Work

Validation and optimization of dual-energy CT for accurate dose calculation in photon radiotherapy of brain metastases

Year: 2025

Type: article

Abstract: Purpose Dual-energy computed tomography (DECT) offers enhanced tissue contrast and the potential for direct electron density mapping, but its role in accurate dose calculation for radiotherapy remains underexplored. This study evaluates whether DECT-derived reconstructions can reproduce conventional single-energy CT (SECT)-based dose distributions for stereotactic brain radiotherapy while minimizing workflow disruption. Methods A Gammex Advanced Electron Density Phantom (Sun Nuclear, Melbourne, FL, USA) was scanned using a twin-spiral DECT protocol on a Somatom go.Open Pro CT scanner (Siemens Healthineers, Forchheim, Germany). Calibration curves were derived for virtual monoenergetic images (VMI) across 40–190 keV and for direct electron density (Rho) reconstructions. An optimal VMI energy (VMI opt ) was identified to match the SECT calibration. Retrospective dose calculations were performed on 24 stereotactic treatment plans for brain metastases, comparing standard SECT, VMI opt , and Rho datasets. Dose-volume histogram (DVH) metrics and gamma analyses assessed accuracy. Results An optimal VMI energy of 72 keV closely replicated the standard SECT calibration. VMI opt -based dose calculations showed minimal dosimetric deviations, with mean differences of 0.27%, 0.37%, and 0.41% for D 95% , D 98% , and D 99% , respectively. Rho-based reconstructions exhibited slightly larger variations, averaging 0.75%, 0.92%, and 1.02% for the same parameters. Gamma analysis demonstrated pass rates above 99.6% for VMI opt and above 99.1% for Rho using 1%/1 mm criteria, even in high-dose and steep-gradient regions. Conclusion Both 72 keV VMI and Rho-based DECT reconstructions enable accurate, clinically acceptable dose calculations for stereotactic radiotherapy of brain metastases. These findings support DECT's integration into clinical workflows, enhancing planning precision without increasing operational complexity.

Source: Physica Medica

DOI: 10.1016/j.ejmp.2025.105702