Intratumoral and peritumoral radiomics combined with computed tomography features for predicting the invasiveness of lung adenocarcinoma presenting as a subpleural ground-glass nodule with a consolidation-to-tumor ratio ≤50.

Background: Preoperative accurate judgment of the degree of invasiveness in subpleural ground-glass lung adenocarcinoma (LUAD) with a consolidation-to-tumor ratio (CTR) ≤50% is very important for the choice of surgical timing and planning. This study aims to investigate the performance of intratumoral and peritumoral radiomics combined with computed tomography (CT) features for predicting the invasiveness of LUAD presenting as a subpleural ground-glass nodule (GGN) with a CTR ≤50%. Methods: A total of 247 patients with LUAD from our hospital were randomly divided into two groups, i.e., the training cohort (n=173) and the internal validation cohort (n=74) (7:3 ratio). Furthermore, 47 patients from three other hospitals were collected as the external validation cohort. In the training cohort, the differences in clinical-radiological features were compared using univariate and multivariate analyses. The gross tumor volume (GTV) and gross peritumoral tumor volume (GPTV5, GPTV10, and GPTV15) radiomics models were constructed based on intratumoral and peritumoral (5, 10, and 15 mm) radiomics features. Additionally, the radscore of the best radiomics model and clinical risk factors were used to construct a combined model and the predictive efficacy of the model was evaluated in the validation cohorts. Finally, the receiver operating characteristics (ROC) curve and area under the curve (AUC) value were used to evaluate the discriminative ability of the model. Results: Tumor size and CTR were independent risk factors for predicting the invasiveness of LUAD. The GPTV10 model outperformed the other radiomics models, with AUC values of 0.910, 0.870, and 0.887 in the three cohorts. The AUC values of the combined model were 0.912, 0.874, and 0.892. Conclusions: A nomogram based on GPTV10-radscore, tumor size, and CTR exhibited high predictive efficiency for predicting the invasiveness of LUAD.

CT-based whole lung radiomics nomogram for identification of PRISm from non-COPD subjects.

BACKGROUND: Preserved Ratio Impaired Spirometry (PRISm) is considered to be a precursor of chronic obstructive pulmonary disease. Radiomics nomogram can effectively identify the PRISm subjects from non-COPD subjects, especially when during large-scale CT lung cancer screening. METHODS: Totally 1481 participants (864, 370 and 247 in training, internal validation, and external validation cohorts, respectively) were included. Whole lung on thin-section computed tomography (CT) was segmented with a fully automated segmentation algorithm. PyRadiomics was adopted for extracting radiomics features. Clinical features were also obtained. Moreover, Spearman correlation analysis, minimum redundancy maximum relevance (mRMR) feature ranking and least absolute shrinkage and selection operator (LASSO) classifier were adopted to analyze whether radiomics features could be used to build radiomics signatures. A nomogram that incorporated clinical features and radiomics signature was constructed through multivariable logistic regression. Last, calibration, discrimination and clinical usefulness were analyzed using validation cohorts. RESULTS: The radiomics signature, which included 14 stable features, was related to PRISm of training and validation cohorts (p < 0.001). The radiomics nomogram incorporating independent predicting factors (radiomics signature, age, BMI, and gender) well discriminated PRISm from non-COPD subjects compared with clinical model or radiomics signature alone for training cohort (AUC 0.787 vs. 0.675 vs. 0.778), internal (AUC 0.773 vs. 0.682 vs. 0.767) and external validation cohorts (AUC 0.702 vs. 0.610 vs. 0.699). Decision curve analysis suggested that our constructed radiomics nomogram outperformed clinical model. CONCLUSIONS: The CT-based whole lung radiomics nomogram could identify PRISm to help decision-making in clinic.

Comparison of deep-learning and radiomics-based machine-learning methods for the identification of chronic obstructive pulmonary disease on low-dose computed tomography images.

