Limits...
A comparison of MRI tissue relaxometry and ROI methods used to determine regional brain iron concentrations in restless legs syndrome.

Moon HJ, Chang Y, Lee YS, Song H, Chang HW, Ku J, Allen RP, Earley CJ, Cho YW - Med Devices (Auckl) (2015)

Bottom Line: These differences could account for the variable and inconsistent results found across these studies.The iron concentration as determined by R2 did not correlate with either of the other two methods, while R2* and R2' showed strong correlations, particularly for the substantia nigra and red nucleus.Both the relaxometry and ROI determination methods significantly influenced the outcome of studies that used these methods to estimate regional brain iron concentrations.

View Article: PubMed Central - PubMed

Affiliation: Department of Neurology, Dongsan Medical Center, Keimyung University School of Medicine, Daegu, Republic of Korea.

ABSTRACT

Purpose: Magnetic resonance imaging relaxometry studies differed on the relaxometry methods and their approaches to determining the regions of interest (ROIs) in restless legs syndrome (RLS) patients. These differences could account for the variable and inconsistent results found across these studies. The aim of this study was to assess the relationship between the different relaxometry methods and different ROI approaches using each of these methods on a single population of controls and RLS subjects.

Methods: A 3.0-T magnetic resonance imaging with the gradient-echo sampling of free induction decay and echo pulse sequence was used. The regional brain "iron concentrations" were determined using three relaxometry metrics (R2, R2*, and R2') through two different ROI methods. The substantia nigra (SN) was the primary ROI with red nucleus, caudate, putamen, and globus pallidus as the secondary ROIs.

Results: Thirty-seven RLS patients and 40 controls were enrolled. The iron concentration as determined by R2 did not correlate with either of the other two methods, while R2* and R2' showed strong correlations, particularly for the substantia nigra and red nucleus. In the fixed-shape ROI method, the RLS group showed a lower iron index compared to the control group in the substantia nigra and several other regions. With the semi-automated ROI method, however, only the red nucleus showed a significant difference between the two groups.

Conclusion: Both the relaxometry and ROI determination methods significantly influenced the outcome of studies that used these methods to estimate regional brain iron concentrations.

No MeSH data available.


Related in: MedlinePlus

Correlations between three relaxometry metrics for substantia nigra in each ROI method.Notes: (A–C) Correlations between different metrics in fixed-shape ROI method. (D–F) Correlations between different metrics in semi-automated tracing method.Abbreviation: ROI, region of interest.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4525804&req=5

f2-mder-8-341: Correlations between three relaxometry metrics for substantia nigra in each ROI method.Notes: (A–C) Correlations between different metrics in fixed-shape ROI method. (D–F) Correlations between different metrics in semi-automated tracing method.Abbreviation: ROI, region of interest.

Mentions: The variance accounted for by the linear regression (r2) between the three relaxometry metrics for the SN is shown in Figure 2 and the values for the other four brain regions are shown in Table 2. Although R2 and R2* are purported to be valid measures of tissue iron concentration, the determinant of the coefficient indicates that minimal, if any, relationship exists between these two estimated iron concentration values irrespective of the methods used to define the ROI (Figure 2A: r2=0.13 and Figure 2D: r2=0.01). Although R2′ is determined by the difference between R2* and R2, the correlation between SN values generated by R2′ and R2 techniques were no better than random or chance associations (Figure 2B: r2=0.03 and Figure 2E: r2=0.02). Unlike the relationship between R2 and R2′, R2′ and R2* showed a reasonably close agreement for both the fixed-shape (Figure 2C: r2=0.53, P<0.001) and the semi-automated tracing (Figure 2F: r2=0.66, P<0.001) ROI methods. This would indicate that most of the determinants of R2′ are derived from R2* when used to estimate the iron concentration in the SN. Of the other ROIs (Table 2), the red nucleus using the semi-automated tracing ROI methods showed the strongest R2* to R2′ correlation (r2 = 0.89, P<0.001) and the globus pallidus using the fixed-shape ROI method showed the weakest, though still statistically significant, correlation (r2=0.35, P<0.001). Overall, regardless of the ROI or the method used to determine that ROI, there were significant value agreements between R2′ and R2* with the strongest correlations generally being seen with the semi-automated tracing ROI method.


