The value of MR-based radiomics in identifying residual disease in patients with carcinoma in situ after cervical conization

AbstractPurposeTo develop a magnetic resonance (MR)-based radiomics model for identifying residual disease in patients with carcinoma-in-situ (CIS) after cervical conization.Method110 patients who had CIS after conization and finally underwent hysterectomy were collected to comprise a database to establish an imaging model for predicting the residual status after conization. The imaging features were extracted from the cervical areas around the conization margin, and the performance of the imaging model was compared using different feature selection methods and with that of the pathological positive margins using receiver-operating-characteristic (ROC) analysis. Then, patients who opted for uterine preservation were included and were classified as high-risk or low-risk patients according to the radiomics model. The disease-free survival was compared between the different risk groups using the Kaplan-Meier curve.ResultsThe model with Boruta features achieved an area under the curve (AUC) of 0.889 and an accuracy of 87.3% in the test cohort and significantly outperformed the model created with the random forest method, which had an AUC of 0.701 (p = 0.039), and the positive margins (p = 0.004). Further validation with patients with uterine preservation showed that the patients classified as high risk with the radiomics model were more likely to have tumor recurrence/residual disease in the follow-up period than low-risk patients(p = 0.024).ConclusionsRadiomics can be used to identify residual disease in patients with CIS after cervical conization and could have the potential to predict recurrence in patients who opt for uterine preservation.