Quantitative Elemental Analysis Using Whole Spectral Information (with GA and MLR Methods) of Proton Induced Prompt Gamma-Rays Simulated Using Geant4 Toolkit

Abstract

Purpose: Online determination of the elemental composition of tissues near the Bragg peak is a challenge in proton therapy related studies. In the present work, an analysis method based on the whole spectral information is presented for the quantitative determination of the elemental composition (weight %) of an irradiated target from its emitted Prompt Gamma (PG) spectrum. Materials and Methods: To address this issue, four test phantoms with different weights (%) of 12C, 16O, 20Ca, and 14N elements were considered. The simulated PG spectra were recorded using 3 × 3 inch NaI detectors. A library consisting of the spectra of single-element phantoms as well as the spectra of test-irradiated phantoms was produced for 30, 70, and 150 MeV incident protons using the Geant4 Monte Carlo toolkit. The elemental analysis was performed using the information of the whole spectrum by applying two methods, including the well-known Genetic Algorithm (GA) and Multiple Linear Regression (MLR). Results: The results show that the proposed method estimates the oxygen concentration accurately. Furthermore, the estimated weights of other elements, with both methods, agree well with nominal values in each test phantom, for the considered energies. Conclusion: The proposed quantitative elemental analysis of proton-bombarded phantoms using their induced PG spectrum is expected to be beneficial in treatment planning and treatment verification studies.

Keywords

• Proton Therapy
• Prompt Gamma-Ray Spectrum
• Whole Spectrum Analysis
• Genetic Algorithm
• Multiple Linear Regression