Artificial general intelligence for radiation oncology.

Area of Science:

• Oncology
• Artificial Intelligence
• Medical Imaging

Background:

• Artificial general intelligence (AGI) is rapidly advancing, with large language models (LLMs) and large vision models (LVMs) showing significant potential.
• Radiation oncology faces increasing demands for efficiency and precision in patient care.

Purpose of the Study:

• To explore the comprehensive applications of AGI across the full spectrum of radiation oncology.
• To highlight the synergistic potential of combining LLMs and LVMs for enhanced clinical pattern recognition.
• To provide an overview of AGI's transformative role in personalized, data-driven radiation therapy.

Main Methods:

• Review of current and emerging applications of AGI in radiation oncology.
• Analysis of large language models (LLMs) for text processing and large vision models (LVMs) for imaging data analysis.
• Exploration of multimodal models integrating vision and language data for clinical insights.

Main Results:

• AGI, including LLMs (e.g., GPT-4, PaLM 2) and LVMs (e.g., SAM), can process extensive text and imaging data.
• Applications span initial consultation, simulation, treatment planning, delivery, verification, and follow-up.
• Multimodal models offer novel ways to identify complex clinical patterns from integrated data.

Conclusions:

• AGI promises to significantly enhance the efficiency and precision of radiation therapy.
• The integration of AGI facilitates a shift towards data-driven, personalized radiation oncology.
• AGI's ability to leverage multimodal clinical data at scale is key to elevating patient care standards.