A comparison of MRI tissue relaxometry and ROI methods used to determine regional brain iron concentrations in restless legs syndrome.

Moon HJ, Chang Y, Lee YS, Song H, Chang HW, Ku J, Allen RP, Earley CJ, Cho YW - Med Devices (Auckl) (2015)

Correlations between three relaxometry metrics for substantia nigra in each ROI method.Notes: (A–C) Correlations between different metrics in fixed-shape ROI method. (D–F) Correlations between different metrics in semi-automated tracing method.Abbreviation: ROI, region of interest.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4525804&req=5

f2-mder-8-341: Correlations between three relaxometry metrics for substantia nigra in each ROI method.Notes: (A–C) Correlations between different metrics in fixed-shape ROI method. (D–F) Correlations between different metrics in semi-automated tracing method.Abbreviation: ROI, region of interest.
Mentions: The variance accounted for by the linear regression (r2) between the three relaxometry metrics for the SN is shown in Figure 2 and the values for the other four brain regions are shown in Table 2. Although R2 and R2* are purported to be valid measures of tissue iron concentration, the determinant of the coefficient indicates that minimal, if any, relationship exists between these two estimated iron concentration values irrespective of the methods used to define the ROI (Figure 2A: r2=0.13 and Figure 2D: r2=0.01). Although R2′ is determined by the difference between R2* and R2, the correlation between SN values generated by R2′ and R2 techniques were no better than random or chance associations (Figure 2B: r2=0.03 and Figure 2E: r2=0.02). Unlike the relationship between R2 and R2′, R2′ and R2* showed a reasonably close agreement for both the fixed-shape (Figure 2C: r2=0.53, P<0.001) and the semi-automated tracing (Figure 2F: r2=0.66, P<0.001) ROI methods. This would indicate that most of the determinants of R2′ are derived from R2* when used to estimate the iron concentration in the SN. Of the other ROIs (Table 2), the red nucleus using the semi-automated tracing ROI methods showed the strongest R2* to R2′ correlation (r2 = 0.89, P<0.001) and the globus pallidus using the fixed-shape ROI method showed the weakest, though still statistically significant, correlation (r2=0.35, P<0.001). Overall, regardless of the ROI or the method used to determine that ROI, there were significant value agreements between R2′ and R2* with the strongest correlations generally being seen with the semi-automated tracing ROI method.

Bottom Line: These differences could account for the variable and inconsistent results found across these studies.The iron concentration as determined by R2 did not correlate with either of the other two methods, while R2* and R2' showed strong correlations, particularly for the substantia nigra and red nucleus.Both the relaxometry and ROI determination methods significantly influenced the outcome of studies that used these methods to estimate regional brain iron concentrations.

View Article: PubMed Central - PubMed

Affiliation: Department of Neurology, Dongsan Medical Center, Keimyung University School of Medicine, Daegu, Republic of Korea.

ABSTRACT

Purpose: Magnetic resonance imaging relaxometry studies differed on the relaxometry methods and their approaches to determining the regions of interest (ROIs) in restless legs syndrome (RLS) patients. These differences could account for the variable and inconsistent results found across these studies. The aim of this study was to assess the relationship between the different relaxometry methods and different ROI approaches using each of these methods on a single population of controls and RLS subjects.

Methods: A 3.0-T magnetic resonance imaging with the gradient-echo sampling of free induction decay and echo pulse sequence was used. The regional brain "iron concentrations" were determined using three relaxometry metrics (R2, R2*, and R2') through two different ROI methods. The substantia nigra (SN) was the primary ROI with red nucleus, caudate, putamen, and globus pallidus as the secondary ROIs.

Results: Thirty-seven RLS patients and 40 controls were enrolled. The iron concentration as determined by R2 did not correlate with either of the other two methods, while R2* and R2' showed strong correlations, particularly for the substantia nigra and red nucleus. In the fixed-shape ROI method, the RLS group showed a lower iron index compared to the control group in the substantia nigra and several other regions. With the semi-automated ROI method, however, only the red nucleus showed a significant difference between the two groups.

Conclusion: Both the relaxometry and ROI determination methods significantly influenced the outcome of studies that used these methods to estimate regional brain iron concentrations.

No MeSH data available.


Related in: MedlinePlus