Abstract
Objectives
This study aims to evaluate the predictive value of CT target scan-based radiomics and clinical features in patients with chronic obstructive pulmonary disease (COPD) who also present with malignant pulmonary nodules.
Methods
A retrospective analysis was conducted on 104 patients diagnosed with COPD and pulmonary nodules, treated at our hospital between September 2022 and September 2024. The cohort was stratified into a benign group (n = 48) and a malignant group (n = 66) based on the definitive pathological examination results. All participants underwent spiral CT target scanning. Univariate and multivariate logistic regression analyses were employed to identify influencing factors. In addition, a receiver operating characteristic (ROC) curve was constructed to assess the predictive value.
Results
In comparison with the benign group, the malignant group exhibited a greater aspect ratio, a higher proportion of patients over 45 years of age, and increased occurrences of mixed ground-glass opacities, calcification, spiculation sign, vascular convergence sign, and lobulation sign, along with a significant elevation in the CAT score (P < 0.05). The CAT score, density, spiculation sign, and lobulation sign emerged as independent risk factors for the presence of malignant pulmonary nodules in patients with COPD (P < 0.05). The aforementioned imaging factors were incorporated into a multivariate logistic regression model: Logit (P) = 10.490 + 0.917 × CAT score + 1.547 × density + 0.823 × spiculation sign + 0.736 × lobulation sign. The area under the curve (AUC) values for the CAT score, radiomics model, and combined model were 0.768, 0.826, and 0.909, respectively.
Conclusions
Radiomics features derived from CT target scanning, when combined with clinical features, demonstrate promising potential for differentiating between benign and malignant pulmonary nodules in patients with COPD. This approach may indirectly reflect patients' pulmonary function and serve as a reference for the early diagnosis of COPD combined with lung cancer.
Keywords: CT radiomics, Clinical features, Chronic obstructive pulmonary disease, Benign and malignant pulmonary nodules, Early diagnosis
Introduction
Chronic obstructive pulmonary disease (COPD) is a prevalent clinical lung disease and a risk factor for lung cancer, characterized by airflow limitation, chronic airway inflammation, and persistent respiratory symptoms [1]. With the rapid advancement of imaging technology, a substantial number of lung nodules have been detected, of which 1–12% are ultimately diagnosed as early stage lung cancer [2]. Research has indicated that lung cancer patients with COPD exhibit lower survival rates and poorer prognoses compared to these without COPD [3]. Therefore, for patients with both COPD and pulmonary nodules, accurately predicting whether these nodules will progress to lung cancer is essential for developing effective treatment strategies and improving patient outcomes.
CT target scanning allows for detailed observation of the morphology, density, and other characteristics of lung nodules, providing critical evidence for assessing their benign or malignant nature. In addition, chest CT can assess the extent and distribution of emphysema and detect bronchial wall thickening. A visual evaluation of the entire lung parenchyma and interstitial CT features can be used to identify COPD [4]. Furthermore, CT target scanning can quantify the heterogeneous structural abnormalities associated with COPD, such as emphysema and airway diseases, aiding in the evaluation of the patient's condition and the formulation of treatment plans [5]. Radiomics, an emerging technology, employs various algorithms to extract high-throughput imaging features from medical images for analysis. These features collectively reflect the biological characteristics and pathological alterations of tissues, offering novel insights and evidence for disease diagnosis and treatment [6]. Research [7] indicates that radiomics exhibits high accuracy, sensitivity, specificity, and clinical benefits, particularly in distinguishing between benign and malignant pulmonary nodules, highlighting its significant potential for application. CT radiomics is capable of extracting numerous quantitative imaging features from CT scans and transforming them into feature space data with high recognition rates, facilitating the development of risk models for lung nodule evaluation and thereby enhancing diagnostic and predictive accuracy. However, CT radiomics may also yield false positives, which can not only exacerbate negative emotions in patients but also lead to inefficient use of medical resources [8]. For patients with COPD accompanied by pulmonary nodules, accurate assessment of nodule malignancy is essential for formulating personalized treatment strategies. Existing methodologies for the identification and detection of pulmonary nodules are currently inadequate in simultaneously fulfilling the clinical requirements for non-invasive, rapid, and accurate assessment.
