Split-Beam and Dual-Spiral Dual-Energy CT Protocol Optimization for Radiotherapy Treatment Planning
Abstract
Purpose
To determine optimal scan and reconstruction parameters of a novel dual-energy CT (DECT) scanner for radiotherapy treatment planning across different patient scenarios.
Methods
Single-energy CT (SECT) and DECT scans (split-beam, dual-spiral) were acquired of tissue surrogates in homogeneous phantoms (50mm-330mm), anthropomorphic phantoms (head, pelvis, spine), and five patients. CT number consistency and image quality were evaluated for varying scan and reconstruction settings, including tube voltage (80/Sn150kVp, 100/Sn150kVp, AuSn120kVp, AuSn140kVp), dose level via the vendor-specific IQ metric (170-400), pseudo-monoenergetic energy selection for proton dose calculation and for metal artifact reduction, beam hardening correction for bones (BHCb) and vendor metal artifact reduction (MAR). The respective influence on Hounsfield-Look-Up-Table (HLUT) definition for stopping-power estimation was assessed.
Results
For SECT, BHCb reduced cupping artifacts and improved CT number consistency across phantom sizes from 100HU to <15HU in bone, the same level observed in DECT without BHCb. A pseudo-monoenergetic CT energy of 71keV best agreed with SECT, enabling dose calculation with the same HLUT. DECT-derived HLUTs from phantom tissue surrogate scans agreed well across all four DECT modes, with ΔSPR<0.2 in bone for the 180-330mm phantoms, but ΔSPR≅0.5 in the 50mm phantom. Image noise at matched IQ levels differed between SECT and DECT, limiting the metric’s usefulness for cross-mode comparison. DECT reconstruction at 170keV substantially reduced metal artifacts for smaller metal objects (e.g., spinal screws), outperforming SECT+MAR, with additional MAR even degrading image quality. For larger metal objects (e.g., hip prostheses), results varied substantially depending on object size, with the best image quality with 170keV+MAR in the acetabular shell and 70keV+MAR in the femoral stem, making a case-by-case decision in clinical routine necessary.
Conclusion
DECT provides stable CT numbers across clinically relevant patient scenarios when appropriately configured. SECT-equivalent monoenergetic reconstructions allow reuse of existing clinical HLUTs, while metal artifact reduction requires implant-specific reconstruction choices.