Abstract
Purpose
To establish a correlation between putative iron content using susceptibility weighted imaging (SWI) phase and T2* weighted magnitude values in the basal ganglia as a function of age in healthy human brains.
Materials and methods
100 healthy adults (20-69 yr.; mean = 43 yr) were evaluated for this study using a gradient echo sequence. The original magnitude and high pass filtered phase data were analyzed as proxy variables for iron content in the substantia nigra, red nucleus, globus pallidus, putamen, caudate nucleus, thalamus and pulvinar thalamus. Each structure was broken into two parts, a high iron content region and a low iron content region.
Results
Both magnitude and phase data showed an increase in putative iron content with age. However, the high iron content region revealed two new pieces of information: both the average iron content per pixel and the area of high iron increased with age. Further, significant increase in iron uptake as a function of age was found past the age of 40.
Conclusion
A two region of interest analysis of iron is a much more sensitive means to evaluate iron content change over time. Contrary to the current belief that iron content increases level off with age, the putative iron deposition in region two is seen to increase with age.
Keywords: brain iron deposition, susceptibility weighted imaging, iron changes with age
Introduction
Iron has long been recognized to play a key role in brain function such as in oxygen transport, neurotransmitter synthesis, electron transfer and myelin production (1,2). Despite the positive implications of iron, in its free ferrous form, iron is known to be toxic and lead to free radical production. Recently, there has been a renewed interest in the role of iron in the pathogenesis of many diseases such as aging, multiple sclerosis, Parkinson’s disease, and other diseases (1-5). The disturbance of iron metabolism can occur at several levels: iron uptake and release, storage, intracellular metabolism and regulation. The body has a number of ways to deal with this issue by storing excess iron in the form of ferritin or hemosiderin. Ferritin appears to be the normal form of iron storage while hemosiderin often appears when there has been bleeding and macrophages have helped to consolidate this iron (5-7). Using methods such as diaminobenzidine acid enhanced Perl’s staining revealed iron in many cell types of the central nervous system such as neurons, microglia, oligodendroglia, astrocytes and in some myelin sheaths (8). This iron is also found to be present in higher concentration in gray matter than white matter (9). Deep gray matter structures accumulate ferritin at different rates throughout different ages (8,10-12). It is believed and often quoted (10) that iron increases quickly with age and then levels off or increases more slowly. Further, it has been shown that there are major iron deposits in the form of ferritin associated with the basal ganglia (13,14). This interest in iron has led to attempts to quantify iron using imaging in neurodegenerative diseases in order to understand their pathogenesis, monitor their progress and establish neuroprotective treatments. Developing and improving magnetic resonance imaging (MRI) has been an important task for decades, trying to establish a correlation between its signal and iron concentration. A variety of methods have been employed, including T2, T2* and T2′ (15,16) as well as Field Dependent Relaxation Rate Increase (17) and Susceptibility Weighted Imaging (SWI) (18). Although R2 showed more sensitivity to Gray Matter (GM)/White Matter (WM) differentiation than R2′, it was not able to show specificity for iron (16). On the other hand, R2′ showed a good correlation with iron load, where R2′ was seen to increase for higher iron content structures. Individual T2 relaxation times showed more variation after 45 years of age (19). It has been shown that R2′ and R2* increase in GM but decrease in WM as a function of age (20). However, for spin echo methods, there are confounding factors, such as increases in local water content, which can hide the R2 effects of an increase in iron because the increased water content will decrease R2 while increased iron content will increase R2.
Recently, SWI (21,22) has been used to measure or at least infer iron content in brain parenchyma (18,21). A recent comparison study between SWI and others sequences (23) confirmed that SWI can be used as a tool to quantify iron deposition in the brain. Other studies (24,25) have also tried to probe brain iron with SWI by investigating the pattern of mineralization in the deep gray matter as a function of age. To date, many researchers have studied total iron content in specific regions of the brain such as basal ganglia and thalamus, but no one has yet focused on establishing the regional changes of high iron content as a function of age.
Materials and methods
Participants
In this study, 188 participants were recruited from university staff, students and their relatives. A consent form approved by the hospital’s research ethics committee was provided for all volunteers to read, really understand and sign. Participants included in this study were normal subjects who didn’t have any history of neurological or psychiatric conditions, head trauma with loss of consciousness for more than five minutes, drug and alcohol abuse, or brain surgery. Also, these participants did not show any focal parenchymal loss, infarction, resection, or large hyperintensities in their brain on T2-weighted images.
One hundred (100) normal human subjects with good quality images were chosen from this group. The age interval ranged from 20 to 69 years old (mean age=43 yr., SD=14 yr.). A total of 50 males (mean age=42 yr., SD=14 yr.) and 50 females (mean age=44 yr., SD=14 yr.) were included. All volunteers were right-handed; this fact was confirmed by the enquiry of a neurologist before the MR scan.
MRI Data Aquisition
MR imaging was performed using a 1.5 T MR scanner (HD, General Electric Co., Milwaukee, USA) equipped with an 8-channel phased array head coil. The imaging plane was oriented parallel to the anterior–posterior commissural (AC–PC) line using a three-dimensional gradient-echo sequence with the following parameters: TR/TE = 53 ms/40 ms, flip angle = 20°, slice thickness = 3mm, bandwidth=31.25 Hz/pixel, field-of-view (FOV) = 24 cm and matrix size (Nx × Ny) = 384 × 320, yielding an in-plane resolution of 0.60 mm by 0.75 mm. The total scan time was 6 minutes and 28 seconds.
Data Analysis
Our in-house software SPIN (Signal Processing in NMR, MRI Institute for Biomedical Research, Detroit, MI, USA) was used for the image processing. Two graduate students were trained by a neuroradiologist to outline the structures of interest. Phase images were filtered using a 64×64 high-pass filter to remove the low-spatial-frequency components (26). The 64×64 filter was used because it has a minimal effect on contrast for small structures (27). The high-pass filtered phase images were analyzed as a means to quantify iron content.