In this study, we explored the predictive efficacy of radiomics derived from CT target scans, alongside clinical features, in patients with COPD who also present with malignant pulmonary nodules. The objective was to identify radiomic and clinical feature indicators that were highly significant in predicting nodule malignancy, thereby equipping clinicians with more precise and personalized evaluation tools. This approach aims to enhance diagnostic and therapeutic strategies for COPD patients with pulmonary nodules and improve the early detection rate of lung cancer.
Results
Comparison of clinical characteristics between two groups of patients
No significant differences were observed between the two patient groups regarding gender, smoking history, smoking index, allergy history, proportion of acute exacerbations, and BMI (P > 0.05). However, the proportion of patients over 45 years of age and the CAT score were significantly higher in the malignant group compared to the benign group (P < 0.05, Table 1).
Table 1.
Comparison of clinical characteristics between two groups of patients [cases (%), ( ± s)]
| Groups | The benign group (n = 48) | The malignant group (n = 66) | t/χ2 | P |
|---|---|---|---|---|
| Gender | 1.404 | 0.236 | ||
| Female | 26 (54.17) | 43 (65.15) | ||
| Male | 22 (45.83) | 23 (34.85) | ||
| Smoking history | 11 (22.92) | 18 (27.27) | 0.278 | 0.598 |
| Smoking index (number/year) | 0.343 | 0.558 | ||
| ≤ 400 | 42 (87.50) | 60 (90.91) | ||
| > 400 | 6 (12.50) | 6 (9.09) | ||
| Allergy history | 22 (45.83) | 24 (36.36) | 1.035 | 0.309 |
| Age (year) | 5.184 | 0.023 | ||
| ≤ 45 | 14 (29.17) | 8 (12.12) | ||
| > 45 | 34 (70.83) | 58 (87.88) | ||
| BMI (kg/m2) | 22.71 ± 2.57 | 22.94 ± 2.93 | 0.435 | 0.664 |
| Number of acute exacerbations (Number) | 0.308 | 0.579 | ||
| ≤ 3 | 45 (93.75) | 60 (90.91) | ||
| > 3 | 3 (6.25) | 6 (9.09) | ||
| CAT score (score) | 9.11 ± 3.42 | 13.59 ± 3.85 | 6.425 | < 0.001 |
Comparison of radiomics features between two groups of patients
There were no significant differences in morphology, location, or the presence of vacuolar signs between the benign and malignant groups (P > 0.05). In contrast, the malignant group exhibited a greater aspect ratio and a higher prevalence of mixed ground-glass opacities, calcification, spiculation sign, vascular convergence sign, and lobulation sign compared to the benign group (P < 0.05, Table 2).
Table 2.
Comparison of radiomics features between two groups of patients [cases (%), ( ± s)]
| Groups | The benign group (n = 48) | The malignant group (n = 66) | t/χ2 | P |
|---|---|---|---|---|
| Aspect ratio | 1.22 ± 0.16 | 1.31 ± 0.22 | 2.324 | 0.022 |
| Morphology | 1.610 | 0.205 | ||
| Regular | 7 (14.58) | 16 (24.24) | ||
| Irregular | 41 (85.42) | 50 (75.76) | ||
| Location | 0.439 | 0.508 | ||
| Left lung | 16 (33.33) | 26 (39.39) | ||
| Right lung | 32 (66.67) | 40 (60.61) | ||
| Density | 20.814 | < 0.001 | ||
| Solid | 4 (8.33) | 1 (1.52) | ||
| Mixed ground glass | 13 (27.08) | 46 (69.70) | ||
| Pure ground glass | 31 (64.58) | 19 (28.79) | ||
| Calcification | 6 (12.50) | 19 (28.79) | 4.306 | 0.038 |
| Spiculation sign | 9 (18.75) | 30 (45.45) | 8.805 | 0.003 |
| Vascular aggregation sign | 22 (45.83) | 56 (84.85) | 19.578 | < 0.001 |
| Lobulation sign | 12 (25.00) | 36 (54.55) | 9.951 | 0.002 |
| Vacuole Sign | 2 (4.17) | 8 (12.12) | 2.197 | 0.138 |
Multivariate analysis of factors affecting COPD complicated with malignant pulmonary nodules
The variable indicating COPD combined with pulmonary nodules was used as the dependent variable, while the indicators showing statistical differences in Tables 1 and 2 were used as independent variables, with their assigned values detailed in Table 3. The multivariate logistic regression analysis identified the CAT score, density, spiculation sign, and lobulation sign as independent risk factors for COPD complicated by malignant pulmonary nodules (P < 0.05, Table 4 and Fig. 1).
