Rapid dynamic radial MRI via reference image enforced histogram constrained reconstruction

J Magn Reson. 2014 Mar:240:1-7. doi: 10.1016/j.jmr.2013.12.011. Epub 2014 Jan 9.

Abstract

Exploiting spatio-temporal redundancies in sub-Nyquist sampled dynamic MRI for the suppression of undersampling artifacts was shown to be of great success. However, temporally averaged and blurred structures in image space composite data poses the risk of false information in the reconstruction. Within this work we assess the possibility of employing the composite image histogram as a measure of undersampling artifacts and as basis of their suppression. The proposed algorithm utilizes a histogram, computed from a composite image within a dynamically acquired interleaved radial MRI measurement as reference to compensate for the impact of undersampling in temporally resolved data without the incorporation of temporal averaging. In addition an image space regularization utilizing a single frame low-resolution reconstruction is implemented to enforce overall contrast fidelity. The performance of the approach was evaluated on a simulated radial dynamic MRI acquisition and on two functional in vivo radial cardiac acquisitions. Results demonstrate that the algorithm maintained contrast properties, details and temporal resolution in the images, while effectively suppressing undersampling artifacts.

Keywords: Dynamic MRI; Histogram entropy; Iterative reconstruction; Non-Cartesian; Undersampling.

MeSH terms

  • Algorithms
  • Artifacts
  • Heart / anatomy & histology
  • Humans
  • Image Enhancement / methods
  • Image Processing, Computer-Assisted / methods*
  • Magnetic Resonance Imaging / methods*
  • Phantoms, Imaging