Search

ObjectivesTo build and evaluate a deep learning radiomics nomogram (DLRN) for preoperative prediction of lung metastasis (LM) status in patients with soft tissue sarcoma (STS).MethodsIn total, 242 patients with STS (training set, n=116; external validation set, n=126) who underwent magnetic resonance imaging were retrospectively enrolled in this study. We identified independent predictors for LM-status and evaluated their performance. The minimum redundancy maximum relevance (mRMR) method and least absolute shrinkage and selection operator (LASSO) algorithm were adopted to screen radiomics features. Logistic regression, decision tree, random forest, support vector machine (SVM), and adaptive boosting classifiers were compared for their ability to predict LM. To overcome the imbalanced distribution of the LM data, we retrained each machine-learning classifier using the synthetic minority over-sampling technique (SMOTE). A DLRN combining the independent clinical predictors with the best performing radiomics prediction signature (mRMR+LASSO+SVM+SMOTE) was established. Area under the receiver operating characteristics curve (AUC), calibration curves, and decision curve analysis (DCA) were used to assess the performance and clinical applicability of the models.ResultComparisons of the AUC values applied to the external validation set revealed that the DLRN model (AUC=0.833) showed better prediction performance than the clinical model (AUC=0.664) and radiomics model (AUC=0.799). The calibration curves indicated good calibration efficiency and the DCA showed the DLRN model to have greater clinical applicability than the other two models.ConclusionThe DLRN was shown to be an accurate and efficient tool for LM-status prediction in STS.

ObjectivesTo investigate the efficacy of multi-parametric MRI-based radiomics nomograms for preoperative distinction between benign and malignant sinonasal tumors.MethodsData of 244 patients with sinonasal tumor (training set, n=192; test set, n=52) who had undergone pre-contrast MRI, and 101 patients who underwent post-contrast MRI (training set, n=74; test set, n=27) were retrospectively analyzed. Independent predictors of malignancy were identified and their performance were evaluated. Seven radiomics signatures (RSs) using maximum relevance minimum redundancy (mRMR), and the least absolute shrinkage selection operator (LASSO) algorithm were established. The radiomics nomograms, comprising the clinical model and the RS algorithms were built: one based on pre-contrast MRI (RNWOC); the other based on pre-contrast and post-contrast MRI (RNWC). The performances of the models were evaluated with area under the curve (AUC), calibration, and decision curve analysis (DCA) respectively.ResultsThe efficacy of the clinical model (AUC=0.81) of RNWC was higher than that of the model (AUC=0.76) of RNWOC in the test set. There was no significant difference in the AUC of radiomic algorithms in the test set. The RS-T1T2 (AUC=0.74) and RS-T1T2T1C (RSWC, AUC=0.81) achieved a good distinction efficacy in the test set. The RNWC and the RNWOC showed excellent distinction (AUC=0.89 and 0.82 respectively) in the test set. The DCA of the nomograms showed better clinical usefulness than the clinical models and radiomics signatures.ConclusionsThe radiomics nomograms combining the clinical model and RS can be accurately, safely and efficiently used to distinguish between benign and malignant sinonasal tumors.