Abstract
Magnetic resonance imaging (MRI) is an important modality with no radiation exposure, used daily for clinical diagnosis. MRI is originally developed based on discovery the difference of relaxation time of water protons between tumor and normal tissues. Although MRI has rapidly spread clinically, the origin of the difference has not been resolved. Interpretation of signal intensity of MR image is commonly based on three water state model, that is, bulk, structural and bound water. But our recent experiment with nuclear magnetic resonance (NMR) on frog skeletal muscle revealed that tissue water is distinctly classified into 5 groups, whose characteristic relaxation time is T2 > 0.4 s, T2 around 0.15 s, 0.03 s < T2 < 0.06 s, T2 < 0.03 s and the fifth groups of T2 shorter than the observable value with our pulse sequence. The water component with T2 > 0.4 s was revealed as extracellular water, and the other four components were distinctly separated each other by their responses for perturbations of osmotic pressure. Based on our result obtained with frog muscle, we tried to reconstruct signal intensities in every pixel of human brain MR image as summation of signals from those five water components. Reconstructed model brain images were very similar to the original T2-weighted images obtained by turbo spin echo sequence method. This means that water components classified in skeletal muscle are applicable to other human tissues. Analysis of MR image of other human organs will be discussed. [J Physiol Sci. 2007;57 Suppl:S93]