The regions of interest (ROIs) were outlined manually by the two graduate students based on their anatomical locations and the boundaries shown in the filtered phase images. Seven structures were measured bilaterally including: the head of the caudate nucleus (CN), the globus pallidus (GP), the putamen (PUT), the pulvinar thalamus (PT), the substantia nigra (SN), the red nucleus (RN), and the thalamus (THA). One single slice which showed the largest area and the most well-defined border (see Fig. 1a-f) was selected for each structure and analyzed for its iron content. As a further condition, slices which were severely affected by sinus artifacts were not used. The bright boundaries outside the structures were also avoided because this large positive phase shift is most likely associated with the iron-induced dipolar field patterns in the tissue (28,29). Structures were zoomed (interpolated) by a factor of four to make the boundaries easier to determine and the area of the structure more accurately drawn. ROIs were traced first on phase images and then copied to the magnitude images which guaranteed that the boundaries were exactly the same in both images (magnitude and filtered phase images). Separate slices were chosen for the CN (Fig. 1a) and the GP and the PUT together (Fig. 1b). The slice for the PT was chosen to have the highest contrast on the phase images and to avoid the SN or RN structures and the major veins (Fig. 1c,d). The RN and SN were drawn on the same slice that showed the ring-like hyperintensity signal outside the RN (capsule of RN) (27), and showed no fusion of the SN (Fig. 1c) and RN (Fig. 1d). For the thalamus, we chose the slice that showed the boundary clearly without showing the RN or subthalamic nucleus. Also, we avoided the internal capsule while drawing the ROI around the thalamus (Fig.1f).
Figure 1.
The regions of interest drown using high-pass filtered phase images. These structures are: the caudate nucleus (a), the globus pallidus (b, fresh red), the putamen (b, dark red), the substantia nigra (c), the red nucleus (d), the pulvinar thalamus (e) and the thalamus (f). The blue highlighted regions are the high iron content regions (RII).
In this study, we analyzed both a low iron content region (RI) and a high iron content region (RII) because an entire ROI analysis would reduce the sensitivity to more subtle changes in iron content when the region containing the iron is a small fraction of the total area. To automatically separate these two regions, a threshold was used based on several recent papers(27,30) on phase and iron content. For the structures that were not included in these papers, in order to find the appropriate thresholds, we analyzed twenty normal cases with ages ranging from 20 to 39 year old. Thresholds used for phase and magnitude images analysis are shown in Table 1. The thresholds were set using the mean value measured minus two times the standard deviation across an elderly population to be on the conservative side as outlined in reference (27) or the above mentioned 20 normal cases.
Table 1.
Threshold values for the regional separation
| CN | GP | PUT | PT | SN | RN | THA | |
|---|---|---|---|---|---|---|---|
| Phase (rad) | 0.190† | 0.172† | 0.330† | 0.143‡ | 0.356† | 0.261† | 0.049‡ |
| Magnitude (a) | 117‡ | 84‡ | 112‡ | 126‡ | 90‡ | 101‡ | 118‡ |
PUT: Putamen, CN: Caudate Nucleus, GP: Globus Pallidus, RN: Red Nucleus, PT: Pulvinar Thalamus, SN: Substantia Nigra, THA: Thalamus, a: arbitrary unit, rad: radians
Determined from reference 27
Determined from twenty normal cases from 20 to 39 years old
The phase differences due to iron are quoted in radians while the magnitude differences are related directly to iron via changes in R2* or ΔR2*. By taking the average magnitude signals of 10 subjects with the lowest iron content as the baseline magnitude value (S0), the iron content change relative to the lowest iron content subjects could be calculated through the ratio of the individual’s magnitude signal divided by the baseline magnitude which yields:
| [1] |
where TE is the echo time for the sequence, which is 40 ms in this study and S1 is the signal of a given subject. We can find ΔR2* which is proportional to the change of iron content from
| [2] |
Statistical analysis
Relationships between phase or magnitude calculations as a function of age were determined using linear and quadratic regression and Pearson correlation analysis from MATLAB. The differences between left and right hemispheres and the overall mean and standard deviation were calculated for each subject. In addition, the frequency analysis was applied to the difference to see how many subjects showed higher iron content in the left brain, equal amount in both sides and higher iron in right brain. All statistics were performed with two-tailed tests with a 0.05 level of significance.
Results
The seven structures of interest are first hand drawn as shown in Figure 1. For all structures except the globus pallidus, phase and ΔR2* analysis showed an increase in putative iron content with age both for the total region and for the high iron content region (Figure 2-8). In some cases, the iron increase itself increased after the age of 40 years old, as evidenced in Figure 9 and Tables 2 and 3. In the figures that follow, we consider first the different measures of iron that were made: the putative iron in the high iron content region, the average iron per voxel in the high iron content region, the area of the high iron content region, and the average iron per voxel in the total slice. Figure 2 shows all four of these iron measures for both ΔR2* and phase for one structure that best delineates the iron behavior in the putamen. Figures 3 and 4 show the putative iron in the high iron content region for both magnitude and phase for the rest of the structures. Figures 5 and 6 show the average iron per pixel in the total structure for both magnitude and phase. Figures 7 and 8 show the normalized area of high iron content region as a function of age for both magnitude and phase. The normalized percent area of high iron content region relative to the total area was used to avoid the influence by the size change for a single structure between individuals. Quantitative results for iron as a function of age for all the structures in terms of the above eight measures are shown in Table 2.
Figure 2.
Putative iron measurement results as a function of age in Putamen. An increase in iron content - represented here by the increase in phase and ΔR2* - with age for total iron, average iron and the iron area are seen in the PUT. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. r2 is larger for the quadratic fit than just linear fit. L stands for the left hemisphere. R stands for the right hemisphere. Quadratic relationships can be found in plots a, b, e, and f representing the total iron in RII predicted from phase (a) and magnitude (b) and the area in high iron region (phase (e) and magnitude (f)). This effect is not seen in the total region iron for the whole structure (c, d, g, and h).
Figure 8.
Normalized area of high iron content region as a function of age in RN, SN and GP. High iron content area in RN shows significant linear relationship with age. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
Figure 9.
Putative iron content in two age intervals - younger and older than 40 years old subjects - giving a clear idea about iron deposition changes. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
Table 2.
