KEY POINTS
- Spectral CT extends conventional CT by separating energy-dependent attenuation and generating multiple inherently co-registered datasets from a single acquisition, including virtual monoenergetic imaging, relative electron density, effective atomic number and stopping-power ratio maps.
- During simulation, low-energy virtual monoenergetic images can improve iodine contrast and lesion conspicuity, while high-energy reconstructions can reduce beam-hardening and metal artifacts. This may improve target definition without requiring separate acquisitions.
- For photon RT, spectral CT-derived relative electron density can provide more physically direct input for dose calculation. Existing phantom and retrospective studies generally report electron-density deviations around 1% with clinically small dosimetric differences.
- The potential impact is greater for particle therapy, where direct estimation of stopping-power ratio may reduce uncertainty associated with conventional HU-to-SPR conversion and therefore improve range prediction. Clinical outcome benefit, however, has not yet been demonstrated.
- Spectral information could also support adaptive RT by combining longitudinal anatomy with quantitative changes in iodine concentration, effective atomic number and tissue composition. These parameters have shown associations with response in several tumor sites but lack validated treatment-adaptation thresholds.
- Photon-counting CT may further improve spatial resolution, noise characteristics and material discrimination, but most RT evidence remains phantom-based or early clinical, and routine treatment-planning integration is not established.
- Widespread implementation faces practical barriers: scanner and vendor dependence, inconsistent acquisition and reconstruction protocols, incomplete commercial TPS integration, lack of spectral-specific QA standards and limited cross-platform reproducibility.
- The review emphasizes that future progress should focus less on generating additional biomarkers and more on task-specific end-to-end validation showing that spectral information changes planning, adaptation or patient outcomes.
CLINICAL TAKEAWAY
Spectral CT is best viewed as an enabling quantitative extension of conventional CT, not a replacement for it. Its physical advantages are already compelling for applications such as particle range estimation and artifact reduction, but clinical adoption should depend on reproducible workflows, local QA and evidence that the additional information actually changes treatment decisions.