[Development of a new position-recognition system for robotic radiosurgery systems using machine vision]

Area of Science:

• Medical physics
• Radiotherapy technology
• Image-guided interventions

Context:

• CyberKnife radiotherapy utilizes radiography for patient setup but lacks real-time intra-irradiation monitoring.
• Current registration methods are performed pre-irradiation, limiting adjustments during treatment.
• Patient movement during irradiation cannot be detected, potentially compromising treatment accuracy.

Purpose:

• To introduce and evaluate a novel, radiation-free registration system called "Machine Vision" for radiotherapy.
• To enable real-time patient registration during irradiation, improving upon CyberKnife's limitations.
• To assess the precision and suitability of "Machine Vision" as an alternative to current registration techniques.

Summary:

• The "Machine Vision" system uses three charge-coupled device cameras for automatic, radiation-free registration.
• It captures detailed anatomical features (fetal fissure, external ears) for precise patient positioning.
• The system demonstrated suitability for radiotherapy with displacement less than 1.0 mm, offering real-time capabilities.

Impact:

• Provides a safer and more precise radiotherapy treatment modality.
• Eliminates additional radiation exposure for patient registration.
• Has potential for integration with existing systems like CyberKnife for enhanced clinical applications.