Open AccessOptimal virtual monoenergetic image in “TwinBeam” dual‐energy CT for organs‐at‐risk delineation based on contrast‐noise‐ratio in head‐and‐neck radiotherapy
Author(s) -
Wang Tonghe,
Ghavidel Beth Bradshaw,
Beitler Jonathan J.,
Tang Xiangyang,
Lei Yang,
Curran Walter J.,
Liu Tian,
Yang Xiaofeng
Publication year - 2019
Publication title -
journal of applied clinical medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.83
H-Index - 48
ISSN - 1526-9914
DOI - 10.1002/acm2.12539
Subject(s) - hounsfield scale , monochromatic color , nuclear medicine , image noise , physics , computer science , computed tomography , optics , medicine , radiology , image (mathematics) , artificial intelligence
Purpose Dual‐energy computed tomography ( DECT ) using TwinBeam CT ( TBCT ) is a new option for radiation oncology simulators. TBCT scanning provides virtual monoenergetic images which are attractive in treatment planning since lower energies offer better contrast for soft tissues, and higher energies reduce noise. A protocol is needed to achieve optimal performance of this feature. In this study, we investigated the TBCT scan schema with the head‐and‐neck radiotherapy workflow at our clinic and selected the optimal energy with best contrast‐noise‐ratio ( CNR ) in organs‐at‐risks ( OAR s) delineation for head‐and‐neck treatment planning. Methods and materials We synthesized monochromatic images from 40 keV to 190 keV at 5 keV increments from data acquired by TBCT . We collected the Hounsfield unit ( HU ) numbers of OAR s (brainstem, mandible, spinal cord, and parotid glands), the HU numbers of marginal regions outside OAR s, and the noise levels for each monochromatic image. We then calculated the CNR for the different OAR s at each energy level to generate a serial of spectral curves for each OAR . Based on these spectral curves of CNR , the mono‐energy corresponding to the max CNR was identified for each OAR of each patient. Results Computed tomography scans of ten patients by TBCT were used to test the optimal monoenergetic image for the CNR of OAR . Based on the maximized CNR , the optimal energy values were 78.5 ± 5.3 keV for the brainstem, 78.0 ± 4.2 keV for the mandible, 78.5 ± 5.7 keV for the parotid glands, and 78.5 ± 5.3 keV for the spinal cord. Overall, the optimal energy for the maximum CNR of these OAR s in head‐and‐neck cancer patients was 80 keV. Conclusion We have proposed a clinically feasible protocol that selects the optimal energy level of the virtual monoenergetic image in TBCT for OAR delineation based on the CNR in head‐and‐neck OAR . This protocol can be applied in TBCT simulation.