Nigrosome-1 has been identified as a swallow-tail sign with a small and hyperintense region in the substantia nigra (SN) on T2*-weighted or susceptibility-weighted imaging (SWI) sequences [1
]. Nigrosome-1 abnormalities such as patients with Parkinson’s disease (PD) and Lewy body dementia exhibit an increased iron deposition in the pars compacta of the substantia nigra (SNpc) and the feature of swallow-tail sign is lost [2
]. However, most reports have focused on the studies in patients with PD, few studies have evaluated the swallow-tail sign in healthy subjects. For swallow-tail sign imaging to become a useful clinical tool, a reliable technique which is reproducible and easy to use in the clinical environment is required to test the performance in healthy subjects before the tool can be recommended. This assessment can allow accurate delineation of the swallow-tail sign to improve the accuracy of tracking disease severity-related changes.
In addition to a single-echo sequence, SWI data can be acquired using a multiecho gradient-echo sequence, which can provide increased flexibility in selecting TEs, leading to an increase in the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) [4
]. T2* and quantitative susceptibility mapping (QSM) have been reported to be sensitive in detecting the SN iron content when evaluating the pathophysiological processes such as PD progression [1
]. A multiecho gradient-echo sequence which allows the diagnostic gain of T2* and QSM can thus be a potential prerequisite for clinical acceptance. However, most SWI examinations in clinical use are still confined to the conventional single-echo gradient-echo method.
T1-weighted neuromelanin-sensitive MRI (NM-MRI) can reflect the changes of neuromelanin and has shown the potential to detect alterations to SNpc morphology for PD [6
]. The SN hyperintensity in the NM-MRI has been associated with the paramagnetic T1-shortening and magnetization transfer (MT) effects by the melanin-iron complex [7
]. The presence of iron in this complex can also affect the transverse relaxation time [6
]. However, limited data exist describing the relationship between hyperintensities of SNpc in the NM-MRI and its MR relaxation signals for healthy subjects.
In this study, we evaluated the multiecho SWI technique in the delineation of the swallow-tail sign through CNR measurements for healthy subjects. To investigate the possible association between the susceptibility of iron content and neuromelanin signal in the NM-MRI, we conducted exploratory analyses of correlations between MR signal in the neuromelanin with mappings of T1, T2 and quantitative susceptibility.
summarizes the demographic information of the study cohorts. There is a significant age difference between two groups (F = 129.164, p
< 0.0001). No other demographic differences were observed. Figure 3
depicts the representative SWI comparison between single-echo and multiecho SWI techniques. The phase variation artifacts in the frontal and temporal lobe areas were successfully improved using the multiecho SWIave
technique. The SWIave
technique further exhibited a superior imaging quality in visualizing the optic nerve. Figure 4
illustrates a representative slice from a single subject in the native space of the QSM, T2*, SWIave
, and SWI images with their corresponding phase and magnitude images acquired from multiecho and single-echo sequences, respectively. Table 2
lists the results of the CNRs obtained using the different imaging methods. The corresponding quantitative measurements of the phase, T2*, and QSM in nigrosome-1 and neighboring SN area for each subgroup are presented in Table 3
. Figure 5
shows a representative slice from a subject in the native space of the NM-MRI, QSM, T1 and T2 mappings. The mean relaxation times were T1 = 1226.99 ± 221.17 ms and T2 = 81.18 ± 5.63 ms for the nigrosome-1; T1 = 1230.81 ± 238.56 ms and T2 = 75.79 ± 7.43 ms for the neighboring SN area, respectively. We further attached the imaging of multiecho SWIave
for a patient with PD for comparison (Figure 6
). It is obvious that the swallow tail sign is absent for this PD patient as expected.