Background: Radiomics and artificial intelligence approaches have been developed to predict chronic obstructive pulmonary disease (COPD), but it is still unclear which approach has the best performance. Therefore, we established five prediction models that employed deep-learning (DL) and radiomics-based machine-learning (ML) approaches to identify COPD on low-dose computed tomography (LDCT) images and compared the relative performance of the different models to find the best model for identifying COPD. Methods: This retrospective analysis included 1,024 subjects (169 COPD patients and 855 control subjects) who underwent LDCT scans from August 2018 to July 2021. Five prediction models, including models that employed computed tomography (CT)-based radiomics features, chest CT images, quantitative lung density parameters, and demographic and clinical characteristics, were established to identify COPD by DL or ML approaches. Model 1 used CT-based radiomics features by ML method. Model 2 used a combination of CT-based radiomics features, lung density parameters, and demographic and clinical characteristics by ML method. Model 3 used CT images only by DL method. Model 4 used a combination of CT images, lung density parameters, and demographic and clinical characteristics by DL method. Model 5 used a combination of CT images, CT-based radiomics features, lung density parameters, and demographic and clinical characteristics by DL method. The accuracy, sensitivity, specificity, highest negative predictive values (NPVs), positive predictive values, and areas under the receiver operating characteristic (AUC) curve of the five prediction models were compared to examine their performance. The DeLong test was used to compare the AUCs of the different models. Results: In total, 107 radiomics features were extracted from each subject's CT images, 17 lung density parameters were acquired by quantitative measurement, and 18 selected demographic and clinical characteristics were recorded in this study. Model 2 had the highest AUC [0.73, 95% confidence interval (CI): 0.64-0.82], while model 3 had the lowest AUC (0.65, 95% CI: 0.55-0.75) in the test set. Model 2 also had the highest sensitivity (0.84), the highest accuracy (0.81), and the highest NPV (0.36). In the test set, based on the AUC results, Model 2 significantly outperformed Model 1 (P=0.03). Conclusions: The results showed that the identification ability of models that employ CT-based radiomics features combined with lung density parameters, and demographic and clinical characteristics using ML methods performed better than the chest CT image-based DL methods. ML methods are more suitable and beneficial for COPD identification.

CT whole lung radiomic nomogram: a potential biomarker for lung function evaluation and identification of COPD.

BACKGROUND: Computed tomography (CT) plays a great role in characterizing and quantifying changes in lung structure and function of chronic obstructive pulmonary disease (COPD). This study aimed to explore the performance of CT-based whole lung radiomic in discriminating COPD patients and non-COPD patients. METHODS: This retrospective study was performed on 2785 patients who underwent pulmonary function examination in 5 hospitals and were divided into non-COPD group and COPD group. The radiomic features of the whole lung volume were extracted. Least absolute shrinkage and selection operator (LASSO) logistic regression was applied for feature selection and radiomic signature construction. A radiomic nomogram was established by combining the radiomic score and clinical factors. Receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA) were used to evaluate the predictive performance of the radiomic nomogram in the training, internal validation, and independent external validation cohorts. RESULTS: Eighteen radiomic features were collected from the whole lung volume to construct a radiomic model. The area under the curve (AUC) of the radiomic model in the training, internal, and independent external validation cohorts were 0.888 [95% confidence interval (CI) 0.869-0.906], 0.874 (95%CI 0.844-0.904) and 0.846 (95%CI 0.822-0.870), respectively. All were higher than the clinical model (AUC were 0.732, 0.714, and 0.777, respectively, P < 0.001). DCA demonstrated that the nomogram constructed by combining radiomic score, age, sex, height, and smoking status was superior to the clinical factor model. CONCLUSIONS: The intuitive nomogram constructed by CT-based whole-lung radiomic has shown good performance and high accuracy in identifying COPD in this multicenter study.

Quantitative Assessment Characteristics of Small Pulmonary Vessel Remodelling in Populations at High Risk for COPD and Smokers Using Low-Dose CT.

Purpose: To explore the morphological alterations in small pulmonary vessels in populations at high risk for chronic obstructive pulmonary disease (COPD) and smokers based on multiple computed tomography (CT) quantitative parameters. Patients and Methods: A total of 1969 Three Major Chest Diseases Screening Study participants with available demographic data and smoking history who underwent low-dose chest CT from 2018 to 2020 were included. All subjects were divided into normal, high risk for COPD, and COPD groups according to their pulmonary function test (PFT) results. Furthermore, the three groups were further subdivided into never-smokers, current smokers, and former smokers subgroups according to their smoking history. Quantitative parameters, such as the number, area at 6 mm~24 mm subpleura and volume of small pulmonary vessels, were extracted by computer software. Differences in small pulmonary vessel parameters among the groups were compared using two-way ANOVA. Results: The number, area at 6 mm~24 mm subpleura and volume of small pulmonary vessels in the group at high risk for COPD were lower than those in the normal group (P<0.05). The number, area at 6 mm~24 mm subpleura and volume of small pulmonary vessels in the COPD group were higher than those in the normal group (P<0.05). The number, area of small pulmonary vessels at 6 mm~12 mm subpleura in current smokers with high risk for COPD were higher than those in former smokers with high risk for COPD (P<0.05). Conclusion: The number, area, and volume of small pulmonary vessels in populations at high risk for COPD were decreased. Smoking cessation may impede structural changes in small pulmonary vessels in populations at high risk for COPD.

The characteristics of emphysema and pulmonary perfusion derived from spectral CT in smokers.

Background: Tobacco smoking may cause pulmonary perfusion abnormality. Assessment of the lung perfusion characteristics is very significant to timely treatment and prevent disease progression in smokers. The purpose was to investigate the value of iodine maps from spectral dual-layer detector computed tomography (DLCT) in assessing lung perfusion changes in smokers. Methods: Nineteen smokers and 29 non-smokers who underwent dual-phase contrast enhanced scans on a spectral DLCT were retrospectively collected. Emphysema on non-contrast images and perfusion defect (PD) on iodine maps were scored visually at bilateral lung fields of three anatomic levels (on the slice of the aortic arch, the carina, and the aperture of the inferior pulmonary veins). The scores were calculated based on the ratio of the abnormality occupied in the pulmonary field of each slice as described below: point 0, no abnormality; point 1, 0%< abnormality scope ≤25%; points 2, 25%< abnormality scope ≤50%; points 3, 50%< abnormality scope ≤75%; points 4, abnormality scope >75%. The sum of scores for each patient was calculated. The iodine density (ID) of PD and thoracic aorta were measured respectively (IDdefect, IDthoracic aorta), then calculating the ratio as the normalized ID (nID). Emphysema index (EI) was defined as the volume percentage of the lung attenuation below -950 Hounsfield units. The percentage of forced expiratory volume in 1 second (FEV1) to the predicted value (FEV1%) and the ratio of FEV1 to forced vital capacity (FVC) were recorded. The differences of the emphysema and PD visual scores, IDdefect, nID, EI were analyzed by analysis of variance between smokers and non-smokers. Correlations between emphysema, PD and FEV1%, FEV1/FVC were evaluated by Spearman correlation analysis. Results: The PD visual scores on ID images were significantly higher in smokers compared with that in non-smokers (P=0.014), while no significantly difference was found for emphysema visual scores (P=0.402). Both IDdefect and nID were significantly lower in smokers compared with non-smokers (P=0.003; P=0.029), while no significantly difference was found for EI (P=0.061). Besides, PD visual scores were negatively correlated with FEV1% (r=-0.61, P=0.025) and FEV1/FVC (r=-0.62, P=0.024) for smokers. Conclusions: Compared with emphysema, the iodine map derived from spectral DLCT showed higher sensitivity for the evaluation of the pulmonary abnormalities of smokers.

Extracellular volume fraction as a potential predictor to differentiate lung cancer from benign lung lesions with dual-layer detector spectral CT.

