Abstract
Purpose: This investigation sought to develop and verify an iodine maps-based radiomics nomogram to assess mucosal healing (MH) in Crohn’s disease (CD) patients.
Methods: A total of 81 CD patients were randomly divided into a training set (n=60) and a testing set (n=21) in a 7:3 ratio. Significant demographic features were selected to construct a clinical model. Radiomics features were extracted from the dual-energy CT iodine density map, and the least absolute shrinkage and selection operator (LASSO)was used to select the most important feature to calculate the radiomics score (Rad-score). Combine radiomics and clinical features to construct a clinical radiomics model by logistic regression analysis and draw a clinical radiomics nomogram. Receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were used to evaluate the effectiveness, fit, and clinical net benefit of the three model.
Results: 25 patients achieved MH among 81 patients. The area under the ROC curve (AUC) of clinical radiomics nomogram based on ESR, CRP, and Rad-score for distinguishing MH from non-MH was 0.838 (95% CI:0.765 to 0.911) and 0.817 (95% CI：0.685 to 0.950) in the training and testing set. The DCA showed good clinical net benefits for all three models when the threshold probability was greater than 0.12. The calibration curve shows that the predicted probability is consistent with the actual probability, within the range of 0.78~0.90.
Conclusion: In summary, this study suggests that the radiomics model based on iodine maps has potential value in evaluating MH in CD patients.