Limits...
Quantifying brain volumes for Multiple Sclerosis patients follow-up in clinical practice - comparison of 1.5 and 3 Tesla magnetic resonance imaging.

Lysandropoulos AP, Absil J, Metens T, Mavroudakis N, Guisset F, Van Vlierberghe E, Smeets D, David P, Maertens A, Van Hecke W - Brain Behav (2016)

Bottom Line: The results are obtained with MSmetrix, which is developed especially for use in the MS clinical care path, and compared to Siena (FSL), a widely used software for research purposes.When data of the same scanner are compared, the error is in the order of 0.06-0.08% for both MSmetrix and Siena.MSmetrix appears robust on both the 1.5T and 3T systems and the measurement error becomes an order of magnitude higher between scanners with different field strength.

View Article: PubMed Central - PubMed

Affiliation: Department of Neurology Hôpital Erasme Université Libre de Bruxelles Anderlecht Belgium.

ABSTRACT

Introduction: There is emerging evidence that brain atrophy is a part of the pathophysiology of Multiple Sclerosis (MS) and correlates with several clinical outcomes of the disease, both physical and cognitive. Consequently, brain atrophy is becoming an important parameter in patients' follow-up. Since in clinical practice both 1.5Tesla (T) and 3T magnetic resonance imaging (MRI) systems are used for MS patients follow-up, questions arise regarding compatibility and a possible need for standardization.

Methods: Therefore, in this study 18 MS patients were scanned on the same day on a 1.5T and a 3T scanner. For each scanner, a 3D T1 and a 3D FLAIR were acquired. As no atrophy is expected within 1 day, these datasets can be used to evaluate the median percentage error of the brain volume measurement for gray matter (GM) volume and parenchymal volume (PV) between 1.5T and 3T scanners. The results are obtained with MSmetrix, which is developed especially for use in the MS clinical care path, and compared to Siena (FSL), a widely used software for research purposes.

Results: The MSmetrix median percentage error of the brain volume measurement between a 1.5T and a 3T scanner is 0.52% for GM and 0.35% for PV. For Siena this error equals 2.99%. When data of the same scanner are compared, the error is in the order of 0.06-0.08% for both MSmetrix and Siena.

Conclusions: MSmetrix appears robust on both the 1.5T and 3T systems and the measurement error becomes an order of magnitude higher between scanners with different field strength.

No MeSH data available.


Related in: MedlinePlus

Visualization of MSmetrix segmentation results for the 1.5T (A–D) and 3T (E–H) scan of a randomly selected MS patient. T1 scans are shown in (A) and (E), FLAIR scans in (B) and (F). Lesion and GM segmentations are superimposed on the T1 scan, as displayed in (C) and (G). Finally, lesion and WM segmentations are visualized on the FLAIR scan in (D) and (H).
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4834931&req=5

brb3422-fig-0002: Visualization of MSmetrix segmentation results for the 1.5T (A–D) and 3T (E–H) scan of a randomly selected MS patient. T1 scans are shown in (A) and (E), FLAIR scans in (B) and (F). Lesion and GM segmentations are superimposed on the T1 scan, as displayed in (C) and (G). Finally, lesion and WM segmentations are visualized on the FLAIR scan in (D) and (H).

Mentions: In Figure 2, some visual results of the MSmetrix segmentations on a 1.5T and 3T scan of the same randomly selected subject are displayed. In Figure 2A and B, an axial slice of the 1.5T 3D T1 and 3D FLAIR are shown, respectively. For visualization purposes, the GM and lesion segmentation are visualized on the T1 (c) and the WM and lesion masks on the FLAIR (d). A similar slice was selected for the 3T scan, as shown in Figure 2E and F, for the 3D T1 and 3D FLAIR, respectively. Similar as in Figure 2C and D, the GM, lesions, and WM segmentations of the 3T MRI are displayed in Figure 2G and H. These lesion segmentations are then used to fill the 3D T1 with normal‐appearing white matter, as explained in Figure 1. The cross‐sectional brain tissue segmentations that are shown in Figure 2 will be used as input for the longitudinal pipeline, to calculate the Jacobian of the deformation fields between both scans, resulting in a measure of brain and GM PBVC.