Slopes of phase and magnitude iron measures as a function of age in the two hemispheres
| Age(yr.) | Iron-Phase-RII (rad×mm3/yr.) | Iron-Phase-Total (rad×mm3/yr.) | Average iron-Phase-Total (10−3 rad/yr.) |
Average iron-Phase-RII (10−3 rad/yr.) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Slope |
Slope |
Slope |
Slope |
||||||||||
| Left | Right | Total | Left | Right | Total | Left | Right | Total | Left | Right | Total | ||
| PUT | 20-39 | 0.16±0.08 | 0.27±0.11 | 0.21±0.07 | 2.24±0.81 | 2.3±0.84 | 2.27±0.58 | 2.03±0.63 | 2.03±0.65 | 2.03±0.45 | 0.14±0.93 | 0.84±1.43 | 0.49±0.85 |
| 40-69 | 2.07±0.52 | 2.21±0.63 | 2.14±0.41 | 2.80±0.76 | 3.00±0.82 | 2.90±0.56 | 2.65±0.67 | 2.60±0.67 | 2.63±0.47 | 2.63±0.67 | 2.65±0.75 | 2.64±0.50 | |
| CN | 20-39 | 0.79±0.32 | 0.64±0.34 | 0.71±0.23 | 0.85±0.52 | 0.38±0.53 | 0.62±0.37 | 1.27±0.73 | 0.79±0.78 | 1.03±0.53 | 0.85±0.53 | 0.42±0.40 | 0.64±0.33 |
| 40-69 | 2.00±0.33 | 1.35±0.29 | 1.68±0.22 | 1.72±0.37 | 1.19±0.37 | 1.46±0.27 | 2.37±0.50 | 1.77±0.54 | 2.07±0.37 | 1.92±0.41 | 1.56±0.51 | 1.74±0.33 | |
| GP | 20-39 | −0.21±0.66 | −0.03±0.64 | − 0.12±0.45 |
−0.20±0.75 | −0.38±0.68 | −0.29±0.50 | − 0.07±0.69 |
− 0.24±0.61 |
− 0.16±0.46 |
−0.29±0.94 | 0.35±0.71 | 0.03±0.59 |
| 40-69 | 1.16±0.91 | 1.03±0.86 | 1.10±0.62 | 0.36±0.68 | 0.08±0.62 | 0.22±0.46 | − 0.14±0.68 |
− 0.10±0.61 |
− 0.12±0.46 |
1.89±1.57 | 2.88±1.60 | 2.39±1.11 | |
| RN | 20-69 | 0.25±0.03 | 0.19±0.03 | 0.22±0.02 | 0.55±0.07 | 0.56±0.07 | 0.56±0.05 | 2.23±0.34 | 2.19±0.33 | 2.21±0.24 | 0.97±0.18 | 0.84±0.16 | 0.91±0.12 |
| PT | 20-69 | 0.30±0.06 | 0.36±0.06 | 0.33±0.04 | 0.31±0.08 | 0.32±0.08 | 0.31±0.08 | 1.02±0.21 | 1.09±0.22 | 1.10±0.15 | 0.51±0.09 | 0.55±0.09 | 0.53±0.07 |
| SN | 20-69 | 0.10±0.06 | 0.12±0.06 | 0.11±0.04 | 0.19±0.12 | 0.20±0.12 | 0.19±0.08 | 1.50±0.39 | 1.43±0.37 | 1.46±0.27 | 0.31±0.21 | 0.52±0.21 | 0.41±0.15 |
| THA | 20-69 | 0.78±0.19 | 0.65±0.19 | 0.72±0.13 | 0.67±0.28 | 0.70±0.27 | 0.69±0.19 | 0.34±0.13 | 0.35±0.13 | 0.34±0.09 | 0.62±0.08 | 0.50±0.08 | 0.56±0.06 |
| Age(yr.) | Iron-Magnitude-RII (sec−1×mm3/yr.) | Iron-Magnitude-Total (sec−1×mm3/yr.) | Average iron-Magnitude-Total (10−3ms−1/yr.) | Area-Phase-RII (10−3 %/yr.) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Slope |
Slope |
Slope |
Slope |
||||||||||
| Left | Right | Total | Left | Right | Total | Left | Right | Total | Left | Right | Total | ||
| PUT | 20-39 | 0.06±0.02 | 0.06±0.02 | 0.06±0.01 | 0.11±0.04 | 0.13±0.04 | 0.12±0.03 | 0.10±0.03 | 0.10±0.03 | 0.10±0.02 | 0.33±0.17 | 0.56±0.22 | 0.45±0.14 |
| 40-69 | 0.21±0.05 | 0.19±0.05 | 0.20±0.04 | 0.14±0.04 | 0.16±0.05 | 0.15±0.03 | 0.15±0.04 | 0.14±0.04 | 0.14±0.03 | 3.97±1.01 | 3.76±1.29 | 3.87±0.84 | |
| CN | 20-39 | 0.04±0.01 | 0.02±0.01 | 0.03±0.01 | 0.06±0.02 | 0.03±0.02 | 0.04±0.02 | 0.07±0.03 | 0.06±0.03 | 0.06±0.02 | 4.21±1.72 | 3.51±1.95 | 3.86±1.28 |
| 40-69 | 0.11±0.03 | 0.08±0.03 | 0.10±0.02 | 0.09±0.02 | 0.07±0.02 | 0.08±0.02 | 0.12±0.02 | 0.09±0.03 | 0.10±0.02 | 8.75±1.50 | 6.01±1.48 | 7.38±1.06 | |
| GP | 20-39 | 0.02±0.02 | 0.02±0.02 | 0.02±0.01 | 0.07±0.06 | 0.06±0.05 | 0.07±0.04 | 0.07±0.05 | 0.07±0.04 | 0.07±0.03 | −0.75±2.06 | −0.45±1.90 | −0.6±1.39 |
| 40-69 | 0.13±0.06 | 0.10±0.05 | 0.11±0.04 | 0.10±0.05 | 0.07±0.05 | 0.09±0.04 | 0.07±0.05 | 0.07±0.05 | 0.07±0.03 | −0.03±1.50 | −0.65±1.38 | −0.34±1.01 | |
| RN | 20-69 | 0.03±0.00 | 0.03±0.00 | 0.03±0.00 | 0.03±0.01 | 0.02±0.00 | 0.02±0.00 | 0.13±0.01 | 0.12±0.01 | 0.13±0.01 | 3.02±0.52 | 2.14±0.44 | 2.58±0.34 |
| PT | 20-69 | 0.03±0.00 | 0.03±0.00 | 0.03±0.00 | 0.02±0.00 | 0.02±0.00 | 0.02±0.00 | 0.07±0.01 | 0.08±0.01 | 0.08±0.01 | 4.40±0.80 | 5.64±0.85 | 5.02±0.59 |
| SN | 20-69 | 0.02±0.00 | 0.01±0.00 | 0.02±0.00 | 0.02±0.01 | 0.02±0.00 | 0.02±0.00 | 0.07±0.01 | 0.07±0.01 | 0.07±0.01 | 0.71±0.42 | 0.76±0.42 | 0.73±0.30 |
| THA | 20-69 | 0.06±0.01 | 0.06±0.01 | 0.06±0.01 | 0.10±0.02 | 0.10±0.01 | 0.10±0.01 | 0.05±0.01 | 0.05±0.01 | 0.05±0.00 | 1.92±0.72 | 1.72±0.71 | 1.82±0.50 |
PUT: Putamen. CN: Caudate Nucleus. GP: Globus Pallidus. RN: Red Nucleus. PT: Pulvinar Thalamus. SN: Substantia Nigra. THA: Thalamus
Table 3.
