Analysis of time resolution in dynamic computed tomography for perfusion studies

Image quality In dynamic tomographic imaging can be improved by taking into account the dynamic nature of the object directly in the reconstruction process. In this article we present a theoretical analysis of the time resolution in dynamic CT. Data sets acquired over several full-turns are interpreted as samples of a continuous signal. Consequently data at any time instant can be achieved by interpolation techniques. A n-th order B-spline interpolation approach is proposed. In order to be prepared for motion correction techniques, interpolation is performed in image space in the partial block backprojection framework. In simulation studies the interpolation method is compared with a linear regression technique and a conventional short-scan reconstruction algorithm. The interpolation approach achieves best time resolution at similar noise properties than the conventional short-scan reconstruction. The linear regression algorithm shows significant denoising with detriment to time resolution