Computer-aided detection and diagnosis (CAD) systems have the potential to improve robustness and efficiency compared to traditional radiological reading of MRI in prostate cancer. Fully automated segmentation of the prostate is a crucial step of CAD. With the advent of the deep learning-based (DL) methods in medical imaging, series of networks have been developed to segment the prostate. Automated detection of poorly segmented cases would therefore be a useful supplement. Therefore, we proposed a quality control (QC) system to detect the cases that will result in poor prostate segmentation. The performance results shows that the proposed QC system is promising.