The orders of high to low iron deposition over time
| PUT | RN | CN | PT | SN | GP | THA | ||
|---|---|---|---|---|---|---|---|---|
| Phase | Average iron (whole structure) | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Iron in RII | 1 | 6 | 2 | 5 | 7 | 3 | 4 | |
| Average iron in RII | 1 | 4 | 3 | 6 | 7 | 2 | 5 | |
| Magnitude | Average iron (whole structure) | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Iron in RII | 1 | 5 | 3 | 5 | 6 | 2 | 4 | |
| Average iron in RII | 3 | 2 | 4 | 5 | 6 | 1 | 7 |
Figure 3.
Putative iron content - represented here by phase and ΔR2* - in high iron content region in CN, PT, and THA where iron is seen to increase with age. Iron in CN and PT shows quadratic relationship with age. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
Figure 4.
Putative iron content - represented here by total phase and ΔR2* - in high iron content region in RN, SN, and GP. In all these structures, a significant increase with age is seen. For GP, iron predicted from magnitude shows quadratic relationship with age. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
Figure 5.
Average iron - represented here by average phase and ΔR2* - of the whole structure as a function of age in CN, PT and THA. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
Figure 6.
Average iron - represented here by average phase and ΔR2* - of the whole structure as a function of age in RN, SN and GP. Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
Figure 7.
Normalized area of high iron content region as a function of age in CN, PT, and THA. High iron content area in CN and PT show strong quadratic relationship with age (increase). Stars and dashdot line are the values and fit of the right hemisphere. Triangles and solid line are the values and fit of the left hemisphere. r2 is the coefficient of determination. L stands for the left hemisphere. R stands for the right hemisphere.
For PUT, CN, PT and GP, the measures related to the high iron content region showed an impressive quadratic relation with age (Fig. 2a, 2b, 2e, 2f, 3a, 3b, 3d, 4f). Fitting the data to a quadratic function increased the coefficient of determination (r2) over that from the linear regression fit. For example, for the putative iron content in RII of the putamen, r2 from the linear fit was 0.294 and 0.278 for the left and right hemispheres respectively, while r2 was 0.354 and 0.316 for the quadratic fit, respectively (Figure 2a). Generally, the results seemed to be fairly flat in the first two decades (20 to 40 years old) but the local iron content dramatically increased after 40 years old. Increases occurred not only in the total area of the iron deposition but also in the average iron content per pixel. Fitting the data to two separate curves for 20 to 40 years old and 40 to 70 years old shows more clearly the dichotomous iron behavior in these two age ranges (Fig. 9). This property of iron build-up cannot be detected easily in the whole structure analysis which showed a linear relationship with age (Fig. 2g, 2h, 5 and 6).
The order of high to low iron deposition changes depending on whether one examines the average iron deposition in the whole structure; iron in the high iron content region; or average iron in region II. The orders are shown in Table 3. Each of these behaviors may tell us something different about the iron deposition in each structure.
The high iron content region reveals a different effect that cannot be seen from the total iron. Even though total iron can also increase with age quite independently of the high iron content region, because the area of the high iron content region is initially so small, its effect is overwhelmed by the total area or volume of the normal iron containing region, which can be seen comparing Fig. 2a and 2e. After the age of 40 years old, the area of high iron content begins to increase more rapidly (Fig. 2c).
We found some phase and magnitude related measures of iron content showed significant differences between the two hemispheres. Most of the time, the left hemisphere contains more iron than the right hemisphere (Tables 4 and 5) as evidenced by the number of subjects with higher iron in the left brain versus the right. Statistically, the PUT, CN, GP and RN showed the largest effect. In addition, from Table 2, it appears that the slope of the linear fit of these measurements in the left hemisphere was always higher than that in the right hemisphere except for the PT and SN.
Table 4.
