127 - Improved patient-specific optimization of the stopping power calibration for proton therapy planning using a single optimized proton radiography

Abstract

Purpose: We present an improved method to calculate patient-specific calibration curves to convert X-ray computed tomography (CT) Hounsfield Unit (HU) to relative stopping powers (RSP) for proton therapy treatment planning. We introduce also an optimization method to improve the spatial resolution and the water equivalent thickness (WET) accuracy of proton radiographies [1]. Materials/Methods: By optimizing the HU-RSP calibration curve, the difference between a proton radiographic image and a digitally reconstructed X-ray radiography (DRR) is minimized. The feasibility of this approach has previously been demonstrated [2,3]. This scenario assumes that all discrepancies between proton radiography and DRR originate from uncertainties in the HU-RSP curve. In reality, external factors cause imperfections in the proton radiography, such as misalignment compared to the DRR and unfaithful representation of geometric structures (“blurring”). We analyze these effects based on synthetic datasets of anthropomorphic phantoms and suggest an extended optimization scheme which explicitly accounts for these effects. Performance of the method is been tested for various simulated irradiation parameters. We also investigate the use of a special optimization method to enhance the spatial resolution and the WET accuracy of proton radiographies prior to the HU-RSP re-calibration. The optimization method is designed for imaging systems measuring only the residual range of protons without relying on tracker detectors to determine the beam trajectory before and after the target. The ultimate purpose of the optimization is to minimize uncertainties in the HU-RSP calibration curve. We therefore suggest and perform a thorough statistical treatment to quantify the accuracy of the optimized HU-RSP curve. Results: We demonstrate that without extending the optimization scheme, spatial blurring (equivalent to FWHM=3mm convolution) in the proton radiographies can cause up to 10% deviation between the optimized and the ground truth HU-RSP calibration curve. Instead, results obtained with our extended method reach 1% or better correspondence, as shown in Figure 1. We have further calculated gamma index maps for different acceptance levels. With DTA=0.5mm and RD=0.5%, a passing ratio of 100% is obtained with the extended method, while an optimization neglecting effects of spatial blurring only reach ~90%. Conclusions: Our contribution underlines the potential of a single optimized proton radiography to generate a patientspecific calibration curve and to improve dose delivery by optimizing the HU-RSP calibration curve as long as all sources of systematic incongruence are properly modeled.