Improving single-event proton CT by removing nuclear interaction events within the energy/range detector

• Published in: Physics in medicine & biology 2019-08, Vol.64 (15), p.15NT01-15NT01
• Main Authors: Volz, Lennart, Piersimoni, Pierluigi, Johnson, Robert P, Bashkirov, Vladimir A, Schulte, Reinhard W, Seco, Joao
• Format: Article
• Language: English
• Subjects: energy telescope; Monte Carlo Method; nuclear interactions; proton computed tomography; proton imaging; Protons; single-event imaging; stopping power; Tomography - instrumentation; Tomography - methods
• ISSN: 0031-9155; 1361-6560; 1361-6560
• DOI: 10.1088/1361-6560/ab2671

Data filtering is crucial for accurate relative stopping power (RSP) reconstruction in proton CT (pCT). In this work, we assess different filters and their performance for the US pCT collaboration prototype pCT system in Monte Carlo (MC) simulations. The potential of using the recently proposed -E filter for removing nuclear interactions that occurred in the energy/range detector of the pCT system is investigated. Full pCT scans were acquired with the TOPAS MC simulated version of the prototype scanner that comprises two tracking detectors and a five stage energy/range detector. An ideal water cylinder and a water cylinder with five tissue inserts were investigated. Before image reconstruction, a WEPL filter was applied as the only filter, or in addition to filters acting on the energy deposit in each of the energy detector stages, as done currently with the prototype. The potential of the -E filter that was recently proposed for helium imaging was assessed. The results were compared to simulations for which nuclear interactions were disabled representing ground truth. The WEPL filter alone was not sufficient to filter out all nuclear interaction events and systematic fluctuations in the form of ring artifacts were present in the pCT reconstructed images. Applying energy filters currently used with the device prior to the WEPL filter only slightly improved the image quality. A WEPL filter improved the mean RSP accuracy, but could not fully remove the systematic fluctuations. The -E filter in addition to the current reconstruction procedure efficiently removed the systematic fluctuations and the achieved RSP accuracy closely matched the simulation without nuclear interactions. This study demonstrates the dependence of the accuracy of the usual WEPL filter on uncertainties arising within the energy detector. By enabling to remove such uncertainties, the -E method proved to yield some potential for improving the accuracy of pCT.