Deep learning for interpretable end-to-end survival (E-E Surv) prediction in gastrointestinal cancer histopathology

Key Points

• Survival prediction accuracy significantly improves with interpretable deep learning models, indicating better clinical applicability.
• Performance metrics showcase over 80% accuracy rates for survival predictions in gastrointestinal cancer cases.
• Using deep learning techniques on histopathology images allows for end-to-end analysis of survival outcomes.
• Potential implications for clinical settings include enhanced patient management and tailored treatment strategies based on interpretable results.

Abstract