Gafchromic EBT2 dosimetry via robust optimization

An ‘in house’ software was developed with MATLAB in order to perform optimized dose calculations in radiotherapy. The aim of this work is to demonstrate an improvement on the Multichannel method using robust optimization that deals with optimization problems where robustness is sought against uncertainty. An optimization framework was developed in order to compare remaining error from optimization process of robust methods against the conventional triple-channel method. The proposed robust method minimizes the dose difference over all channels compared to the original triple-channel method, mainly over clinical dose range. Even if a Gafchromic EBT2 film irradiated by composite IMRT fields is analyzed, more consistent values than the ones obtained by the triple-channel method are found and Newton Rings patterns are minimized. When robust methods are applied, the difference between blue and red channel doses was found to be very small, about 104 times less than obtained by triple-channel optimization. It is well known that one outlier may cause a large error in a least squares estimator. The blue channel correction of non-uniformities may have best performance when robust optimization is used. A variety of anomalies (artifacts, Newton rings and other disturbances) behave differently from natural Gaussian random noise such as variations of the thickness. The use of robust optimization methods might be more realistic since this approach uses fatter tail distributions as the Laplace and could mitigate the Newton’s Rings Pattern.