Prediction of Mutations and Outcome in Gastrointestinal Stromal Tumors with Deep Learning: A Multicenter, Multinational Study

Key Points

• Deep learning predicts treatment sensitivity and molecular alterations in gastrointestinal stromal tumors, enhancing precision medicine.
• The study demonstrates that deep learning models perform comparably to established risk scores for evaluating recurrence-free survival.
• Analysis conducted across multiple international cohorts showcases the utility of whole slide images in predicting patient outcomes.
• The findings suggest that integrating deep learning could lead to improved multimodal predictors for gastrointestinal stromal tumors.

Abstract