Table 3.
Assignment table of independent variables
| Variables | Assignments |
|---|---|
| Age | 0 = ≤ 45 years, 1 = > 45 years |
| CAT score | Continuous variable |
| Aspect ratio | Continuous variable |
| Density | 0 = Solid and Pure ground glass, 1 = mixed ground glass |
| Calcification | 0 = No, 1 = Yes |
| Spiculation sign | 0 = No, 1 = Yes |
| Lobulation sign | 0 = No, 1 = Yes |
| Vascular aggregation sign | 0 = No, 1 = Yes |
| The property COPD combined with pulmonary nodules | 0 = benign, 1 = malignant |
Table 4.
Multivariate analysis of factors affecting COPD complicated with malignant pulmonary nodules
| Indicators | B value | SE value | Wald value | P value | OR value | 95% CI |
|---|---|---|---|---|---|---|
| Age | 0.315 | 0.128 | 6.056 | 0.077 | 1.370 | 1.066 ~ 1.761 |
| CAT score | 0.917 | 0.302 | 9.220 | < 0.001 | 1.502 | 1.384 ~ 4.522 |
| Aspect ratio | 0.926 | 0.376 | 6.065 | 0.079 | 2.524 | 1.208 ~ 5.275 |
| Density | 1.547 | 0.454 | 11.611 | < 0.001 | 4.697 | 1.929 ~ 11.439 |
| Calcification | 0.364 | 0.240 | 2.300 | 0.566 | 1.439 | 0.899 ~ 2.303 |
| Spiculation sign | 0.823 | 0.214 | 14.790 | < 0.001 | 2.277 | 1.498 ~ 3.463 |
| Lobulation sign | 0.736 | 0.229 | 10.330 | < 0.001 | 2.088 | 1.332 ~ 3.271 |
| Vascular aggregation sign | 0.752 | 0.358 | 4.412 | 0.137 | 2.121 | 1.051 ~ 4.280 |
Fig. 1.
Forest plot of multifactorial analysis affecting COPD complicated with malignant pulmonary nodules
The predictive value of radiomics and clinical features based on CT target scanning for COPD complicated with malignant pulmonary nodules
These imaging factors were incorporated into a multivariate logistic regression model: Logit (P) = 10.490 + 0.917 × CAT score + 1.547 × density + 0.823 × spiculation sign + 0.736 × lobulation sign. A radiomics model was subsequently developed based on the logistic integration of these indicators. The diagnostic efficacy of the CAT score, the radiomics model, and a combined model for predicting COPD with malignant pulmonary nodules was evaluated using the ROC curve, yielding AUC values of 0.768, 0.826, and 0.909, respectively (Fig. 2, and Table 5).
Fig. 2.
ROC curve analysis chart. A CAT score; B radiomics model; C joint model
Table 5.
Predictive value of radiomics and clinical features based on CT target scanning for COPD complicated with malignant pulmonary nodules
| Indicators | AUC | Positive predictive value | Negative predictive value | Sensitivity | Specificity | Accuracy index | 95% CI |
|---|---|---|---|---|---|---|---|
| CAT score | 0.768 | 74.18% | 66.88% | 90.61% | 57.65% | 0.483 | 0.627 ~ 0.839 |
| Radiomics model | 0.826 | 75.78% | 79.38% | 96.71% | 59.27% | 0.560 | 0.882 ~ 0.935 |
| Joint model | 0.909 | 96.81% | 80.86% | 88.74% | 90.59% | 0.793 | 0.879 ~ 0.996 |
Discussion
Numerous studies have indicated a potential association between COPD and the onset and progression of lung cancer, suggesting a possible causal relationship. Lung cancer is frequently accompanied by impaired lung function and pulmonary ventilation dysfunction, which can elevate both the incidence and mortality rates associated with the disease [9]. However, the early clinical manifestations of lung cancer often lack specificity, with pulmonary nodules serving as a significant imaging indicator of early stage lung cancer [10]. Consequently, it is imperative to screen COPD patients for potential lung cancer development. Early detection and differentiation of benign versus malignant pulmonary nodules, followed by prompt intervention, may be crucial strategies for reducing lung cancer incidence and enhancing patient outcomes.
