Rethinking Integration of Prediction and Planning in Deep Learning-Based Automated Driving Systems: A Review