Purpose: To develop a new design of an x-ray beam shaper for helical computed tomography (CT) that increases the dose utilization.
Methods: For typical reconstruction algorithms in helical CT, different data are utilized with different weights during back-projection. In particular, data of the outer detector rows, i.e., data acquired at larger cone-angles, are used with smaller weights than data from the central detector rows. Given this spatial variation of the back-projection weights, a beam shaper is designed that creates a spatial variation of the noise variance across the detector such that the used back-projection weights are the statistically optimal weights. The effect of the beam shaper on the reconstructed images are studied using simulated data and analytical as well as iterative reconstruction algorithms.
Results: For a particular analytical reconstruction algorithm, we obtain an average reduction of the noise by 12% within the object. In combination with iterative reconstruction, the beam shaper creates an insensitivity to patient motion without introducing any heuristic data weighting.
Conclusions: The demonstrated noise reduction by 12% is equivalent to a possible dose saving of 25%. This dose saving can be achieved by a relatively minor hardware change in the CT system and it does not require any changes to the reconstruction algorithm.