Development and validation of a multimodal clinical-radiomics-deep learning nomogram based on automated chest CT segmentation for classifying COPD severity: A multicenter study

Development and validation of a multimodal clinical-radiomics-deep learning nomogram based on automated chest CT segmentation for classifying COPD severity: A multicenter study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Development and validation of a multimodal clinical-radiomics-deep learning nomogram based on automated chest CT segmentation for classifying COPD severity: A multicenter study Qiang Fei, Xiaoli Mei, Jiacheng Zhao, Yili Xing, Yuanbin Li, Mei Yang, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9272720/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 11 You are reading this latest preprint version Abstract Background and Objective: Chronic obstructive pulmonary disease (COPD) is a widespread and severely disabling respiratory disorder that places a substantial burden on global healthcare systems. Precise determination of COPD severity staging is essential for effective patient management and treatment planning. This study seeks to develop and validate a comprehensive nomogram that combines clinical characteristics, whole-lung computed tomography (CT) radiomic features, and deep learning-derived features to classify COPD severity. Method A retrospective analysis included 1,794 patients from three hospitals, spanning January 1, 2021, to May 30, 2025. Following fully automated segmentation of the entire lungs, radiomic features and three-dimensional deep learning features were extracted. A comprehensive nomogram was developed and validated, integrating radiomics features, deep learning features, and independent clinical factors. Model performance was assessed and compared using receiver operating characteristic (ROC) curves, the area under the ROC curve (AUC), decision curve analysis (DCA), and the DeLong test. Result In the training, internal validation, and external validation cohorts, the area under the receiver operating characteristic curve (AUC) values for the clinical model were 0.634, 0.630, and 0.616, respectively; for the radiomics (Rad) model, 0.760, 0.729, and 0.704, respectively; for the deep learning (DL) model, 0.805, 0.757, and 0.752, respectively; for the radiomics-deep learning combined (DLR) model, 0.822, 0.740, and 0.759, respectively; and for the logistic regression model, 0.839, 0.759, and 0.767, respectively. The logistic regression model outperformed the individual clinical, radiomics, and three-dimensional deep learning models. Conclusion This study constructed and validated a novel combined logistic regression model for identifying the severity of COPD by integrating the clinical characteristics of independent risk factors, the full lung radiomics features, and the deep learning features. It demonstrated the additional value of chest CT in evaluating lung structure and lung function status. Chronic Obstructive Pulmonary Disease Radiomics Deep Learning Prediction Model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Chronic obstructive pulmonary disease (COPD) is a prevalent, preventable, and treatable condition marked by persistent respiratory symptoms and airflow limitation. It has emerged as the third leading cause of death worldwide, imposing a substantial burden on health systems and society [ 1 – 3 ]. Research indicates that increased COPD severity correlates with a higher frequency of exacerbations, leading to elevated mortality rates [ 4 ]. Furthermore, COPD frequently coexists with other chronic conditions, such as cardiovascular disease, osteoporosis, diabetes, lung cancer, and cachexia [ 5 – 8 ]. These comorbidities collectively diminish patients' quality of life and increase mortality. Treatment strategies for COPD vary according to disease severity [ 9 ], highlighting the importance of accurately assessing COPD severity to improve patient prognosis and reduce mortality. Pulmonary function testing (PFT) remains the gold standard for diagnosing and staging COPD, categorizing severity into four stages: mild, moderate, severe, and very severe [ 10 ]. Despite its status, PFT utilization remains low; in China, only 6.7% of individuals aged 40 and above have undergone PFT [ 11 ]. Moreover, PFT has notable limitations: it reflects only a single dimension of airflow limitation and fails to capture regional lung tissue changes, such as emphysema, small airway disease, and airway remodeling [ 12 ]. Meanwhile, detection tools such as the Lung Function Questionnaire (LFQ), the COPD Diagnosis Questionnaire (CDQ), and the COPD Population Screening Tool (COPD-PS) have demonstrated their practicality in identifying individuals at high risk of COPD. However, these tools mainly rely on self-reported symptoms and risk factors, which may lead to recall bias, and they cannot directly assess potential structural changes in the lungs. With the widespread adoption of lung cancer screening programs, chest computed tomography (CT) has become increasingly accessible, offering detailed anatomical information. Studies have shown that COPD of varying severities exhibits distinct CT morphological features. Quantitative imaging has been employed to assess COPD severity and predict coronary heart disease risk [ 14 ]. Cho et al. [ 15 ] demonstrated that quantitative pulmonary vascular characteristics correlate with the severity and extent of emphysema in COPD. Leveraging CT images for COPD staging could thus enable timely "one-stop" medical interventions, yielding greater benefits [ 16 ]. However, traditional CT assessment relies heavily on radiologists' subjective visual evaluation of manifestations such as emphysema and airway wall thickening, making it challenging to detect complex image texture patterns imperceptible to the human eye. In recent years, artificial intelligence has been increasingly applied in COPD research. Radiomics techniques quantify the heterogeneity, texture, and morphological features of lesion regions by extracting numerous quantitative features from medical images [ 17 ]. Given COPD's diffuse and heterogeneous nature, the entire lung is designated as the region of interest. Meanwhile, deep learning (DL) technology has made significant advances in medical image analysis. DL models automatically learn hierarchical deep feature representations from raw images, overcoming the limitations of manual feature engineering and excelling in image classification, segmentation, and prediction tasks [ 18 ]. Combining deep learning with radiomics offers complementary advantages: deep learning mines deep implicit features automatically, while radiomics provides interpretable quantitative features, jointly enhancing predictive model performance. Our team has found that a fusion model integrating radiomics and deep learning via one-stop chest CT can effectively predict the risk of concurrent coronary heart disease in COPD patients [ 19 ]. We hypothesize that a fusion model combining whole-lung radiomics from chest CT with deep learning can identify COPD severity in routine clinical practice. This approach would enhance the value of chest CT, particularly for patients unable to undergo PFT, by providing both morphological information and pulmonary function assessment. To facilitate clinical use, we integrate automatic deep learning features with interpretable radiomics features to comprehensively explore effective information in CT images and combine it with key clinical factors to construct a nomogram model. This provides clinicians with a visual and quantitative individualized assessment tool. Methods and Materials Patients and clinical data This study has been approved by the Institutional Review Board of The First Affiliated Hospital of Huzhou University (Approval Number: 2025KYLL002-01). Given the retrospective nature of the study, the requirement for informed consent was waived. This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Clinical data and tomographic images of eligible patients were collected from three hospitals from January 1, 2021, to May 30, 2025. During this period, a total of 1,794 patients diagnosed with chronic obstructive pulmonary disease (COPD) through pulmonary function tests (PFTs) at these three centers were included in this study. The inclusion criteria were as follows: (1) Diagnosed with chronic obstructive pulmonary disease confirmed by pulmonary function tests; (2) Completion of pulmonary function tests and chest CT scans within 2 weeks; (3) Possession of complete thin-slice (1 mm) chest CT images. The exclusion criteria were as follows: (1) Incomplete clinical data, concurrent other thoracic diseases (such as pneumonia, atelectasis, pulmonary nodules or masses larger than 6 mm, and pleural effusion); (2) Concurrent any malignancy; (3) Presence of spinal implants or significant image artifacts, with poor image quality affecting diagnosis; (4) Lack of thin-slice chest CT images. We randomly assigned 1,313 patients from the first center to the training cohort (n = 919) or the internal validation cohort (n = 394) in a ratio of 7:3, and allocated 434 patients from the second center and 47 patients from the third center to the external validation cohort (n = 481). Clinical information included age, body mass index (BMI), gender, smoking status, as well as laboratory test indicators such as C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and procalcitonin (PCT). For some missing values, we used the mean to fill them in, and for data with missing values accounting for more than 30% of the information, we did not include them in the final study. CT image acquisition and pulmonary function examination Participants underwent non-enhanced CT scans using equipment from manufacturers such as Aquilion ONE TSX-301C and Brilliance CT 16. Axial CT images of the entire thorax were acquired during full inspiration (scanning parameters are detailed in Supplementary Data 1). Pulmonary function testing was conducted using the Ganshorn PowerCube. The diagnosis and classification of chronic obstructive pulmonary disease (COPD) adhered to the criteria outlined by the Global Initiative for Chronic Obstructive Lung Disease (GOLD) [ 20 ]. COPD was defined as a post-bronchodilator ratio of forced expiratory volume in the first second (FEV1) to forced vital capacity (FVC) below 0.7. In this study, COPD severity was categorized into two stages according to GOLD criteria: mild to moderate airflow limitation (GOLD stages 1–2, FEV1 ≥ 50% of predicted value) and severe to very severe airflow limitation (GOLD stages 3–4, FEV1 < 50% of predicted value). Whole lung automatic segmentation A deep learning model ( https://github.com/OnekeyAI-Platform/onekey ) was employed to segment the left and right lungs, and the extracted regions were merged into a single region of interest (ROI). The specific algorithm process is detailed in Supplementary Data 2. To assess the consistency between automatic and manual segmentation results, 100 samples were randomly selected and independently evaluated by two chest radiologists, each with over 10 years of experience, using ITK-SNAP (version 3.8.0, www.itksnap.org ). The Dice index was then calculated to objectively quantify the spatial overlap of contours and determine the agreement between fully automatic and manual segmentation results. Finally, automatic segmentation was applied to the remaining samples. Extraction and selection of radiomics features We employed Pyradiomics ( http://pyradiomics.readthedocs.io ) to extract radiomic features from segmented lung regions. Before feature extraction, the images underwent a three-step preprocessing procedure to standardize them. First, the images were resampled to a resolution of 1 mm × 1 mm × 1 mm. Second, the gray-level values were adjusted to a 25 Gy level to mitigate the effects of varying slice thicknesses and reduce noise interference. Subsequently, features were extracted using Z-score normalization, yielding three categories of features: first-order, shape, and texture features. Statistical significance was evaluated using the t-test, and only features with a p-value less than 0.05 were retained. To address collinearity, we assessed the correlation between features using the Pearson correlation coefficient and excluded features with a correlation coefficient exceeding 0.9. Furthermore, within a 10-fold cross-validation framework, we used Lasso regression to optimize the feature set by determining the optimal regularization parameter λ and selecting features with non-zero coefficients. Finally, the radiomic score (Rad Score, RS) was computed as a linear combination of the retained features and their corresponding coefficients. Extraction and selection of deep learning features In this study, the ResNet50 architecture was employed as a convolutional neural network (CNN) to extract deep learning features. Image intensity distributions were standardized using Z-score normalization, and the standardized images were used as inputs to the deep learning model. During training, real-time data augmentation techniques—such as random cropping and flipping—were applied to improve model robustness. For test images, preprocessing was restricted to normalization to ensure consistency during evaluation. A pre-trained CNN model was used to extract deep transfer learning (DTL) features from the largest region of interest (ROI) in each image, specifically from the penultimate layer. In the proposed model, the output probabilities computed by the CNN were defined as deep learning feature signatures. Given the complexity of the CNN, principal component analysis (PCA) was applied to reduce these features to a 512-dimensional space for improved manageability. Model construction and evaluation Following feature selection, five distinct models were constructed. Logistic regression analysis was initially applied to clinical features to identify statistically significant predictors, which were then used to develop a clinical model. Using the selected radiomic and deep learning features, an imaging biomarker (Rad) model and a deep learning (DTL) model were constructed separately. To create a deep learning imaging biomarker (DLR) model, a fusion algorithm was employed to integrate deep learning features with radiomic features. To improve clinical applicability, univariate and stepwise multivariate analyses were performed on all clinical features to identify significant predictors. These selected clinical features were subsequently combined with the predictive results of the DLR model to develop a logistic regression (LR)-based linear model, termed the comprehensive model, which was effectively visualized through a nomogram. The models were evaluated using the area under the curve (AUC), accuracy, sensitivity, specificity, recall, and F1 score. Model performance was further assessed using positive predictive value (PPV) and negative predictive value (NPV). The AUCs of different models were compared using the DeLong test, and calibration was analyzed using calibration curves (Hosmer-Lemeshow test) to verify reliability. Decision curve analysis (DCA) was also conducted to evaluate the clinical utility of the predictive models, providing insights into their potential benefits in clinical settings. Figure 1 presents a flowchart of the entire study design. Statistical analysis The Shapiro-Wilk test was employed to evaluate the normality of clinical characteristics. Continuous variables were analyzed using the t-test or the Mann-Whitney U test, depending on their distribution. Categorical variables were assessed using Chi-square (χ²) tests. All data analyses were performed using Python 3.7.12. Statistical analyses utilized