Impact of deep learning reconstructions on image quality and liver lesion detectability in dual‐energy CT: An anthropomorphic phantom study

Compared to the routinely used level of iterative reconstruction, DLIR reduces noise without consequential noise texture modification, and may improve the detectability of hypervascular liver lesions while enabling the use of lower energy virtual monoenergetic images. The optimal energy level and DLIR level may depend on the lesion enhancement.