Introduction: The Ki67 expression is associated with the malignant characteristics of bladder cancer (BCa) and can only be postoperatively detected by immunohistochemistry. We aimed to develop and validate Magnetic Resonance Imaging (MRI)-based radiomics signatures to preoperatively predict the Ki67 expression status and survival outcome in BCa.
Methods: We retrospectively collected 179 BCa patients (41 low Ki67 expression, 138 high Ki67 expression) with preoperative MRI. Patients were randomly allocated into training and validation sets based on a 7:3 ratio. Axial T2-weighted images (T2WI) and dynamic contrast-enhancement (DCE) were used for radiomics features extraction. The synthetic minority over-sampling technique (SMOTE) was used to balance the low Ki67 expression group. After feature filtering and feature selection, four radiomic signatures were developed in the training set and assessed in the validation set. The predictive performance of radiomics signatures was assessed by the area under the curve (AUC) and accuracy. The decision curve analyses and calibration curve and were used to investigate the clinical usefulness and calibration of the radiomics signature, respectively. Univariate and multivariate Cox regression analyses were used to obtain independent predictors for Ki67 expression status prediction. Kaplan-Meier test and Cox analyses were performed to investigate the prognostic value of radiomics-predicted Ki67 expression status.
Results: A 38-feature-based support vector machine model achieved the best predictive performance in the SMOTE-balanced training (AUC, 0.978; accuracy, 91.19%) and validation sets (AUC, 0.951; accuracy, 84.34%) with a good calibration performance and clinical usefulness. Multivariate logistic regression showed that radiomics signature was an independent predictor for Ki67 expression prediction. Immunohistochemistry-based high Ki67 expression and radiomics-predicted high Ki67 expression were significantly associated with poor disease-free survival (P=0.004, 0.026, respectively).
Conclusions: A radiomics signature could predict the Ki67 expression status and the prognoses of the BCa patients, thereby may aid in clinical decision-making.

Magnetic Resonance Imaging -based radiomics signature for preoperative prediction of Ki67 expression and outcome in bladder cancer

For 2021, our focus remains: turning new data into your daily practice. The programme will consist of state-of-the-art lectures, case discussions and panel debates. Our faculty is a group of international & multidisciplinary experts.

We will take you on our journey to Ghent and beyond. Prior the congress, we send out case challenges, where you can compare your opinion with peers & experts. These cases will be part of the Global Congresses on Prostate & Bladder Cancer programme during the congress days (12 & 13 October for PROSCA and 14 & 15 October for BLADDR). After these days, you will be updated on prostate & bladder cancer management.