Multiple regression method for pulmonary apparent diffusion coefficient measurement by hyperpolarized 3 He MRI

Purpose To develop and validate a new multiple regression technique for the separation of flip angle effect in pulmonary apparent diffusion coefficient (ADC) measurement. Materials and Methods Hyperpolarized 3 He MRI (HP 3 He MRI) ADC measurements were performed on phantom, pig, and human models. The diffusion‐sensitization sequence is modified from a standard gradient echo (GRE) sequence with a nonlinear progression in the bipolar gradient amplitude with each image. In the self‐diffusion phantom experiment, four images were acquired with base gradient factor b 0 = 0.15 second/cm 2 ; in the pig and human experiment, six images were acquired with base gradient factor b 0 = 1.4 second/cm 2 . Results The self‐diffusion coefficient measured in the phantom experiment was 1.98 ± 0.16 cm 2 /second. The measured uncertainty curve was consistent with the theoretically predicted curve. The measured in vivo ADC values (three coronal slices in the supine direction) were 0.20/0.16/0.13 cm 2 /second and 0.20/0.18/0.16 cm 2 /second for pig and human experiments, respectively. Conclusion With the introduction of a nonlinear progression in the diffusion‐sensitization gradients, the multiple regression technique is capable of separating the flip angle effect in ADC measurement. In addition, this technique can perform a rigorous measurement uncertainty analysis and provide the optimal scan parameters that yield best noise performance. J. Magn. Reson. Imaging 2007;25:982–991. © 2007 Wiley‐Liss, Inc.