Multimodal Co-Attention Transformer for Survival Prediction in Gigapixel Whole Slide Images | Synapse