An Online Dynamic Radiomics–Clinical Nomogram to Predict Recurrence in Patients with Spontaneous Intracerebral Hemorrhage

Abstract

ObjectiveRadiomics can reflect the heterogeneity within the focus. We aim to explore whether radiomics can predict recurrent intracerebral hemorrhage (RICH) and develop an online dynamic nomogram to predict it. MethodsThis retrospective study collected the clinical and radiomics features of spontaneous intracerebral hemorrhage (ICH) patients seen in our hospital from October 2013 to October 2016. We used the minimum redundancy maximum relevancy and the least absolute shrinkage and selection operator methods to screen radiomics features and calculate Rad-score. We use the univariate and multivariate analyses to screen clinical predictors. Optimal clinical features and Rad-score were used to construct different logistics regression models called the clinical model, radiomics model, and combined-LR model. The Delong testing was performed to compare performance between different models. The model with the best predictive performance was used to construct an online dynamic nomogram. ResultsOverall, 304 ICH patients were enrolled in this study. 14 radiomics features was selected to calculate the Rad-score. The patients with RICH had a significantly higher Rad-score than without (0.5 VS. -0.8, p < 0.001). The predictive performance of the combined-LR model with Rad-score was better than that of the clinical model for both the training (AUC, 0.81 vs. 0.71, p = 0.02) and testing (AUC, 0.65 vs. 0.58, p = 0.04) cohorts in statistically. ConclusionsRadiomics features were determined related to RICH. Adding Rad-score into conventional clinical models significantly improves the prediction efficiency. We developed a online dynamic nomogram to accurately and conveniently evaluate RICH.