AUTHOR=Yue Xiaoning , He Xiaoyu , He Shuaijie , Wu Jingjing , Fan Wei , Zhang Haijun , Wang Chengwei 

TITLE=Multiparametric magnetic resonance imaging-based radiomics nomogram for predicting tumor grade in endometrial cancer

JOURNAL=Frontiers in Oncology

VOLUME=Volume 13 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2023.1081134

DOI=10.3389/fonc.2023.1081134

ISSN=2234-943X

ABSTRACT=Background: Tumor grade is bound up with the treatment and prognosis of endometrial carcinoma (EC) patients. Preoperative accurate prediction of tumor grade is essential for EC risk stratification. We aim to assess the performance of multiparametric MRI-based radiomics nomogram in predicting high-grade EC.
Methods: 143 (mean age: 55.5±10.5 [SD] years; range: 32-84 years) patients with a histopathologic diagnosis of EC underwent preoperative pelvic MRI were retrospectively collected and divided into a training set and a test set. Radiomic features were extracted based on T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI) and dynamic contrast-enhanced MRI (DCE-MRI). The minimum absolute contraction selection operator (LASSO) was implemented to obtain the optimal radiomics features. The clinical risk factors and radiomics features were determined by stepwise multiple logistic regression. The receiver operating characteristic (ROC) curve was used to evaluate the performance of nomogram in prediction of high-grade EC. The clinical net benefit of nomogram was assessed by decision curve analysis (DCA), net reclassification index (NRI) and integrated discrimination index (IDI).
Results: High-grade EC was confirmed pathologically in 35 patients. The radiomics nomogram was developed by combining clinical risk factors (patient age, human epididymal protein 4, deep myometrial invasion and cervical stromal invasion reported by MRI) and 11 robust radiomics features. The AUCs of the radiomics nomogram were 0.920 (95%CI: 0.867-0.974) in the training set and 0.896 (95%CI: 0.799-0.994) in the validation set. The DCA showed a good net benefit of nomogram with NRI of 0.43 and IDI of 0.05 compared with dilation and curettage (D&C).
Conclusions: The radiomics nomogram using MRI and clinical data yielded excellent performance in predicting high-grade EC, which can optimize the clinical treatment plan.