Statsmodels version 0.13.2, while PyRadiomics version 3.0.1 was used for radiomics feature extraction. Scikit-learn version 1.0.2 facilitated machine learning tasks, and PyTorch version 1.11.0 was employed for deep learning framework development, with performance optimized using CUDA version 11.3.1 and cuDNN version 8.2.1. Result Baseline characteristics and clinical feature screening of the patients Figure 2 presents the flowchart illustrating patient selection. As of May 30, 2025, a total of 2,658 patients diagnosed with COPD were included in the study. Following screening based on inclusion and exclusion criteria, the final cohort comprised 1,794 patients, including 276 females and 1,518 males. The mean age of the cohort was 74.02 years. The number of patients from the three hospitals was 1,313, 434, and 47, respectively. Among these patients, 989 had COPD classified as GOLD I–II, and 805 had COPD classified as GOLD III–IV. Table 1 provides details on the baseline clinical characteristics of the study cohort (standardized units for all indicators in Table 1 are available in Supplementary Data 5). Table 1 Baseline characteristics of the study population Clinical factors Training cohort(n = 919) Internal validation cohort (n = 394) External vaidation cohort(n = 481) GOLD I–II (N = 554) GOLD III–IV (N = 365) p value GOLD I–II (N = 228) GOLD III–IV (N = 166) p value GOLD I–II (N = 207) GOLD III–IV (N = 274) p value Age 74.38 ± 8.15 73.76 ± 8.11 0.387 75.04 ± 8.11 73.84 ± 8.35 0.127 73.33 ± 8.02 73.35 ± 7.49 0.96 BMI 508.90 ± 246.13 472.64 ± 256.58 0.045 545.96 ± 228.39 431.58 ± 239.93 < 0.001 495.43 ± 230.98 473.70 ± 276.15 0.315 Albumin 39.66 ± 14.07 38.59 ± 4.08 0.012 39.13 ± 4.16 38.66 ± 4.26 0.275 38.05 ± 5.06 37.45 ± 4.32 0.19 Globulin 104.55 ± 42.37 98.36 ± 43.21 0.011 104.65 ± 46.47 94.98 ± 41.30 0.063 100.43 ± 39.72 90.53 ± 40.16 0.01 Triglyceride 1.16 ± 0.69 1.04 ± 0.69 < 0.001 1.16 ± 0.69 1.41 ± 5.16 0.016 2.01 ± 12.97 1.24 ± 4.52 < 0.001 Alkaline Phosphatase 89.31 ± 67.48 83.30 ± 41.92 0.056 87.33 ± 48.09 87.18 ± 37.23 0.801 80.39 ± 59.90 76.89 ± 27.77 0.761 White blood cell count 6.74 ± 3.63 7.31 ± 3.62 0.018 6.48 ± 2.35 6.95 ± 3.18 0.469 6.68 ± 3.30 7.27 ± 3.84 0.179 Neutrophilic granulocyte percentage 248.91 ± 105.51 286.46 ± 109.46 < 0.001 252.74 ± 107.78 274.40 ± 114.88 0.032 216.81 ± 108.32 247.69 ± 110.54 0.002 Percentage of lymphocytes 26.95 ± 118.66 18.65 ± 9.84 < 0.001 21.75 ± 9.43 21.20 ± 17.72 0.049 23.80 ± 10.91 20.16 ± 9.29 < 0.001 Absolute eosinophil count 15.12 ± 16.50 12.59 ± 15.68 0.001 14.09 ± 12.82 13.52 ± 15.98 0.087 17.08 ± 16.77 17.32 ± 18.99 0.365 Plateletcrit 0.19 ± 0.09 0.19 ± 0.06 0.516 0.97 ± 11.78 0.19 ± 0.06 0.44 0.25 ± 0.86 0.19 ± 0.06 0.524 Red blood cell distribution width 13.48 ± 1.45 13.43 ± 1.35 0.967 13.57 ± 1.61 13.27 ± 1.53 0.136 13.71 ± 7.08 13.36 ± 1.19 0.377 Mean platelet volume 10.48 ± 5.61 10.11 ± 1.26 0.734 10.25 ± 1.30 9.92 ± 1.17 0.024 10.61 ± 1.31 10.42 ± 1.25 0.101 Platelet Distribution Width 16.05 ± 1.52 15.91 ± 1.34 0.639 16.16 ± 1.17 15.93 ± 1.23 0.008 12.93 ± 2.97 12.97 ± 2.83 0.472 C-Reactive Protein 187.51 ± 172.49 210.22 ± 176.06 0.022 205.12 ± 168.28 196.15 ± 175.59 0.565 325.33 ± 208.07 332.22 ± 211.67 0.736 Arterial Oxygen Partial Pressure 89.26 ± 22.12 87.04 ± 21.31 0.019 87.39 ± 18.84 90.20 ± 25.21 0.947 84.18 ± 9.92 82.86 ± 15.81 0.053 Arterial Carbon Dioxide Partial Pressure 38.85 ± 3.80 41.32 ± 5.86 < 0.001 39.05 ± 3.61 42.13 ± 7.69 < 0.001 40.13 ± 5.69 58.63 ± 277.77 < 0.001 Gender 0.008 0.119 0.778 Female 80(14.44) 78(21.37) 38(16.67) 39(23.49) 19(9.18) 22(8.03) Male 474(85.56) 287(78.63) 190(83.33) 127(76.51) 188(90.82) 252(91.97) Smoke 0.956 0.292 1 No 189(34.12) 126(34.52) 75(32.89) 64(38.55) 89(43.00) 118(43.07) Yes 365(65.88) 239(65.48) 153(67.11) 102(61.45) 118(57.00) 156(56.93) Univariate and multivariate analyses revealed age, gender, and arterial partial pressure of carbon dioxide (PaCO2) as predictors in the clinical model; these variables also independently predicted the severity of chronic obstructive pulmonary disease (COPD), as detailed in Table 2 . The area under the receiver operating characteristic curve (AUC) values for the clinical model were 0.634 (95% confidence interval: 0.596–0.671) in the training set, 0.63 (95% confidence interval: 0.572–0.688) in the internal validation set, and 0.616 (95% confidence interval: 0.566–0.666) in the external validation set, as presented in Table 3 . Table 2 Univariable and multivariable analysis of clinical features. Variable Univariable analysis Multivariable analysis OR[95%CI] p value OR[95%CI] p value Plateletcrit 0.128[0.073,0.224] < 0.01 0.399[0.067,2.373] 0.397 Gender 0.606[0.535,0.685] < 0.01 0.485[0.34,0.693] 0.001 Smoke 0.655[0.571,0.751] < 0.01 1.032[0.77,1.383] 0.858 Triglyceride 0.699[0.638,0.765] < 0.01 0.887[0.732,1.075] 0.305 Mean platelet volume 0.96[0.949,0.97] < 0.01 0.97[0.924,1.017] 0.288 White blood cell count 0.961[0.947,0.975] < 0.01 1.025[0.987,1.065] 0.277 Red blood cell distribution width 0.969[0.962,0.977] < 0.01 0.964[0.891.1.043] 0.442 Platelet Distribution Width 0.974[0.968,0.98] < 0.01 0.925[0.856,1.001] 0.102 Percentage of lymphocytes 0.977[0.972,0.982] < 0.01 0.971[0.94,1.002] 0.12 Absolute eosinophil count 0.981[0.975,0.986] < 0.01 0.996[0.987,1.005] 0.484 albumin 0.989[0.986,0.992] < 0.01 0.996[0.982,1.01] 0.661 Arterial Carbon Dioxide Partial Pressure 0.991[0.988,0.994] < 0.01 1.15[1.115,1.185] 0 age 0.994[0.993,0.996] < 0.01 0.981[0.967,0.995] 0.028 Arterial Oxygen Partial Pressure 0.995[0.994,0.996] < 0.01 0.994[0.989,1] 0.111 Alkaline Phosphatase 0.996[0.994,0.997] < 0.01 0.998[0.995,1.001] 0.228 Globulin 0.996[0.995,0.997] < 0.01 0.997[0.994,1] 0.105 BMI 0.999[0.999,0.999] < 0.01 0.999[0.999,1] 0.067 C-Reactive Protein 0.999[0.999,1] < 0.01 1.001[1,1.002] 0.056 Neutrophilic granulocyte percentage 0.999[0.999,0.999] < 0.01 1[0.997,1.004] 0.818 Table 3 presents the performance of four models across different cohorts. The models include the Clinic (clinical model signature), Rad (radiomics signature), DTL (deep learning signature),DLR (deep learning radiomics signature), and Combined (a combination of clinical, deep learning, and radiomics signatures). Clinic AUC Accuracy 95% CI Sensitivity Specificity PPV NPV Precision Recall F1 Threshold Cohort 0.634 0.656 0.5963–0.6708 0.471 0.778 0.583 0.691 0.583 0.471 0.521 0.402 train Rad 0.76 0.712 0.7274–0.7918 0.652 0.751 0.633 0.766 0.633 0.652 0.642 0.374 train DTL 0.805 0.737 0.7775–0.8333 0.69 0.767 0.661 0.79 0.661 0.69 0.676 0.468 train DLR 0.822 0.743 0.7940–0.8491 0.775 0.722 0.648 0.83 0.648 0.775 0.706 0.379 train Combined 0.839 0.777 0.8134–0.8652 0.679 0.841 0.738 0.799 0.738 0.679 0.708 0.493 train Clinic 0.63 0.637 0.5720–0.6875 0.506 0.732 0.579 0.671 0.579 0.506 0.54 0.39 val Rad 0.729 0.69 0.6783–0.7792 0.627 0.737 0.634 0.73 0.634 0.627 0.63 0.396 val DTL 0.757 0.68 0.7088–0.8043 0.711 0.658 0.602 0.758 0.602 0.711 0.652 0.391 val DLR 0.74 0.688 0.6912–0.7894 0.681 0.693 0.617 0.749 0.617 0.681 0.648 0.386 val Combined 0.759 0.706 0.7100–0.8074 0.735 0.684 0.629 0.78 0.629 0.735 0.678 0.326 val Clinic 0.616 0.593 0.5663–0.6659 0.467 0.758 0.719 0.518 0.719 0.467 0.566 0.398 test Rad 0.704 0.674 0.6571–0.7506 0.701 0.638 0.719 0.617 0.719 0.701 0.71 0.356 test DTL 0.752 0.711 0.7091–0.7957 0.785 0.614 0.729 0.683 0.729 0.785 0.756 0.38 test DLR 0.759 0.699 0.7169–0.8021 0.664 0.744 0.774 0.626 0.774 0.664 0.715 0.39 test Combined 0.767 0.701 0.7247–0.8097 0.646 0.773 0.79 0.623 0.79 0.646 0.711 0.382 test Feature selection and model construction Radiological features and three-dimensional deep learning features were extracted from images based on the automatic segmentation of chest CT scans. Reproducibility was assessed using the intraclass correlation coefficient (ICC). Following t-tests, Spearman correlation analyses, and LASSO screening, 20 optimal radiomic features with non-zero coefficients were selected, and a radiological model was constructed. The coefficients and average standard errors derived from five-fold cross-validation are presented in Figs. 3 A and 3 B, while Fig. 3 C shows the values of the selected features with non-zero coefficients. The specific formula for screening radiomic features is detailed in Supplementary Data 3. The ResNet50 architecture, a convolutional neural network (CNN), was employed to extract deep learning features, yielding a total of 2048 features. After compression, 32 deep learning features were retained to construct the deep learning model (DTL). To enhance the accuracy of predicting chronic obstructive pulmonary disease (COPD) severity, radiological and deep learning features were integrated. Through t-tests, Spearman correlation coefficients, and LASSO regression analysis, 19 features with non-zero coefficients were selected (Figs. 3 D, E, and F), enabling the construction of a deep learning radiological model (DLR). The formula used to screen DLR features is provided in Supplementary Data 4. Finally, age, gender, and partial pressure of carbon dioxide (PaCO2) were combined with the DLR model, and a nomogram model was constructed using multivariate logistic regression. Figure 4 presents a nomogram designed specifically for clinical use. Comparison of clinical models, imaging-based models, deep learning models, DLR models and combined models Table 3 presents the performance of clinical, radiomics, deep learning (DTL), deep learning radiomics (DLR), and combined models. In the training cohort (Fig. 5 A), the combined model achieved the highest area under the curve (AUC) of 0.839, followed by the DLR model (0.822), the DTL model (0.805), and the radiomics (Rad) model (0.76). The clinical model had a relatively low AUC of 0.634. In the internal validation cohort (Fig. 5 B), the combined model again outperformed others with an AUC of 0.759, followed by the DTL model (0.757), the DLR model (0.74), and the Rad model (0.729). The clinical model had an AUC of 0.63. In the external validation cohort (Fig. 5 C), the combined model had the highest AUC (0.767), followed by the DLR model (0.759), the DTL model (0.752), and the Rad model (0.704). The clinical model had an AUC of 0.616. These results consistently demonstrate the low performance of the clinical model across cohorts, highlighting its limitations for standalone use. The Rad and DTL models showed moderate performance, with one slightly outperforming the other in different contexts. The combined model generally exhibited the best AUC performance, demonstrating its robustness and generalizability. These findings underscore the importance of advanced modeling techniques and multimodal data fusion in predicting chronic obstructive pulmonary disease (COPD) staging. Figure 6 displays decision curve analysis (DCA) results, which were used to assess the clinical applicability of the models. The combined model demonstrated a higher net benefit compared to the others, indicating a superior benefit-risk ratio in clinical decision-making. Figure 7 presents calibration curves for each cohort, showing the agreement between model-predicted probabilities and observed outcomes. Additionally, Fig. 8 shows the results of discriminative performance comparisons among the models using the DeLong test. Based on the analysis, the combined nomogram model exhibited superior predictive efficacy, performing well in terms of both discrimination and calibration. Discussion We developed and validated a composite nomogram that integrates full-lung radiomics features derived from chest CT scans, deep learning features, and clinical independent predictors to assess COPD severity. Radiomics and deep learning features independently stratify COPD severity, demonstrating that chest CT can evaluate both lung structure and functional status. Mild COPD patients are often overlooked due to asymptomatic or mild symptoms [ 22 , 23 ], whereas severe COPD significantly impairs quality of life and escalates treatment costs [ 24 ]. Differentiated treatment strategies based on COPD severity are critical for patient prognosis [ 9 ], necessitating an early, rapid, and accessible diagnostic method for staging. Prior studies have identified morphological CT changes—such as bronchial wall thickening, tracheal shape alterations, and total lung emphysema percentage—as correlates of severe COPD [ 25 ]. Machine learning advancements have enabled novel staging approaches: Chen TT [ 26 ] and Rueda R [ 27 ] combined machine learning models with clinical and biochemical markers, while Gu YF et al. developed a COPD severity prediction model using biochemical and immunological parameters, though its 62.7% accuracy limited clinical utility [ 28 ]. Given COPD’s heterogeneous lung involvement, leveraging CT to detect abnormal textures and assess disease status is crucial [ 29 , 30 ]. Sui H et al. improved severity classification by integrating lung parenchyma shape, size, distribution, and airway morphology into a machine learning framework [ 31 ]. However, selecting and capturing relevant images from large datasets remains challenging despite machine learning advancements. Radiomics enables high-throughput extraction of quantitative features from medical images, with Li Z [ 32 ] and Zhou T [ 33 ] demonstrating its efficacy in COPD severity classification. Some researchers refined grading by analyzing radiomics features across individual lung lobes, capturing localized CT characteristics more effectively than whole-lung approaches [ 34 ]. These studies underscore radiomics’ feasibility for COPD staging. With artificial intelligence progress, deep learning has gained traction: Sara Rezvanjou et al. combined 2D tMPR images and 3D lung views to classify COPD using airway data [ 35 ], while Zou X et al. developed a cost-effective but less sensitive deep learning model for COPD staging using chest X-rays and clinical parameters [ 36 ]. Our study integrates chest CT images, clinical parameters, radiomics, and deep learning features into a clinically applicable nomogram, enhancing CT’s functional assessment value. Our model extracts texture, edge, structural, and shape information from CT images, yielding richer feature descriptions for disease prediction. Adding deep learning features significantly improved diagnostic performance compared to models using only clinical or radiomics features. The final nomogram incorporates clinical factors (age, gender, PaCO2), radiomics features, and deep learning features. Decision curve analysis (DCA) confirmed its superior net benefit over clinical models across diagnostic thresholds, supporting its clinical utility. By improving COPD severity classification accuracy, our nomogram aids in managing acute exacerbations, personalizing treatment interventions, controlling symptoms, and slowing disease progression. This study has limitations. First, as a retrospective analysis from three institutions, the third center’s smaller sample size may limit generalizability; a prospective multicenter study is planned to optimize the cohort. Second, variables with > 30% missing values were excluded, warranting closer attention in future research. Third, the single-timepoint design did not assess longitudinal lung changes in the same patients. Future directions include integrating traditional CT features and quantitative parameters to enhance model efficacy and using delta radiomics to study COPD dynamics. Conclusion This study utilized automatic segmentation technology to extract the entire lung parenchyma from CT images. By integrating clinical characteristics of independent risk factors, whole-lung radiomic features, and deep learning features, a novel combined nomogram was constructed to identify COPD severity. The study also demonstrated the added value of chest CT in evaluating lung structure and functional status. Abbreviations COPD Chronic obstructive pulmonary disease ROI Region of Interest ROC Receiver Operating Characteristic PPV Positive Predictive Value NPV Negative Predictive Value LASSO Least Absolute Shrinkage and Selection Operator ICC Intraclass Correlation Coefficient DL Deep Learning CNN Convolutional Neural Network AUC Area Under the Receiver Operating Characteristic Curve DCA Decision Curve Analysis PFT Pulmonary Function Test CRP C-Reactive Protein ESR Erythrocyte Sedimentation Rate PCT Procalcitonin PDW Platelet Distribution Width PaCO2 Arterial Partial Pressure of Carbon Dioxide PaO2 Arterial Partial Pressure of Oxygen RS Radiomics Score DLR Deep Learning radiomics CT Computed Tomography Declarations Acknowledgements The authors would like to thank all patients, nurses, and physicians who participated in the study. Authors’ contributions BHP, and FQ conceived the study. XJN, QLY, YM, MXLand LYB collected data. XYL, ZJC and ZYF analysed data and drafted the manuscript. BHP, ZHX and FQ revised the manuscript. All authors contributed to the article and approved the submitted version. Funding This work was supported by the Science and Technology Project of Huzhou City, Zhejiang Province (2023GY33) and Postgraduate Research and Innovation Project of Huzhou University (2025KYCX99). Availability of data and materials The data are available from the corresponding author on reasonable request. Ethics approval and consent to participate This retrospective study using anonymized medical record data received an Institutional Review Board (IRB) waiver of informed consent from the First People’s Hospital of Huzhou, as it involves no direct patient intervention and adheres to the Declaration of Helsinki. Patient privacy is fully protected, and ethical standards are strictly maintained. Consent for publication Not applicable Competing interests The authors declare no competing interests. References Yin P, Wu J, Wang L, et al. The Burden of COPD in China and Its Provinces: Findings From the Global Burden of Disease Study 2019. Front Public Health. 