Multi-Day Flash and Conventional Dosimetry on a Clinical Truebeam Linac: Film, Toroid Charge, and Grid-Voltage Stability Analysis
Abstract
Purpose
To characterize the reproducibility and dosimetric stability of a 5-day fractionated preclinical irradiation protocol delivered on a clinical Varian TrueBeam linac operated under FLASH and conventional (CONV) settings.
Methods
We assessed day-to-day dose consistency, accuracy of toroid charge as a surrogate for dose, and the impact of grid-voltage (grid-V) tuning on linearity and reproducibility. Daily fractions of 4 Gy and 5 Gy were delivered in both modes, and the machine was returned to standard clinical configuration between sessions. Toroid charge, pulse count, and grid-V were recorded for each irradiation and converted to derived dose via a daily film calibration. Correlations between film dose, toroid charge, and grid-V-based predictions were analyzed, with emphasis on the effect of intra-session grid-V changes.
Results
Dose delivery remained stable across all groups (CONV 4±0.1 Gy, CONV 5±0.2 Gy, FLASH 4±0.1Gy, FLASH 5±0.1Gy) despite daily resets to clinical mode. Toroid-derived dose showed strong accordance with film dose (overall R² = 0.93). Experimentally, dose was varied using the grid-V and number of pulses. Voltage adjustments produced a transient effect: an initial increase in output followed by a gradual drift downward over several minutes until stabilization. We accounted for this by delaying beam-release after grid-V changes and using the averaged toroid-dose calibration across multiple days. Days without intra-session grid-V changes showed the tightest agreement between derived and film-measured dose (<3% variation), whereas sessions involving grid-V and other adjustments exhibited greater scatter (<10%).
Conclusion
Both FLASH and CONV fractions delivered on a clinical TrueBeam showed excellent multi-day reproducibility. Toroid charge functions as a reliable real-time surrogate for delivered dose especially when grid-voltage is held constant, as intra-session grid voltage adjustments induce temporary non-linear machine output. This work supports the long-term objective of developing a clinical FLASH dosimetry kit with validated, reliable real-time monitoring capabilities.