MRI of lung parenchyma and pulmonary embolism (PE) remains challenging
. ''Ferrum,'' a ferric hydroxide sucrose complex used clinically for i
ron deficiency anemia for more than 40 years, was investigated as a ne
gative MRI contrast agent in five rabbits bearing experimental PE as w
ell as in five normal volunteers. Clots were prepared by spontaneous c
oagulation of 0.1 ml In-111 labeled autologous red blood cells and int
roduced through the jugular vein. Scintigraphic imaging permitted anat
omical localization of PE in vivo and thereby served as a control for
MR imaging. MRI was performed on a 1.5 T GE Signa scanner before and a
fter induction of PE, and before and after the injection of Ferrum. T1
-weighted images were obtained continuously for up to 90 min using var
ying doses of Ferrum. In five normal human volunteers, a single dose o
f 100 mg each was administered. T1- and T2-weighted spin-echo and grad
ient-echo images of lung parenchyma were repeatedly obtained before an
d after agent administration. In rabbit, Ferrum remained in circulatio
n for several hours where it shortened both T1 and T2 of blood, improv
ing the contrast between PE and lung parenchyma (i.e., intravascular c
ompartment). A dose of 3 mg/kg was enough to increase the contrast-to-
noise ratio (CNR) between PE and lung parenchyma by almost three fold,
substantially improving lesion detectability. CNR increased up to fiv
e-fold when the dose was increased up to 20 mg/kg at which point CNR r
eached a plateau. In humans, T2-weighted spin-echo sequence appeared t
o be most sensitive to changes in signal-to-noise ratio (SNR) of norma
l lung parenchyma. Within 60 min after injection of 100 mg of iron, SN
R dropped by 34% (p < .025). However, 24 hr later, SNR returned to alm
ost normal. Ferrum increased the contrast between PE and lung parenchy
ma in the rabbit and decreased the parenchymal SNR in humans in nontox
ic doses. These results suggest that Ferrum is worthy of further inves
tigation of PE imaging in humans.