The convergence of various information sciences has facilitated the advancement of imaging technologies in medical applications, evolving traditional radiology into the realm of imaging medicine. Radiomics, as an innovative technique for tumor characterization, plays a pivotal role in supporting clinical decision-making. Radiomics technology is capable of transforming medical images into high-dimensional data, allowing for a quantitative characterization of regions of interest and facilitating effectively correlations with clinical information. Moreover, radiomics exhibits rapid and robust post-processing capabilities, which significantly enhance image quality and improve the accuracy of pulmonary nodule diagnoses [11, 12]. A study examining the clinical and imaging data of colorectal cancer patients demonstrated that CT-based radiomic features could predict disease-free survival and offer prognostic insights [13]. Reports indicate that CT radiomics possesses superior diagnostic value compared to conventional CT in detecting in lung lymphoma and lung adenocarcinoma, with combined detection methods offering enhanced differential diagnostic value [14]. In addition, research has shown that CT-based radiomics is instrumental in differentiating between benign and malignant thyroid nodules, achiving high sensitivity (89.4%), specificity (86.8%), and accuracy (87.6%) [15]. CT target scanning is an advanced imaging technique that integrates thin-layer scanning with overlapping reconstruction, using a reduced scanning field. This study revealed that patients with malignant tumors exhibited a significant increase in aspect ratio, as well as in the proportions of calcification, spiculation sign, vascular aggregation sign, and lobulation sign, while the proportion of pure ground-glass opacity was markedly decreased. Consequently, it is posited that the radiomic features derived from CT target scanning are instrumental in assessing malignant pulmonary nodules. Malignant lung nodules frequently display irregular morphologies, with aspect ratios potentially more pronounced than those of benign nodules. Calcification in pulmonary nodules is a complex phenomenon; although certain calcification patterns (such as central, layered, or explosive) are typically indicative of benign nodules, calcification can also manifest in malignant nodules, particularly in cases of tumor necrosis or inadequate blood supply [16]. This study suggested that the observed increase in calcification within malignant nodules may reflect specific patterns or degrees of calcification within the nodules. CT target scanning, as a specialized CT imaging technique, enables the detailed visualization of lung tissue's intricate structures through ultra-thin scanning and reconstructed images, thereby enhancing the clarity and diagnostic capability for detecting small pulmonary nodules. In the assessment of malignant pulmonary nodules, target scanning not only provides the aforementioned radiomic features but also evaluates the nodules’ margins and infiltration, investigates their internal structure, and assesses vascular growth status, thereby offering more precise diagnostic information for clinical application [17].
Furthermore, multivariate logistic analysis in this study identified the CAT score, spiculation sign, lobulation sign, and density as independent risk factors for COPD patients with malignant pulmonary nodules. The CAT score is a critical tool for evaluating lung function and plays a vital role in quantifying symptoms in COPD patients. Previous research has demonstrated a significant correlation between the CAT score and lung function in COPD patients, effectively predicting pulmonary ventilation and gas exchange function [18]. Previous studies have demonstrated that the presence of pulmonary nodules is associated with alterations in lung function, manifesting as small airway dysfunction and restrictive ventilatory impairment. Notably, lung dysfunction is significantly more pronounced in patients with malignant nodules compared to those with benign nodules [19]. The spiculation sign is a prevalent imaging characteristic of malignant pulmonary nodules, identified by thin, short linear shadows extending from the nodule’s edge to its periphery without branching. This feature may result from the invasive growth of tumor cells toward the periphery, indicative of their aggressive nature [20]. Another critical feature of malignant pulmonary nodules is the lobulation sign, characterized by irregular, lobulated contours at the nodule edges, which may arise from the tumor's uneven growth in various directions, reflecting cellular heterogeneity. In addition, pure ground-glass nodules are defined as those that appear slightly denser on CT images but do not obscure the underlying pulmonary parenchymal blood vessels and bronchial structures. Although ground-glass nodules may be malignant, existing research suggests that a reduction in the proportion of pure ground-glass nodules in malignant cases may indicate a tendency for malignant nodules to exhibit more pronounced solid components or more intricate internal structures [21, 22]. Reports indicate that the proportion of mixed ground-glass nodules is significantly higher in patients with invasive adenocarcinoma compared to those with adenomatous hyperplasia. Consistent with the findings of the present study, CT radiomics features may aid in with the qualitative diagnosis of nodules [23].
