ESTRO 36 Abstract Book

S399 ESTRO 36 _______________________________________________________________________________________________

PO-0757 Variation of mean dose output from 204 UK linacs (Jan-June 2015) and its potential clinical impact. M. Bolt 1 , A. Nisbet 1 , C. Clark 2 , T. Chen 3 , R. Jena 4 1 St. Luke's Cancer Centre Royal Surrey County Hosp, Radiotherapy Physics, Guildford, United Kingdom 2 National Physical Laboratory, Radiation Dosimetry, Teddington, United Kingdom 3 University of Surrey, Chemical and Process Engineering, Guildford, United Kingdom 4 Addenbrooke's Hospital, Department of Oncology, Cambridge, United Kingdom Purpose or Objective Variation in dose delivered to patients directly impacts the effectiveness of radiotherapy treatments. The drift and daily fluctuations in linac beam calibration (output) is a contributing factor to the cumulative dose received by the patient. Knowledge of the variation in measured outputs on a national scale provides an insight into the uncertainties in dose delivery and its clinical impact. Material and Methods A request for 6MV output measurement data was sent to all UK radiotherapy centres. In total, data was provided for 204 linacs situated at 52 cancer centres across the UK. The data spans 6 months from January to June 2015, totalling almost 25,000 data points. Additional data collected includes: linac model, year of install, measurement equipment and recording method. The dose response parameter, gamma, is the percentage change in treatment response caused by a percentage change in dose. Gamma values of 2.3 and 5.2 (representative for Head and Neck cancers) were used to estimate the effect on TCP and NTCP respectively [1]. Results Based on the collated data, the UK linac outputs had a mean of -0.01% with one standard deviation of 0.88%. There was a wide variety of recording methods, with 8 centres having no form of electronic record for daily checks. Measurement data for both constancy devices and ionisation chambers was provided for 29 linacs. Of these, 8 (28%) had a discrepancy between measurement devices of greater than 0.5%, with 3 linacs (10%) having greater than 1%. The greatest variation in the mean output of an individual linac was -2.1%, with 90% of linacs having a mean output within 1%of the national mean. No significant variations were observed based on the age of the linac. The maximum range within a single centre for the mean output for each linac was 2.3% (min:-1.1%, max:+1.2%) – see figure. Assuming patients are treated on a single linac for their treatment duration, this indicates a variation in TCP of 5.3% and a variation in NTCP of 12% dependent on which linac they are treated on at that centre. Conclusion The data collection process indicates that many departments still rely heavily on paper QC records. The variation in treatment outcomes caused by dose variation alone indicates the importance of accurate QC. Output adjustment is one of the simplest ways of maintaining treatment consistency between individual patients, and its significance should not be forgotten with the introduction of more advanced techniques. This variation in dose should be considered when participating in clinical trials. This applies both to small scale local trials in which the technique used may determine the treatment linac, and therefore the dose delivered, as well as large multi- centre trials where the dose variation should be considered for the trial power calculations. [1] Bentzen, S.M et al. (2000), Eur J Cancer, Mar: 36(5): pp.615-620. PO-0758 Development of patient-specific phantoms for verification of SBRT planning using 3D printer C.S. Hong 1 , D. Oh 1 , S.G. Ju 1 , M. Kim 1 , B.Y. Koo 1 , H.C. Park 1 , D.H. Choi 1 , H. Pyo 1

tissue/tumor equivalent materials show greater PDD curve agreement than lung equivalent materials. While all PDDs showed expected curve shapes, the smallest field size (3x3 cm 2 ) showed a deviation of 11.9% and 4.6%, respectively for lung and soft tissue PDD curves. At B=0 and at 6.5cm in depth of the mini phantom, beam profiles were extracted from the film and were corrected and normalized to the TLDs. It is expected that beam profiles will show similar beam profile results with a 1.5T magnetic field.

Conclusion It was determined that the four testing materials were MR/CT compatible, showed expected PDD curves, and preliminary data show expected beam profiles with and without 1.5T on the MR-Linac.