Sci—Thurs PM: Planning—11: An Epid‐Based Monte Carlo Approach to In‐Vivo Dosimetry for Intensity‐Modulated Radiation Therapy Treatments

As intensity‐modulated‐radiation therapy (IMRT) becomes more prevalent, the importance of determining, and accounting for, treatment planning errors and patient setup errors, which may result in discrepancies between the calculated and actual delivereddose concurrently increases. Such errors include Multi‐Leaf Collimator (MLC) mis‐calibration, organ motion, the tongue‐and‐groove effect, etc. Precise Monte Carlo‐based modeling of radiation transport through all components of IMRT‐capable linacs overcomes some of these deficiencies, but the complexity of transport through the MLC is an impediment to clinically‐timely implementation. Also, not all delivery errors can be accounted for with greater care in MLC simulation e.g. differences in MLC calibration. However, use of the amorphous‐silicon detector (or EPID, for Electronic Portal Imaging Device), available on many linacs, provides a solution; appropriately deconvolving EPID‐captured beam images provides the particle fluence at the EPID. Back‐projected to MLC height, this array can be used to modulate an existing phase space file, scored at the same height, such that direct simulation of these particles through the MLC is not required. This method was compared against our centre's treatment planning system, based on Pencil Beam Convolution (PBC), for various MLC configurations. The new method is promising, in that it matched well with film measurements, resulting in approximately 90% of pixels having Gamma‐Dose (3mm Distance‐To‐Agreement, 3% Dose‐Difference criteria) less than 1, versus 72% to 84% for the PBC‐based dose distributions. By utilizing Cone Beam CT to account for any setup errors or physiological changes, we believe this method will provide for more accurate 3D in‐vivo dosimetry.