• Gu-Mu-Yang Zhang, Yu-Qi Han, Jing-Wei Wei, Ya-Fei Qi, Dong-Sheng Gu, Jing Lei, Wei-Gang Yan, Yu Xiao, Hua-Dan Xue, Feng Feng, Hao Sun, Zheng-Yu Jin, and Jie Tian.
• Department of Radiology, Peking Union Medical College Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences, Beijing, China.
• J Magn Reson Imaging. 2020 Oct 1; 52 (4): 1239-1248.

BackgroundBiopsy Gleason score (GS) is crucial for prostate cancer (PCa) treatment decision-making. Upgrading in GS from biopsy to radical prostatectomy (RP) puts a proportion of patients at risk of undertreatment.PurposeTo develop and validate a radiomics model based on multiparametric magnetic resonance imaging (mp-MRI) to predict PCa upgrading.Study TypeRetrospective, radiomics.PopulationA total of 166 RP-confirmed PCa patients (training cohort, n = 116; validation cohort, n = 50) were included.Field Strength/Sequence3.0T/T2 -weighted (T2 W), apparent diffusion coefficient (ADC), and dynamic contrast enhancement (DCE) sequences.AssessmentPI-RADSv2 score for each tumor was recorded. Radiomic features were extracted from T2 W, ADC, and DCE sequences and Mutual Information Maximization criterion was used to identify the optimal features on each sequence. Multivariate logistic regression analysis was used to develop predictive models and a radiomics nomogram and their performance was evaluated.Statistical TestsStudent's t or chi-square were used to assess the differences in clinicopathologic data between the training and validation cohorts. Receiver operating characteristic (ROC) curve analysis was performed and the area under the curve (AUC) was calculated.ResultsIn PI-RADSv2 assessment, 67 lesions scored 5, 70 lesions scored 4, and 29 lesions scored 3. For each sequence, 4404 features were extracted and the top 20 best features were selected. The radiomics model incorporating signatures from the three sequences achieved better performance than any single sequence (AUC: radiomics model 0.868, T2 W 0.700, ADC 0.759, DCE 0.726). The combined mode incorporating radiomics signature, clinical stage, and time from biopsy to RP outperformed the clinical model and radiomics model (AUC: combined model 0.910, clinical model 0.646, radiomics model 0.868). The nomogram showed good performance (AUC 0.910) and calibration (P-values: training cohort 0.624, validation cohort 0.294).Data ConclusionRadiomics based on mp-MRI has potential to predict upgrading of PCa from biopsy to RP.Level Of Evidence3 TECHNICAL EFFICACY: Stage 5 J. Magn. Reson. Imaging 2020;52:1239-1248.© 2020 International Society for Magnetic Resonance in Medicine.

14 keywords...

hide…