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

Abstract

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.

Keywords: Chronic obstructive pulmonary disease (COPD); Computed tomography (CT); Radiomic.

Fig. 1

Diagram showing the patient inclusion and exclusion process. Center 1: Tongji Hospital, School of Medicine, Tongji University; Center 2: Zhejiang Province People’s Hospital; Center 3: Sir Run Run Shaw Hospital; Center 4: the First Affiliated Hospital of Nanchang Medical College; Center 5: the Second Affiliated Hospital of Naval Medical University. COPD chronic obstructive pulmonary disease, PFT pulmonary function disease, CT computed tomography

Fig. 2

Original chest HRCT images (a–c) and segmentation results (d–f) of typical lung regions in transverse, coronal, and sagittal planes based on the original chest HRCT images, respectively. The red mask is the right lung parenchyma, and the green one is the left lung parenchyma. HRCT high-resolution computed tomography

Fig. 3

LASSO coefficients of radiomic features. a The LASSO coefficient profiles of the 283 radiomics features. A vertical line was generated at the log (λ) value by using tenfold cross-validation, where the optimal λ value resulted in 18 radiomics features. The optimal λ value of 0.00057 was selected. The X-axis on the top indicates the number of nonzero coefficient features in the model. b The black vertical line was drawn at the value selected using tenfold cross-validation in (a). The X-axis on the top indicates the number of nonzero coefficient features in the model. c Histogram of the Radscore: the Y-axis indicates the selected 18 radiomic features, and the X-axis represents the coefficient of the radiomic features. LASSO least absolute shrinkage and selection operator

Fig. 4

ROC curves of the radiomic model, clinical model, and combined model in predicting COPD in the training cohort (a), internal validation cohort (b), and external validation cohort (c). ROC receiver operating characteristic, COPD chronic obstructive pulmonary disease

Fig. 5

Development and performance of radiomic nomogram. a Radiomic nomogram developed to predict COPD. b Calibration curve between the predicted and actual incidences of COPD. c Decision curve analysis compares the net benefits of four scenarios in predicting the risk of COPD: Combined model (red line), Clinical model (blue line), All (green line, refers to the assumption that all patients have COPD) and None (horizontal solid black line, represents the assumption that no patient has COPD). COPD chronic obstructive pulmonary disease

Fig. 6

The risk scores of COPD in two patients were calculated by using the nomogram. a Thin-slice chest CT images of non-COPD in a 45-year-old woman with height 152 cm, non-smoker, Radscore -2.08. b Lung density analysis diagram showed no emphysema area in both lungs. c The nomogram shows that the total score was 44.8 points, corresponding to the probability of developing COPD is approximately 8.0%. Lung function examination showed that FEV1/FVC = 0.8. d Thin-slice chest CT image of COPD in an 82-year-old female subject. She is 152 cm tall, non-smoker, and has a Radscore of 3.17. e Lung density analysis diagram showed that both lungs are mostly scattered in the emphysema area (red). f The total score of the nomogram was 48.2, corresponding to the probability of developing COPD of approximately 96.9%. Pulmonary function examination showed that FEV1/FVC = 0.6. COPD chronic obstructive pulmonary disease, CT computed tomography, FEV1/FVC ratio of forced expiratory volume in 1 s to forced vital capacity