Artificial intelligence for response prediction and personalisation in radiation oncology

Area of Science:

• Oncology
• Medical Physics
• Radiotherapy

Background:

• Radiotherapy personalization is crucial for optimizing treatment efficacy and minimizing side effects.
• Artificial intelligence (AI) offers advanced capabilities for analyzing complex patient data in radiotherapy.

Purpose of the Study:

• To describe three generations of AI systems for radiotherapy personalization.
• To discuss challenges in the clinical translation of AI for radiotherapy.

Main Methods:

• Review of AI systems for radiotherapy personalization.
• Categorization of AI systems into pretreatment data-based, response-driven, and dynamically optimized approaches.

Main Results:

• AI systems can personalize radiotherapy through data assessment and response prediction.
• Three distinct generations of AI in radiotherapy are identified: pretreatment-based, response-driven, and dynamic optimization.

Conclusions:

• AI holds significant potential to personalize radiotherapy by predicting treatment outcomes.
• Clinical translation of AI in radiotherapy faces challenges that require further investigation.