ESTRO 2024 - Abstract Book

S4838

Physics - Quality assurance and auditing

ESTRO 2024

IAEA TRS398 (3) was implemented as an end-to-end simulation of the calibration procedure required to determine correction factors to apply to the ionisation chamber (reference quality) calibration factor required for the Mega Voltage x-ray qualities available on the VERT virtual Linacs. VERT provides a fully articulated virtual C-arm Linac (e.g. Truebeam or VersaHD) faithfully simulating the clinical counterparts. It has nominal hardcoded 6MV and 15MV flattened beam qualities, the calibration configuration of which are user defined. The default configuration reflects the UK standard (4): 1Gy per 100MU at 5cm depth in water, isocentrically, for a 10cm x 10cm field. All ionisation chamber factors are user configurable: the default quality correction factors (k QQo ) are taken from TRS398 and ion recombination, polarisation factors and absorbed dose to water (reference Cobalt-60) factor were taken from clinical examples. Temperature and pressure corrections are required to correct for the random environmental conditions simulated for each measurement session. The user/ trainee is able to work through the exact measurements required to create the Quality Index for the available qualities, and determine the polarity and ion recombination correction factors. These are determined using the formalisms within TRS398. Measurement conditions are set by the user using a virtual water phantom positioned under the linac. The depth of effective point of measurement of the chamber, field size and beam quality are chosen by the user and set in ‘room simulation view’. The Linac/beam parameters are set using simple menus or by integrated Linac hand pendants. The chamber bias voltage is set by the user. The workflow follows the template in the TRS document and is provided on a web-interface where measurements are displayed and recorded. Realistic charge measurements are simulated for the general depth, field size and SSD, using a simple dose calculation and assigned chamber factors. All measurements are subject to measurement uncertainty, the magnitude of which is estimated using a Gaussian distribution with user defined standard deviation. When an individual measurement is ‘recorded’ for a particular quantity (depth d for a general field size and FSD, 6MV or 15MV, Bias Voltage), it contributes to a running average of that quantity.

Results:

In a worked example, a virtual Farmer thimble chamber was used with a bias potential of -250V. A 6MV flattened beam was characterised as having a QI of 0.667 by measurement of a TPR 20/10 ratio using a 10cm x 10cm beam, resulting in a look up k QQo of 0.994. The Polarity correction factor was measured as 1.000 and the Ion Recombination was found to be 1.006 using a V1/V2 ratio of 5. Correcting an isocentric measurement at 5cm depth for a 10cm x 10cm 6MV beam, using a chamber with a (reference) Absorbed Dose to Water factor of 4.784 x10 7 Gy/C (ambient conditions: 22.9 o C/1010.5mbar), gave an output of 1.001 Gy for the virtual Linac.

Conclusion:

Where measurements are made incorrectly, errors propagate through the process resulting in incorrect factors and ultimately errant calibration factor/ dose measurements. This makes the simulation system a valuable experimental platform which broadens experience by ‘learning from the experience of making mistakes’, as opposed to a passive system that simply illustrates steps required for a required process. The comprehensive TRS398 workflow within the VERT simulation training system was found to be realistic and able to reproduce the expected calibrated output of the virtual machine. These VERT tools have been used on an ESTRO Dosimetry Audit course (5) and the UK NPL Primary Calibration Reference Dosimetry courses in 2023, a key usage being the exploration of common errors and mistakes. VERT may provide a valuable tool to aid Medical Physics