Artificial intelligence for radiotherapy dose prediction: A comprehensive review

Area of Science:

• Medical Physics
• Radiotherapy
• Artificial Intelligence

Background:

• Radiation treatment planning is crucial for patient outcomes.
• Deep learning offers automation potential in dose prediction.
• Convolutional neural networks are a key AI tool in this field.

Purpose of the Study:

• To critically analyze deep learning-based dose prediction methods in radiotherapy.
• To focus on the application of convolutional neural networks in this domain.
• To assess the potential of AI for improving radiation treatment planning.

Main Methods:

• Comprehensive literature search of Scopus, Medline, and Web of Science databases.
• Analysis of papers published between 2018 and 2024.
• Focus on deep learning techniques, specifically convolutional neural networks.

Main Results:

• Deep learning methods show significant promise for automating dose prediction in radiotherapy.
• Analysis provides insights into the effectiveness of current AI approaches.
• Convolutional neural networks are a primary focus for dose prediction.

Conclusions:

• Deep learning can significantly improve radiation treatment planning automation.
• Findings guide future research for AI integration in clinical workflows.
• Safe and effective clinical integration of AI in radiotherapy is feasible.