MR imaging of the lung is handicapped by three negative influences. Fi
rst, the low proton spin density in lung tissue results in a low signa
l-to-noise ratio relative to the surrounding structures. Second, cardi
ac and respiratory motion induce artifacts that tend to obscure fine s
tructural detail in the lung. Third, a considerable magnetic susceptib
ility gradient, arising from the large surface areas of air and tissue
interfaces, produces a very low value for T2. MR imaging can be used
to stage the activity of interstitial lung disease and for the diagno
sis of lipoid pneumonia and pulmonary infarction. In combination with
MR angiography, perfusion MR imaging might eventually become a test fo
r pulmonary embolisms.