Empirical Determination of Output Factors for Proton Therapy Fields Using a Uniform Scanning Proton Beam and 10cm Snout
E Aaron1*, C Cheng2, M Lamba3, (1) University of Cincinnati, Cincinnati, OH (2) Indiana University- School of Medicine, Bloomington, IN, (3) University of Cincinnati, Cincinnati, OHSU-E-T-180 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: he output factor of a proton field is affected by energy, width of spread-out Bragg Peak (SOBP), source to measurement point distance, shape and size of the aperture, and the thickness of the compensator. It is generally measured in a water phantom to determine MU needed for each treatment field. This is time consuming and labor intensive. Previous studies employed empirical fits to measured data and then applied correction factors to account for various parameters. In this study, we have developed an empirical model to determine the output factors for proton fields with a 10cm snout using a cubic equation.
Methods: Measured output factors with (OFclosed) and without (OFopen) a compensator for 693 fields delivered with the uniform scanning beam were analyzed. The measurements were made for various ranges, SOBP widths, air gaps, aperture shapes and sizes using a 10cm snout. 3D empirical equations for predicting OFclosed and OFopen were determined by fitting the data to several 3D curves and evaluating each fit.
Results: The proposed model uses a simplified cubic fit. The distribution of closed measured OF vs predicted OF for this model has a mean of 1.000 and 0.013 the standard deviation and agrees within 3% of measured data 88.9% of all fields. The current model being used has a mean of 0.994 and a standard deviation of 0.009 and agrees within 3% of measured data 98.2% of all fields.
Conclusion: We have shown that a simplified cubic fitting of the measured data allows for calculation of output factor with a single equation of the form OF=f(Inverse Square Law Correction, SOBP, Range). This empirical equation may be used to double check the OFclosed obtained with the separate model that is currently used. The model will significantly reduce the QA time required for measurements