Background: Extracellular volume (ECV) fraction has been used in cardiovascular diseases, pancreatic fibrosis, and hepatic fibrosis. The diagnostic value of ECV for focal lung lesions remains to be explored. The aim of this study was to evaluate the feasibility of ECV derived from a dual-layer detector computed tomography (DLCT) to differentiate lung cancer (LC) from benign lung lesions (BLLs). Methods: Retrospectively, 128 consecutive patients with pathologically confirmed LC (n=86) or BLLs (n=42) were included. Conventional computed tomography (CT) characteristics and spectral CT parameters were assessed. All patients' hematocrits were measured to correct contrast volume distributions in blood while calculating ECV. After performing logistic regression analysis, a conventional CT-based model (Model A), DLCT-based model (Model B), combined diagnostic models (Model C), and an ECV-based model (Model D) were developed. The diagnostic effectiveness of each model was examined using the receiver operating characteristic (ROC) curve and their corresponding 95% confidence intervals (CIs). The area under the curve (AUC) of each model was compared using the DeLong test. Results: Certain conventional CT features (such as lesion size, lobulation, spiculation, pleural indentation, and enlarged lymph nodes) differed significantly between the LC and BLL groups (all P<0.05). Statistical differences were found in the following DLCT parameters (all P<0.05): effective atomic number (Zeff) (non-enhancement), electron density (ED) (non-enhancement), ECV, iodine concentration (IC), and normalized iodine concentration (NIC). Models A, B, C, and D had AUCs of 0.801 [95% confidence interval (CI): 0.721-0.866], 0.805 (95% CI: 0.726-0.870), 0.925 (95% CI: 0.865-0.964), and 0.754 (95% CI: 0.671-0.826), respectively. The AUC of Model D (ECV) showed no significant difference from that of Models A and B (DeLong test, P>0.05). Conclusions: The ECV derived from DLCT may be a potential new method to differentiate LC from BLLs, broadening the scope of ECV in clinical research.

Ultra-low-dose spectral-detector computed tomography for the accurate quantification of pulmonary nodules: an anthropomorphic chest phantom study.

PURPOSE: To assess the quantification accuracy of pulmonary nodules using virtual monoenergetic images (VMIs) derived from spectral-detector computed tomography (CT) under an ultra-low-dose scan protocol. METHODS: A chest phantom consisting of 12 pulmonary nodules was scanned using spectral-detector CT at 100 kVp/10 mAs, 100 kVp/20 mAs, 120 kVp/10 mAs, and 120 kVp/30 mAs. Each scanning protocol was repeated three times. Each CT scan was reconstructed utilizing filtered back projection, hybrid iterative reconstruction, iterative model reconstruction (IMR), and VMIs of 40-100 keV. The signal-to-noise ratio and air noise of images, absolute differences, and absolute percentage measurement errors (APEs) of the diameter, density, and volume of the four scan protocols and ten reconstruction images were compared. RESULTS: With each fixed reconstruction image, the four scanning protocols exhibited no significant differences in APEs for diameter and density (all P > 0.05). Of the four scan protocols and ten reconstruction images, APEs for nodule volume had no significant differences (all P > 0.05). At 100 kVp/10 mAs, APEs for density using IMR were the lowest (APE-mean: 6.69), but no significant difference was detected between VMIs at 50 keV (APE-mean: 11.69) and IMR (P = 0.666). In the subgroup analysis, at 100 kVp/10 mAs, there were no significant differences between VMIs at 50 keV and IMR in diameter and density (all P > 0.05). The radiation dose at 100 kVp/10 mAs was reduced by 77.8% compared with that at 120 kVp/30 mAs. CONCLUSION: Compared with IMR, reconstruction at 100 kVp/10 mAs and 50 keV provides a more accurate quantification of pulmonary nodules, and the radiation dose is reduced by 77.8% compared with that at 120 kVp/30 mAs, demonstrating great potential for ultra-low-dose spectral-detector CT.

Radiomics-based Machine Learning Methods for Volume Doubling Time Prediction of Pulmonary Ground-glass Nodules With Baseline Chest Computed Tomography.

PURPOSE: Reliable prediction of volume doubling time (VDT) is essential for the personalized management of pulmonary ground-glass nodules (GGNs). We aimed to determine the optimal VDT prediction method by comparing different machine learning methods only based on the baseline chest computed tomography (CT) images. MATERIALS AND METHODS: Seven classical machine learning methods were evaluated in terms of their stability and performance for VDT prediction. The VDT, calculated by the preoperative and baseline CT, was divided into 2 groups with a cutoff value of 400 days. A total of 90 GGNs from 3 hospitals constituted the training set, and 86 GGNs from the fourth hospital served as the external validation set. The training set was used for feature selection and model training, and the validation set was used to evaluate the predictive performance of the model independently. RESULTS: The eXtreme Gradient Boosting showed the highest predictive performance (accuracy: 0.890±0.128 and area under the ROC curve (AUC): 0.896±0.134), followed by the neural network (NNet) (accuracy: 0.865±0.103 and AUC: 0.886±0.097). While regarding stability, the NNet showed the highest robustness against data perturbation (relative SDs [%] of mean AUC: 10.9%). Therefore, the NNet was chosen as the final model, achieving high accuracy of 0.756 in the external validation set. CONCLUSION: The NNet is a promising machine learning method to predict the VDT of GGNs, which would assist in the personalized follow-up and treatment strategies for GGNs reducing unnecessary follow-up and radiation dose.

Nomograms Using CT Morphological Features and Clinical Characteristics to Identify COPD in Patients with Lung Cancer: A Multicenter Study.

Purpose: This study aimed to screen out computed tomography (CT) morphological features and clinical characteristics of patients with lung cancer to identify chronic obstructive pulmonary disease (COPD). Further, we aimed to develop and validate different diagnostic nomograms for predicting whether lung cancer is comorbid with COPD. Patients and Methods: This retrospective study examined data from 498 patients with lung cancer (280 with COPD, 218 without COPD; 349 in training cohort, 149 in validation cohort) from two centers. Five clinical characteristics and 20 CT morphological features were evaluated. Differences in all variables were assessed between COPD and non-COPD groups. Models were developed using multivariable logistic regression to identify COPD, including clinical, imaging, and combined nomograms. Receiver operating characteristic curves were used to evaluate and compare the performance of nomograms. Results: Age, sex, interface, bronchus cutoff sign, spine-like process, and spiculation sign were independent predictors of COPD in patients with lung cancer. In the training and validation cohorts, the clinical nomogram showed good performance to predict COPD in lung cancer patients (areas under the curves [AUCs] of 0.807 [95% CI, 0.761-0.854] and 0.753 [95% CI, 0.674-0.832]); while the imaging nomogram showed slightly better performance (AUCs of 0.814 [95% CI, 0.770-0.858] and 0.780 [95% CI, 0.705-0.856]). For the combined nomogram generated with clinical and imaging features, the performance was further improved (AUC=0.863 [95% CI, 0.824-0.903], 0.811 [95% CI, 0.742-0.880] in the training and validation cohort). At 60% risk threshold, there were more true negative predictions (48 vs 44) and higher accuracy (73.15% vs 71.14%) for the combined nomogram compared with the clinical nomogram in the validation cohort. Conclusion: The combined nomogram developed with clinical and imaging features outperformed clinical and imaging nomograms; this provides a convenient method to detect COPD in patients with lung cancer using one-stop CT scanning.

CT-Based Radiomic Nomogram for the Prediction of Chronic Obstructive Pulmonary Disease in Patients with Lung cancer.

RATIONALE AND OBJECTIVES: To develop and validate a model for predicting chronic obstructive pulmonary disease (COPD) in patients with lung cancer based on computed tomography (CT) radiomic signatures and clinical and imaging features. MATERIALS AND METHODS: We retrospectively enrolled 443 patients with lung cancer who underwent pulmonary function test as the primary cohort. They were randomly assigned to the training (n = 311) or validation (n = 132) set in a 7:3 ratio. Additionally, an independent external cohort of 54 patients was evaluated. The radiomic lung nodule signature was constructed using the least absolute shrinkage and selection operator algorithm, while key variables were selected using logistic regression to develop the clinical and combined models presented as a nomogram. RESULTS: COPD was significantly related to the radiomics signature in both cohorts. Moreover, the signature served as an independent predictor of COPD in the multivariate regression analysis. For the training, internal, and external cohorts, the area under the receiver operating characteristic curve (ROC, AUC) values of our radiomics signature for COPD prediction were 0.85, 0.85, and 0.76, respectively. Additionally, the AUC values of the radiomic nomogram for COPD prediction were 0.927, 0.879, and 0.762 for the three cohorts, respectively, which outperformed the other two models. CONCLUSION: The present study presents a nomogram that incorporates radiomics signatures and clinical and radiological features, which could be used to predict the risk of COPD in patients with lung cancer with one-stop chest CT scanning.

