S2046
Physics - Image acquisition and processing
ESTRO 2026
evaluates the accuracy of PCCT in material differentiation compared to Dual-energy CT using an Energy Integrating Detector (EID). Material/Methods: A Gammex Model 467 cheese phantom with eleven inserts was scanned with a Siemens NAEOTOM Alpha PCCT scanner at 120 kVp and 140 kVp. The material composition of the phantom was obtained from previous literature. Six virtual monoenergetic images (VMIs) were reconstructed, with three energies below or equal to 80 keV and three above or equal to 120 keV. Following the approach by [1], for each VMI pairs, the ρ e, Zeff, and SPR under a 250 MeV proton beam were calculated. The same protocol was applied to dual-energy (80/Sn140 kVp) EID scans on a Siemens SOMATOM Drive CT. Material differentiation performance was quantified by residuals of individual inserts and root-mean-square error (RMSE) of the estimated physical quantities, including all inserts. Results: PCCT and EID both demonstrated similar RMSE for material-specific parameters, as summarized in Table 1. Measured ρ e has a lower RMSE compared with that of Zeff. It may due to a linear relation of CT number with ρ e and a non-linear relation with Zeff. A Mayneord exponent was involved in the estimation. Optimal VMI pairs for accurate SPR prediction were 60/120 keV (120 kVp), 40/120 keV (140 kVp), and 80/120 keV (EID 80/Sn140). 140 kVp-derived images generally yielded lower RMSEs than those of 120 kVp. As illustrated in Fig.1, the largest discrepancies in Zeff were observed in lung-equivalent inserts, attributed to volume heterogeneity. Lower ρ e will lead to a larger discrepancy of Zeff between scans. Nevertheless, PCCT-140kVp scan still demonstrated a better material differentiation from Fig.1. This trend may attribute to higher penetration and less noise by PCCT. Bone- tissues were all successfully modeled for all 3 types of scans. Overall, PCCT demonstrated slightly lower SPR residuals than EID. Outliers likely come from limitations in fitting accuracy for specific materials.
Conclusion: PCCT provides material differentiation capability comparable to, and sometimes better than, dual- energy EID CT. The improvement in SPR estimation is modest and appears dependent on calibration methods. Reduced electronic noise in PCCT may explain its improved performance for some
parameters. References:
1. Hünemohr N., et al, 2013 Experimental verification of ion stopping power prediction from dual energy CT data in tissue surrogates. Phys. Med. Biol. 59 8 Keywords: Photo-counting CT, material differentation, SPR Poster Discussion 833 First experience of Photon-Counting Computed Tomography for lung tumor delineation: assessment of optimal reconstruction parameters Thiele Kroes-Kobus 1 , Nicolas Martz 2,3 , Panagiotis Balermpas 2 , Jens von der Grün 2 , Joris B.W. Elbers 1 , Pim J.J. Damen 1 , Marieke van Zwienen 1 , Thomas Frauenfelder 4 , Lotte Wilke 2 , Serena Psoroulas 2 , Stephanie Tanadini-Lang 2 , Linda Rossi 1 , Mischa Hoogeman 1 1 Department of Radiotherapy, Erasmus MC Cancer Institute, Rotterdam, Netherlands. 2 Department of Radiation Oncology, University Hospital Zürich, Zürich, Switzerland. 3 Department of Radiation Oncology, Institut de Cancérologie de Lorraine, Vandœuvre-Lès- Nancy, France. 4 Institute for Diagnostic and Interventional Radiology, University Hospital Zürich, Zürich, Switzerland Purpose/Objective: Photon-counting computed tomography (PCCT) enables acquisition of intrinsic spectral data that can be used to generate virtual mono-energetic images (VMI) and iodine maps. These features might improve tumor delineation in radiotherapy. However, the surplus of generated data makes the optimal choice of reconstructions challenging and current software
Made with FlippingBook - Share PDF online