Zhang et al. [24] investigated the application of radiomics models based on CT target scanning and conventional CT imaging to predict the growth trajectory of ground-glass nodules in the lungs over a 2-year period. In this study, the combined AUC of the radiomics model and clinical features was 0.909, surpassing the AUC of 0.880 reported by Zhang et al. [25], who utilized a radiomics model based on CT target scanning to differentiate benign from malignant solitary pulmonary nodules. Zhao et al. [26] employed a CT target scanning radiomics model to predict the invasion level ground-glass adenocarcinoma in the lungs, achieving a training group AUC of 0.911. However, their study categorized carcinoma in situ and microinvasive adenocarcinoma into a low infiltration group, while invasive adenocarcinoma was classified as a high infiltration group. This classification approach differs from that used in our current study. In our research, the spiculation sign, lobulation sign, and density observed in clinical CT images were identified as independent risk factors for malignant pulmonary nodules. Conversely, the study by Chen et al. [27] excluded parameters with a Spearman correlation coefficient greater than 0.6 in CT signs and ultimately developed a clinical model using three CT signs: vascular aggregation, spiculation, and lobulation, which contrasts with the statistical method we employed to screen clinical models. However, the diagnostic efficacy of the nomogram constructed by integrating CT features with the Rad-score was comparable, aligning with the conclusion of our present study that the combined model exhibited the highest predictive efficacy for malignant pulmonary nodules. The aforementioned studies indicated that the joint model effectively integrated high-dimensional radiomic features with clinical data, thereby enhancing the model's diagnostic efficacy. This model equation holds substantial significance for practical applications, offering a quantitative prediction tool that enables clinicians to calculate the probability of malignant pulmonary nodules in patients. By inputting CT-derived radiomic and clinical features into the model equation, it aids in clinical decision-making. Furthermore, each independent variable within the model equation possesses clear clinical relevance, facilitating a deeper understanding of the mechanisms underlying malignant pulmonary nodules and aiding in the development of targeted treatment strategies.
Conclusion
Overall, the integration of CT-based radiomic features with clinical data demonstrates promising potential in distinguishing between benign and malignant pulmonary nodules in COPD patients. This approach may also indirectly reflect pulmonary function and serve as a reference for the early diagnosis of COPD in conjunction with lung cancer.
Research limitations and future directions: This investigation was conducted as a single-center retrospective study, which may not fully capture the diversity of the broader population of COPD patients with pulmonary nodules, thereby introducing potential selection bias. In addition, the study's sample size was relatively small (n = 104), which could affect the precision and stability of the model estimates and increase the risk of overfitting. Furthermore, while the model demonstrated excellent performance (AUC of the joint model = 0.909), it has only undergone internal validation and has not been tested in an independent external cohort, particularly a multi-center, prospectively collected cohort. This limitation raises questions about the model's generalizability and reliability for routine clinical application. Finally, the radiomics model developed in this study primarily focused on the morphological characteristics of the nodules and did not incorporate additional background information, such as quantitative emphysema parameters and pulmonary vascular changes, which are critical for the diagnosis and classification of COPD. This limitation may restrict the model's ability to provide targeted interpretations within the complex context of COPD.
