Accurate segmentation of the prostate and its substructures consist the most important component for reliable localization and characterization of prostate cancer. In this study a Spatial Attention Residual U-Net (Spatial ResU-Net) deep learning (DL) network is proposed for segmenting the transitional zone of the prostate, by leveraging the learning capacity of spatial attention modules and residual connections. Spatial attention modules efficiently extract features in intra-channel manner and boost the performance of encoder and decoder while residual connections facilitate the information flow within the different network's levels. The proposed model was compared against 8 state-of-the-art DL segmentation models demonstrating a superior performance. The improvement in terms of Sensitivity, Dice Score, Hausdorff distance and Average surface distance was at least 1%, 1%, 0.05 mm and 0.09 mm, respectively.