Machine Learning in Radiation Oncology: Opportunities, Requirements, and Needs

Area of Science:

• Radiation Oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Machine learning (ML) offers significant potential to advance radiation oncology.
• Current applications of ML in radiotherapy are emerging but not yet widespread.
• A comprehensive workflow analysis is needed to identify ML integration opportunities.

Purpose of the Study:

• To analyze the radiotherapy workflow from a data-centric perspective.
• To identify specific areas within radiotherapy where ML can improve quality and efficiency.
• To guide clinicians and researchers on prioritizing ML development and implementation.

Main Methods:

• Workflow analysis of the radiotherapy process.
• Literature review of existing ML applications in radiation oncology.
• Identification of high-impact areas for future ML investment.

Main Results:

• Specific stages in the radiotherapy workflow amenable to ML-driven improvements were identified.
• Existing ML applications and their impact were cataloged.
• Key areas requiring further research and resource allocation were pinpointed.

Conclusions:

• A data-centric, ML-informed approach can significantly enhance radiotherapy.
• Targeted investment in specific ML applications is crucial for advancing patient care.
• This work provides a roadmap for collaborative efforts between clinicians and researchers.