# Evaluation of a fast kV switching dual energy CT in deriving relative stopping power using tissue equivalent phantoms of various sizes

## TL;DR

This study evaluates a fast kV switching dual energy CT for accurately predicting relative stopping power in tissue-equivalent phantoms, which could be useful for proton therapy planning.

## Contribution

The study introduces a fast kV switching dual energy CT system and evaluates its accuracy in predicting relative stopping power for clinical applications.

## Key findings

- The largest differences in RSP were observed in low-density inserts with a maximum error of 6.3% in sinus material.
- The mean absolute percent error for RSP was 1.13% ± 1.17%, which is non-inferior to clinical range uncertainties.
- No significant difference was found between DECT-derived values using Body versus Head SFOV (p = 0.61).

## Abstract

Dual energy CT (DECT) has been demonstrated to improve relative stopping power (RSP) estimation with knowledge of the effective atomic number (Zeff) and relative electron density (RED). A fast kV switching DECT has been developed and it is necessary to quantify the accuracy of RSP prior to clinical use.

This study evaluates the accuracy of a fast kV switching DECT to calculate Zeff, RED, and RSP in tissue‐equivalent phantoms of various sizes.

Five different phantoms containing tissue‐mimicking inserts were scanned using the Body Scanning Field of View (SFOV) of a fast kV switching DECT. Head phantoms were rescanned with the Head SFOV. Data from the DECT was used to obtain voxel‐matched Zeff and RED volumetric images. From these data, a volumetric image of RSP was calculated for each voxel in the volume. Derived values were compared to their corresponding reference values.

The largest differences were seen in the low‐density inserts. The mean absolute percent error (MAPE) of Zeff and RED was 2.02% ± 2.48% and 1.29% ± 1.86% respectively. RSP MAPE was 1.13% ± 1.17% with the largest difference in sinus material at 6.3%. When comparing DECT‐derived values obtained using Body versus Head SFOV, a paired t‐test showed no significant difference (p = 0.61).

Ultra‐fast kV switching DECT can be used to predict RSP in phantom with an accuracy that is non‐inferior to the range uncertainties commonly used in the clinical environment and has the potential to be used for proton therapy planning.