Meeting Banner
Abstract #4115

Optimistaion of B-Value Distribution in Biexponential Modelling

Peter Gibbs1, Martin D. Pickles1, Lindsay W. Turnbull1

1University of Hull, Hull, East Yorkshire, United Kingdom

Bi-exponential fitting of DWI data to account for microperfusion effects is becoming increasingly prevalent both in the brain and other organs. Increased sampling of low b-values to adequately define the bi-exponential curve appears necessary. However, little work has attempted to assess the optimum b-value sampling strategy. Using synthetic data this study demonstrates that a highly non-linear sampling scheme is necessary for tissues with a rapid and large perfusive component and that high SNR data is required to appropriately define tissues with a small perfusive fraction.

Abstract PDF E-Poster Video

Keywords

accurate acquisition adapting added adequate adequately adoption aims annual appears assess attempt better biomedicine brain breast coefficient combination component consistently contribute contributions created curve decay decaying define desired determination determined developed difficult diffusion distribution east efforts equation evident exponential extrapolation feasible fitted fitting formula fraction future generated gradient gradients highest hull ignoring initial inputted investigated involves journal kidney kingdom linear literature liver loss made martin medicine meeting model mono necessary noise nominal occurred often optimal organ organs otherwise outperformed overestimated paramount perfusion peter pickles potentially power prevented proceedings produce protocol pseudo pure quantify radiology ranging rapid rapidly recent regime regimes relatively riches sampling significantly simplex simulated simulating spacing strategy synthetic system systems technically tissue tissues traditionally true types uncertainties varied various wherein widespread