Encrypted login | home

Program Information

Implementation of a Correction-Based Output Model for a Compact Passively Scattered Proton Therapy System

no image available
S Ferguson

S Ferguson1*, S Ahmad1 , Y Chen1 , C Ferreira1 , M Islam1 , V Keeling2 , A Lau1 , H Jin1 , (1) University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, (2) CARTI, Inc., Little Rock, AR


SU-F-T-143 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To commission and investigate the accuracy of an output (cGy/MU) prediction model for a compact passively scattered proton therapy system.

Methods: A previously published output prediction model (Sahoo et al, Med Phys, 35, 5088-5097, 2008) was commissioned for our Mevion S250 proton therapy system. This model is a correction-based model that multiplies correction factors (d/MUwnc=ROFxSOBPFxRSFxSOBPOCFxOCRxFSFxISF). These factors accounted for changes in output due to options (12 large, 5 deep, and 7 small), modulation width M, range R, off-center, off-axis, field-size, and off-isocenter. In this study, the model was modified to ROFxSOBPFxRSFxOCRxFSFxISF_OCFxGACF by merging SOBPOCF and ISF for simplicity and introducing a gantry angle correction factor (GACF). To commission the model, outputs over 1,000 data points were taken at the time of the system commissioning. The output was predicted by interpolation (1D for SOBPF, FSF, and GACF; 2D for RSF and OCR) with inverse-square calculation (ISF_OCR). The outputs of 273 combinations of R and M covering total 24 options were measured to test the model. To minimize fluence perturbation, scattered dose from range compensator and patient was not considered. The percent differences between the predicted (P) and measured (M) outputs were calculated to test the prediction accuracy ([P-M]/Mx100%).

Results: GACF was required because of up to 3.5% output variation dependence on the gantry angle. A 2D interpolation was required for OCR because the dose distribution was not radially symmetric especially for the deep options. The average percent differences were -0.03±0.98% (mean±SD) and the differences of all the measurements fell within ±3%.

Conclusion: It is concluded that the model can be clinically used for the compact passively scattered proton therapy system. However, great care should be taken when the field-size is less than 5x5 cm² where a direct output measurement is required due to substantial output change by irregular block shape.

Contact Email: