Author: Ross I. Berbeco, Vera Birrer, Raphael Bruegger, Pablo Corral Arroyo, Roshanak Etemadpour, Dianne M. Ferguson, Rony Fueglistaller, Thomas C. Harris, Yue-Houng Hu, Matthew W. Jacobson, Mathias Lehmann, Marios Myronakis 👨🔬
Affiliation: Brigham and Women's Hospital, Harvard Medial School, Dana-Farber Cancer Institute, Department of Radiation Oncology, Dana Farber/Brigham and Women's Cancer Center, Department of Radiation Oncology, Brigham and Women’s Hospital, Dana-Farber Cancer Institute, Harvard Medical School, Brigham and Women's Hospital, Varian Imaging Laboratory, Dana-Farber Cancer Institute 🌍
Purpose: Dual-energy imaging offers benefits, including metal artifact reductions (MAR) and improved contrast, that enhance visualization. Currently, dual-energy imaging is largely limited to CT scanners. This study investigates the clinical utility of spectral imaging on a linac’s onboard imaging system.
Methods: A novel dual-layer kV detector was installed on a clinical linac. Patient data was obtained under IRB supervision. Clinical CBCT projections were collected from five head and neck and two pelvis patients to date. CBCT energies ranged from 100 to 140kVp. Decompositions of the scatter- and log-corrected projections into two basis materials (adipose and titanium) were made using an iterative least squares solution to the poly-energetic Beer’s Law equations, regularized by Huber roughness penalties. The decomposed projections were then FDK reconstructed to obtain material basis volumes, which in turn were combined to form virtual monoenergetic images (VMIs). 120keV reconstructions were created for MAR and 40keV reconstructions for contrast.
Change in uniformity (pixel standard deviation) was used to assess MAR changes. Regions affected by metal such as dental work, as well as unaffected regions, were segmented identically on standard CBCTs and 120keV scans. Soft tissue contrast (head and neck: adipose vs. muscle; pelvis: prostate vs. adipose) was calculated via segmented regions on standard CBCTs and 40keV scans.
Results: Uniformity increased by an average of 41.4% (metal artifact regions) and 12.1% (clear regions). Head and neck soft tissue contrast improved on average by 1.6x, with no significant reduction in CNR. Pelvic (Prostate-adipose) contrast improved by 2.7x, with a reduction in CNR of 26.4%.
Conclusion: Dual-energy imaging was successfully added to linac imaging using a prototype dual-layer kV detector. MAR and contrast can be enhanced with the detector. Uniformity may have increased in clear regions due to an effectively higher energy reducing noise. Future work will investigate other dual-energy capabilities.