Alexander L. Sukstanskii1, Jim D. Quirk1, Mark S. Conradi2, Dmitriy A. Yablonskiy1,2
1Mallinckrodt Institute of Radiology, Washington University, St. Louis, MO, USA; 2Physics, Washington University, St. Louis, MO, USA
Lung morphometry with hyperpolarized 3He diffusion MRI allows estimation of lung microstructural parameters. Herein Bayesian analysis is used to analyze how the accuracy of this technique depends on data sampling strategy, generalizing the previous theory of equidistant sampling. We demonstrate that the optimum strategy requires three samples with b-values of 0, 2 and 8 s/cm2. This 3b-value sequence results in the relative errors in the parameter estimate smaller than that can be achieved by means of a longer 6b-value approach with equidistant sampling. This reduced imaging time can be harnessed to increase the number of acquired slices.