2022 Jun 3;10:859499. doi: 10.3389/fpubh.2022.859499. Gomes F, Cheng SL. Pathophysiology, Therapeutic Targets, and Future Therapeutic Alternatives in COPD: Focus on the Importance of the Cholinergic System. Biomolecules. 2023 Mar 5;13(3):476. doi: 10.3390/biom13030476. Qin J, Ran B, Wu Y, et al. Association of hs-CRP/HDL, AIP and NHHR with chronic obstructive pulmonary disease: a cross-sectional NHANES study. Ann Med. 2025 Dec;57(1):2522317. doi: 10.1080/07853890.2025.2522317. Decramer M, Janssens W, Miravitlles M. Chronic obstructive pulmonary disease. Lancet. 2012 Apr 7;379(9823):1341-51. doi: 10.1016/S0140-6736(11)60968-9. Kahnert K, Jörres RA, Behr J, Welte T. The Diagnosis and Treatment of COPD and Its Comorbidities. Dtsch Arztebl Int. 2023 Jun 23;120(25):434-444. doi: 10.3238/arztebl.m2023.027. Bian H, Zhu S, Zhang Y, et al. Artificial Intelligence in Chronic Obstructive Pulmonary Disease: Research Status, Trends, and Future Directions --A Bibliometric Analysis from 2009 to 2023. Int J Chron Obstruct Pulmon Dis. 2024 Aug 21;19:1849-1864. doi: 10.2147/COPD.S474402. Magrì D, Fiori E, Agostoni P, et al. Heart failure and chronic obstructive pulmonary disease. A combination not to be underestimated. Heart Fail Rev. 2025 Dec;30(6):1525-1538. doi: 10.1007/s10741-025-10566-3. Bian H, Zhu S, Xing W, et al. Research Status and Direction of Chronic Obstructive Pulmonary Disease Complicated with Coronary Heart Disease: A Bibliometric Analysis from 2005 to 2024. Int J Chron Obstruct Pulmon Dis. 2025 Jan 7;20:23-41. doi: 10.2147/COPD.S495326. Guo C, Yu T, Chang LY, et al. Mortality risk attributable to classification of chronic obstructive pulmonary disease and reduced lung function: a 21-year longitudinal cohort study. Respir Med. 2021;184:106471. doi: 10.1016/j.rmed.2021.106471. Kahnert K, Jörres RA, Behr J, Welte T. The Diagnosis and Treatment of COPD and Its Comorbidities. Dtsch Arztebl Int. 2023 Jun 23;120(25):434-444. doi: 10.3238/arztebl.m2023.027. Tong H, Cong S, Fang LW, et al. [Performance of pulmonary function test in people aged 40 years and above in China, 2019-2020]. Zhonghua Liu Xing Bing Xue Za Zhi. 2023 May 10;44(5):727-734. Chinese. doi: 10.3760/cma.j.cn112338-20230202-00051. Wang R, Huang C, Yang W, et al. Respiratory microbiota and radiomics features in the stable COPD patients. Respir Res. 2023 May 12;24(1):131. doi: 10.1186/s12931-023-02434-1. Yu X, Xiong Z, Cao J, et al. Pulmonary arterial morphological markers on non-contrast CT predicted acute exacerbations and disease progression in chronic obstructive pulmonary disease: a longitudinal cohort study. Jpn J Radiol. 2025 Aug 4. doi: 10.1007/s11604-025-01841-2. Su L, Qian C, Yu C, et al. Quantitatively Assessed Emphysema Severity on HRCT Independently Predicts Coronary Artery Disease in COPD: A Retrospective Cohort Study. Int J Chron Obstruct Pulmon Dis. 2025 Sep 10;20:3147-3161. doi: 10.2147/COPD.S540503. Cho YH, Lee SM, Seo JB, et al. Quantitative assessment of pulmonary vascular alterations in chronic obstructive lung disease: associations with pulmonary function test and survival in the KOLD cohort. Eur J Radiol. 2018;108:276–282. doi: 10.1016/j.ejrad.2018.09.013. Lin X, Zhou T, Ni J, et al. CT-based radiomics nomogram of lung and mediastinal features to identify cardiovascular disease in chronic obstructive pulmonary disease: a multicenter study. BMC Pulm Med. 2025;25(1):121. doi: 10.1186/s12890-025-03568-2. Qi YJ, Su GH, You C, et al. Radiomics in breast cancer: Current advances and future directions. Cell Rep Med. 2024 Sep 17;5(9):101719. doi: 10.1016/j.xcrm.2024.101719. Wang Y, Zhang L, Xie H, et al. Predicting response and survival of lung adenocarcinoma under anti-programmed death-1 therapy using biological deep learning. Brief Bioinform. 2025 Aug 31;26(5):bbaf479. doi: 10.1093/bib/bbaf479. Bian H, Qian H, Zhu S, et al. Nomogram Model for Identifying the Risk of Coronary Heart Disease in Patients with Chronic Obstructive Pulmonary Disease Based on Deep Learning Radiomics and Clinical Data: A Multicenter Study. Int J Chron Obstruct Pulmon Dis. 2025 Sep 2;20:3045-3057. doi: 10.2147/COPD.S539307. Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global strategy for prevention, diagnosis and management of chronic obstructive pulmonary disease 2025. 2025. Available from: https://goldcopd.org/2025-gold-report/. Accessed May 13, 2025. Brown KH, Ghita-Pettigrew M, Kerr BN, et al. Characterisation of quantitative imaging biomarkers for inflammatory and fibrotic radiation-induced lung injuries using preclinical radiomics. Radiother Oncol. 2024 Mar;192:110106. doi: 10.1016/j.radonc.2024.110106. Zhong N, Wang C, Yao W, et al. Prevalence of chronic obstructive pulmonary disease in China: a large, population-based survey. Am J Respir Crit Care Med. 2007;176(8):753–60. doi: 10.1164/rccm.200612-1749oc. Mapel DW, Dalal AA, Blanchette CM, et al. Severity of COPD at initial spirometry-confirmed diagnosis: data from medical charts and administrative claims. Int J Chron Obstruct Pulmon Dis. 2011;6:573-81. doi: 10.2147/COPD.S16975. Bellamy D, Smith J. Role of primary care in early diagnosis and effective management of COPD. Int J Clin Pract. 2007;61(8):1380–9. doi: 10.1111/j.1742-1241.2007.01447.x. Pu Y, Zhou X, Zhang D, et al. Re-defining high risk COPD with parameter response mapping based on machine learning models. Int J Chron Obstruct Pulmon Dis. 2022;17:2471–2483. doi: 10.2147/COPD.S369904. Chen TT, Cheng TY, Liu IJ,et al. Leveraging Subjective Parameters and Biomarkers in Machine Learning Models: The Feasibility of lnc-IL7R for Managing Emphysema Progression. Diagnostics (Basel). 2025 May 3;15(9):1165. doi: 10.3390/diagnostics15091165. Rueda R, Fabello E, Silva T, et al. Machine learning approach to flare-up detection and clustering in chronic obstructive pulmonary disease (COPD) patients. Health Inf Sci Syst. 2024 Oct 23;12(1):50. doi: 10.1007/s13755-024-00308-4. Gu YF, Chen L, Qiu R, et al. Development of a model for predicting the severity of chronic obstructive pulmonary disease. Front Med (Lausanne). 2022 Dec 16;9:1073536. doi: 10.3389/fmed.2022.1073536. Xu C, Qi S, Feng J, et al. DCT-MIL: Deep CNN transferred multiple instance learning for COPD identification using CT images. Phys Med Biol. 2020;65:145011. doi: 10.1088/1361-6560/ab857d. Matsumura K, Ito S. Novel biomarker genes which distinguish between smokers and chronic obstructive pulmonary disease patients with machine learning approach. BMC Pulm Med. 2020;20(1):29. doi: 10.1186/s12890-020-1062-9. Sui H, Mo Z, Wei Y, et al. Diagnosis and Severity Assessment of COPD Based on Machine Learning of Chest CT Images. Int J Chron Obstruct Pulmon Dis. 2025 Aug 14;20:2853-2867. doi: 10.2147/COPD.S528988. PMID: 40831903; PMCID: PMC12360390. Li Z, Liu L, Zhang Z, et al. A Novel CT-Based Radiomics Features Analysis for Identification and Severity Staging of COPD. Acad Radiol. 2022 May;29(5):663-673. doi: 10.1016/j.acra.2022.01.004. Zhou T, Zhou X, Ni J, et al. A CT-Based Lung Radiomics Nomogram for Classifying the Severity of Chronic Obstructive Pulmonary Disease. Int J Chron Obstruct Pulmon Dis. 2024 Dec 11;19:2705-2717. doi: 10.2147/COPD.S483007. Zhao M, Wu Y, Li Y, et al. Learning and depicting lobe-based radiomics feature for COPD Severity staging in low-dose CT images. BMC Pulm Med. 2024 Jun 24;24(1):294. doi: 10.1186/s12890-024-03109-3. Rezvanjou S, Moslemi A, Peterson S, et al. Classifying chronic obstructive pulmonary disease status using computed tomography imaging and convolutional neural networks: comparison of model input image types and training data severity. J Med Imaging (Bellingham). 2025 May;12(3):034502. doi: 10.1117/1.JMI.12.3.034502. Zou X, Ren Y, Yang H, et al. Screening and staging of chronic obstructive pulmonary disease with deep learning based on chest X-ray images and clinical parameters. BMC Pulm Med. 2024 Mar 26;24(1):153. doi: 10.1186/s12890-024-02945-7. Additional Declarations No competing interests reported. Supplementary Files SupplementaryData.docx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 18 May, 2026 Reviews received at journal 15 May, 2026 Reviewers agreed at journal 05 May, 2026 Reviews received at journal 23 Apr, 2026 Reviewers agreed at journal 14 Apr, 2026 Reviewers agreed at journal 13 Apr, 2026 Reviewers invited by journal 13 Apr, 2026 Editor assigned by journal 13 Apr, 2026 Editor invited by journal 06 Apr, 2026 Submission checks completed at journal 04 Apr, 2026 First submitted to journal 04 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9272720","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":625003247,"identity":"e5aae44a-8f21-42a1-a57b-15ef94bc97e3","order_by":0,"name":"Qiang Fei","email":"","orcid":"","institution":"The Linghu People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qiang","middleName":"","lastName":"Fei","suffix":""},{"id":625003248,"identity":"ad513e9b-e7c9-4180-be91-ca604d7cf0a6","order_by":1,"name":"Xiaoli Mei","email":"","orcid":"","institution":"The Linghu People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xiaoli","middleName":"","lastName":"Mei","suffix":""},{"id":625003249,"identity":"a9e76441-e47f-4d31-ab30-7bc1db5d4c76","order_by":2,"name":"Jiacheng Zhao","email":"","orcid":"","institution":"Tokyo Denki University","correspondingAuthor":false,"prefix":"","firstName":"Jiacheng","middleName":"","lastName":"Zhao","suffix":""},{"id":625003251,"identity":"b2a1711e-3200-4497-9466-263e38d0354a","order_by":3,"name":"Yili Xing","email":"","orcid":"","institution":"The Linghu People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yili","middleName":"","lastName":"Xing","suffix":""},{"id":625003252,"identity":"ba560481-f445-4caa-a828-fd86ba4acf42","order_by":4,"name":"Yuanbin Li","email":"","orcid":"","institution":"The First Affiliated Hospital of Huzhou University","correspondingAuthor":false,"prefix":"","firstName":"Yuanbin","middleName":"","lastName":"Li","suffix":""},{"id":625003254,"identity":"6a87db76-5bb3-4ebb-b68d-c3a14de1933a","order_by":5,"name":"Mei Yang","email":"","orcid":"","institution":"The First Affiliated Hospital of Huzhou University","correspondingAuthor":false,"prefix":"","firstName":"Mei","middleName":"","lastName":"Yang","suffix":""},{"id":625003256,"identity":"36ac5c49-c172-441d-8979-314e66c246e4","order_by":6,"name":"Jingnan Xue","email":"","orcid":"","institution":"The First Affiliated Hospital of Huzhou University","correspondingAuthor":false,"prefix":"","firstName":"Jingnan","middleName":"","lastName":"Xue","suffix":""},{"id":625003260,"identity":"9df68f66-092f-407b-81b3-ce9e3a6484a1","order_by":7,"name":"Luying Qi","email":"","orcid":"","institution":"The First Affiliated Hospital of Huzhou University","correspondingAuthor":false,"prefix":"","firstName":"Luying","middleName":"","lastName":"Qi","suffix":""},{"id":625003262,"identity":"df84964e-5d75-45f5-b96e-8c9b0e785412","order_by":8,"name":"Yifeng Zheng","email":"","orcid":"","institution":"Huzhou Central Hospital affiliated to Huzhou University","correspondingAuthor":false,"prefix":"","firstName":"Yifeng","middleName":"","lastName":"Zheng","suffix":""},{"id":625003263,"identity":"e123adf9-a233-4409-afd5-1c40e7c3fa2e","order_by":9,"name":"Hongxing Zhao","email":"","orcid":"","institution":"The First Affiliated Hospital of Huzhou University","correspondingAuthor":false,"prefix":"","firstName":"Hongxing","middleName":"","lastName":"Zhao","suffix":""},{"id":625003264,"identity":"b06af858-b3e2-43a1-9450-9e64d899643c","order_by":10,"name":"Hupo Bian","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2UlEQVRIiWNgGAWjYBACfvbmAwcS/7DJ8TMcPkCcFsmeY4kHPjbwGUs2HksgTovBDR/lgzMb5BI3HD5jQKTLbvAwHObdYZa44diZjzfeMNjJ6TYQ0ME4u/fAYd4zacYzz5zdbDmHIdnY7AABLcwy5xIO87Adk+27cXabNA/DgcRthLSwSeQYALX8Z2y4/+YZcVp4gFoOzmxjU5xw4AwbcVokeI4lHPhwhs1YsuGYseUcAyL8Yn+8+fCHhApwVD688abCTo6gFjQriY0aJC2k6hgFo2AUjIIRAQDw0k61LTRXRgAAAABJRU5ErkJggg==","orcid":"","institution":"The First Affiliated Hospital of Huzhou University","correspondingAuthor":true,"prefix":"","firstName":"Hupo","middleName":"","lastName":"Bian","suffix":""}],"badges":[],"createdAt":"2026-03-31 01:08:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9272720/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9272720/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107378062,"identity":"59c70889-9744-44c6-9d32-7b45acbe0737","added_by":"auto","created_at":"2026-04-21 01:33:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":302558,"visible":true,"origin":"","legend":"\u003cp\u003eThe flowchart of the entire research design.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/ebbbfc0f7f049b7be2e755f7.png"},{"id":107488299,"identity":"1d8bf158-e6d7-47e3-83b0-12984461632b","added_by":"auto","created_at":"2026-04-22 02:44:10","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":274090,"visible":true,"origin":"","legend":"\u003cp\u003eThe flowchart of the patient's selection.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/1e3582eb0b7a7e6309475834.png"},{"id":107378063,"identity":"fa40502e-239c-4f51-86a7-3c2bd3fd6738","added_by":"auto","created_at":"2026-04-21 01:33:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":308203,"visible":true,"origin":"","legend":"\u003cp\u003eA and D represent the LASSO for radiomics and DLR features. B and E represent the MSE for radiomics and DLR features. C and F represent the feature weights for radiomics and DLR features.