Chia-Wen Chiang1, Ezequiel Farrher2, Kuan-Hung Cho1, Shih-Yen Lin1,3, Kuo-Jen Wu4, Yun Wang4, Teh-Chen Wang5, Yi-Ping Chao6, Yeun-Chung Chang7, Chang-Hoon Choi2, and Li-Wei Kuo1,8
1Institute of Biomedical Engineering and Nanomedicine, National Health Research Institutes, Miaoli, Taiwan, 2Institute of Neuroscience and Medicine – 4, Medical Imaging Physics, Forschungszentrum Jülich, Jülich, Germany, 3Department of Computer Science, National Chiao Tung University, Hsinchu, Taiwan, 4Center for Neuropsychiatric Research, National Health Research Institutes, Miaoli, Taiwan, 5Department of Medical Imaging, Taipei City Hospital, Taipei, Taiwan, 6Department of Computer Science and Information Engineering, Chang Gung University, Taoyuan, Taiwan, 7Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan, 8Institute of Medical Device and Imaging, National Taiwan University College of Medicine, Taipei, Taiwan
Synopsis
Brain swelling typically occurs in acute stroke.1
Mannitol, as a hyperosmolar agent, enables to effectively treat the
increased intraocular pressure and cerebral edema in brain injury.2,3
Diffusion kurtosis imaging with free water elimination (DKI-FWE)4
has been recently reported the ability to assess diffusion indices by
separating free water compartment in simulations and healthy volunteers. The purpose
of the study was to examine the effect of mannitol infusion using a stroke rat
model at acute and chronic stages assessed by DKI-FWE and DKI. Our preliminary
results revealed that mean kurtosis (MK) was sensitive to reflect
mannitol-treated dehydration in acute stroke rat.
Introduction
Brain swelling typically occurs in acute
stroke.1 The administration of hyperosmolar agents, such as mannitol,
allows to effectively reduce intraocular pressure and cerebral edema by extracting both intracellular
and interstitial water in brain injury.2,3 Recently, Diffusion Kurtosis Imaging with free water elimination (DKI-FWE),4
extending DKI,5 has been reported in simulation and healthy
volunteers, showing the capability of separating the signal contributions of
free water and tissue for accurate estimates of diffusion characteristics and free
water fraction. Although potentially useful, the capability of DKI-FWE on
neurological diseases has not been investigated yet. Therefore, this study aims
to employ DKI-FWE and DKI techniques to explore the mannitol-treated dehydration
within brain lesions in acute and chronic stages of stroke. We attempted to
examine if the mean kurtosis (MK) index could reveal the dehydration in lesions
and compare the sensitivities of conventional and FWE models of DKI technique.Methods
Three adult male Sprague–Dawley rats,
weighting 350-450 g, were underwent the right middle cerebral artery occlusion for 90 minutes as described.6 Rat was treated
with a single dose of 20 % mannitol solution7 through intravenous
injection. The first MRI scan was performed before mannitol injection and under
the same experimental setup, the second scan starts at 20 minutes after
mannitol injection. Experiments were performed on an in-house 3T MRI system.8
An ultra-high-strength gradient coil with a maximum strength of 675 mT/m
(Resonance Research Inc., MA, USA) and a custom transmit/receive surface coil
were utilized. Diffusion-weighted images (DWI) and T2-weighted images (T2WI)
were acquired. For DWI, a
Stejskal-Tanner EPI pulse sequence was implemented with the b-value (s/mm2)/number
of diffusion directions: 500/12, 1000/26, and 2000/40. Eight b=0 images were
obtained. The sequence parameters were TR/TE of 2000/50 ms, Δ/δ of 24/3
ms, 1 average, slice thickness of 1 mm, FOV of 25 × 25 mm2, matrix
size of 96 × 96, and 20 slices. T2WI was acquired using fast spin-echo sequence
with in-plane resolution of 130 × 130 μm2 for examination of brain
infarction. All datasets were first for denoising.9 Diffusion
tensors and diffusion kurtosis tensors were reconstructed with DKI5 and
DKI-FWE4 model analyses in Matlab. For DKI-FWE, the diffusivity of free
water was fixed at 3 μm2/ms. Non-linear least-squares fitting was
used. Regions-of-interest (ROIs) of lesion were determined on hyper-intensity
regions in T2WI. ROI analysis was performed on MK and mean diffusivity (MD)
maps from both conventional and FWE models. The percentage change of indices were
quantified to evaluate the sensitivities to detect the dehydration in lesions.Results
Figure 1 shows cerebral infarction of stroke rats in T2WI. For
each time point, there was no obvious decrease of lesion size in stroke after
mannitol infusions. Figure 2 shows, with time course, both DKI and DKI-FWE
analyses indicated MD of lesions was decreased on day 1, close to normal on day
5, and elevated on day 26. MD was slightly decreased by using DKI-FWE model,
compared to conventional DKI model. With mannitol administration, our results indicated
that MD was apparently decreased on day 5 (~6%) from both analyses (h and k,
arrows), compared to that on day 1 and 26 (~1%). Figure 3 shows both analyses
indicated that MK of lesions was apparently increased particularly on day 1 and day
5. MK then became much smaller at chronic stage. DKI-FWE, compared to DKI,
indicated an incremental decrease of MK on day 1 (~3% for stroke, a vs. d; ~6%
for stroke after mannitol, g vs. j), but revealed similar or slightly increased
MK on day 5 (b vs. e and h vs. k) and day 26 (c vs. f and i vs. l). In stroke lesions,
DKI-FWE revealed different patterns of MK on day 1 (a vs. d), but not MD (a vs.
