MR imaging of epithelial ovarian cancer: a combined model to predict histologic subtypes
- 13 June 2020
- journal article
- research article
- Published by Springer Science and Business Media LLC in European Radiology
- Vol. 30 (11), 5815-5825
- https://doi.org/10.1007/s00330-020-06993-5
Abstract
Objective To compare the performance of clinical features, conventional MR image features, ADC value, T2WI, DWI, DCE-MRI radiomics, and a combined multiple features model in predicting the type of epithelial ovarian cancer (EOC). Methods In this retrospective analysis, 61 EOC patients were confirmed by histology. Significant features (p < 0.05) by multivariate logistic regression were retained to establish a clinical model, conventional MRI morphological model, ADC model, and traditional model. The radiomics model included FS-T2WI, DWI, and DCE-MRI, and also, a multisequence model was established. A total of 1070 radiomics features of each sequence were extracted; then, univariate analysis and LASSO were used to select important features. Traditional models were combined with a combined radiomics model to establish a mixed model. The predictive performance was validated by receiver operating characteristic curve (ROC) analysis, calibration curve, and decision curve analysis (DCA). A stratified analysis was conducted to compare the differences between the combined radiomics model and the traditional model in identifying early- and late-stage EOC. Results Traditional models showed the highest performance (AUC = 0.96). The performance of the mixed model (AUC = 0.97) was not significantly different from that of the traditional model. The calibration curve showed that the traditional model had the highest reliability. Stratified analysis showed the potential of the combined radiomics model in the early distinction of the two tumor types. Conclusion The traditional model is an effective tool to distinguish EOC type I/II. Combined radiomics models have the potential to better distinguish EOC types in early FIGO stage disease. Key Points • The combined radiomics model resulted in a better predictive model than that from a single sequence model. • The traditional model showed higher classification accuracy than the combined radiomics model. • Combined radiomics models have the potential to better distinguish EOC types in early FIGO stage disease.Keywords
Funding Information
- Natural Science Foundation of Inner Mongolia (2017MS(LH)0837)
This publication has 33 references indexed in Scilit:
- Radiomics: Images Are More than Pictures, They Are DataRadiology, 2016
- Diffusion-weighted MRI of epithelial ovarian cancers: Correlation of apparent diffusion coefficient values with histologic grade and surgical stageEuropean Journal of Radiology, 2015
- Adnexal Masses: Development and Preliminary Validation of an MR Imaging Scoring SystemRadiology, 2013
- Clinical and Ultrasound Features of Type I and Type II Epithelial Ovarian CancerInternational Journal of Gynecologic Cancer, 2013
- Radiomics: Extracting more information from medical images using advanced feature analysisEuropean Journal Of Cancer, 2012
- The Origin and Pathogenesis of Epithelial Ovarian Cancer: A Proposed Unifying TheoryThe American Journal of Surgical Pathology, 2010
- Effect of quality of gynaecological ultrasonography on management of patients with suspected ovarian cancer: a randomised controlled trialThe Lancet Oncology, 2008
- Comparative evaluation of multidetector CT and MR imaging in the differentiation of adnexal massesEuropean Radiology, 2008
- MR imaging compared with intraoperative frozen-section examination for the diagnosis of adnexal tumors; correlation with final histologyEuropean Radiology, 2006
- Indeterminate Ovarian Mass at US: Incremental Value of Second Imaging Test for Characterization—Meta-Analysis and Bayesian AnalysisRadiology, 2005