Future research endeavors should focus on implementing multi-center, large-sample prospective cohort studies to reduce selection bias and ensure rigorous external validation. Concurrently, it is imperative to investigate COPD-related quantitative imaging parameters, such as the emphysema index and airway wall thickness, to develop a more comprehensive and precise personalized prediction tool.
Materials and methods
General materials
A retrospective analysis was conducted involving 104 patients with COPD and concurrent pulmonary nodules, treated at our hospital from September 2022 to September 2024. The study encompassed a total of 114 pulmonary nodules, comprising 94 cases of single pulmonary nodules and 10 cases of double pulmonary nodules. Patients were categorized into two groups based on the definitive pathological examination results [28]: a benign group (n = 48), comprising individuals whose nodule tissue was identified as a benign adenoma, hamartoma, or other benign lesions during biopsy, and a malignant group (n = 66), consisting of individuals whose biopsy results indicated non-small cell lung cancer, such as malignant cells or adenocarcinoma. Inclusion criteria: (1) all patients satisfied the diagnostic and therapeutic criteria for COPD [29]. (2) Chest CT scans revealed the presence of regular or irregular dense shadows in the lungs, with a diameter not exceeding 30 mm. (3) Patient’s possessed complete clinical, imaging, and pathological data. Exclusion criteria: (1) patients with concomitant coagulation dysfunction or respiratory system abnormalities. (2) Patients who had previously undergone lung radiotherapy, chemotherapy, or immunosuppressive therapy. (3) Patients presenting with pulmonary lymph node enlargement, atelectasis, pleural effusion, etc. (4) Patients with malignant tumors in other regions. The patient inclusion process was illustrated in Fig. 3.
Fig. 3.
Process diagram for patient inclusion
Clinical data collection
Clinical data from selected cases were systematically collected, encompassing variables, such as gender, smoking history, smoking index, allergy history, number of acute exacerbations, body mass index (BMI), and age.
Spiral CT target scanning
Imaging was conducted using the NeuroLogica CereTom spiral CT scanner, acquired from Huanxi Medical Equipment Co., Ltd. Prior to scanning, patients were instructed to remove any metal jewelry and were positioned supine for the procedure. The protocol involved a complete lung scan performed during inhalation with breath-holding, with the patient's arms positioned flat at their sides. The chest scan was centered at the sternal angle, with the head oriented forward. The scanning equipment utilized was the GE64-row CT scanner (manufacturer: GE, Japan; model: Optima CT660; registration certificate number: 132101033). The multi-slice spiral CT scan covered the area from the thoracic inlet to the diaphragm, specifically at the level of the costophrenic angles bilaterally. Instrument parameters were configured as follows: a tube current of 120mAs, tube voltage of 120kVp, a pitch of 1.2, a matrix size of 512 × 512, a collimation of 0.5 mm × 128, collimation rotation speed of 0.33 s per revolution, and image reconstruction with a conventional slice thickness of 5 mm. Subsequently, a multi-slice spiral CT perfusion scan was conducted. Prior to scanning, a high-pressure injector was utilized to administer a contrast agent, specifically 90 mL of a 300 mL/L iodine propamide injection (manufactured by Chengdu Beite Pharmaceutical Co., Ltd.; approval number: H20233136; specifications: 100 mL: 62.34 g), at a rate of 3.00 mL/s, followed by enhanced scanning. The resulting images were uploaded and transmitted to the image processing workstation for further analysis. Two experienced physicians independently reviewed the images to identify suspected areas while avoiding calcification, cavities, and adjacent blood vessels within the lesion. Key parameters such as blood volume, peak enhancement, surface permeability, and peak perfusion time at lesion site were calculated. Each measure was performed three, and the mean value was recorded.
For three-dimensional reconstruction, a GE64 CT scanner (X-ray CT device, model: Optima CT660CT) was employed to perform horizontal scans from the chest entrance to the intercostal angle. Subsequently, the acquired image data were uploaded to the workstation for multi-planar reconstruction of the nodule lesion, including 1 mm above and below the lesion. Simultaneously, within the reconstructed sagittal and coronal lung and mediastinal window regions, the lung nodule was positioned as the focal point and rotated along the Y-axis, X-axis, and Z-axis to identify the optimal lesion view and display the maximum cross-sectional image. The resulting data were subjected to analysis, extraction, and documentation by a minimum of three expert physicians to assess radiomic features, including morphology, location, vacuolar sign, aspect ratio, calcification, spiculation sign, vascular aggregation sign, lobulation sign, and density.
