Susceptibility weighted imaging (SWI) of the kidney at 3T--initial results.

Abstract

Susceptibility weighted imaging provides diagnostic information in strokes, hemorrhages, and cerebral tumors and has proven to be a valuable tool in imaging venous vessels in the cerebrum. The SWI principle is based on the weighting of T(2)* weighted magnitude images with a phase mask, therewith improving image contrast of veins or neighbouring structures of different susceptibility, in general. T(2)* weighted MRI is already used for assessment of kidney function. In this paper, the feasibility of SWI on kidneys was investigated. Translation of SWI from the brain to the kidneys comes along with two main challenges: (i) organ motion due to breathing and (ii) a higher oxygenation level of renal veins compared to the brain. To handle these problems, the acquisition time has been cut down to allow for breath-hold examinations, and different post-processing methods including a new phase mask were investigated to visualize renal veins. Results showed that by a new post-processing strategy SWI contrast was enhanced on average by a factor of 1.33 compared to the standard phase mask. In summary, initial experiences of SWI on the kidneys demonstrated the feasibility. However, further technical developments have to be performed to make this technology applicable in clinical abdominal MRI.

Full text

Page 1

Susceptibility Weighted Imaging (SWI) of the Kidney at 3T – Initial Results

M. Mie1, J. C. Nissen2, F. G. Zöllner1, M. Heilmann1, H. J. Michaely2, S. O. Schönberg2, and L. R. Schad1
1Department of Computer Assisted Clinical Medicine, Heidelberg University, Mannheim, Germany, 2Institute of Clinical Radiology and Nuclear Medicine,
Heidelberg University, Mannheim, Germany

Introduction:
The kidney is important for the control of the blood pressure as well as for the homeostasis by selectively excreting or retaining various substances
[1]. There are many diseases leading to an impaired function of the kidney particularly longstanding diabetes mellitus, renal artery stenosis but also
acute and chronic rejection after renal transplantation [2]. Magnetic resonance imaging is a tool in the diagnostic work-up of kidney diseases also
allowing for assessment of the organ function. The susceptibility differences in the kidney can be represented in susceptibility weighted imaging
(SWI) by using phase imaging in addition to the magnitude imaging [3]. Usually, SWI is performed in imaging of the cerebrum, primarily in
enhancing the venous vessels, but also in strokes, hemorrhages and cerebral tumours [4]. Implementation of SWI on the kidney requires different
problems to be solved: Respiratory and bulk motion due to peristaltics leading to pronounced artifacts and require acquisition time as short as
possible. On the other hand, the kidney shows a different oxygen consumption with a higher oxygen fraction in the venous vessel system than in the
cerebral veins. Thus, the susceptibility difference between the venous vessels and the surrounding tissue is not as pronounced as in the brain. The
aim of this work was to implement SWI on the kidney assuming that through variation of the phase mask small susceptibility difference might be
fostered.

Methods:
In a first step a comparison between abdominal and brain images has been performed with well established imaging parameters
(TE/TR/FA=20ms/28ms/15°, slice thickness=1.5mm, FoV=162.5x200mm2, matrix size=260x320, BW=120Hz/px, slices per slab=56, TA=9 min). To
transfer this imaging protocol to abdominal imaging several parameters had to be adapted to reduce motion artifacts in the image and therefore,
allowing for breath-hold acquisition. To reduce scan time, parallel imaging (GRAPPA acceleration factor 2) and partial Fourier-imaging (6/8) were
applied, resulting in a total scan time of TA=67s. In addition, to even more shorten the acquisition time, the field of view (FoV) was kept as small as
possible covering one kidney and avoiding infolding artifacts. Other important imaging parameters (TE, FA) were kept constant. In summary, the
abdominal SWI protocol comprises the following parameters: TE/TR/FA=20ms/28ms/15°, slice thickness=1.8mm, FoV=220x220mm², matrix
size=256x256, BW=120Hz/px, slices per slab=52, TA=67s.

Results:
Figure 1 shows that there are larger phase changes in the brain (Fig. 1a) than in the kidney data (Fig. 1b), because of a lower oxygen extraction
fraction (OEF) in the kidney (0.08-0.1, [5]) than in the brain (white matter 0.25-0.35, gray matter 0.45-0.5, [5]). The corresponding abdominal
magnitude image and the resulting SW image can be seen in Figures 2a and 2b. The SW contrast enhanced abdominal images created with the
standard (brain) phase mask were of poor image quality, as hardly any obvious and visually detectable contrast enhancement was observed. It is
either useful to change the phase mask, or to increase the number of multiplications n with the phase mask M. In this work, both possibilities were
investigated. At first, different mask multiplications (n=4, n=8) were compared. Fig. 2b and 2c show the different weighted SW images. It can be
observed that a higher number of multiplications visually increases the contrast in the image. Next, we created a new phase mask altering the slope
of the previously used mask. The given corresponding magnitude image and SW images are shown in Figure 2a and 2d. The SW image shows
some areas where venous structures can be seen. To evaluate the contrast enhancement by our new post-processing, we calculated the contrast-
to-noise ratio (CNR) in carefully selected regions-of-interest. The CNR of the original post-processing was 15.31, whereas for the higher number of
multiplications a CNR of 24.13 was obtained. The modified phase mask leads to almost the same value. Small black lines enhanced by the phase
mask could be identified and interpreted as small renal veins.

Discussion:
CNR analysis showed that within the
kidney structures an increase in contrast
of a factor 1.5 could be obtained.The
acquisition time of 67s may be acceptable
for certain healthy volunteers and was
sufficient for the aim of this pilot study but
of course is not clinically feasible. The
acquisition time must be further shortened
to allow for breath-hold examinations in
patients, or navigator
techniques must be implemented to
reduce motion artifacts in images at
longer scan times. In summary, the
results represent initial experiences with
the SWI for abdominal imaging which
proof the principal feasibility. However,
further research has do be performed to
make this technology applicable in clinical
abdominal MRI.








References:
[1] Schoenberg et al. Der Radiologe, 37(8):651-62 (1997) [2] Michaely et al. Der Radiologe,
48(2):185-200; (2008), [3] Haacke et al. Magn Reson Med., 52(3):612-8 (2004), [4] Barth et al.
Invest Radiol., 38(7):409-14 (2003), [5] Vaupel et al. Cancer Res., 1;49(23):6449-65 (1989)
or trigging
Figure 1: (a) Filtered phase image of the brain (left)
and corresponding phase values along the black line.
The red part of the plot shows the phase values of the
brain. The arrow marks a venous vessel (b) Filtered
phase image and phase values of the abdomen. The
red part of the curve shows the phase values of the
left kidney.
Figure 2: (a) Magnitude image and (b)
corresponding SW image reconstructed
with standard phase mask using n=4, (c)
n=8 and (d) modified phase mask with
n=4. The red arrows mark the spots
where venous structures can be seen. (e)
Minimum intensity projection (mIP) of 3
slices.