MRI Radiomics Predicts Progression-free Survival In Prostate Cancer

Objective:The aim of this study was to investigate the predictive value of biparametric magnetic resonance imaging(bp-MRI)based radiomics for progression-free survival(PFS)in prostate cancer(Pca).Methods:Retrospective analysis of 373 patients with PCa who underwent MRI examination and confirmed by puncture biopsy or surgical pathology at our hospital from January 2016 to December 2018,with 191 patients were finally enrolled,including 133 in the training group and 58 in the validation group.All patients underwent T2 WI and DWI serial scans.The region of interest(ROI)was outlined along the edge of the lesion on each patient’s T2-weightedimaging(T2WI),apparent diffusion coefficients(ADC)image using ITK-snap software.A total of 1037 features were extracted using Py Radiomics software,and the T2 WI sequence and ADC sequence were selected by one-way logistic regression and Gradient Boosting Decision Tree(GBDT).The T2 WI model,ADC model and combined T2WI-ADC model were developed.Clinical independent predictors were screened using Cox risk regression,a mixed radiomics and clinical model was built,and a nomogram was plotted to predict the risk of progression in PCa patients.The diagnostic efficacy of the model was assessed using receiver operating charactsteristic(ROC)curves,calibration curves and decision curves.Radiomics scores were calculated for each patient,patients were divided into high and low risk groups and Kaplan-Meier curves were applied to compare PFS between risk groups.Statistical analysis was performed using R software and P < 0.05 was considered statistically significant.Results:There were 35 patients who showed progression and 156 patients who did not show progression in this study.Among the radiomics models,the combined T2WI-ADC model had higher efficacy than the T2 WI model and ADC model alone,with an AUC of 0.904 in the training group and 0.870 in the validation group.the mixed model consisting of histological features and clinical data performed best in predicting PFS,with AUCs of 0.926 and 0.917 respectively.the survival curves showed that the high-risk group had PFS was lower than that of the low-risk group.Conclusion:The hybrid model constructed from radiomics and clinical data has shown excellent performance in predicting PFS in prostate cancer patients.The nomogram provides a non-invasive diagnostic tool for clinical risk stratification of patients.