Multiparametric MRI–based Radiomics to Predict the Depth of Myometrial Invasion of Endometrial Carcinoma

Motivation: Myometrial invasion depth of endometrial cancer (EC) is one of the important prognostic factors and is crucial for clinical treatment planning. Goal(s): To evaluate the predictive value and potential of radiomics features based on multiparameter MRI for myometrial invasion depth of EC. Approach: Radiomics was based on multiparametric MRI. Results: The combined predictive models had a better performance than the clinical model. In the training set and the test set, the NRI and IDI of combined predictive models was higher than that of clinical model. The nomogram achieved good net benefit by DCA for myometrial invasion depth of EC. Impact: The performance of MRI appears strongly dependent on radiologist's experience with a relatively high interobserver variability. 3D segmentation can better reflect tumor heterogeneity. Preoperative radiomics based on multiparameter MRI showed higher diagnostic performance, may enhanced decision-making capabilities of radiologists.