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/ee7ca90061657d09fe268b46.png"},{"id":107487795,"identity":"8e3f6bd6-4ce7-467e-9144-7b6784f23fa8","added_by":"auto","created_at":"2026-04-22 02:42:49","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":137728,"visible":true,"origin":"","legend":"\u003cp\u003eThe clinical application of nomogram in the prediction of COPD grades.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/d95675fb925e3cf49c8a8e99.png"},{"id":107485856,"identity":"9ac0b1b4-fd1e-4513-8050-28c40d7d1ed4","added_by":"auto","created_at":"2026-04-22 02:36:37","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":165031,"visible":true,"origin":"","legend":"\u003cp\u003eROC curve of different models in the (A) train, (B) internal validation, and (C) external validation cohort, respectively.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/a8a8cae3e9ffa115a2694804.png"},{"id":107378067,"identity":"02e19a00-54ec-4148-966e-93cec7d14831","added_by":"auto","created_at":"2026-04-21 01:33:13","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":180280,"visible":true,"origin":"","legend":"\u003cp\u003eDCA curve of different models in the (A train, (B) internal validation, and (C) external validation cohort, respectively.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/d1a9c3075cef49c444f6fc8a.png"},{"id":107485857,"identity":"f5fffeb2-7e0a-4810-9652-55bf23099ef8","added_by":"auto","created_at":"2026-04-22 02:36:37","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":156217,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration curve of different models in the (A) train, (B) internal validation, and (C) external validation cohort, respectively\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/2b72e2682ebbdf648ad086bd.png"},{"id":107378069,"identity":"332f53b4-5833-4f91-9aef-5dc2eaea6c3a","added_by":"auto","created_at":"2026-04-21 01:33:13","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":170919,"visible":true,"origin":"","legend":"\u003cp\u003epresents the Delong test results for different models evaluated in (A) the training cohort, (B) the internal validation cohort, and (C) the external validation cohort.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/323a3f00556bf06c502159cc.png"},{"id":107705623,"identity":"bb3797de-7a29-4b02-afba-24060153e6be","added_by":"auto","created_at":"2026-04-24 09:13:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2294787,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/879075b6-2dbc-44d4-9886-53e906af5912.pdf"},{"id":107378061,"identity":"afb8788a-60d3-472c-bdd5-74df3c11682b","added_by":"auto","created_at":"2026-04-21 01:33:13","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":20631,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryData.docx","url":"https://assets-eu.researchsquare.com/files/rs-9272720/v1/ee8a7a9be024e9eeef33463e.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and validation of a multimodal clinical-radiomics-deep learning nomogram based on automated chest CT segmentation for classifying COPD severity: A multicenter study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChronic obstructive pulmonary disease (COPD) is a prevalent, preventable, and treatable condition marked by persistent respiratory symptoms and airflow limitation. It has emerged as the third leading cause of death worldwide, imposing a substantial burden on health systems and society [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Research indicates that increased COPD severity correlates with a higher frequency of exacerbations, leading to elevated mortality rates [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Furthermore, COPD frequently coexists with other chronic conditions, such as cardiovascular disease, osteoporosis, diabetes, lung cancer, and cachexia [\u003cspan additionalcitationids=\"CR6 CR7\" citationid=\"CR4\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. These comorbidities collectively diminish patients' quality of life and increase mortality. Treatment strategies for COPD vary according to disease severity [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e9\u003c/span\u003e], highlighting the importance of accurately assessing COPD severity to improve patient prognosis and reduce mortality.\u003c/p\u003e \u003cp\u003ePulmonary function testing (PFT) remains the gold standard for diagnosing and staging COPD, categorizing severity into four stages: mild, moderate, severe, and very severe [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Despite its status, PFT utilization remains low; in China, only 6.7% of individuals aged 40 and above have undergone PFT [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Moreover, PFT has notable limitations: it reflects only a single dimension of airflow limitation and fails to capture regional lung tissue changes, such as emphysema, small airway disease, and airway remodeling [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Meanwhile, detection tools such as the Lung Function Questionnaire (LFQ), the COPD Diagnosis Questionnaire (CDQ), and the COPD Population Screening Tool (COPD-PS) have demonstrated their practicality in identifying individuals at high risk of COPD. However, these tools mainly rely on self-reported symptoms and risk factors, which may lead to recall bias, and they cannot directly assess potential structural changes in the lungs. With the widespread adoption of lung cancer screening programs, chest computed tomography (CT) has become increasingly accessible, offering detailed anatomical information. Studies have shown that COPD of varying severities exhibits distinct CT morphological features. Quantitative imaging has been employed to assess COPD severity and predict coronary heart disease risk [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Cho et al. [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e15\u003c/span\u003e] demonstrated that quantitative pulmonary vascular characteristics correlate with the severity and extent of emphysema in COPD. Leveraging CT images for COPD staging could thus enable timely \"one-stop\" medical interventions, yielding greater benefits [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. However, traditional CT assessment relies heavily on radiologists' subjective visual evaluation of manifestations such as emphysema and airway wall thickening, making it challenging to detect complex image texture patterns imperceptible to the human eye.\u003c/p\u003e \u003cp\u003eIn recent years, artificial intelligence has been increasingly applied in COPD research. Radiomics techniques quantify the heterogeneity, texture, and morphological features of lesion regions by extracting numerous quantitative features from medical images [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Given COPD's diffuse and heterogeneous nature, the entire lung is designated as the region of interest. Meanwhile, deep learning (DL) technology has made significant advances in medical image analysis. DL models automatically learn hierarchical deep feature representations from raw images, overcoming the limitations of manual feature engineering and excelling in image classification, segmentation, and prediction tasks [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Combining deep learning with radiomics offers complementary advantages: deep learning mines deep implicit features automatically, while radiomics provides interpretable quantitative features, jointly enhancing predictive model performance.\u003c/p\u003e \u003cp\u003eOur team has found that a fusion model integrating radiomics and deep learning via one-stop chest CT can effectively predict the risk of concurrent coronary heart disease in COPD patients [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. We hypothesize that a fusion model combining whole-lung radiomics from chest CT with deep learning can identify COPD severity in routine clinical practice. This approach would enhance the value of chest CT, particularly for patients unable to undergo PFT, by providing both morphological information and pulmonary function assessment. To facilitate clinical use, we integrate automatic deep learning features with interpretable radiomics features to comprehensively explore effective information in CT images and combine it with key clinical factors to construct a nomogram model. This provides clinicians with a visual and quantitative individualized assessment tool.\u003c/p\u003e"},{"header":"Methods and Materials","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatients and clinical data\u003c/h2\u003e \u003cp\u003e This study has been approved by the Institutional Review Board of The First Affiliated Hospital of Huzhou University (Approval Number: 2025KYLL002-01). Given the retrospective nature of the study, the requirement for informed consent was waived. This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Clinical data and tomographic images of eligible patients were collected from three hospitals from January 1, 2021, to May 30, 2025. During this period, a total of 1,794 patients diagnosed with chronic obstructive pulmonary disease (COPD) through pulmonary function tests (PFTs) at these three centers were included in this study.\u003c/p\u003e \u003cp\u003eThe inclusion criteria were as follows: (1) Diagnosed with chronic obstructive pulmonary disease confirmed by pulmonary function tests; (2) Completion of pulmonary function tests and chest CT scans within 2 weeks; (3) Possession of complete thin-slice (1 mm) chest CT images. The exclusion criteria were as follows: (1) Incomplete clinical data, concurrent other thoracic diseases (such as pneumonia, atelectasis, pulmonary nodules or masses larger than 6 mm, and pleural effusion); (2) Concurrent any malignancy; (3) Presence of spinal implants or significant image artifacts, with poor image quality affecting diagnosis; (4) Lack of thin-slice chest CT images. We randomly assigned 1,313 patients from the first center to the training cohort (n\u0026thinsp;=\u0026thinsp;919) or the internal validation cohort (n\u0026thinsp;=\u0026thinsp;394) in a ratio of 7:3, and allocated 434 patients from the second center and 47 patients from the third center to the external validation cohort (n\u0026thinsp;=\u0026thinsp;481). Clinical information included age, body mass index (BMI), gender, smoking status, as well as laboratory test indicators such as C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and procalcitonin (PCT). For some missing values, we used the mean to fill them in, and for data with missing values accounting for more than 30% of the information, we did not include them in the final study.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCT image acquisition and pulmonary function examination\u003c/h3\u003e\n\u003cp\u003eParticipants underwent non-enhanced CT scans using equipment from manufacturers such as Aquilion ONE TSX-301C and Brilliance CT 16. Axial CT images of the entire thorax were acquired during full inspiration (scanning parameters are detailed in Supplementary Data 1). Pulmonary function testing was conducted using the Ganshorn PowerCube. The diagnosis and classification of chronic obstructive pulmonary disease (COPD) adhered to the criteria outlined by the Global Initiative for Chronic Obstructive Lung Disease (GOLD) [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. COPD was defined as a post-bronchodilator ratio of forced expiratory volume in the first second (FEV1) to forced vital capacity (FVC) below 0.7. In this study, COPD severity was categorized into two stages according to GOLD criteria: mild to moderate airflow limitation (GOLD stages 1\u0026ndash;2, FEV1\u0026thinsp;\u0026ge;\u0026thinsp;50% of predicted value) and severe to very severe airflow limitation (GOLD stages 3\u0026ndash;4, FEV1\u0026thinsp;\u0026lt;\u0026thinsp;50% of predicted value).\u003c/p\u003e\n\u003ch3\u003eWhole lung automatic segmentation\u003c/h3\u003e\n\u003cp\u003eA deep learning model (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/OnekeyAI-Platform/onekey\u003c/span\u003e\u003cspan address=\"https://github.com/OnekeyAI-Platform/onekey\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was employed to segment the left and right lungs, and the extracted regions were merged into a single region of interest (ROI). The specific algorithm process is detailed in Supplementary Data 2. To assess the consistency between automatic and manual segmentation results, 100 samples were randomly selected and independently evaluated by two chest radiologists, each with over 10 years of experience, using ITK-SNAP (version 3.8.0, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ewww.itksnap.org\u003c/span\u003e\u003c/span\u003e). The Dice index was then calculated to objectively quantify the spatial overlap of contours and determine the agreement between fully automatic and manual segmentation results. Finally, automatic segmentation was applied to the remaining samples.\u003c/p\u003e\n\u003ch3\u003eExtraction and selection of radiomics features\u003c/h3\u003e\n\u003cp\u003eWe employed Pyradiomics (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://pyradiomics.readthedocs.io\u003c/span\u003e\u003cspan address=\"http://pyradiomics.readthedocs.io\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to extract radiomic features from segmented lung regions. Before feature extraction, the images underwent a three-step preprocessing procedure to standardize them. First, the images were resampled to a resolution of 1 mm \u0026times; 1 mm \u0026times; 1 mm. Second, the gray-level values were adjusted to a 25 Gy level to mitigate the effects of varying slice thicknesses and reduce noise interference. Subsequently, features were extracted using Z-score normalization, yielding three categories of features: first-order, shape, and texture features. Statistical significance was evaluated using the t-test, and only features with a p-value less than 0.05 were retained. To address collinearity, we assessed the correlation between features using the Pearson correlation coefficient and excluded features with a correlation coefficient exceeding 0.9. Furthermore, within a 10-fold cross-validation framework, we used Lasso regression to optimize the feature set by determining the optimal regularization parameter λ and selecting features with non-zero coefficients. Finally, the radiomic score (Rad Score, RS) was computed as a linear combination of the retained features and their corresponding coefficients.\u003c/p\u003e\n\u003ch3\u003eExtraction and selection of deep learning features\u003c/h3\u003e\n\u003cp\u003eIn this study, the ResNet50 architecture was employed as a convolutional neural network (CNN) to extract deep learning features. Image intensity distributions were standardized using Z-score normalization, and the standardized images were used as inputs to the deep learning model. During training, real-time data augmentation techniques\u0026mdash;such as random cropping and flipping\u0026mdash;were applied to improve model robustness. For test images, preprocessing was restricted to normalization to ensure consistency during evaluation. A pre-trained CNN model was used to extract deep transfer learning (DTL) features from the largest region of interest (ROI) in each image, specifically from the penultimate layer. In the proposed model, the output probabilities computed by the CNN were defined as deep learning feature signatures. Given the complexity of the CNN, principal component analysis (PCA) was applied to reduce these features to a 512-dimensional space for improved manageability.