Transverse relaxation of solvent protons induced by magnetized spheres: application to ferritin, erythrocytes, and magnetite.

Since 1/T2 of protons of tissue water is generally much greater than 1/T1 at typical imaging fields, small single-ion contrast agents--such as Gd(DTPA), which make comparable incremental contributions and therefore smaller fractional contributions to 1/T2 compared to 1/T1--are not as desirable for contrast-enhancement as agents that could enhance 1/T2 preferentially. In principle, such specialized agents will only be effective at higher fields because the field dependence (dispersion) of 1/T1 is such that it approaches zero at high fields whereas 1/T2 approaches a constant value. The residual 1/T2 is called the "secular" contribution and arises from fluctuations in time--as sensed by the protons of diffusing solvent or tissue water molecules--of the component of the magnetic field parallel to the static applied field. For solutions or suspensions of sufficiently large paramagnetic or ferromagnetic particles (greater than or equal to 250 A diameter), the paramagnetic contributions to the relaxation rates satisfy 1/T2 much greater than 1/T1 at typical imaging fields. We examine the theory of secular relaxation in some detail, particularly as it applies to systems relevant to magnetic resonance imaging, and then analyze the data for solutions, suspensions, or tissue containing ferritin, erythrocytes, agar-bound magnetite particles, and liver with low-density composite polymer-coated magnetite. In most cases we can explain the relaxation data, often quantitatively, in terms of the theory of relaxation of protons (water molecules) diffusing in the outer sphere environments of magnetized particles. The dipolar field produced by these particles has a strong spatial dependence, and its apparent fluctuations in time as seen by the diffusing protons produce spin transitions that contribute to both 1/T1 and /T2 comparably at low fields; for the larger particles, because of dispersion, the secular term dominates at fields of interest. On the basis of the agreement of theory with data for solutions of small paramagnetic complexes, large magnetite particles, and liver containing low-density polymer-coated magnetite agglomerates, it is argued that the theory is sufficiently reliable so that, e.g., for ferritin--for which 1/T2 is unexpectedly large--the source of its large relaxivity must reside in nonideal chemistry of the ferritin core. For blood, it appears that diffusion through intracellular gradients determines 1/T2.