Deep Learning Radiomics Analysis of CT Imaging for Differentiating Between Crohn's Disease and Intestinal Tuberculosis

Abstract

This study aimed to develop and evaluate a CT-based deep learning radiomics model for differentiating between Crohn's disease (CD) and intestinal tuberculosis (ITB). A total of 330 patients with pathologically confirmed as CD or ITB from the First Affiliated Hospital of Zhengzhou University were divided into the validation dataset one (CD: 167; ITB: 57) and validation dataset two (CD: 78; ITB: 28). Based on the validation dataset one, the synthetic minority oversampling technique (SMOTE) was adopted to create balanced dataset as training data for feature selection and model construction. The handcrafted and deep learning (DL) radiomics features were extracted from the arterial and venous phases images, respectively. The interobserver consistency analysis, Spearman's correlation, univariate analysis, and the least absolute shrinkage and selection operator (LASSO) regression were used to select features. Based on extracted multi-phase radiomics features, six logistic regression models were finally constructed. The diagnostic performances of different models were compared using ROC analysis and Delong test. The arterial-venous combined deep learning radiomics model for differentiating between CD and ITB showed a high prediction quality with AUCs of 0.885, 0.877, and 0.800 in SMOTE dataset, validation dataset one, and validation dataset two, respectively. Moreover, the deep learning radiomics model outperformed the handcrafted radiomics model in same phase images. In validation dataset one, the Delong test results indicated that there was a significant difference in the AUC of the arterial models (p = 0.037), while not in venous and arterial-venous combined models (p = 0.398 and p = 0.265) as comparing deep learning radiomics models and handcrafted radiomics models. In our study, the arterial-venous combined model based on deep learning radiomics analysis exhibited good performance in differentiating between CD and ITB.

Keywords: Crohn’s disease; Deep learning; Diagnosis; Intestinal tuberculosis; Radiomics.

Fig. 1

The patient enrollment workflow

Fig. 2

The boxplot for Radscores from deep learning and handcrafted radiomics models and their statistical differences between CD and ITB in the SMOTE dataset. A–C The respective distribution of Radscore_AP__DL, Radscore_VP__DL, and Radscore_{AP_VP}__DL in the CD and ITB groups in the SMOTE dataset. D–F The respective distribution of Radscore_{AP_HC}, Radscore_VP__HC, and Radscore_{AP_VP}__HC in the CD and ITB groups in the SMOTE dataset

Fig. 3

The boxplot for Radscores from deep learning and handcrafted radiomics models and their statistical differences between CD and ITB in validation dataset one. A–C The respective distribution of Radscore_AP__DL, Radscore_VP__DL, and Radscore_{AP_VP}__DL in the CD and ITB groups in validation dataset one. D–F The respective distribution of Radscore_{AP_HC}, Radscore_VP__HC, and Radscore_{AP_VP}__HC in the CD and ITB groups in validation dataset one

Fig. 4

The ROC curves for deep learning and handcrafted radiomics models. The ROC curves of artery (blue), venous (dark sea green), and artery-venous (red) based on deep learning radiomics models in SMOTE dataset (A), the validation dataset one (B), and the validation dataset two (C). The ROC curves of artery (blue), venous (dark sea green), and artery-venous (red) based on handcrafted radiomics models in SMOTE dataset (D), validation dataset one (E), and validation dataset two (F)

Fig. 5

The calibration curves for deep learning and handcrafted radiomics models. The calibration curves of artery (blue), venous (dark sea green), and artery-venous (red) based on deep learning radiomics models in SMOTE dataset (A), validation dataset one (B), and validation dataset two (C). The ROC curves of artery (blue), venous (dark sea green), and artery-venous (red) based on handcrafted radiomics models in SMOTE dataset (D), validation dataset one (E), and validation dataset two (F)

Fig. 6

The decision curves for deep learning and handcrafted radiomics models. The decision curves for deep learning radiomics models in SMOTE dataset (A), validation dataset one (B), and validation dataset two (C). The decision curves for handcrafted radiomics models in SMOTE dataset (D), validation dataset one (E), and validation dataset two (F). The gray line represents hypothesis that all patients are intestinal tuberculosis (all). Black line represents hypothesis that all patients are Crohn’s disease (None). The colored lines of each model respectively illustrate the net benefit brought to each patient based on artery (blue), venous (dark sea green), and artery-venous combined (red) radiomics models. The closer the decision curves to the black and gray curves, the lower the clinical decision net benefit of the model