Integrating Temporal and Spatial Learning for Prostate Cancer Prediction using a Pruned 1D-CNN BiLSTM Hybrid Model