Quantifying brain volumes for Multiple Sclerosis patients follow-up in clinical practice - comparison of 1.5 and 3 Tesla magnetic resonance imaging.

Lysandropoulos AP, Absil J, Metens T, Mavroudakis N, Guisset F, Van Vlierberghe E, Smeets D, David P, Maertens A, Van Hecke W - Brain Behav (2016)

Visualization of MSmetrix segmentation results for the 1.5T (A–D) and 3T (E–H) scan of a randomly selected MS patient. T1 scans are shown in (A) and (E), FLAIR scans in (B) and (F). Lesion and GM segmentations are superimposed on the T1 scan, as displayed in (C) and (G). Finally, lesion and WM segmentations are visualized on the FLAIR scan in (D) and (H).
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4834931&req=5

brb3422-fig-0002: Visualization of MSmetrix segmentation results for the 1.5T (A–D) and 3T (E–H) scan of a randomly selected MS patient. T1 scans are shown in (A) and (E), FLAIR scans in (B) and (F). Lesion and GM segmentations are superimposed on the T1 scan, as displayed in (C) and (G). Finally, lesion and WM segmentations are visualized on the FLAIR scan in (D) and (H).
Mentions: In Figure 2, some visual results of the MSmetrix segmentations on a 1.5T and 3T scan of the same randomly selected subject are displayed. In Figure 2A and B, an axial slice of the 1.5T 3D T1 and 3D FLAIR are shown, respectively. For visualization purposes, the GM and lesion segmentation are visualized on the T1 (c) and the WM and lesion masks on the FLAIR (d). A similar slice was selected for the 3T scan, as shown in Figure 2E and F, for the 3D T1 and 3D FLAIR, respectively. Similar as in Figure 2C and D, the GM, lesions, and WM segmentations of the 3T MRI are displayed in Figure 2G and H. These lesion segmentations are then used to fill the 3D T1 with normal‐appearing white matter, as explained in Figure 1. The cross‐sectional brain tissue segmentations that are shown in Figure 2 will be used as input for the longitudinal pipeline, to calculate the Jacobian of the deformation fields between both scans, resulting in a measure of brain and GM PBVC.

Bottom Line: The results are obtained with MSmetrix, which is developed especially for use in the MS clinical care path, and compared to Siena (FSL), a widely used software for research purposes.When data of the same scanner are compared, the error is in the order of 0.06-0.08% for both MSmetrix and Siena.MSmetrix appears robust on both the 1.5T and 3T systems and the measurement error becomes an order of magnitude higher between scanners with different field strength.

View Article: PubMed Central - PubMed

Affiliation: Department of Neurology Hôpital Erasme Université Libre de Bruxelles Anderlecht Belgium.

ABSTRACT

Introduction: There is emerging evidence that brain atrophy is a part of the pathophysiology of Multiple Sclerosis (MS) and correlates with several clinical outcomes of the disease, both physical and cognitive. Consequently, brain atrophy is becoming an important parameter in patients' follow-up. Since in clinical practice both 1.5Tesla (T) and 3T magnetic resonance imaging (MRI) systems are used for MS patients follow-up, questions arise regarding compatibility and a possible need for standardization.

Methods: Therefore, in this study 18 MS patients were scanned on the same day on a 1.5T and a 3T scanner. For each scanner, a 3D T1 and a 3D FLAIR were acquired. As no atrophy is expected within 1 day, these datasets can be used to evaluate the median percentage error of the brain volume measurement for gray matter (GM) volume and parenchymal volume (PV) between 1.5T and 3T scanners. The results are obtained with MSmetrix, which is developed especially for use in the MS clinical care path, and compared to Siena (FSL), a widely used software for research purposes.

Results: The MSmetrix median percentage error of the brain volume measurement between a 1.5T and a 3T scanner is 0.52% for GM and 0.35% for PV. For Siena this error equals 2.99%. When data of the same scanner are compared, the error is in the order of 0.06-0.08% for both MSmetrix and Siena.

Conclusions: MSmetrix appears robust on both the 1.5T and 3T systems and the measurement error becomes an order of magnitude higher between scanners with different field strength.

No MeSH data available.


Related in: MedlinePlus