Generalizability of deep learning-based dose conversion model in proton beam therapy.

Key Points

• To evaluate the accuracy and generalizability of a deep learning-based dose conversion model in proton beam therapy.
• Developed a deep learning model for dose conversion in proton beam therapy.
• Tested the model on various tumor sites, including untrained ones.
• Assessed the model's adaptability to biases in disease data collection.
• The model achieved high accuracy for dose conversion across different tumor sites.
• Showed strong generalizability even for untrained sites and rare diseases.