Groupwise Non-Rigid Registration on Multiparametric Abdominal DWI Acquisitions for Robust ADC Estimation: Comparison with Pairwise Approaches and Different Multimodal Metrics

Registration of diffusion weighted datasets remains a challenging task in the process of quantifying diffusion indexes. Respiratory and cardiac motion, as well as echo-planar characteristic geometric distortions, may greatly limit accuracy on parameter estimation, specially for the liver. This work proposes a methodology for the non-rigid registration of multiparametric abdominal diffusion weighted imaging by using different well-known metrics under the groupwise paradigm. A three-stage validation of the methodology is carried out on a computational diffusion phantom, a watery solution phantom and a set of voluntary patients. Diffusion estimation accuracy has been directly calculated on the computational phantom and indirectly by means of a residual analysis on the real data. On the other hand, effectiveness in distortion correction has been measured on the phantom. Results have shown statistical significant improvements compared to pairwise registration being able to cope with elastic deformations.