In this work, we aim at a GAN-based method which can concurrently meet these three requirements and synthesize high quality mp-MRI images of CS PCa to meaningfully augment mp-MRI data of CS PCa for improving the performance of CNN-based PCa detection and classification. In the following, we start with a survey of related work and summarize their limitations.Prior work: If we refer different modalities of MRI as different domains, mp-MRI data synthesis can then be more broadly formulated as a multidomain or crossdomain data synthesis problem, which has been widely studied in recent years for synthesizing both natural and medical images. Existing multi-/cross- domain data synthesis methods can be categorized into three major classes: (1) cross-domain image translation which, given a real image sampled from one domain (e.g., a T1 image), synthesizes its counterpart in another domain (e.g., a T2 image); (2) direct multi- domain image synthesis which generates two images of different domains with a constraint on the paired relationship based on a common low-dimensional vector; (3) sequential multi-domain image synthesis, which first generates images in one domain based on low-dimensional vectors, followed by cross-domain image translation that maps them to their counterparts in another domain.