Contributions
Abstract: EP1457
Type: ePoster
Abstract Category: Pathology and pathogenesis of MS - Imaging
Objective: To assess the reliability of cortical thickness (CTh) measures from 1.5T and 3T MRI in patients with multiple sclerosis (MS) and normal controls (NC).
Background: The cerebral cortex is a common and clinically-relevant site of atrophy in MS. MRI-derived CTh measurements have the potential to capture gray matter tissue loss. We have previously shown that 3T scans detected more areas of cortical thinning than 1.5T in patients with MS vs. NC.
Methods: We studied 15 MS patients [11 relapsing and 4 progressive, 12 (80%) women, age (mean±SD) 50.3±8 years, disease duration 18.4±10.4 years, Expanded Disability Status Score 3.1±2, timed 25-foot walk 5.6±1.8 seconds] and 15 NC [age 37.7±9.6 years, 10 (67%) women]. All subjects underwent a cerebral MRI scan-rescan at 1.5T (GE Signa) and 3T (Siemens Skyra) using 3D-T1-weighted high resolution sequences. The fully automated FreeSurfer pipeline was applied to obtain CTh measures from 34 regions in each hemisphere. Suboptimal segmentation cases were discarded. The agreement between the scan-rescan measurements at both 1.5T and 3T were compared using Pearson"s correlation coefficient, the one-way intraclass correlation coefficient (ICC), and the standard deviation of the differences.
Results: For the scan-rescan measurements at 1.5T, correlations of 0.8 or greater were observed for 22 of the 68 regions, and only seven regions had correlations below 0.5. In addition, an ICC of 0.8 or greater was observed for 21 of the 68 regions, and only five regions had an ICC below 0.5. For the scan-rescan measurements at 3T, correlations of 0.8 or greater were observed for 46 of the 68 regions, and only two regions had correlations below 0.5. In addition, an ICC of 0.8 or greater was observed for 42 of the 68 regions, and only two regions had an ICC below 0.5. When the two field strengths were compared in terms of the magnitude of the estimated correlation, the correlation was higher at 3T for 47 out of the 68 regions. For the ICC, the estimated ICC was higher at 3T for 46 out of the 68 regions. Finally, the estimated standard deviation of the differences was smaller for the 3T measurements vs. 1.5T for 65 of the 68 regions.
Conclusions: 3T MRI may provide better reliability compared to 1.5T in measuring CTh.
Disclosure: Dr. Healy received consulting fees from Biogen and grant support from Genzyme, Merck Serono and Novartis. Dr. Bakshi received consulting fees from AbbVie, EMD Serono, Genentech, Novartis, and Sanofi-Genzyme. Dr. Bakshi received research support from Biogen, EMD-Serono, Novartis, and Sanofi-Genzyme. Dr. Bakshi serves as Editor-in-Chief of the Journal of Neuroimaging. Dr. Tummala has nothing to disclose. Dr. Chu has nothing to disclose. Dr. Tauhid has nothing to disclose. Dr. Khalid has nothing to disclose. Sheena Dupuy has nothing to disclose.
Abstract: EP1457
Type: ePoster
Abstract Category: Pathology and pathogenesis of MS - Imaging
Objective: To assess the reliability of cortical thickness (CTh) measures from 1.5T and 3T MRI in patients with multiple sclerosis (MS) and normal controls (NC).
Background: The cerebral cortex is a common and clinically-relevant site of atrophy in MS. MRI-derived CTh measurements have the potential to capture gray matter tissue loss. We have previously shown that 3T scans detected more areas of cortical thinning than 1.5T in patients with MS vs. NC.
Methods: We studied 15 MS patients [11 relapsing and 4 progressive, 12 (80%) women, age (mean±SD) 50.3±8 years, disease duration 18.4±10.4 years, Expanded Disability Status Score 3.1±2, timed 25-foot walk 5.6±1.8 seconds] and 15 NC [age 37.7±9.6 years, 10 (67%) women]. All subjects underwent a cerebral MRI scan-rescan at 1.5T (GE Signa) and 3T (Siemens Skyra) using 3D-T1-weighted high resolution sequences. The fully automated FreeSurfer pipeline was applied to obtain CTh measures from 34 regions in each hemisphere. Suboptimal segmentation cases were discarded. The agreement between the scan-rescan measurements at both 1.5T and 3T were compared using Pearson"s correlation coefficient, the one-way intraclass correlation coefficient (ICC), and the standard deviation of the differences.
Results: For the scan-rescan measurements at 1.5T, correlations of 0.8 or greater were observed for 22 of the 68 regions, and only seven regions had correlations below 0.5. In addition, an ICC of 0.8 or greater was observed for 21 of the 68 regions, and only five regions had an ICC below 0.5. For the scan-rescan measurements at 3T, correlations of 0.8 or greater were observed for 46 of the 68 regions, and only two regions had correlations below 0.5. In addition, an ICC of 0.8 or greater was observed for 42 of the 68 regions, and only two regions had an ICC below 0.5. When the two field strengths were compared in terms of the magnitude of the estimated correlation, the correlation was higher at 3T for 47 out of the 68 regions. For the ICC, the estimated ICC was higher at 3T for 46 out of the 68 regions. Finally, the estimated standard deviation of the differences was smaller for the 3T measurements vs. 1.5T for 65 of the 68 regions.
Conclusions: 3T MRI may provide better reliability compared to 1.5T in measuring CTh.
Disclosure: Dr. Healy received consulting fees from Biogen and grant support from Genzyme, Merck Serono and Novartis. Dr. Bakshi received consulting fees from AbbVie, EMD Serono, Genentech, Novartis, and Sanofi-Genzyme. Dr. Bakshi received research support from Biogen, EMD-Serono, Novartis, and Sanofi-Genzyme. Dr. Bakshi serves as Editor-in-Chief of the Journal of Neuroimaging. Dr. Tummala has nothing to disclose. Dr. Chu has nothing to disclose. Dr. Tauhid has nothing to disclose. Dr. Khalid has nothing to disclose. Sheena Dupuy has nothing to disclose.