Three-dimensional dosimetry using multiple-energy delivery and a single-layer detector for quality assurance in proton pencil beam scanning

Abstract

Background

In Proton Therapy, the presence of implants along the beam path is known to potentially affect the dose distribution. The way such implants are managed in the planning process can vary in the different treatment planning systems (TPSs) and different centers. A specific validation procedure should be accomplished to verify the accuracy of TPS computation in these conditions and accept the applied process before treating patients.

Purpose

The aim of this study is to present a quality assurance (QA) tool in pencil beam scanning proton therapy by a method based on multiple-energy delivery and a single-layer two-dimensional detector and to apply it for verifying three-dimensional dose computation and correcting CT calibration in the presence of implants.

Methods

Multiple-energy delivery with a single-layer detector (MESL) acquisitions were performed for 80 energy layers (70-150MeV), composed of equally weighted pencil beam spots. MESL measures were acquired using a two-dimensional MatriXX-IBA detector. A transformation of the energy modulation to spatial modulation was obtained by using the power-law relationship of initial energy and range. The setup design involved a reference configuration, allowing to compensate for potential offsets, and the same configuration with an additional phantom to be measured. Both setups were imaged by a CT scanner, and the dose was computed by the TPS. The comparison of TPS-computed and MESL-measured data of the phantom was performed by producing a 2D map of range-error. For testing the procedure, plastic slabs and rods made of tissue equivalent materials (TEMs), with known water equivalent path length (WEPL), were used. Range error mapping was then applied to verify dose computation with a titanium cylinder and a titanium implant. Numerical procedures were obtained by modifying at the TPS the segmented volume, or the value in the CT calibration curve for the titanium objects. The optimal values were then determined by identifying the one that minimizes residual range error.

Results

The results of the consistency test on the plastic slabs and the TEM rods showed differences between measured and expected WEPL below 1%, confirming the reliability of the method and the energy-spatial transformation. In the titanium cylinder, the optimal volume and the point in the calibration curve (relative to the titanium saturated value), to be used for TPS simulation is about the real size of the cylinder and the tabulated stopping power value. However, the optimal value to be assigned to the CT calibration curve might depend on the type and shape of the object, as they were different for the cylinder and the implant with screws.

Conclusions

The availability of a QA tool, like the one presented, paves the way for systematic studies of all the parameters that impact computation accuracy, and the methods to improve the accuracy of TPS computation.