Prediction of the Prognostic Factors of Breast Cancer Using MRI with a Continuous-Time Random-Walk Diffusion Model

Breast cancer is a heterogeneous tumor, and its treatment is customized determined by the pathologic and tumor molecular features. This study used a continuous-time random-walk (CTRW) diffusion model to characterize prognostic factors status of breast lesions. The parameter β showed significantly differences in ER and PR status and showed significant correlation with ER and PR expression, and the combination of Dm, α, and β produced the best performance in identifying ER and PR status. Our study demonstrated that the CTRW parameters as potential biomarkers were superior to conventional ADC in predicting prognostic factors of breast cancer.