Evaluating photon-counting computed tomography for quantitative material characteristics and material differentiation in radiotherapy

Abstract Objective: 
Photon-counting computed tomography (PCCT) counts the individual photons and measures their energy, which allows for energy binning and thereby multi-energy CT imaging. It is expected that quantitative data can be accurately extracted from the images and enable accurate material separation, yet its potential in radiotherapy is mostly unexplored. In this study, PCCT was assessed by evaluating estimation accuracies for relative electron density (RED), effective atomic number (Zeff), and proton stopping-power ratio (SPR), as well as the potential for material differentiation.

Approach: 
PCCT images of a Gammex Advanced Electron Density phantom (Sun Nuclear) with tissue-equivalent materials were acquired in a small and large phantom setup on a NAEOTOM Alpha PCCT scanner (Siemens Healthineers). The scans were performed at 120 and 140 kVp, and virtual monoenergetic images (VMIs) were generated. These VMIs were used to estimate RED, Zeff, and SPR based on two calibration methods for each of the two phantom sizes. These results were compared to findings obtained based on dual-energy CT (DECT) scans acquired on a SOMATOM Confidence scanner (Siemens Healthineers) at 80 and 140 kVp, by using the low and high energy pair and VMIs. Calibration accuracy was quantified by the root-mean-squared error (RMSE). Additional, material differentiation was assessed for both tissue-equivalent and calcium/iodine inserts by creating [RED/Zeff]-space plots.

Main results: 
There was minimal differences between the two PCCT x-ray spectra, with SPR errors below 0.8% for the large phantom and 0.7% for the small phantom, which was comparable to DECT using VMIs. Material differentiation showed similar results for DECT and PCCT using VMIs, and resulted in less Zeff spread, than the regular DECT kVp pair, possibly due to denoising.

Significance: 
This study showed the ability of PCCT to retrieve material characteristics and possibility for material differentiation between tissue-equivalent material and calcium/iodine, with results comparable to DECT.