d in Figure 2). After mannitol administration, MK was obviously decreased on
day 1 (~5% from DKI-FWE and ~2% from DKI; g and j, arrows) and on day 5 (~11%
from DKI-FWE and ~10% from DKI; h and k, arrows), but almost unchanged on day
26. Table 1 lists the percentage changes of MD and MK before and after mannitol
injection.Discussion and Conclusion
DKI-FWE, compared to conventional DKI, revealed
decreases of MD and MK in stroke lesions, reflecting the removal of water contamination,
consistent with Collier et al., 2018. With
mannitol administration, our results indicated that MK was apparently decreased
on day 1 and day 5 from DKI-FWE model. However, the conventional DKI analysis
only showed the decrease on day 5. This may suggest that DKI-FWE has the higher
sensitivity to reflect dehydration in stroke lesions. Interestingly, at chronic
stage, the percentage changes of MD and MK show a very minor change before and
after mannitol injection. One possible explanation is that, due to blood-brain
barrier (BBB) integrity at chronic stage, mannitol may remain in vascular
compartment and show a slower dehydration, compared to acute stage with damaged
BBB. In summary, our preliminary results show MK could reflect the dehydration
in stroke lesion and the DKI-FWE model could provide higher sensitivity than
conventional DKI model. For verifying our current observations, the number of
rats has to be increased and a more sophisticated analysis is needed in further
studies. Acknowledgements
This work was supported by the National Health Research Institutes (108-PP-06) and Ministry of Science and Technology (108-2221-E-400-002 and 108-2911-I-400-502) in Taiwan.References
1. Bansal S et al., Drug
Treatment of Acute Ischemic Stroke. Am J Cardiovasc Drugs. 2013; 13(1):
10.1007/s40256-013-0007-6.
2. Knapp JM, Hyperosmolar therapy in the treatment of
severe head injury in children: mannitol and hypertonic saline. AACN Clin Issues. 2005; 16(2):199-211.
3. Diringer MN et al.,
Osmotic therapy. Neurocritical Care. 2004; 1(2):219-233.
4. Collier Q et al.,
Diffusion Kurtosis Imaging With Free Water Elimination: A Bayesian Estimation Approach. Magnetic Resonance in Medicine. 2018; 80:802–813.
5. Jensen JH et al.,
MRI Quantification of Non-Gaussian Water Diffusion by Kurtosis Analysis.
NMR Biomed. 2010; 23(7): 698–710.
6. Wu KJ et al.,
Transplantation of Human Placenta-Derived Multipotent Stem Cells Reduces Ischemic Brain Injury in
Adult Rats. Cell Transplantation. 2015; 24(3): 459–470.
7. Paczynski RP et
al., Multiple-Dose Mannitol Reduces Brain Water
Content in a Rat Model of Cortical Infarction. Stroke. 1997; 28(7):1437–1444.
8. Cho
KH et al., Development,
integration and use of an ultrahigh-strength gradient system on a human
size 3T magnet for small animal MRI. PLoS ONE. 2019; 14(6):e0217916.
9. Manjon JV et al., Adaptive Non-Local
Means Denoising of MR Images with Spatially Varying Noise Levels. Journal
of Magnetic Resonance Imaging. 2010; 31(1):192–203.