Therapeutic radiation beam output and energy variations analysis across clinics, people, technology, and time

M. M. Zaini,¹ E. I. Marshall¹, K. M. Hedrick,¹ J. M. Fagerstrom,¹ D. Zaks,¹ T. M. Fitzgerald¹

¹Northwest Medical Physics Center, Lynnwood, WA 

IUPESM World Congress on Medical Physics and Biomedical Engineering, Singapore (2022).

Our group of about 50 clinical physicists has been engaged in measuring various beam characteristics of more than 50 therapeutic linear accelerators in more than 30 clinics manufactured by the main two radiation oncology machine manufacturers in North America over many decades. For the past decade, we have been using the same measurement setup for all our machines every month. These linac radiation output measurements employ a set of acrylic slabs, Farmer ion chambers, electrometers, and temp/pressure measuring gadgets. Such charge readings in acrylic phantoms are precisely linked to AAPM’s TG-51 formalism. A unique parameter, called, kacrylic, has been used for converting the raw readings of such radiation measurements to the actual machine output calibration values. Statistical analysis of this parameter, radiation output, and beam energy of our therapeutic linacs over time and among machine types by various physicists is presented.

The number of Varian and Elekta linacs included in this work are 42 and 9, respectively. The two linac types showed similar coefficient of variations (COV) for kacrylic for all radiation beams across different linacs at different clinics. For Varian machines, COV of kacrylic varied from 0.1% to 1.6% for 16 photon and electrons beams and that for Elekta machines ranged from 0.1% to 1.5% for 11 radiation beams. However, COV of kacrylic for single photons beams on individual linacs across time ranged from 0% to 1.8% for Varian machines, whereas that for Elekta linacs varied from 0% to 0.8%. This variation for individual electron beams of each linac spanned the range of 0% to 2.0% and 0% to 1.1% for Varian and Elekta linacs, respectively. Histograms of radiation beam energies and outputs for all the machines throughout the network are depicted. The impact of slight deviations in the measurement apparatus as well as observational disparities among different physicists are discussed.