Comparison between the two hemispheres of phase related iron measures
| Iron-Phase-RII | Average iron-Phase-RII | Iron-Phase-Total | Average iron-Phase-Total | Area-Phase-RII | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean (rad×mm3) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (10−3 rad) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (rad×mm3) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (10−3 rad) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (10−3%) |
SD | Number of L>R, L=R, L<R |
P value |
|
| PUT | −3.18 | 19.14 | 40/14/46 | 0.09 | 2.11 | 44.43 | 49/16/35 | 0.64 | −2.59 | 27.25 | 44/0/56 | 0.33 | 0.76 | 20.12 | 50/0/50 | 0.70 | −0.77% | 4.98% | 41/14/45 | 0.12 |
| CN | 3.54 | 13.71 | 61/0/39 | 0.01 | 0.66 | 26.39 | 51/0/49 | 0.8 | 7.28 | 14.84 | 70/0/30 | <0.01 | 7.19 | 19.26 | 59/5/36 | <0.01 | 1.28% | 6.56% | 60/0/40 | 0.05 |
| GP | 0.87 | 26.76 | 54/0/46 | 0.75 | −5.68 | 52.52 | 47/6/47 | 0.29 | 4.81 | 26.24 | 58/0/42 | 0.07 | 4.83 | 26.78 | 57/3/40 | 0.07 | 0.68% | 5.99% | 55/0/45 | 0.26 |
| RN | 1.13 | 3.19 | 60/18/22 | <0.01 | 6.58 | 19.34 | 51/22/27 | <0.01 | 0.73 | 5.18 | 55/0/45 | 0.16 | 5.56 | 25.95 | 59/1/40 | 0.03 | 1.67% | 4.86% | 56/20/24 | <0.01 |
| PT | −1.36 | 4.34 | 34/0/66 | <0.01 | −1.20 | 10.57 | 42/3/55 | 0.27 | −1.06 | 5.90 | 40/0/60 | 0.08 | −6.58 | 15.56 | 31/3/66 | <0.01 | −2.68% | 6.60% | 35/0/65 | <0.01 |
| SN | −0.89 | 5.76 | 35/10/55 | 0.12 | −2.54 | 28.24 | 37/13/50 | 0.32 | −3.24 | 10.13 | 40/0/60 | <0.01 | −6.21 | 33.18 | 45/1/54 | 0.07 | −0.49% | 3.82% | 36/10/54 | 0.19 |
| THA | 0.65 | 12.20 | 53/0/47 | 0.61 | 0.13 | 6.76 | 40/8/52 | 0.86 | −0.72 | 14.40 | 48/0/52 | 0.63 | −0.44 | 7.06 | 39/8/53 | 0.55 | −0.17% | 4.04% | 46/0/54 | 0.68 |
PUT: Putamen, CN: Caudate Nucleus, GP: Globus Pallidus, RN: Red Nucleus, PT: Pulvinar Thalamus, SN: Substantia Nigra, THA: Thalamus, Mean: mean of the two hemispheres difference from left minus right hand side, SD: standard deviation, L>R: the number of subjects that show larger value in the left hemisphere. L=R: the number of subjects that show equal values in both hemispheres. L<R: the number of subjects that show larger value in the right hemisphere.
Table 5.
Comparison between the two hemispheres of magnitude related iron measures
| Iron-Magnitude-RII | Average iron-Magnitude-RII | Iron-Magnitude-Total | Average iron-Magnitude-Total | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean (sec−1×mm3) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (10−3ms−1) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (sec−1×mm3) |
SD | Number of L>R, L=R, L<R |
P value |
Mean (10−3ms−1) |
SD | Number of L>R, L=R, L<R |
P value |
|
| PUT | 0.42 | 1.02 | 70/0/30 | <0.01 | 0.07 | 0.68 | 50/14/36 | 0.3 | 0.51 | 1.23 | 70/0/30 | <0.01 | 0.65 | 0.83 | 76/7/17 | <0.01 |
| CN | 0.46 | 0.82 | 80/2/18 | <0.01 | 0.17 | 0.48 | 55/20/25 | <0.01 | 0.66 | 0.80 | 85/100 | <0.01 | 0.81 | 0.93 | 83/100 | <0.01 |
| GP | 0.29 | 1.13 | 60/5/35 | 0.01 | 0.04 | 1.60 | 51/16/33 | 0.81 | 0.57 | 1.41 | 70/0/30 | <0.01 | 0.55 | 1.20 | 66/5/29 | <0.01 |
| RN | 0.02 | 0.25 | 58/3/39 | 0.43 | 0.11 | 0.62 | 49/21/30 | 0.08 | 0.00 | 0.21 | 54/0/46 | 0.97 | 0.17 | 1.01 | 59/7/34 | 0.09 |
| PT | 0.04 | 0.30 | 47/4/49 | 0.23 | −0.03 | 0.53 | 41/17/42 | 0.6 | 0.05 | 0.36 | 55/0/45 | 0.17 | −0.03 | 0.79 | 44/9/47 | 0.76 |
| SN | 0.08 | 0.31 | 55/0/45 | <0.01 | 0.18 | 0.83 | 48/19/33 | 0.03 | 0.20 | 0.54 | 68/0/32 | <0.01 | 0.67 | 1.53 | 62/10/28 | <0.01 |
| THA | 0.20 | 1.47 | 50/0/50 | 0.08 | 0.06 | 0.45 | 43/28/29 | 0.12 | 0.20 | 1.30 | 52/0/48 | 0.13 | 0.07 | 0.57 | 43/13/44 | 0.3 |
PUT: Putamen, CN: Caudate Nucleus, GP: Globus Pallidus, RN: Red Nucleus, PT: Pulvinar Thalamus, SN: Substantia Nigra, THA: Thalamus, Mean: mean of the two hemispheres difference from left minus right hand side, SD: standard deviation, L>R: the number of subjects that show larger value in the left hemisphere. L=R: the number of subjects that show equal values in both hemispheres. L<R: the number of subjects that show larger value in the right hemisphere.
Discussion
It is well known that iron deposition in the brain increases with age (9,10,18,25,31). The work by Hallgren and Sourander has been considered the standard reference for brain iron quantification (10) presenting a wide belief that iron in the human brain levels off after a certain age (30-40 years old) for many structures. For instance, many studies have shown that iron levels are low in newborn brain and quickly increase during the rapid brain growth period which concurs with the peak of myelinogenesis (6). By the age of 20 to 30 years, it is believed that it then reaches the maximum values (10,32). However, this was not seen in all brain structures. Hallgren’s and Sourander’s work (10) also suggests that iron continues to increase in the putamen and caudate nucleus over time and reaches a maximum value around the age of 60 years old. On the other hand, the iron level in the thalamus reaches its peak at the age of 30 years old, and then it appears to decline after that age.
Recent MR SWI studies used phase information to evaluate iron content as a function of age in the human brain (25,28,33). For instance, Xu et al. (25) showed that iron uptake remained age dependent in the putamen, red nucleus and frontal white matter but not in the GP, SN and caudate nucleus where, after a rapid increase in the first two decades, iron in these regions reaches a plateau. However, they did not consider breaking the structures into low and high iron content regions.
