Reference beam data may streamline commissioning but cannot replace local validation

Reference datasets can support linear accelerator beam modelling, but machine-specific measurements remain essential for small fields, complex delivery, and long-term verification.

KEY POINTS

  • This point-counterpoint examines whether manufacturer-provided reference beam datasets should replace machine-specific measurements during linear accelerator commissioning.
  • The supporting position argues that modern matched linear accelerators often reproduce percentage depth doses, profiles, and output factors within approximately 0.5%, allowing vetted reference data to reduce measurement errors and commissioning time.
  • The opposing position emphasizes that clinically important differences may persist in small fields, penumbrae, leakage, and highly modulated delivery. Reported small-field output-factor differences can reach approximately 5%-7% even between beam-matched systems.
  • Larger inter-institutional variation has been reported before standardized detector corrections, including differences above 20% for very small fields with high-definition multileaf collimators.
  • Both positions support local validation. The practical disagreement is whether reference data should form the primary beam model for well-matched modern systems or remain only a starting point and cross-check within full machine-specific commissioning.

CLINICAL TAKEAWAY

Manufacturer reference data can be a useful starting point for standardized modern linear accelerators, particularly for conventional fields and resource-limited commissioning workflows. They should not be treated as ground truth: local validation, small-field measurements, multileaf collimator modelling, end-to-end testing, periodic verification, and recommissioning when drift occurs remain necessary.

SOURCE

Medical Physics