Surformer: An interpretable pattern-perceptive survival transformer for cancer survival prediction from histopathology whole slide images

Keyphrases

• Transformer 100%
• Histological Slides 100%
• Cancer Survivability Prediction 100%
• Cross Attention Module 50%
• Ratio Factor 50%
• Local Features 33%
• Survival Prediction 33%
• Histological Pattern 33%
• High-resolution 16%
• Distinct Patterns 16%
• Loss Function 16%
• State-of-the-art Techniques 16%
• Existing State 16%
• Feature Enhancement 16%
• Low Risk 16%
• Decision-making Process 16%
• Image Data 16%
• Cellular Level 16%
• High Risk 16%
• Patient Survival 16%
• Auxiliary Information 16%
• Histopathology Images 16%
• Data-driven Analysis 16%
• Global Features 16%
• Self-attention Module 16%
• Computational Pathology 16%
• Whole Slide Image 16%
• Survival State 16%
• Multi-head Attention 16%
• Histopathological Features 16%
• TCGA Database 16%
• Cancer Prognosis 16%
• Feature Quantification 16%
• Factors Distribution 16%
• Interpretable Neural Networks 16%

Computer Science

• local feature 100%
• Interpretability 50%
• Decision-Making 50%
• Neural Network 50%
• Global Feature 50%
• Distinct Pattern 50%
• Auxiliary Information 50%