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eModel construction and evaluation\u003c/h2\u003e \u003cp\u003eFollowing feature selection, five distinct models were constructed. Logistic regression analysis was initially applied to clinical features to identify statistically significant predictors, which were then used to develop a clinical model. Using the selected radiomic and deep learning features, an imaging biomarker (Rad) model and a deep learning (DTL) model were constructed separately. To create a deep learning imaging biomarker (DLR) model, a fusion algorithm was employed to integrate deep learning features with radiomic features. To improve clinical applicability, univariate and stepwise multivariate analyses were performed on all clinical features to identify significant predictors. These selected clinical features were subsequently combined with the predictive results of the DLR model to develop a logistic regression (LR)-based linear model, termed the comprehensive model, which was effectively visualized through a nomogram. The models were evaluated using the area under the curve (AUC), accuracy, sensitivity, specificity, recall, and F1 score. Model performance was further assessed using positive predictive value (PPV) and negative predictive value (NPV). The AUCs of different models were compared using the DeLong test, and calibration was analyzed using calibration curves (Hosmer-Lemeshow test) to verify reliability. Decision curve analysis (DCA) was also conducted to evaluate the clinical utility of the predictive models, providing insights into their potential benefits in clinical settings. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents a flowchart of the entire study design.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eThe Shapiro-Wilk test was employed to evaluate the normality of clinical characteristics. Continuous variables were analyzed using the t-test or the Mann-Whitney U test, depending on their distribution. Categorical variables were assessed using Chi-square (χ\u0026sup2;) tests.\u003c/p\u003e \u003cp\u003eAll data analyses were performed using Python 3.7.12. Statistical analyses utilized Statsmodels version 0.13.2, while PyRadiomics version 3.0.1 was used for radiomics feature extraction. Scikit-learn version 1.0.2 facilitated machine learning tasks, and PyTorch version 1.11.0 was employed for deep learning framework development, with performance optimized using CUDA version 11.3.1 and cuDNN version 8.2.1.\u003c/p\u003e \u003c/div\u003e"},{"header":"Result","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eBaseline characteristics and clinical feature screening of the patients\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e presents the flowchart illustrating patient selection. As of May 30, 2025, a total of 2,658 patients diagnosed with COPD were included in the study. Following screening based on inclusion and exclusion criteria, the final cohort comprised 1,794 patients, including 276 females and 1,518 males. The mean age of the cohort was 74.02 years. The number of patients from the three hospitals was 1,313, 434, and 47, respectively. Among these patients, 989 had COPD classified as GOLD I\u0026ndash;II, and 805 had COPD classified as GOLD III\u0026ndash;IV. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e provides details on the baseline clinical characteristics of the study cohort (standardized units for all indicators in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e are available in Supplementary Data 5).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics of the study population\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"13\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eClinical factors\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e \u003cp\u003eTraining cohort(n\u0026thinsp;=\u0026thinsp;919)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003eInternal validation cohort (n\u0026thinsp;=\u0026thinsp;394)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003eExternal vaidation cohort(n\u0026thinsp;=\u0026thinsp;481)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGOLD I\u0026ndash;II (N\u0026thinsp;=\u0026thinsp;554)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGOLD III\u0026ndash;IV (N\u0026thinsp;=\u0026thinsp;365)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eGOLD I\u0026ndash;II (N\u0026thinsp;=\u0026thinsp;228)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eGOLD III\u0026ndash;IV (N\u0026thinsp;=\u0026thinsp;166)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eGOLD I\u0026ndash;II (N\u0026thinsp;=\u0026thinsp;207)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eGOLD III\u0026ndash;IV (N\u0026thinsp;=\u0026thinsp;274)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e74.38\u0026thinsp;\u0026plusmn;\u0026thinsp;8.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73.76\u0026thinsp;\u0026plusmn;\u0026thinsp;8.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.387\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e75.04\u0026thinsp;\u0026plusmn;\u0026thinsp;8.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e73.84\u0026thinsp;\u0026plusmn;\u0026thinsp;8.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.127\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e73.33\u0026thinsp;\u0026plusmn;\u0026thinsp;8.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e73.35\u0026thinsp;\u0026plusmn;\u0026thinsp;7.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.96\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e508.90\u0026thinsp;\u0026plusmn;\u0026thinsp;246.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e472.64\u0026thinsp;\u0026plusmn;\u0026thinsp;256.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e545.96\u0026thinsp;\u0026plusmn;\u0026thinsp;228.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e431.58\u0026thinsp;\u0026plusmn;\u0026thinsp;239.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e495.43\u0026thinsp;\u0026plusmn;\u0026thinsp;230.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e473.70\u0026thinsp;\u0026plusmn;\u0026thinsp;276.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.315\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlbumin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39.66\u0026thinsp;\u0026plusmn;\u0026thinsp;14.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38.59\u0026thinsp;\u0026plusmn;\u0026thinsp;4.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e39.13\u0026thinsp;\u0026plusmn;\u0026thinsp;4.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e38.66\u0026thinsp;\u0026plusmn;\u0026thinsp;4.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.275\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e38.05\u0026thinsp;\u0026plusmn;\u0026thinsp;5.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e37.45\u0026thinsp;\u0026plusmn;\u0026thinsp;4.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlobulin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e104.55\u0026thinsp;\u0026plusmn;\u0026thinsp;42.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e98.36\u0026thinsp;\u0026plusmn;\u0026thinsp;43.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.011\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e104.65\u0026thinsp;\u0026plusmn;\u0026thinsp;46.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e94.98\u0026thinsp;\u0026plusmn;\u0026thinsp;41.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.063\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e100.43\u0026thinsp;\u0026plusmn;\u0026thinsp;39.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e90.53\u0026thinsp;\u0026plusmn;\u0026thinsp;40.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriglyceride\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.16\u0026thinsp;\u0026plusmn;\u0026thinsp;0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.04\u0026thinsp;\u0026plusmn;\u0026thinsp;0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.16\u0026thinsp;\u0026plusmn;\u0026thinsp;0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.41\u0026thinsp;\u0026plusmn;\u0026thinsp;5.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e2.01\u0026thinsp;\u0026plusmn;\u0026thinsp;12.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.24\u0026thinsp;\u0026plusmn;\u0026thinsp;4.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlkaline Phosphatase\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e89.31\u0026thinsp;\u0026plusmn;\u0026thinsp;67.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e83.30\u0026thinsp;\u0026plusmn;\u0026thinsp;41.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.056\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e87.33\u0026thinsp;\u0026plusmn;\u0026thinsp;48.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e87.18\u0026thinsp;\u0026plusmn;\u0026thinsp;37.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.801\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e80.39\u0026thinsp;\u0026plusmn;\u0026thinsp;59.90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e76.89\u0026thinsp;\u0026plusmn;\u0026thinsp;27.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.761\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWhite blood cell count\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.74\u0026thinsp;\u0026plusmn;\u0026thinsp;3.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.31\u0026thinsp;\u0026plusmn;\u0026thinsp;3.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6.48\u0026thinsp;\u0026plusmn;\u0026thinsp;2.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e6.95\u0026thinsp;\u0026plusmn;\u0026thinsp;3.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.469\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e6.68\u0026thinsp;\u0026plusmn;\u0026thinsp;3.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e7.27\u0026thinsp;\u0026plusmn;\u0026thinsp;3.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.179\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeutrophilic granulocyte percentage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e248.91\u0026thinsp;\u0026plusmn;\u0026thinsp;105.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e286.46\u0026thinsp;\u0026plusmn;\u0026thinsp;109.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e252.74\u0026thinsp;\u0026plusmn;\u0026thinsp;107.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e274.40\u0026thinsp;\u0026plusmn;\u0026thinsp;114.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.032\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e216.81\u0026thinsp;\u0026plusmn;\u0026thinsp;108.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e247.69\u0026thinsp;\u0026plusmn;\u0026thinsp;110.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePercentage of lymphocytes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26.95\u0026thinsp;\u0026plusmn;\u0026thinsp;118.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18.65\u0026thinsp;\u0026plusmn;\u0026thinsp;9.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e21.75\u0026thinsp;\u0026plusmn;\u0026thinsp;9.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e21.20\u0026thinsp;\u0026plusmn;\u0026thinsp;17.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.049\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e23.80\u0026thinsp;\u0026plusmn;\u0026thinsp;10.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e20.16\u0026thinsp;\u0026plusmn;\u0026thinsp;9.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbsolute eosinophil count\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15.12\u0026thinsp;\u0026plusmn;\u0026thinsp;16.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.59\u0026thinsp;\u0026plusmn;\u0026thinsp;15.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e14.09\u0026thinsp;\u0026plusmn;\u0026thinsp;12.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e13.52\u0026thinsp;\u0026plusmn;\u0026thinsp;15.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.087\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e17.08\u0026thinsp;\u0026plusmn;\u0026thinsp;16.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e17.32\u0026thinsp;\u0026plusmn;\u0026thinsp;18.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.365\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlateletcrit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.516\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.97\u0026thinsp;\u0026plusmn;\u0026thinsp;11.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.25\u0026thinsp;\u0026plusmn;\u0026thinsp;0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.524\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRed blood cell distribution width\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.48\u0026thinsp;\u0026plusmn;\u0026thinsp;1.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.43\u0026thinsp;\u0026plusmn;\u0026thinsp;1.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.967\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e13.57\u0026thinsp;\u0026plusmn;\u0026thinsp;1.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e13.27\u0026thinsp;\u0026plusmn;\u0026thinsp;1.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.136\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e13.71\u0026thinsp;\u0026plusmn;\u0026thinsp;7.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e13.36\u0026thinsp;\u0026plusmn;\u0026thinsp;1.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.377\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean platelet volume\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.48\u0026thinsp;\u0026plusmn;\u0026thinsp;5.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.11\u0026thinsp;\u0026plusmn;\u0026thinsp;1.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.734\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e10.25\u0026thinsp;\u0026plusmn;\u0026thinsp;1.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e9.92\u0026thinsp;\u0026plusmn;\u0026thinsp;1.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e10.61\u0026thinsp;\u0026plusmn;\u0026thinsp;1.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e10.42\u0026thinsp;\u0026plusmn;\u0026thinsp;1.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.101\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet Distribution Width\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.05\u0026thinsp;\u0026plusmn;\u0026thinsp;1.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15.91\u0026thinsp;\u0026plusmn;\u0026thinsp;1.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.639\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e16.16\u0026thinsp;\u0026plusmn;\u0026thinsp;1.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e15.93\u0026thinsp;\u0026plusmn;\u0026thinsp;1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e12.93\u0026thinsp;\u0026plusmn;\u0026thinsp;2.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e12.97\u0026thinsp;\u0026plusmn;\u0026thinsp;2.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eC-Reactive Protein\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e187.51\u0026thinsp;\u0026plusmn;\u0026thinsp;172.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e210.22\u0026thinsp;\u0026plusmn;\u0026thinsp;176.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e205.12\u0026thinsp;\u0026plusmn;\u0026thinsp;168.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e196.15\u0026thinsp;\u0026plusmn;\u0026thinsp;175.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.565\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e325.33\u0026thinsp;\u0026plusmn;\u0026thinsp;208.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e332.22\u0026thinsp;\u0026plusmn;\u0026thinsp;211.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.736\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArterial Oxygen Partial Pressure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e89.26\u0026thinsp;\u0026plusmn;\u0026thinsp;22.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e87.04\u0026thinsp;\u0026plusmn;\u0026thinsp;21.