Computed Tomography Screening for Early Lung Cancer, COPD and Cardiovascular Disease in Shanghai: Rationale and Design of a Population-based Comparative Study.

RATIONALE AND OBJECTIVES: To describe the rational and design of a population-based comparative study. The objective of the study is to assess the screening performance of volume-based management of CT-detected lung nodule in comparison to diameter-based management, and to improve the effectiveness of CT screening for chronic obstructive pulmonary disease (COPD) and cardiovascular disease (CVD), in addition to lung cancer, based on quantitative measurement of CT imaging biomarkers in a Chinese screening setting. MATERIALS AND METHODS: A population-based comparative study is being performed, including 10,000 asymptomatic participants between 40 and 74 years old from Shanghai urban population. Participants in the intervention group undergo a low-dose chest and cardiac CT scan at baseline and 1 year later, and are managed according to NELCIN-B3 protocol. Participants in the control group undergo a low-dose chest CT scan according to the routine CT protocol and are managed according to the clinical practice. Epidemiological data are collected through questionnaires. In the fourth year from baseline, the diagnosis of the three diseases will be collected. RESULTS: The unnecessary referral rate will be compared between NELCIN-B3 and standard protocol for managing early-detected lung nodules. The effectiveness of quantitative measurement of CT imaging biomarkers for early detection of lung cancer, COPD and CVD will be evaluated. CONCLUSION: We expect that the quantitative assessment of the CT imaging biomarkers will reduce the number of unnecessary referrals for early detected lung nodules, and will improve the early detection of COPD and CVD in a Chinese urban population. TRIAL REGISTRATION: ClinicalTrials.gov, NCT03988322. Registered on 14 June 2019.

Radiomics Signature: A Biomarker for the Preoperative Distant Metastatic Prediction of Stage I Nonsmall Cell Lung Cancer.

OBJECTIVES: To evaluate the predictive value of radiomics features on the distant metastasis (DM) of stage I nonsmall cell lung cancer (NSCLC) preoperatively, by comparing with clinical characteristics and CT morphological features, and to screen the important prognostic predictors. METHODS: One hundred ninety-four stage I NSCLC patients were retrospectively enrolled, DM free survival (DMFS) was evaluated. The consensus clustering analysis was used to build the radiomics signatures in the primary cohort and validated in the validation cohort. The univariate survival analysis was performed in clinical characteristics, CT morphological features and radiomics signatures, respectively. Cox model was performed and C-index was calculated. RESULTS: There were 25 patients (12.9%) with DM. The median DMFS was 15 months. Three hundred thirteen radiomics features were selected, then classified into five groups, two subtypes (I and II) with each group. The RS1 showed the best prognostic ability with C-index of 0.355(95% confidence interval [CI], 0.269-0.442; p < 0.001). The histological type exhibited a good prognostic ability with C-index of 0.123 (95% CI, 0.000-0.305; p < 0.001) for DMFS. Cox model showed RS1(hazard ratio [HR] 18.025, 95% CI 2.366-137.340), pleural indentation sign (HR 2.623, 95% CI 1.070-6.426) and histological type (HR 4.461, 95% CI 1.783-11.162) were the independent prognostic factors (p < 0.05). CONCLUSION: Radiomics provided a new modality for the distant metastatic prediction of stage I NSCLC. Patients with type II of RS1, pleural indentation sign and nonadenocarcinoma indicated the high probability of postsurgical DM.