Contributions
Abstract: 26
Type: Young Scientific Investigators' Session
Abstract Category: Pathology and pathogenesis of MS - MRI and PET
Introduction: Multiple sclerosis (MS) is a complex inflammatory disorder of the central nervous system. Advanced structural magnetic resonance imaging (MRI) has been suggested to be a sensitive tool for disease staging, prognosis and monitoring in a clinical MS setting. Here, combining longitudinal and sensitive measures of MRI-based brain morphometry and brain-age prediction using machine learning, we tested the hypothesis that MS patients have higher estimates of brain age than healthy controls (HC), and that higher or accelerated rate of brain aging over time has impact on the clinical course in MS patients.
Methods: A longitudinal cohort of MS patients (n=76, mean age 35.1 years (±7.2), mean EDSS 2.0 (±0.9)) was examined with brain MRI in the same scanner at three time points with a mean follow up period of 4.3 years. For the MRI data from the MS sample, we used the longitudinal stream in FreeSurfer 5.3 to define sensitive features of brain morphometry. In addition, we obtained cross-sectional MRI data from 235 HC collected locally. An independent set of brain scans from 3452 HC collected from publicly available MRI datasetswere used as training set to build a model for brain-age estimation in R using the xgboost package, utilizing 1118 global and regional structural brain features. For all scans, we calculated the brain age gap (BAG, difference between chronological age and brain age) as well as the annual change in BAG (brain age slope, BAS). We tested for associations between BAG and BAS compared to clinical and MRI variables within the MS group using multiple regression analysis. We also performed cross-sectional comparisons of BAG between MS patients and HC.
Results: On average,MS patients showed higher BAG compared to HC (4.4 years, p=4.0 x 10-6). In the MS cohort we found an annual increase in BAG of 0.41 years (p=0.008). We found significant correlations with BAG and BAS and MRI measurements for brain atrophy, brain volume, white matter lesion load and change in white matter lesion load. After correcting for multiple testing, no significant correlations with BAG or BAS and disability progression or treatment were found.
Conclusion: MS patients had higher BAG compared to matched HC. Within the MS cohort we found an accelerated rate of brain aging in the period. Brain age estimation is a promising method for evaluation of structural brain changes in MS, with potential for predicting future outcome and guide choice of MS treatment.
Disclosure: The paper was supported by grants from The Research Council of Norway (NFR, grant number 240102) and the South-Eastern Health Authorities of Norway (grant number 257955). Einar A. Høgestøl has received honoraria for lecturing from Merck. Tobias Kaufmann, Gro O. Nygaard and Lars T. Westlye: Nothing to disclose. Mona K. Beyer has received honoraria for lecturing from Novartis and Biogen Idec. Hanne F. Harbo has received travel support, honoraria for advice and lecturing from Biogen, Genzyme, Merck, Novartis, Roche, Sanofi-Aventis and Teva. Hanne F. Harbo has received unrestricted research grants for research from Novartis.
Abstract: 26
Type: Young Scientific Investigators' Session
Abstract Category: Pathology and pathogenesis of MS - MRI and PET
Introduction: Multiple sclerosis (MS) is a complex inflammatory disorder of the central nervous system. Advanced structural magnetic resonance imaging (MRI) has been suggested to be a sensitive tool for disease staging, prognosis and monitoring in a clinical MS setting. Here, combining longitudinal and sensitive measures of MRI-based brain morphometry and brain-age prediction using machine learning, we tested the hypothesis that MS patients have higher estimates of brain age than healthy controls (HC), and that higher or accelerated rate of brain aging over time has impact on the clinical course in MS patients.
Methods: A longitudinal cohort of MS patients (n=76, mean age 35.1 years (±7.2), mean EDSS 2.0 (±0.9)) was examined with brain MRI in the same scanner at three time points with a mean follow up period of 4.3 years. For the MRI data from the MS sample, we used the longitudinal stream in FreeSurfer 5.3 to define sensitive features of brain morphometry. In addition, we obtained cross-sectional MRI data from 235 HC collected locally. An independent set of brain scans from 3452 HC collected from publicly available MRI datasetswere used as training set to build a model for brain-age estimation in R using the xgboost package, utilizing 1118 global and regional structural brain features. For all scans, we calculated the brain age gap (BAG, difference between chronological age and brain age) as well as the annual change in BAG (brain age slope, BAS). We tested for associations between BAG and BAS compared to clinical and MRI variables within the MS group using multiple regression analysis. We also performed cross-sectional comparisons of BAG between MS patients and HC.
Results: On average,MS patients showed higher BAG compared to HC (4.4 years, p=4.0 x 10-6). In the MS cohort we found an annual increase in BAG of 0.41 years (p=0.008). We found significant correlations with BAG and BAS and MRI measurements for brain atrophy, brain volume, white matter lesion load and change in white matter lesion load. After correcting for multiple testing, no significant correlations with BAG or BAS and disability progression or treatment were found.
Conclusion: MS patients had higher BAG compared to matched HC. Within the MS cohort we found an accelerated rate of brain aging in the period. Brain age estimation is a promising method for evaluation of structural brain changes in MS, with potential for predicting future outcome and guide choice of MS treatment.
Disclosure: The paper was supported by grants from The Research Council of Norway (NFR, grant number 240102) and the South-Eastern Health Authorities of Norway (grant number 257955). Einar A. Høgestøl has received honoraria for lecturing from Merck. Tobias Kaufmann, Gro O. Nygaard and Lars T. Westlye: Nothing to disclose. Mona K. Beyer has received honoraria for lecturing from Novartis and Biogen Idec. Hanne F. Harbo has received travel support, honoraria for advice and lecturing from Biogen, Genzyme, Merck, Novartis, Roche, Sanofi-Aventis and Teva. Hanne F. Harbo has received unrestricted research grants for research from Novartis.