Linear regression analyses comparing the CNR values obtained between the two groups using different methods indicated the considerable influence of the selected method (F = 902.896, p
< 0.0001) and the group (F = 30.076, p
< 0.0001). Post-hoc pairwise comparisons indicated that the older group exhibited a considerably higher swallow-tail CNR than the young group when using the SWIave
= 0.003), Magave
= 0.007), SWI (p
= 0.004), Mag (p
= 0.005), and T2*(p
= 0.008) methods. SWIave
demonstrated the most significant group difference as compared with other methods. For both the older and young groups, the SWIave
method generated the highest CNR. On average, the SWI method exhibited a higher CNR of the swallow-tail sign compared with the corresponding magnitude, phase, T2*, and QSM methods (Table 2
). Moreover, we found that compared with the young subjects, the older subjects exhibited considerably higher mean values of the quantitative measurements of QSM in the nigrosome-1 (p
< 0.001) and neighboring substantia nigra (p
< 0.001) (Table 3
). For the comparison of different coils, we did not find any significant difference in all of the CNR measurements between the 20- and 64-channel array coils (Figure 7
There were no correlations of NM-MRI signal with quantitative values of susceptibility and T2 in the nigrosome-1(p
> 0.05). However, significant correlations between neuromelanin signal in NM-MRI (normalized to cerebral crus) and CNRs in QSM (Rho = −0.60, p
= 0.003) and T2 mapping (Rho = 0.65, p
= 0.001) of the nigrosome-1 were identified. We also found a significant correlation between neuromelanin signal and T1 values in the nigrosome-1 (Rho = −0.60, p
= 0.003) (Figure 8
). An exploratory analysis further showed a significant correlation between QSM and T2 values in the nigrosome-1(Rho = −0.76, p
< 0.0005) (Figure 8
This study was conducted to assess the potential of using a simple and practical method of multiecho SWIave
in the delineation of the swallow-tail sign of nigrosome-1 and compare with the single-echo SWI which is the most commonly used in routine clinical practice [2
]. The major findings of this study are: first, compared with the conventional single-echo SWI, the simple method using multiecho SWIave
can improve the CNR of the swallow-tail sign; second, significant correlations between neuromelanin signal and CNRs in QSM and T2 mapping, and T1 value of the nigrosome-1 were identified; third, the older subjects exhibited increased CNR values compared with the young adults.
Seeking a novel approach through a multiecho gradient-echo sequence to map for the evaluation of nigrosome-1 has been carried out in various studies [12
]. MR transverse relaxation and QSM have also been employed to characterize iron deposition within the SN in the diagnosis of PD [20
]. QSM is theoretically more accurate in iron quantification and has attracted considerable interest as a potential means of quantifying neuronal loss and iron accumulation in the SNpc for patients with PD [21
]. However, our results demonstrated that SWIs yielded a superior and distinct CNR in visualizing the anatomical details of the swallow-tail sign in the diagnosis of the nigrosome-1 compared with QSM and T2*. QSM is more computationally expensive since it requires several processing steps, including phase unwrapping, brain extraction, background phase removal and overcoming the ill-posed problem [10
]. An accurate estimation of the QSM is thus essential for approaches using QSM [21
] or QSM-generated mask [18
] in detecting the swallow-tail sign. This may need computational expertise, especially given that QSM has not been widely used for clinical routine as compared with SWI so far. Nigrosome-1 is hyperintense in SWI because of its low iron content than the surrounding tissues [23
]. The multiecho SWIave
method in the present study is relatively simple and practical for clinical use with less computationally demanding and can provide much higher CNR than QSM (more than three times) and conventional single-echo SWI in detecting the swallow-tail sign. In a multiecho gradient-echo sequence, because several echoes can be collected within one TR, the acquisition time can be potentially not prolonged, while the SNR can increase [4
]. The considerably improved CNR with the SWIave
technique shown in the present study could be attributed to the increased SNR.