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e87.39\u0026thinsp;\u0026plusmn;\u0026thinsp;18.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e90.20\u0026thinsp;\u0026plusmn;\u0026thinsp;25.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.947\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e84.18\u0026thinsp;\u0026plusmn;\u0026thinsp;9.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e82.86\u0026thinsp;\u0026plusmn;\u0026thinsp;15.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.053\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArterial Carbon Dioxide Partial Pressure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.85\u0026thinsp;\u0026plusmn;\u0026thinsp;3.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e41.32\u0026thinsp;\u0026plusmn;\u0026thinsp;5.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e39.05\u0026thinsp;\u0026plusmn;\u0026thinsp;3.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e42.13\u0026thinsp;\u0026plusmn;\u0026thinsp;7.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e40.13\u0026thinsp;\u0026plusmn;\u0026thinsp;5.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e58.63\u0026thinsp;\u0026plusmn;\u0026thinsp;277.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.119\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.778\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e80(14.44)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e78(21.37)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e38(16.67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e39(23.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e19(9.18)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e22(8.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e474(85.56)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e287(78.63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e190(83.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e127(76.51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e188(90.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e252(91.97)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmoke\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.956\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.292\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e189(34.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e126(34.52)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e75(32.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e64(38.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e89(43.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e118(43.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e365(65.88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e239(65.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e153(67.11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e102(61.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e118(57.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e156(56.93)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eUnivariate and multivariate analyses revealed age, gender, and arterial partial pressure of carbon dioxide (PaCO2) as predictors in the clinical model; these variables also independently predicted the severity of chronic obstructive pulmonary disease (COPD), as detailed in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The area under the receiver operating characteristic curve (AUC) values for the clinical model were 0.634 (95% confidence interval: 0.596\u0026ndash;0.671) in the training set, 0.63 (95% confidence interval: 0.572\u0026ndash;0.688) in the internal validation set, and 0.616 (95% confidence interval: 0.566\u0026ndash;0.666) in the external validation set, as presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnivariable and multivariable analysis of clinical features.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eUnivariable analysis\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eMultivariable analysis\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOR[95%CI]\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOR[95%CI]\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlateletcrit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.128[0.073,0.224]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.399[0.067,2.373]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.397\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.606[0.535,0.685]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.485[0.34,0.693]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmoke\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.655[0.571,0.751]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.032[0.77,1.383]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.858\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriglyceride\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.699[0.638,0.765]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.887[0.732,1.075]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.305\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean platelet volume\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.96[0.949,0.97]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.97[0.924,1.017]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.288\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWhite blood cell count\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.961[0.947,0.975]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.025[0.987,1.065]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.277\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRed blood cell distribution width\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.969[0.962,0.977]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.964[0.891.1.043]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.442\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet Distribution Width\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.974[0.968,0.98]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.925[0.856,1.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.102\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePercentage of lymphocytes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.977[0.972,0.982]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.971[0.94,1.002]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.12\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbsolute eosinophil count\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.981[0.975,0.986]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.996[0.987,1.005]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.484\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ealbumin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.989[0.986,0.992]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.996[0.982,1.01]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.661\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArterial Carbon Dioxide Partial Pressure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.991[0.988,0.994]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.15[1.115,1.185]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.994[0.993,0.996]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.981[0.967,0.995]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.028\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArterial Oxygen Partial Pressure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.995[0.994,0.996]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.994[0.989,1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.111\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlkaline Phosphatase\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.996[0.994,0.997]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.998[0.995,1.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.228\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlobulin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.996[0.995,0.997]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.997[0.994,1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.105\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.999[0.999,0.999]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.999[0.999,1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.067\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eC-Reactive Protein\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.999[0.999,1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.001[1,1.002]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.056\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeutrophilic granulocyte percentage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.999[0.999,0.999]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1[0.997,1.004]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.818\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003epresents the performance of four models across different cohorts. The models include the Clinic (clinical model signature), Rad (radiomics signature), DTL (deep learning signature),DLR (deep learning radiomics signature), and Combined (a combination of clinical, deep learning, and radiomics signatures).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"13\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eClinic\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e95% CI\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ePrecision\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eRecall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eThreshold\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eCohort\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.656\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.5963\u0026ndash;0.6708\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.471\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.778\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.583\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.691\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.583\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0.471\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.521\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.402\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003etrain\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRad\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.712\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7274\u0026ndash;0.7918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.751\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.766\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.642\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.374\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etrain\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDTL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.805\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.737\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7775\u0026ndash;0.8333\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.767\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.661\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.661\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.468\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etrain\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDLR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.822\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.743\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7940\u0026ndash;0.8491\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.775\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.722\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.775\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.706\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.379\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etrain\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCombined\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.839\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.777\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.8134\u0026ndash;0.8652\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.679\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.841\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.738\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.799\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.738\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.679\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.708\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.493\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etrain\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinic\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.637\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.5720\u0026ndash;0.6875\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.506\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.732\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.579\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.671\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.579\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.506\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003eval\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRad\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.729\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.6783\u0026ndash;0.7792\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.627\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.737\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.627\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.396\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003eval\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDTL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7088\u0026ndash;0.8043\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.711\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.658\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.602\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.758\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.602\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.711\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.391\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003eval\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDLR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.688\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.6912\u0026ndash;0.7894\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.681\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.693\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.749\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.681\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.386\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003eval\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCombined\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.759\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.706\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7100\u0026ndash;0.8074\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.735\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.684\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.629\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.629\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.735\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.678\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.326\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003eval\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinic\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.616\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.593\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.5663\u0026ndash;0.6659\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.467\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.758\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.719\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.518\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.719\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.467\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.566\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.398\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etest\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRad\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.704\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.674\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.6571\u0026ndash;0.7506\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.638\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.719\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.719\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.356\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etest\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDTL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.752\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.711\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7091\u0026ndash;0.7957\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.614\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.729\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.683\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.729\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.756\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etest\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDLR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.759\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.699\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7169\u0026ndash;0.8021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.664\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.744\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.626\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.664\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.715\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etest\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCombined\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.767\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7247\u0026ndash;0.8097\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.646\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.773\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.623\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.646\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.711\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003etest\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eFeature selection and model construction\u003c/h2\u003e \u003cp\u003eRadiological features and three-dimensional deep learning features were extracted from images based on the automatic segmentation of chest CT scans. Reproducibility was assessed using the intraclass correlation coefficient (ICC). Following t-tests, Spearman correlation analyses, and LASSO screening, 20 optimal radiomic features with non-zero coefficients were selected, and a radiological model was constructed. The coefficients and average standard errors derived from five-fold cross-validation are presented in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, while Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC shows the values of the selected features with non-zero coefficients. The specific formula for screening radiomic features is detailed in Supplementary Data 3.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe ResNet50 architecture, a convolutional neural network (CNN), was employed to extract deep learning features, yielding a total of 2048 features. After compression, 32 deep learning features were retained to construct the deep learning model (DTL). To enhance the accuracy of predicting chronic obstructive pulmonary disease (COPD) severity, radiological and deep learning features were integrated. Through t-tests, Spearman correlation coefficients, and LASSO regression analysis, 19 features with non-zero coefficients were selected (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, E, and F), enabling the construction of a deep learning radiological model (DLR). The formula used to screen DLR features is provided in Supplementary Data 4. Finally, age, gender, and partial pressure of carbon dioxide (PaCO2) were combined with the DLR model, and a nomogram model was constructed using multivariate logistic regression. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e presents a nomogram designed specifically for clinical use.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eComparison of clinical models, imaging-based models, deep learning models, DLR models and combined models\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the performance of clinical, radiomics, deep learning (DTL), deep learning radiomics (DLR), and combined models. In the training cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA), the combined model achieved the highest area under the curve (AUC) of 0.839, followed by the DLR model (0.822), the DTL model (0.805), and the radiomics (Rad) model (0.76). The clinical model had a relatively low AUC of 0.634. In the internal validation cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), the combined model again outperformed others with an AUC of 0.759, followed by the DTL model (0.757), the DLR model (0.74), and the Rad model (0.729). The clinical model had an AUC of 0.63. In the external validation cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC), the combined model had the highest AUC (0.767), followed by the DLR model (0.759), the DTL model (0.752), and the Rad model (0.704). The clinical model had an AUC of 0.616. These results consistently demonstrate the low performance of the clinical model across cohorts, highlighting its limitations for standalone use. The Rad and DTL models showed moderate performance, with one slightly outperforming the other in different contexts. The combined model generally exhibited the best AUC performance, demonstrating its robustness and generalizability. These findings underscore the importance of advanced modeling techniques and multimodal data fusion in predicting chronic obstructive pulmonary disease (COPD) staging. Figure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e displays decision curve analysis (DCA) results, which were used to assess the clinical applicability of the models. The combined model demonstrated a higher net benefit compared to the others, indicating a superior benefit-risk ratio in clinical decision-making. Figure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e presents calibration curves for each cohort, showing the agreement between model-predicted probabilities and observed outcomes. Additionally, Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e shows the results of discriminative performance comparisons among the models using the DeLong test. Based on the analysis, the combined nomogram model exhibited superior predictive efficacy, performing well in terms of both discrimination and calibration.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe developed and validated a composite nomogram that integrates full-lung radiomics features derived from chest CT scans, deep learning features, and clinical independent predictors to assess COPD severity. Radiomics and deep learning features independently stratify COPD severity, demonstrating that chest CT can evaluate both lung structure and functional status.\u003c/p\u003e \u003cp\u003eMild COPD patients are often overlooked due to asymptomatic or mild symptoms [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e23\u003c/span\u003e], whereas severe COPD significantly impairs quality of life and escalates treatment costs [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Differentiated treatment strategies based on COPD severity are critical for patient prognosis [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e9\u003c/span\u003e], necessitating an early, rapid, and accessible diagnostic method for staging. Prior studies have identified morphological CT changes\u0026mdash;such as bronchial wall thickening, tracheal shape alterations, and total lung emphysema percentage\u0026mdash;as correlates of severe COPD [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Machine learning advancements have enabled novel staging approaches: Chen TT [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e26\u003c/span\u003e] and Rueda R [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e27\u003c/span\u003e] combined machine learning models with clinical and biochemical markers, while Gu YF et al. developed a COPD severity prediction model using biochemical and immunological parameters, though its 62.7% accuracy limited clinical utility [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Given COPD\u0026rsquo;s heterogeneous lung involvement, leveraging CT to detect abnormal textures and assess disease status is crucial [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Sui H et al. improved severity classification by integrating lung parenchyma shape, size, distribution, and airway morphology into a machine learning framework [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e31\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eHowever, selecting and capturing relevant images from large datasets remains challenging despite machine learning advancements. Radiomics enables high-throughput extraction of quantitative features from medical images, with Li Z [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e32\u003c/span\u003e] and Zhou T [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e33\u003c/span\u003e] demonstrating its efficacy in COPD severity classification. Some researchers refined grading by analyzing radiomics features across individual lung lobes, capturing localized CT characteristics more effectively than whole-lung approaches [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. These studies underscore radiomics\u0026rsquo; feasibility for COPD staging. With artificial intelligence progress, deep learning has gained traction: Sara Rezvanjou et al. combined 2D tMPR images and 3D lung views to classify COPD using airway data [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e35\u003c/span\u003e], while Zou X et al. developed a cost-effective but less sensitive deep learning model for COPD staging using chest X-rays and clinical parameters [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Our study integrates chest CT images, clinical parameters, radiomics, and deep learning features into a clinically applicable nomogram, enhancing CT\u0026rsquo;s functional assessment value.\u003c/p\u003e \u003cp\u003eOur model extracts texture, edge, structural, and shape information from CT images, yielding richer feature descriptions for disease prediction. Adding deep learning features significantly improved diagnostic performance compared to models using only clinical or radiomics features. The final nomogram incorporates clinical factors (age, gender, PaCO2), radiomics features, and deep learning features. Decision curve analysis (DCA) confirmed its superior net benefit over clinical models across diagnostic thresholds, supporting its clinical utility. By improving COPD severity classification accuracy, our nomogram aids in managing acute exacerbations, personalizing treatment interventions, controlling symptoms, and slowing disease progression.\u003c/p\u003e \u003cp\u003eThis study has limitations. First, as a retrospective analysis from three institutions, the third center\u0026rsquo;s smaller sample size may limit generalizability; a prospective multicenter study is planned to optimize the cohort. Second, variables with \u0026gt;\u0026thinsp;30% missing values were excluded, warranting closer attention in future research. Third, the single-timepoint design did not assess longitudinal lung changes in the same patients. Future directions include integrating traditional CT features and quantitative parameters to enhance model efficacy and using delta radiomics to study COPD dynamics.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study utilized automatic segmentation technology to extract the entire lung parenchyma from CT images. By integrating clinical characteristics of independent risk factors, whole-lung radiomic features, and deep learning features, a novel combined nomogram was constructed to identify COPD severity. The study also demonstrated the added value of chest CT in evaluating lung structure and functional status.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003e\u003cstrong\u003eCOPD\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eChronic obstructive pulmonary disease\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eROI\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRegion of Interest\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eROC\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eReceiver Operating Characteristic\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePPV\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePositive Predictive Value\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNPV\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNegative Predictive Value\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLASSO\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLeast Absolute Shrinkage and Selection Operator\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eICC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Intraclass Correlation Coefficient\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDL\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDeep Learning\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCNN\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConvolutional Neural Network\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAUC\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eArea Under the Receiver Operating Characteristic Curve\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDCA\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDecision Curve Analysis\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePFT\u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePulmonary Function Test\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCRP\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eC-Reactive Protein\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eESR\u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eErythrocyte Sedimentation Rate\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePCT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eProcalcitonin\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePDW\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePlatelet Distribution Width\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePaCO2\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eArterial Partial Pressure of Carbon Dioxide\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePaO2\u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eArterial Partial Pressure of Oxygen\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRadiomics Score\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDLR\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDeep Learning radiomics\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCT\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eComputed Tomography\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to thank all patients, nurses, and physicians who participated in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBHP, and FQ conceived the study. XJN, QLY, YM, MXLand LYB collected data. XYL, ZJC and ZYF analysed data and drafted the manuscript. BHP, ZHX and FQ revised the manuscript. All authors contributed to the article and approved the submitted version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Science and Technology Project of Huzhou City, Zhejiang Province (2023GY33) and Postgraduate Research and Innovation Project of Huzhou University (2025KYCX99).