Our work introduces the concept of a high iron content region and normal iron content region. This makes it possible to study not only smaller increases in iron content more confidently but also to study both the average iron content per pixel and the area of increased iron content with age. Our results agree in part with these previous studies but also reveal new information. Overall we found that our conclusions from the magnitude images were mimicked by phase as well. We discuss each of the structures below relative to this new approach of examining low and high iron content regions.
Putamen
Our results show that iron content in the putamen remained age dependent throughout the 5 decades we studied and are in agreement with Xu et al.’s results and Aquino et al’s results who show a slow rate of increase and a variability after the age of 50 (34). This partly agrees with Hallgren’s and Sourander’s work (10) which showed that iron in the putamen was age dependent and did not reach its maximum until the 6th decade. On the other hand, the small area of increased iron for most structures in the early decades as seen in our work and its rapid increase after the age of 40 years is in stark contrast to the previous held belief that iron content increases rapidly only in the first 2 decades of life. We found that both the volume of the affected tissue with high iron content and the average iron content per pixel increases with age and in some cases dramatically changes the rate of iron deposition after the age of 40 years, not just the putamen but also in the caudate nucleus, globus pallidus and pulvinar thalamus.
Red nucleus
Our results show that the iron content in the red nucleus also remained age dependent throughout the 5 decades we studied and these too are in agreement with Xu et al.’s results. However, Hallgren and Sourander’s work showed a rapid increase in the first 2 decades, whereas no conclusions were drawn regarding iron in the red nucleus after the age of 20 due to technical difficulties (10).
Globus pallidus
In several studies (10,34), iron in the globus pallidus was shown to increase rapidly and almost reach constant value after the age of 30 years old, whereas our results as well as Xu et al.’s results show a linear decrease as a function of the age range studied. From Table 3, it can be seen that the iron concentration in region II grows very fast, but the iron concentration in the whole structure grows slowly. This has not been seen as well with other studies because of the larger volume of the normal iron containing region. Thus, the two region analysis is necessary for detecting brain iron changes in this structure (as well as all other structures).
Substantia nigra
None of these studies (including our study) showed significant changes in SN iron content as a function of age. However, Hallgren and Sourander showed a rapid increase in the first 2 decades and inconclusive results for older ages. Aquino et al. (34) showed a rapid increase until the age of 15, and then the iron content reached a plateau. No correlation with our study can be drawn since the first 2 decades of life data were not available.
Caudate nucleus
In Hallgren and Sourander’s study (10), the iron in the caudate nucleus was seen to behave in a similar fashion to the putamen, showing a slow increase reaching a maximum value at age 60 years old. Xu et al. showed that iron in the caudate nucleus increased slightly with age. Although we also see an increase in iron, where we found that iron increased slowly in the early ages, our results indicate a more rapid rise in iron deposition after the age of 40 years old (Fig. 9e, f). On the other hand, Aquino et al. (34) showed almost a flat curve for iron as a function of age.
Thalamus
For our results, the behavior of iron content in the thalamus was in distinct contradiction to these two studies (10,25). In Hallgren and Sourander, iron in the thalamus was shown to increase in the early stages, it reached its maximum at the age of 30, and then it started declining after that age. Xu et al. (25) showed that iron in thalamus is almost constant with respect to age. However, our results showed that iron in this structure increases linearly with age.
Pulvinar thalamus
The pulvinar thalamus has not yet been analyzed by others for its iron content. We chose this structure specifically because of new results indicating that iron in the pulvinar thalamus increased significantly in patients with multiple sclerosis (unpublished work). We found that iron concentration increased linearly with age for normal subjects. Establishing this baseline is critical to draw any future conclusions to what is abnormal iron content in assessing patients with multiple sclerosis.
Iron quantification: T2* versus phase
As discussed in the introduction, both T2* and phase have been used to study iron in the brain(15,16,25). Several studies where these two approaches (T2* and phase methods) were used showed good agreement in some structures, but not such good agreement in other structures(15,16,34). Which approach is better likely depends on the amount of iron present and possibly some further complications related to the mechanism of signal loss in T2* and the generation of phase for microscopic structures. The manifestation of the signal changes in each approach is quite different.
For large amounts of iron where strong dephasing across a voxel occurs, changes in iron content will cause the appropriate decreases in T2*. However, for smaller amounts of iron where phase dephasing effects across a voxel are small, there may be little noticeable effect on T2* even in the presence of what could be physiologically a dangerously high level of iron. On the other hand, phase is known to be very sensitive to subtle changes in magnetic field or frequency shift and therefore is much more sensitive to iron in smaller concentrations. As the amount of iron increases, phase and magnitude information start to correlate more strongly showing an almost linear correlation. However, when the iron content is too large, phase aliasing occurs and the linear correlation again disappears. This problem may be avoidable by using a multiple echo SWI sequence in the future (35,36) and using the short echo data to unwrap the long echo data. From this discussion, the following conclusion can be drawn: T2* underestimates iron content when the amount of iron is small, whereas phase underestimates iron content if iron content is too high.
Converting phase and magnitude measures into the amount of iron in tissue
In order to make these results available for comparison with other papers, Haacke et al (27) suggested that one could correlate the measured phase changes in Siemens Phase Units (SPU) to μg Fe/g tissue (18,27). Moreover, phase changes can also be expressed in terms of parts per million (ppm) and radians. The conversion factors between these units are provided below for a left-handed system (for a right handed system Δφ becomes – Δφ) :
| [3] |
| [4] |
| [5] |
Equation 5 is derived by taking 180 SPU to be equivalent to 60 μg Fe/g tissue (18). However, based on the T2* changes, our results suggest that a phase change of 180 SPU (0.276 radians) could represent as much as 240 μg Fe/g fresh tissue. If this number is used, and we assume a density of 1g/cm3 for tissue, then the total iron per slice can be calculated as follows:
| [6] |
For the magnitude prediction of iron content, we suggest using 67/s for ΔR2* corresponding to 1mg Fe/g tissue (18,27).