It is worth noting that both QSM and phase images displayed negative values of the CNR in nigrosome-1 compared with other imaging methods. These results provide evidence that nigrosome-1 in the SNpc is less paramagnetic than the surrounding tissues, such as substantia nigra pars reticulate (SNpr) in which the brain region appeared to have a high iron content. The less paramagnetic susceptibility of the nigrosome-1 than the surrounding tissues could be resulted from its known low iron content than the surrounding tissues as mentioned above [23
]. This is further supported by the observation that the neighboring SN area exhibited lower T2* values than the nigrosome-1 in the quantitative T2* data analysis in our study (Table 3
The source of NM-sensitive contrast has been suggested as a combination of MT and T1-shortening effects, in which the latter is associated with the paramagnetic nature of melanin-iron complex [24
]. The neuromelanin pigment has been reported to act as a T1-shortening agent when bound with iron molecules [7
]. Our observation of negative correlation between the NM-MRI signal and T1 values added support to this speculation (Figure 8
C). We found a significant correlation between NM-MRI signal and CNR of T2 mapping in the nigrosome-1 (Figure 8
B). It has been reported that T2 values tend to increase as the size of paramagnetic perturbers increases and melanin-iron complex is known to have a significantly larger effective size as compared with the ferric iron [26
]. The observed positive correlation between these two measures may be associated with the quantity of the melanin-iron complex-containing neurons in the nigrosome-1, in which the increased signal intensity of the NM-MRI and CNR in the T2 mapping could be attributed to the increased melanin-iron content in this brain region. This finding also coincides with the previous report that the presence of iron in the melanin-iron complex can affect the proton transverse relaxation time [6
], which is also supported by the result of significant correlation between susceptibility measurements (i.e., QSM) and T2 values of the melanin-iron complex in the nigrosome1 (Figure 8
D). Due to the problem of magic angle in QSM, the calculated susceptibility values are therefore relative rather than absolute quantification [27
]. We speculate that this quantification error along with the one in calculation of the absolute T2 values may lead to the limited sensitivity in correlating with the neuromelanin signal. Our finding may imply that CNR preserves the information of the relationship between neuromelanin signal, the content of paramagnetic susceptibility and T2 relaxation for the melanin-iron complex-containing neurons in the nigrosome-1. We also observed a significant correlation between NM-MRI signal and CNR of QSM in the nigrosome-1 (Figure 8
A). The correlation between these two measures implies that the content of paramagnetic susceptibility in the nigrosome-1 have significant influences on the hyperintensity of SNpc in the T1 and MT-related contrast of NM-MRI.
We demonstrated the increased CNR of the swallow-tail sign in older subjects. This result is concordant with the age-related iron deposits for older subjects [28
], especially in the SNpr, which is one of the brain areas that have high iron concentration. Exploratory measurements of QSM and T2* provide further evidence showing significantly higher paramagnetic susceptibility in the nigrosome-1 and the surrounding area in older subjects. The observed increased paramagnetic susceptibility of the nigrosome-1 in the older group was also corroborated by the known age effect of the gradual increase of NM [26
], which is known with the paramagnetic nature of melanin-iron complex.
We ran the scans of CNR measurements in the swallow-tail sign using a 20-channel coil that had similar results to 64-channels. Although a comparison of 96- and 32-channel arrays has been reported in a previous study [30
], no studies have compared the difference of swallow-tail sign between 64- and 20-channels in clinical practice. The more complementary information provided by the collection and analysis of different phased-array MR imaging data can enable more complete exploitation of current technologies using phased-array coils in clinical systems, especially for some clinical institutes using the 20-channel coil instead of 64-channel one as a routine practice. The comparison between the 64- and 20-channel coils revealed negligible differences of the CNR in the delineation of SNpc. It has been reported that the central SNR is nearly identical and the peripheral SNR is higher in the coil with a higher channel count compared with a lower channel one [30
]. The sensitivity in the center for accelerated acquisitions can be affected only for more highly accelerated imaging [31
]. We attributed our results to the similar SNR measurements in the central brain regions between different array coils because the acceleration (factor = 2) we used in the present study was not high.
Our study had limitations. First, even though the quality of registration was carefully checked by visual inspection with particular care to manually adjust the image registration if there is a presence of coregistration error, we cannot exclude possible confounds caused by the misregistration of the small size for nigrosome-1. Second, a 3D acquisition of several echoes with a relatively long scan time may still lead to practical limitations for subjects who may have difficulty maintaining a stable head position. Third, this study was conducted mainly in healthy subjects. Future studies of a prospective evaluation showing a comparison between subjects with PD and without PD are needed to investigate in detail and support the advantage of this technique.
In conclusion, we used a multiecho SWIave technique at 3T to demonstrate its advantages in obtaining the CNR and other quantitative information for delineating the nigrosome-1 of the swallow-tail sign. A multiecho gradient-echo sequence can also benefit from quantitative QSM and T2* measurements for assessing the brain iron content with the capabilities of SWIave. Our results of correlations between CNRs in QSM, T2 mapping and NM-MRI provide a new understanding of the contrast behavior of the nigrosome-1, indicating that using these techniques may provide more comprehensive information of the melanin-iron complex in nigrosome-1.