT images were accessed from the hospital’s picture archiving and communication system (PACS) archive. Details of the software and equipment used in this study are provided in the Table of Materials. No physical reagents or laboratory materials were used; all analyses were performed using existing clinical and radiological data. All patient data were reviewed by two independent investigators. Missing data were handled using the ‘missForest’ non-parametric imputation method implemented in R, to minimize distortion in multivariate analyses.

HRCT diagnostic criteria for mucus plugs

All patients underwent HRCT using standardized institutional imaging protocols. Mucus plugs were defined radiologically on axial CT slices as identified as tubular or branching soft-tissue attenuation structures occupying an airway lumen, visible on at least two contiguous axial slices, consistent with published diagnostic criteria. Only cases with clearly demarcated, segmental or subsegmental airway opacities with soft tissue attenuation similar to soft tissue and not attributable to artifacts or bronchiectasis alone were labeled as mucus-plug positive. HRCT imaging was performed using a Siemens SOMATOM Definition AS (128-slice) CT scanner with the following acquisition parameters: slice thickness 1.0 mm, reconstruction interval 0.75 mm, and use of the B70f high-resolution kernel. Images were reviewed in standard lung window settings (window width : 1600 Hounsfield units [HU]; Window level: 600 HU. Two board-certified thoracic radiologists with over 8 years of experience independently reviewed all scans. Cases with interpretive discrepancies were resolved by consensus discussion. Diagnostic criteria were applied uniformly across all cases to ensure classification consistency.

Nomogram construction, evaluation, and validation

A nomogram was developed to predict CT-detected mucus plugs in COPD patients based on multivariate logistic regression results. The final model incorporated the following independent predictors: bronchiectasis, chronic rhinosinusitis (CRS), acute exacerbations (AE), BMI, FEF25–75%pred, RV/TLC ratio, and serum 25(OH)D levels. Each predictor is assigned a score on a horizontal points scale; individual scores are summed to yield a total score, which corresponds to a predicted probability of mucus plug presence on the output probability scale. The nomogram was subjected to internal validation via bootstrap resampling (1000 iterations) to assess predictive accuracy and discrimination using calibration curves (AUC and ROC).

Statistical analyses

All statistical analyses were performed using R version 4.1.2 and IBM SPSS Statistics version 25.0. Categorical data were expressed as frequencies and percentages; comparisons between groups were performed using the chi-squared test or Fisher’s exact test, as appropriate. Continuous data with normal distribution were expressed as mean ± standard deviation (SD) and compared using the independent samples t-test; non-normally distributed continuous data were expressed as median (interquartile range (IQR) and compared using the Mann-Whitney U test. Variables with P < 0.1 in the univariate logistic regression analysis were included in the model, consistent with standard practice in predictive model development. The R packages used were “rms”, “mstate”, “data.table”, “pROC”, “rmada”, “rio”, “boot”, and “missForest”. Nomogram construction was implemented using the lrm and nomogram functions from the rms package. ROC curves and AUC values were computed using the roc and auc functions from the pROC package. Calibration curves were generated with the calibrate function in RMS. Decision curve analysis (DCA) was performed using the decision curve function from the rmda package. Missing data imputation was performed using the missForest function. Bootstrap internal validation (1000 iterations) was conducted using the boot package. A fixed random seed (set.seed[240708] was applied at the start of the analysis to ensure reproducibility. A P-value of < 0.05 was considered statistically significant. The logistic regression model formula was:

glm(mucus_status ~ bronchiectasis + CRS + BMI + FEF25_75 + RV_TLC + VitD, family = "binomial")

Results

Baseline characteristics

The study comprised a cohort of 212 patients with COPD, divided into two groups: 47 with mucus plugs (MP) and 165 without mucus plugs (NMP). The occurrence of mucus plugs in this COPD population was found to be 28.33%. Statistical analysis, detailed in Table 1, identified significant differences between the MP and NMP groups in several key metrics. These included body mass index (BMI), the frequency of acute exacerbations (AE), preval...

Discussion

In this study, the prevalence of CT-detected mucus plug formation among hospitalized COPD patients was 22.16%, consistent with estimates reported in prior literature²⁷. Mucus plugs in COPD are clinically significant due to their association with accelerated pulmonary function decline, increased acute exacerbation frequency, and higher mortality risk²⁸. Despite this, a validated predictive tool for identifying at-risk patients was previously lacking. This analysis identified...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
HRCT Scans	Shenzhen Second People's Hospital		Used for diagnosing small airway mucus obstruction in COPD patients
SPSS 25.0 Software	1BM		Statistical software used for data analysis, including t-tests and logistic regression.
R Software (Packages: mms, mstate, etc.)	R Foundation for Statistical Computing		Used for statistical analysis and model validation, including calculation of the C-index.
Electronic Medical Record System	Shenzhen Second People's Hospital		Data source for clinical and laboratory variables, including patient history and diagnostic parameters.
Logistic Regression Equation	Custom (Applied via SPSS and R)		Used to screen for independent risk factors related to small airway mucus obstruction in COPD patients.