Synopsis
To
quantify the iron loading in the whole body among transfusion-dependent
patients, 32 transfusion-dependent patients and 32 healthy volunteers were
recruited to participate in this study. The quantitative susceptibility mapping
was processed to get the susceptibility of the ROIs in the brain, and T2*values
of their livers, pancreas and myocardium. Significantly higher iron levels in
the putamen were found in transfusion-dependent patients
(right/left=0.147±0.066/0.149±0.811ppm) compared with
healthy controls(right/left=0.064±0.037/0.060±0.326ppm) (P=0.021/0.011). A ROC curve was performed, and the results suggested that
liver T2* and pancreas T2* values can be great predictors to diagnose the iron
overload in the brain ( AUC=0.877, 0.974, P＜0.01).Target audience
Those
interested in hematological disease and radiologists, scientist and MRI
researchers.
PURPOSE
Transfusion-dependent
disease (TDD) is a kind of chronic disease which includes β-thalassemia
major, myelodysplastic syndromes (MDS), and aplastic anemia
(AA). Patients
with this disease depend on regular transfusions of packed
red blood cells (PRBC) during the course of treatment. Long-term transfusion treatment can lead to iron deposits in
different organs, including the brain, heart, liver and pancreas, and causes organ
dysfunction. Serum ferritin (SF), despite being the biomarker of choice for
estimating blood iron concentration, cannot measure the organs’ iron deposition
and can be easily affected by inflammation, alcohol-related cirrhosis or the
onset of viral infection. Needle biopsy cannot be widely used because it is
invasive and carries a high risk of bleeding among patients with TDD. In this
study, we used quantitative susceptibility mapping (QSM)
1 in brain
iron evaluation, and quantified iron loading in the different brain areas among
patients with TDD We found a relationship between the brain iron and pancreas,
heart or liver iron. Furthermore, we want to find the best
predictor for brain iron among liver, pancreas or myocardiumT2* value.
METHODS
32 TDD patients and 32 healthy volunteers participated in this study. The
groups were matched for age. Patients in the acute inflammation or infection stage
and patients with tumors were excluded. Venous blood was collected to test SF
on the day of the MR scan. The MRI exam
included 6-echo gradient-echo SWI sequence covering the full brain, 8-echo Dixon sequence in the abdominal and 8-echo FLASH sequence in the heart. All data were collected
on a Siemens MAGNETOM Trio Tim system 3.0T MR scanner (Siemens
Healthcare, Erlangen, Germany). The parameters are as follows: Brain SWI: TR=28ms,
TE=20ms, flip angle=15°, 64 slices, slice
thickness=2 mm, distance factor=20%, FOV=230×172.5 mm
2, Voxel size: 1.0×0.5×2.0 mm
3,
measurements=1. Abdominal 8echo-Dixon sequence: TR=11.4 ms, TE= 1.17/2.39/3.68/4.92/6.14/7.36/8.58/9.80 ms,
flip angle=10°, 40 slices, slice thickness=5 mm, distance factor=20%, FOV=420×315 mm
2,
Voxel size: 2.4×1.6×5.0mm
3, measurements=1, Resp. control:
Breath-hold, Dimension=3D; Myocardium FLASH sequence: TR=730ms, TE= 2.7/5/7.3/9.6/11.9/14.2/16.5/18.80ms,
flip angle=18°, 1 slices, slice thickness=8.0 mm, distance factor=20%, FOV=380×285mm
2,
Voxel size: 2.5×1.5×8.0mm
3, measurements=1, ECG/Trigger. The T2*(R2*) values of liver
pancreas and myocardium were analyzed from R2*map using the Siemens syngo workstation.
MEDI Toolbox (Cornell MRI Research Lab, USA) and MATLAB R2014a (MathWorks, Inc,
US) were used to reconstruct the Quantitative Susceptibility Map (QSM) in figure1. Then the susceptibility of
brain areas were analyzed from QSM using ImageJ 1.48v software (National
Institutes of Health,USA). The ROIs of brain included head of nucleus caudate (HNC)，putamen (PT), globuspallidus (GP), red nucleus(RN),
substantia nigra (SN) and corpora dentatum (CD).
Two sampled t-test was used
to compare between TDD and healthy controls. Then we investigated the diagnostic
value of liver, pancreas and myocardium T2* (figure2) cutoff points based on
ROC curve.
RESULTS
Significantly higher levels of PT were present in transfusion-dependent
patients (right/left=0.147±0.066/0.149±0.811ppm) compared with healthy controls
(right/left=0.064±0.037/0.060±0.326ppm)
(
P=0.021/0.011). No differences were found in HNC, GP, RN, SN and CD between
TDD patient group and the control group (
P>0.05). ROC curve analysis was
performed (Figure 3) and suggested that liver T2* and pancreas T2* values can
be great predictors to diagnose the iron overload in the brain (AUC=0.877,
0.974,
P＜0.01). The T2* values of
liver and pancreas at the best cut-off point are 11.9 ms and 34.8ms, respectively.
DISCUSSION
MRI was gradually accepted
for clinical use as a non-invasive method and can be used for monitoring the
changes of iron concentration in different organs. In this study, we compared
the iron loading in different organs. The results could show the iron loading process
in the whole system.
CONCLUSION
TDD can lead to iron overload in the brain, especially in the putamen. The T2* values of liver and pancreas can be a great predictor for iron overload in the brain.
Acknowledgements
No acknowledgement found.References
1. Wang Y, Liu T. Quantitative susceptibility
mapping (QSM): decoding MRI data for a tissue magnetic biomarker[J]. Magnetic
Resonance in Medicine, 2015, 73(1): 82-101.