Statistical analysis
Statistical analysis was performed using SPSS version 24.0. The enumeration data were presented as [cases (%)], and intergroup comparisons were conducted using the χ2 test. Measurement data following a normal distribution were expressed as ( ± s), and comparisons between two groups were made using the independent sample t test. The predictive value was evaluated through Receiver Operating Characteristic (ROC) curve analysis. Univariate and multivariate logistic regression analyses were employed to explore the factors influencing COPD complicated with malignant pulmonary nodules. An ROC curve was employed to assess the sensitivity and specificity of the combined detection of clinical and radiomic features, as well as its diagnostic efficacy for COPD complicated by malignant pulmonary nodules. The statistical significance threshold was set at P < 0.05.
Acknowledgements
Not applicable.
Author contributions
Fang Liu confirmed the authenticity of all the raw data and edited the manuscript, Mingjing Yuan and Jun Luo collected data and processed the data. Fang Liu and Rong Luo conducted the statistics. Rong Luo reviewed and revised the article. All authors read and approved the final manuscript.
Funding
This work was supported by the Natural Science Foundation Project of the Science and Technology Bureau of Yongchuan District, Chongqing City (NO.2023yc-jckx20052).
Availability of data and materials
The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by The Ethics Committee of Yongchuan Hospital of Chongqing Medical University. Written informed consent was obtained from participants for the participation in the study. All procedures performed in studies involving human participants were in accordance with the standards upheld with those of the 1964 Helsinki Declaration and its later amendments for ethical research involving human subjects.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interest.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Qi C, Sun SW, Xiong XZ. From COPD to lung cancer: mechanisms linking, diagnosis, treatment, and prognosis. Int J Chron Obstruct Pulmon Dis. 2022;17:2603–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.MacMahon H, Naidich DP, Goo JM, et al. Guidelines for management of incidental pulmonary nodules detected on CT images: from the Fleischner Society 2017. Radiology. 2017;284(1):228–43. [DOI] [PubMed] [Google Scholar]
- 3.Lim JU, Yeo CD, Rhee CK, et al. Comparison of clinical characteristics and overall survival between spirometrically diagnosed chronic obstructive pulmonary disease (COPD) and non-COPD never-smoking stage I-IV non-small cell lung cancer patients. Int J Chron Obstruct Pulmon Dis. 2019;14:929–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Singhvi D, Bon J. CT imaging and comorbidities in COPD: beyond lung cancer screening. Chest. 2021;159(1):147–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Huang X, Xue Y, Deng B, et al. Predicting pathological grade of stage I pulmonary adenocarcinoma: a CT radiomics approach. Front Oncol. 2024;14:1406166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.HajiEsmailPoor Z, Kargar Z, Baradaran M, et al. Prognostic value of CT scan-based radiomics in intracerebral hemorrhage patients: a systematic review and meta-analysis. Eur J Radiol. 2024;178:111652. [DOI] [PubMed] [Google Scholar]
- 7.Xiong S, Fu Z, Deng Z, et al. Machine learning-based CT radiomics enhances bladder cancer staging predictions: a comparative study of clinical, radiomics, and combined models. Med Phys. 2024;51(9):5965–77. [DOI] [PubMed] [Google Scholar]
- 8.Kang YK, Ha S, Jeong JB, et al. The value of PET/CT radiomics for predicting survival outcomes in patients with pancreatic ductal adenocarcinoma. Sci Rep. 2024;14(1):28958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang XZ, Wang JY, Meng T, et al. Non-malignant pathological results from CT-guided biopsy for pulmonary nodules: a predictive model for identifying false-negative results. J Cardiothorac Surg. 2024;19(1):386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hopkinson NS, Molyneux A, Pink J, et al. Chronic obstructive pulmonary disease: diagnosis and management: summary of updated NICE guidance. BMJ (Clinical research ed). 2019;29(366):l4486. [DOI] [PubMed] [Google Scholar]