This study has a number of limitations. First, besides iron, there are also some other elements that can contribute to phase changes such as calcium(37), magnesium, copper and zinc. However, they would tend to reduce the effects of iron. Also, we cannot separate these effects from those of iron especially if these other elements are randomly mixed in with the iron. Second, since there was only a single gradient echo acquired, it was not possible to calculate an R2* map or to create directly a ΔR2* map. Therefore, we had to assume that we could use the original signal from a younger population as a means to normalize the data and create an effective ΔR2* map. These variations might account for some of the spread of ΔR2* as a function of age. Although this appears to have worked reasonably well, there can be problems associated with confounding factors such rf coil loading, T1 values and water content from person to person. Third, a limitation on drawing further conclusions regarding the effects of iron changes with age come from the fact that the age range was only from 20 year-olds to 69 year-olds. Including younger and older subjects would have been of interest as well.
In summary, both SWI magnitude and phase data provide a new approach to measuring brain putative iron as a function of age. When a two compartment model (low iron and high iron regions) is used, significant changes in putative iron content for the high iron content region as a function of age are seen, particularly for subjects over 40 years old. The high iron content region shows major changes with age especially in the putamen, pulvinar thalamus, caudate nucleus and globus pallidus. Also the normalized area, the average iron changes and the product of both provide further information on the amount of iron deposited and its spatial distribution. The data presented here may serve as a means to assess normal versus abnormal iron content in a variety of iron-related diseases.
Acknowledgements
This work was supported in part by National Science Foundation of China (No. 30870699) and National Institutes of Health Award Number NHLBI R01 HL62983. The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, And Blood Institute or the National Institutes of Health.”
Grant Support: This work was supported in part by National Science Foundation of China (No. 30870699) and National Institutes of Health Award Number NHLBI R01HL062983. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, And Blood Instituteor the National Institutes of Health.”
References
- 1.Quintana C, Bellefqih S, Laval JY, et al. Study of the localization of iron, ferritin, and hemosiderin in Alzheimer’s disease hippocampus by analytical microscopy at the subcellular level. J Struct Biol. 2006;153(1):42–54. doi: 10.1016/j.jsb.2005.11.001. [DOI] [PubMed] [Google Scholar]
- 2.Stankiewicz J, Panter SS, Neema M, Arora A, Batt CE, Bakshi R. Iron in chronic brain disorders: imaging and neurotherapeutic implications. Neurotherapeutics. 2007;4(3):371–386. doi: 10.1016/j.nurt.2007.05.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Haacke EM, Makki M, Ge Y, et al. Characterizing iron deposition in multiple sclerosis lesions using susceptibility weighted imaging. J Magn Reson Imaging. 2009;29(3):537–544. doi: 10.1002/jmri.21676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Schenck JF, Zimmerman EA. High-field magnetic resonance imaging of brain iron: birth of a biomarker? NMR Biomed. 2004;17(7):433–445. doi: 10.1002/nbm.922. [DOI] [PubMed] [Google Scholar]
- 5.Connor JR, Snyder BS, Arosio P, Loeffler DA, LeWitt P. A quantitative analysis of isoferritins in select regions of aged, parkinsonian, and Alzheimer’s diseased brains. J Neurochem. 1995;65(2):717–724. doi: 10.1046/j.1471-4159.1995.65020717.x. [DOI] [PubMed] [Google Scholar]
- 6.Beard JL, Connor JR, Jones BC. Iron in the brain. Nutr Rev. 1993;51(6):157–170. doi: 10.1111/j.1753-4887.1993.tb03096.x. [DOI] [PubMed] [Google Scholar]
- 7.Connor JR, Menzies SL. Cellular management of iron in the brain. J Neurol Sci. 1995;134(Suppl):33–44. doi: 10.1016/0022-510x(95)00206-h. [DOI] [PubMed] [Google Scholar]
- 8.Koeppen AH. The history of iron in the brain. J Neurol Sci. 1995;134(Suppl):1–9. doi: 10.1016/0022-510x(95)00202-d. [DOI] [PubMed] [Google Scholar]
- 9.Ogg RJ, Steen RG. Age-related changes in brain T1 are correlated with iron concentration. Magn Reson Med. 1998;40(5):749–753. doi: 10.1002/mrm.1910400516. [DOI] [PubMed] [Google Scholar]
- 10.Hallgren B, Sourander P. The effect of age on the non-haemin iron in the human brain. J Neurochem. 1958;3(1):41–51. doi: 10.1111/j.1471-4159.1958.tb12607.x. [DOI] [PubMed] [Google Scholar]
- 11.Tingey AH. The iron content of the human brain part 2. J Ment Sci. 1938;84:980–984. [Google Scholar]
- 12.Tingey AH. The iron, copper and manganese content of the human brain. J Ment Sci. 1937;83:452–460. [Google Scholar]
- 13.Morris CM, Candy JM, Oakley AE, Bloxham CA, Edwardson JA. Histochemical distribution of non-haem iron in the human brain. Acta Anat (Basel) 1992;144(3):235–257. doi: 10.1159/000147312. [DOI] [PubMed] [Google Scholar]
- 14.Morris CM, Kerwin JM, Edwardson JA. Non-haem iron histochemistry of the normal and Alzheimer’s disease hippocampus. Neurodegeneration. 1994;3(4):267–275. [PubMed] [Google Scholar]
