Use of multiparametric magnetic resonance imaging in prostate cancer: A review

Abstract

Prostate cancer, or PCa, is a prominent male malignancy. For men with prostate cancer, accurate staging is essential for planning treatment and determining the prognosis. The way prostate cancer is currently diagnosed has led to two types of problems: overdiagnosis, which results in overtreatment, and underdiagnosis, which leads to missed diagnoses. Multiparametric magnetic resonance imaging (mpMRI) could assist in reducing prostate cancer diagnosis errors. Automatic prostate cancer techniques often use deep learning or machine learning to identify the lesion or tumor. Even after using these methods, they are not accurate every time in detecting and identifying prostate tumors after giving multiple sequences of mpMRI as input. Due to the absence of a clinically established test dataset, the output of the automatic prostate cancer system is extremely hard to verify. With the help of Explainable Artificial Intelligence (XAI) and expert review, the results of automatic prostate cancer techniques can be verified.

Keywords

• Prostate cancer
• PI-RADS
• Gleason grade group (GGG)
• mpMRI
• CNN
• Deep learning