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study using anonymized medical record data received an Institutional Review Board (IRB) waiver of informed consent from the First People\u0026rsquo;s Hospital of Huzhou, as it involves no direct patient intervention and adheres to the Declaration of Helsinki. Patient privacy is fully protected, and ethical standards are strictly maintained.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eNot applicable\u003cbr\u003e\u0026nbsp;\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eYin P, Wu J, Wang L, et al. The Burden of COPD in China and Its Provinces: Findings From the Global Burden of Disease Study 2019. Front Public Health. 2022 Jun 3;10:859499. doi: 10.3389/fpubh.2022.859499.\u003c/li\u003e\n\u003cli\u003eGomes F, Cheng SL. Pathophysiology, Therapeutic Targets, and Future Therapeutic Alternatives in COPD: Focus on the Importance of the Cholinergic System. Biomolecules. 2023 Mar 5;13(3):476. doi: 10.3390/biom13030476. \u003c/li\u003e\n\u003cli\u003eQin J, Ran B, Wu Y, et al. Association of hs-CRP/HDL, AIP and NHHR with chronic obstructive pulmonary disease: a cross-sectional NHANES study. Ann Med. 2025 Dec;57(1):2522317. doi: 10.1080/07853890.2025.2522317. \u003c/li\u003e\n\u003cli\u003eDecramer M, Janssens W, Miravitlles M. Chronic obstructive pulmonary disease. Lancet. 2012 Apr 7;379(9823):1341-51. doi: 10.1016/S0140-6736(11)60968-9. \u003c/li\u003e\n\u003cli\u003eKahnert K, J\u0026ouml;rres RA, Behr J, Welte T. The Diagnosis and Treatment of COPD and Its Comorbidities. Dtsch Arztebl Int. 2023 Jun 23;120(25):434-444. doi: 10.3238/arztebl.m2023.027. \u003c/li\u003e\n\u003cli\u003eBian H, Zhu S, Zhang Y, et al. Artificial Intelligence in Chronic Obstructive Pulmonary Disease: Research Status, Trends, and Future Directions --A Bibliometric Analysis from 2009 to 2023. Int J Chron Obstruct Pulmon Dis. 2024 Aug 21;19:1849-1864. doi: 10.2147/COPD.S474402. \u003c/li\u003e\n\u003cli\u003eMagr\u0026igrave; D, Fiori E, Agostoni P, et al. Heart failure and chronic obstructive pulmonary disease. A combination not to be underestimated. Heart Fail Rev. 2025 Dec;30(6):1525-1538. doi: 10.1007/s10741-025-10566-3. \u003c/li\u003e\n\u003cli\u003eBian H, Zhu S, Xing W, et al. Research Status and Direction of Chronic Obstructive Pulmonary Disease Complicated with Coronary Heart Disease: A Bibliometric Analysis from 2005 to 2024. Int J Chron Obstruct Pulmon Dis. 2025 Jan 7;20:23-41. doi: 10.2147/COPD.S495326. \u003c/li\u003e\n\u003cli\u003eGuo C, Yu T, Chang LY, et al. Mortality risk attributable to classification of chronic obstructive pulmonary disease and reduced lung function: a 21-year longitudinal cohort study. Respir Med. 2021;184:106471. doi: 10.1016/j.rmed.2021.106471. \u003c/li\u003e\n\u003cli\u003eKahnert K, J\u0026ouml;rres RA, Behr J, Welte T. The Diagnosis and Treatment of COPD and Its Comorbidities. Dtsch Arztebl Int. 2023 Jun 23;120(25):434-444. doi: 10.3238/arztebl.m2023.027.\u003c/li\u003e\n\u003cli\u003eTong H, Cong S, Fang LW, et al. [Performance of pulmonary function test in people aged 40 years and above in China, 2019-2020]. Zhonghua Liu Xing Bing Xue Za Zhi. 2023 May 10;44(5):727-734. Chinese. doi: 10.3760/cma.j.cn112338-20230202-00051.\u003c/li\u003e\n\u003cli\u003eWang R, Huang C, Yang W, et al. Respiratory microbiota and radiomics features in the stable COPD patients. Respir Res. 2023 May 12;24(1):131. doi: 10.1186/s12931-023-02434-1. \u003c/li\u003e\n\u003cli\u003eYu X, Xiong Z, Cao J, et al. Pulmonary arterial morphological markers on non-contrast CT predicted acute exacerbations and disease progression in chronic obstructive pulmonary disease: a longitudinal cohort study. Jpn J Radiol. 2025 Aug 4. doi: 10.1007/s11604-025-01841-2. \u003c/li\u003e\n\u003cli\u003eSu L, Qian C, Yu C, et al. Quantitatively Assessed Emphysema Severity on HRCT Independently Predicts Coronary Artery Disease in COPD: A Retrospective Cohort Study. Int J Chron Obstruct Pulmon Dis. 2025 Sep 10;20:3147-3161. doi: 10.2147/COPD.S540503. \u003c/li\u003e\n\u003cli\u003eCho YH, Lee SM, Seo JB, et al. Quantitative assessment of pulmonary vascular alterations in chronic obstructive lung disease: associations with pulmonary function test and survival in the KOLD cohort. Eur J Radiol. 2018;108:276\u0026ndash;282. doi: 10.1016/j.ejrad.2018.09.013.\u003c/li\u003e\n\u003cli\u003eLin X, Zhou T, Ni J, et al. CT-based radiomics nomogram of lung and mediastinal features to identify cardiovascular disease in chronic obstructive pulmonary disease: a multicenter study. BMC Pulm Med. 2025;25(1):121. doi: 10.1186/s12890-025-03568-2. \u003c/li\u003e\n\u003cli\u003eQi YJ, Su GH, You C, et al. Radiomics in breast cancer: Current advances and future directions. Cell Rep Med. 2024 Sep 17;5(9):101719. doi: 10.1016/j.xcrm.2024.101719. \u003c/li\u003e\n\u003cli\u003eWang Y, Zhang L, Xie H, et al. Predicting response and survival of lung adenocarcinoma under anti-programmed death-1 therapy using biological deep learning. Brief Bioinform. 2025 Aug 31;26(5):bbaf479. doi: 10.1093/bib/bbaf479. \u003c/li\u003e\n\u003cli\u003eBian H, Qian H, Zhu S, et al. Nomogram Model for Identifying the Risk of Coronary Heart Disease in Patients with Chronic Obstructive Pulmonary Disease Based on Deep Learning Radiomics and Clinical Data: A Multicenter Study. Int J Chron Obstruct Pulmon Dis. 2025 Sep 2;20:3045-3057. doi: 10.2147/COPD.S539307.\u003c/li\u003e\n\u003cli\u003eGlobal Initiative for Chronic Obstructive Lung Disease (GOLD). Global strategy for prevention, diagnosis and management of chronic obstructive pulmonary disease 2025. 2025. Available from: https://goldcopd.org/2025-gold-report/. Accessed May 13, 2025.\u003c/li\u003e\n\u003cli\u003eBrown KH, Ghita-Pettigrew M, Kerr BN, et al. Characterisation of quantitative imaging biomarkers for inflammatory and fibrotic radiation-induced lung injuries using preclinical radiomics. Radiother Oncol. 2024 Mar;192:110106. doi: 10.1016/j.radonc.2024.110106. \u003c/li\u003e\n\u003cli\u003eZhong N, Wang C, Yao W, et al. Prevalence of chronic obstructive pulmonary disease in China: a large, population-based survey. Am J Respir Crit Care Med. 2007;176(8):753\u0026ndash;60. doi: 10.1164/rccm.200612-1749oc.\u003c/li\u003e\n\u003cli\u003eMapel DW, Dalal AA, Blanchette CM, et al. Severity of COPD at initial spirometry-confirmed diagnosis: data from medical charts and administrative claims. Int J Chron Obstruct Pulmon Dis. 2011;6:573-81. doi: 10.2147/COPD.S16975. \u003c/li\u003e\n\u003cli\u003eBellamy D, Smith J. Role of primary care in early diagnosis and effective management of COPD. Int J Clin Pract. 2007;61(8):1380\u0026ndash;9. doi: 10.1111/j.1742-1241.2007.01447.x.\u003c/li\u003e\n\u003cli\u003ePu Y, Zhou X, Zhang D, et al. Re-defining high risk COPD with parameter response mapping based on machine learning models. Int J Chron Obstruct Pulmon Dis. 2022;17:2471\u0026ndash;2483. doi: 10.2147/COPD.S369904. \u003c/li\u003e\n\u003cli\u003eChen TT, Cheng TY, Liu IJ,et al. Leveraging Subjective Parameters and Biomarkers in Machine Learning Models: The Feasibility of lnc-IL7R for Managing Emphysema Progression. Diagnostics (Basel). 2025 May 3;15(9):1165. doi: 10.3390/diagnostics15091165.\u003c/li\u003e\n\u003cli\u003eRueda R, Fabello E, Silva T, et al. Machine learning approach to flare-up detection and clustering in chronic obstructive pulmonary disease (COPD) patients. Health Inf Sci Syst. 2024 Oct 23;12(1):50. doi: 10.1007/s13755-024-00308-4. \u003c/li\u003e\n\u003cli\u003eGu YF, Chen L, Qiu R, et al. Development of a model for predicting the severity of chronic obstructive pulmonary disease. Front Med (Lausanne). 2022 Dec 16;9:1073536. doi: 10.3389/fmed.2022.1073536.\u003c/li\u003e\n\u003cli\u003eXu C, Qi S, Feng J, et al. DCT-MIL: Deep CNN transferred multiple instance learning for COPD identification using CT images. Phys Med Biol. 2020;65:145011. doi: 10.1088/1361-6560/ab857d. \u003c/li\u003e\n\u003cli\u003eMatsumura K, Ito S. Novel biomarker genes which distinguish between smokers and chronic obstructive pulmonary disease patients with machine learning approach. BMC Pulm Med. 2020;20(1):29. doi: 10.1186/s12890-020-1062-9.\u003c/li\u003e\n\u003cli\u003eSui H, Mo Z, Wei Y, et al. Diagnosis and Severity Assessment of COPD Based on Machine Learning of Chest CT Images. Int J Chron Obstruct Pulmon Dis. 2025 Aug 14;20:2853-2867. doi: 10.2147/COPD.S528988. PMID: 40831903; PMCID: PMC12360390. \u003c/li\u003e\n\u003cli\u003eLi Z, Liu L, Zhang Z, et al. A Novel CT-Based Radiomics Features Analysis for Identification and Severity Staging of COPD. Acad Radiol. 2022 May;29(5):663-673. doi: 10.1016/j.acra.2022.01.004.\u003c/li\u003e\n\u003cli\u003eZhou T, Zhou X, Ni J, et al. A CT-Based Lung Radiomics Nomogram for Classifying the Severity of Chronic Obstructive Pulmonary Disease. Int J Chron Obstruct Pulmon Dis. 2024 Dec 11;19:2705-2717. doi: 10.2147/COPD.S483007. \u003c/li\u003e\n\u003cli\u003eZhao M, Wu Y, Li Y, et al. Learning and depicting lobe-based radiomics feature for COPD Severity staging in low-dose CT images. BMC Pulm Med. 2024 Jun 24;24(1):294. doi: 10.1186/s12890-024-03109-3. \u003c/li\u003e\n\u003cli\u003eRezvanjou S, Moslemi A, Peterson S, et al. Classifying chronic obstructive pulmonary disease status using computed tomography imaging and convolutional neural networks: comparison of model input image types and training data severity. J Med Imaging (Bellingham). 2025 May;12(3):034502. doi: 10.1117/1.JMI.12.3.034502. \u003c/li\u003e\n\u003cli\u003eZou X, Ren Y, Yang H, et al. Screening and staging of chronic obstructive pulmonary disease with deep learning based on chest X-ray images and clinical parameters. BMC Pulm Med. 2024 Mar 26;24(1):153. doi: 10.1186/s12890-024-02945-7. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-pulmonary-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"pulm","sideBox":"Learn more about [BMC Pulmonary Medicine](http://bmcpulmmed.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/pulm/default.aspx","title":"BMC Pulmonary Medicine","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Chronic Obstructive Pulmonary Disease, Radiomics, Deep Learning, Prediction Model","lastPublishedDoi":"10.21203/rs.3.rs-9272720/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9272720/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground and Objective:\u003c/h2\u003e \u003cp\u003eChronic obstructive pulmonary disease (COPD) is a widespread and severely disabling respiratory disorder that places a substantial burden on global healthcare systems. Precise determination of COPD severity staging is essential for effective patient management and treatment planning. This study seeks to develop and validate a comprehensive nomogram that combines clinical characteristics, whole-lung computed tomography (CT) radiomic features, and deep learning-derived features to classify COPD severity.\u003c/p\u003e\u003ch2\u003eMethod\u003c/h2\u003e \u003cp\u003eA retrospective analysis included 1,794 patients from three hospitals, spanning January 1, 2021, to May 30, 2025. Following fully automated segmentation of the entire lungs, radiomic features and three-dimensional deep learning features were extracted. A comprehensive nomogram was developed and validated, integrating radiomics features, deep learning features, and independent clinical factors. Model performance was assessed and compared using receiver operating characteristic (ROC) curves, the area under the ROC curve (AUC), decision curve analysis (DCA), and the DeLong test.\u003c/p\u003e\u003ch2\u003eResult\u003c/h2\u003e \u003cp\u003eIn the training, internal validation, and external validation cohorts, the area under the receiver operating characteristic curve (AUC) values for the clinical model were 0.634, 0.630, and 0.616, respectively; for the radiomics (Rad) model, 0.760, 0.729, and 0.704, respectively; for the deep learning (DL) model, 0.805, 0.757, and 0.752, respectively; for the radiomics-deep learning combined (DLR) model, 0.822, 0.740, and 0.759, respectively; and for the logistic regression model, 0.839, 0.759, and 0.767, respectively. The logistic regression model outperformed the individual clinical, radiomics, and three-dimensional deep learning models.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThis study constructed and validated a novel combined logistic regression model for identifying the severity of COPD by integrating the clinical characteristics of independent risk factors, the full lung radiomics features, and the deep learning features. It demonstrated the additional value of chest CT in evaluating lung structure and lung function status.\u003c/p\u003e","manuscriptTitle":"Development and validation of a multimodal clinical-radiomics-deep learning nomogram based on automated chest CT segmentation for classifying COPD severity: A multicenter study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-21 01:33:08","doi":"10.21203/rs.3.rs-9272720/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-18T09:54:01+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-15T09:21:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"209605267734748890982465187917508778060","date":"2026-05-05T18:19:04+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-23T14:52:40+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"157159985349529984006603504980081504694","date":"2026-04-14T14:48:47+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"155006354867621069413370871575500684335","date":"2026-04-13T15:26:48+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-13T14:02:48+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-13T13:51:18+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-06T16:39:27+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-04T07:17:38+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Pulmonary Medicine","date":"2026-04-04T07:11:43+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-pulmonary-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"pulm","sideBox":"Learn more about [BMC Pulmonary Medicine](http://bmcpulmmed.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/pulm/default.aspx","title":"BMC Pulmonary Medicine","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3c8e5eca-4a52-4cdd-ad37-c60051158350","owner":[],"postedDate":"April 21st, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-18T09:54:01+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-15T09:21:52+00:00","index":194,"fulltext":""},{"type":"reviewerAgreed","content":"209605267734748890982465187917508778060","date":"2026-05-05T18:19:04+00:00","index":183,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-05-18T10:10:23+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-21 01:33:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9272720","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9272720","identity":"rs-9272720","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}