- 11.Zhou T, Tu W, Dong P, et al. CT-based radiomic nomogram for the prediction of chronic obstructive pulmonary disease in patients with lung cancer. Acad Radiol. 2023;30(12):2894–903. [DOI] [PubMed] [Google Scholar]
- 12.Wang Y, Zhou C, Ying L, et al. Enhancing early lung cancer diagnosis: predicting lung nodule progression in follow-up low-dose CT scan with deep generative model. Cancers. 2024;16(12). [DOI] [PMC free article] [PubMed]
- 13.Refaee T, Wu G, Ibrahim A, et al. The emerging role of radiomics in COPD and lung cancer. Respiration. 2020;99(2):99–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang Y, Lei D. Research progress in CT-based radiomics constructing hypopharyngeal cancer and multisystem tumor prediction model. Lin Chung Er Bi Yan Hou Tou Jing Wai Ke Za Zhi. 2022;36(2):158–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Intarak S, Chongpison Y, Vimolnoch M, et al. Tumor prognostic prediction of nasopharyngeal carcinoma using CT-based radiomics in non-Chinese patients. Front Oncol. 2022;12:775248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fan L, Fang M, Li Z, et al. Radiomics signature: a biomarker for the preoperative discrimination of lung invasive adenocarcinoma manifesting as a ground-glass nodule. Eur Radiol. 2019;29(2):889–97. [DOI] [PubMed] [Google Scholar]
- 17.Xu H, Wang X, Guan C, et al. Value of whole-thyroid CT-based radiomics in predicting benign and malignant thyroid nodules. Front Oncol. 2022;12:828259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zhao C, Xiao R, Jin H, et al. The immune microenvironment of lung adenocarcinoma featured with ground-glass nodules. Thoracic Cancer. 2024;15(19):1459–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Franco G, Debray MP, Anzani N, et al. Cystic lung in sarcoidosis: clinico-radiologic characteristic and evolution. Respirology (Carlton, Vic). 2024;29(12):1067–76. [DOI] [PubMed] [Google Scholar]
- 20.Vaezi A, Mirsaeidi M. Proposing the potential of utilizing the CAT score for early detection of COPD in asymptomatic patients, shifting towards a patient-centered approach: a review. Medicine. 2024;103(15):e37715. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Yao Q, Ji Q, Zhou Y. Pulmonary function in pulmonary sarcoidosis. J Clin Med. 2023;12(21). [DOI] [PMC free article] [PubMed]
- 22.Zhang L, Wan R, Chen J, et al. Analysis of the correlation between clinical and imaging features of malignant lung nodules and pathological types. Front Surg. 2023;10:1321118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Sun L, Zhang B, Li P, et al. Application of target scanning combined with three-dimensional reconstruction in the diagnosis of early-stage lung adenocarcinoma. J Cancer. 2024;15(18):6016–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wen Y, Wu W, Liufu Y, et al. Differentiation of granulomatous nodules with lobulation and spiculation signs from solid lung adenocarcinomas using a CT deep learning model. BMC Cancer. 2024;24(1):875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hou H, Yu S, Xu Z, et al. Prediction of malignancy for solitary pulmonary nodules based on imaging, clinical characteristics and tumor marker levels. Eur J Cancer Prev. 2021;30(5):382–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhang Z, Yin F, Kang S, et al. Dual-layer spectral detector CT (SDCT) can improve the detection of mixed ground-glass lung nodules. J Cancer Res Clin Oncol. 2023;149(9):5901–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zhang W, Cui X, Wang J, et al. The study of plain CT combined with contrast-enhanced CT-based models in predicting malignancy of solitary solid pulmonary nodules. Sci Rep. 2024;14(1):21871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zhao Y, Ye Z, Yan Q, et al. Predicting the invasiveness of ground-glass opacity predominant lung adenocarcinoma with clinical stage Ia: a CT-based semantic and radiomics analysis. J Thorac Dis. 2024;16(10):6713–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Chen H, Wen Y, Wu W, et al. Prediction of malignancy and pathological types of solid lung nodules on CT scans using a volumetric SWIN transformer. J Imaging Inf Med. 2024. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request.