- 15.Gelman N, Gorell JM, Barker PB, et al. MR imaging of human brain at 3.0 T: preliminary report on transverse relaxation rates and relation to estimated iron content. Radiology. 1999;210(3):759–767. doi: 10.1148/radiology.210.3.r99fe41759. [DOI] [PubMed] [Google Scholar]
- 16.Ordidge RJ, Gorell JM, Deniau JC, Knight RA, Helpern JA. Assessment of relative brain iron concentrations using T2-weighted and T2*-weighted MRI at 3 Tesla. Magn Reson Med. 1994;32(3):335–341. doi: 10.1002/mrm.1910320309. [DOI] [PubMed] [Google Scholar]
- 17.Bartzokis G, Tishler TA, Lu PH, et al. Brain ferritin iron may influence age- and gender-related risks of neurodegeneration. Neurobiol Aging. 2007;28(3):414–423. doi: 10.1016/j.neurobiolaging.2006.02.005. [DOI] [PubMed] [Google Scholar]
- 18.Haacke EM, Cheng NY, House MJ, et al. Imaging iron stores in the brain using magnetic resonance imaging. Magn Reson Imaging. 2005;23(1):1–25. doi: 10.1016/j.mri.2004.10.001. [DOI] [PubMed] [Google Scholar]
- 19.Schenker C, Meier D, Wichmann W, Boesiger P, Valavanis A. Age distribution and iron dependency of the T2 relaxation time in the globus pallidus and putamen. Neuroradiology. 1993;35(2):119–124. doi: 10.1007/BF00593967. [DOI] [PubMed] [Google Scholar]
- 20.Siemonsen S, Finsterbusch J, Matschke J, Lorenzen A, Ding XQ, Fiehler J. Age-dependent normal values of T2* and T2′ in brain parenchyma. AJNR Am J Neuroradiol. 2008;29(5):950–955. doi: 10.3174/ajnr.A0951. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Haacke EM, Xu Y, Cheng YC, Reichenbach JR. Susceptibility weighted imaging (SWI) Magn Reson Med. 2004;52(3):612–618. doi: 10.1002/mrm.20198. [DOI] [PubMed] [Google Scholar]
- 22.Reichenbach JR, Venkatesan R, Schillinger DJ, Kido DK, Haacke EM. Small vessels in the human brain: MR venography with deoxyhemoglobin as an intrinsic contrast agent. Radiology. 1997;204(1):272–277. doi: 10.1148/radiology.204.1.9205259. [DOI] [PubMed] [Google Scholar]
- 23.McCrea RP, Harder SL, Martin M, Buist R, Nichol H. A comparison of rapid-scanning X-ray fluorescence mapping and magnetic resonance imaging to localize brain iron distribution. Eur J Radiol. 2008;68(3 Suppl):S109–113. doi: 10.1016/j.ejrad.2008.04.048. [DOI] [PubMed] [Google Scholar]
- 24.Harder SL, Hopp KM, Ward H, Neglio H, Gitlin J, Kido D. Mineralization of the deep gray matter with age: a retrospective review with susceptibility-weighted MR imaging. AJNR Am J Neuroradiol. 2008;29(1):176–183. doi: 10.3174/ajnr.A0770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Xu X, Wang Q, Zhang M. Age, gender, and hemispheric differences in iron deposition in the human brain: an in vivo MRI study. Neuroimage. 2008;40(1):35–42. doi: 10.1016/j.neuroimage.2007.11.017. [DOI] [PubMed] [Google Scholar]
- 26.Wang Y, Yu Y, Li D, et al. Artery and vein separation using susceptibility-dependent phase in contrast-enhanced MRA. J Magn Reson Imaging. 2000;12(5):661–670. doi: 10.1002/1522-2586(200011)12:5<661::aid-jmri2>3.0.co;2-l. [DOI] [PubMed] [Google Scholar]
- 27.Haacke EM, Ayaz M, Khan A, et al. Establishing a baseline phase behavior in magnetic resonance imaging to determine normal vs. abnormal iron content in the brain. J Magn Reson Imaging. 2007;26(2):256–264. doi: 10.1002/jmri.22987. [DOI] [PubMed] [Google Scholar]
- 28.Ogg RJ, Langston JW, Haacke EM, Steen RG, Taylor JS. The correlation between phase shifts in gradient-echo MR images and regional brain iron concentration. Magn Reson Imaging. 1999;17(8):1141–1148. doi: 10.1016/s0730-725x(99)00017-x. [DOI] [PubMed] [Google Scholar]
- 29.Faber C, Heil C, Zahneisen B, Balla DZ, Bowtell R. Sensitivity to local dipole fields in the CRAZED experiment: an approach to bright spot MRI. J Magn Reson. 2006;182(2):315–324. doi: 10.1016/j.jmr.2006.05.002. [DOI] [PubMed] [Google Scholar]
- 30.Miao Y, Liu T, Wu JL. Correlation Study of Susceptibility Weighted Imaging (SWI) Study on Deep Brain Gray Nuclei and Age. Chicago: 2008. [Google Scholar]
- 31.Bartzokis G, Mintz J, Sultzer D, et al. In vivo MR evaluation of age-related increases in brain iron. AJNR Am J Neuroradiol. 1994;15(6):1129–1138. [PMC free article] [PubMed] [Google Scholar]
- 32.Goodman L. Alzheimer’s disease; a clinico-pathologic analysis of twenty-three cases with a theory on pathogenesis. J Nerv Ment Dis. 1953;118(2):97–130. [PubMed] [Google Scholar]
- 33.Wismer GL, Buxton RB, Rosen BR, et al. Susceptibility induced MR line broadening: applications to brain iron mapping. J Comput Assist Tomogr. 1988;12(2):259–265. doi: 10.1097/00004728-198803000-00014. [DOI] [PubMed] [Google Scholar]
- 34.Aquino D, Bizzi A, Grisoli M, et al. Age-related iron deposition in the basal ganglia: quantitative analysis in healthy subjects. Radiology. 2009;252(1):165–172. doi: 10.1148/radiol.2522081399. [DOI] [PubMed] [Google Scholar]
- 35.Deistung A, Dittrich E, Sedlacik J, Rauscher A, Reichenbach JR. ToF-SWI: simultaneous time of flight and fully flow compensated susceptibility weighted imaging. J Magn Reson Imaging. 2009;29(6):1478–1484. doi: 10.1002/jmri.21673. [DOI] [PubMed] [Google Scholar]
- 36.Du YP, Jin Z, Hu Y, Tanabe J. Multi-echo acquisition of MR angiography and venography of the brain at 3 Tesla. J Magn Reson Imaging. 2009;30(2):449–454. doi: 10.1002/jmri.21833. [DOI] [PubMed] [Google Scholar]
- 37.Haacke EM, Mittal S, Wu Z, Neelavalli J, Cheng YC. Susceptibility-weighted imaging: technical aspects and clinical applications, part 1. AJNR Am J Neuroradiol. 2009;30(1):19–30. doi: 10.3174/ajnr.A1400. [DOI] [PMC free article] [PubMed] [Google Scholar]