Contributions
Abstract: P1907
Type: Poster
Abstract Category: Late breaking news
Background: Multiple Sclerosis (MS) phenotypes (Relapsing-Remitting Multiple Sclerosis (RRMS) and Secondary Progressive Multiple Sclerosis (SPMS)) are to date differentiated solely on clinical grounds. Treatment effects differ considerably across phenotypes. Non-clinical measures, such as Magnetic Resonance Imaging (MRI), to support this differentiation could contribute to improved patient management and in selecting patients for clinical trials.
Objective: Primary objective is to compare cerebral grey matter volumes (Cortical Grey Matter volume (CGMv) and Deep Grey Matter volume (DGMv)) across a RRMS and SPMS patient cohort. Secondary objective is to test a logistic regression model fitted to the outcome of the primary objective, to predict clinical phenotype.
Methods: Fully automated segmentation (MSmetrix; icoMetrix NV, Leuven, Belgium) was applied to 108 RRMS and 30 SPMS patient brain MRI scans (acquired at Antwerp University Hospital, Belgium) to quantify CGMv and DGMv. Acquired volumes were compared to MSmetrix normalised healthy control volumes (HCv's), by expressing patient volume in ratio of its respective age, gender and intracranial normalised HCv as denominator, giving rise to CGM-ratio (CGMr) and DGM-ratio (DGMr). Expanded Disability Status Scale equal to or greater than 4.0 was applied as cut-off for SPMS. Differences in ratio means, Pearson-coefficients and predictors of phenotype were evaluated.
Results: Mean CGMr was found to be the most significant difference across phenotypes (RRMS 1.0240, SPMS 1.0016, P=.006). DGMr mean however, did not differ (RRMS 0.8767, SPMS 0.8533, P=.209). Split analysis of total cohort indicated differences in CGMr's were not attributable to active therapy. CGMr correlated with DGMr; in total cohort (r=0.327, P=.000), in RRMS (r=0.308, P=.001), however not in SPMS (r=0.324, P=.080). Binary logistic regression of log-transformed CGMr (logCGMr) and disease duration indicated both covariates as independent predictors of phenotype (P=.038, P=.002 respectively).
Conclusion: CGMr's are increased (>1) in total cohort (70.3%), independent of active therapy. CGMr's are higher in RRMS compared to SPMS, conceivably due to inflammatory and remodelling effects. In contrast, DGMr's are lower (< 1) in total cohort (92,8%), without difference across subgroups. logCGMr and disease duration emerged as independent predictors of phenotype. These findings need to be replicated and validated in a prospective cohort.
Disclosure:
Jeroen Fleer*: nothing to disclose
Sébastien Vermeulen*: nothing to disclose
Paul Parizel: Board member of icometrix
Patrick Cras: nothing to disclose
Barbara Willekens: The institution (Antwerp University Hospital) received fees for speaking, consultancy and travel support from Biogen, Teva, Sanofi, Roche, Novartis, Merck. All fees are used for research purposes.
Source of funding: Antwerp University Hospital (see disclosure above).
*These authors contributed equally to this work
Abstract: P1907
Type: Poster
Abstract Category: Late breaking news
Background: Multiple Sclerosis (MS) phenotypes (Relapsing-Remitting Multiple Sclerosis (RRMS) and Secondary Progressive Multiple Sclerosis (SPMS)) are to date differentiated solely on clinical grounds. Treatment effects differ considerably across phenotypes. Non-clinical measures, such as Magnetic Resonance Imaging (MRI), to support this differentiation could contribute to improved patient management and in selecting patients for clinical trials.
Objective: Primary objective is to compare cerebral grey matter volumes (Cortical Grey Matter volume (CGMv) and Deep Grey Matter volume (DGMv)) across a RRMS and SPMS patient cohort. Secondary objective is to test a logistic regression model fitted to the outcome of the primary objective, to predict clinical phenotype.
Methods: Fully automated segmentation (MSmetrix; icoMetrix NV, Leuven, Belgium) was applied to 108 RRMS and 30 SPMS patient brain MRI scans (acquired at Antwerp University Hospital, Belgium) to quantify CGMv and DGMv. Acquired volumes were compared to MSmetrix normalised healthy control volumes (HCv's), by expressing patient volume in ratio of its respective age, gender and intracranial normalised HCv as denominator, giving rise to CGM-ratio (CGMr) and DGM-ratio (DGMr). Expanded Disability Status Scale equal to or greater than 4.0 was applied as cut-off for SPMS. Differences in ratio means, Pearson-coefficients and predictors of phenotype were evaluated.
Results: Mean CGMr was found to be the most significant difference across phenotypes (RRMS 1.0240, SPMS 1.0016, P=.006). DGMr mean however, did not differ (RRMS 0.8767, SPMS 0.8533, P=.209). Split analysis of total cohort indicated differences in CGMr's were not attributable to active therapy. CGMr correlated with DGMr; in total cohort (r=0.327, P=.000), in RRMS (r=0.308, P=.001), however not in SPMS (r=0.324, P=.080). Binary logistic regression of log-transformed CGMr (logCGMr) and disease duration indicated both covariates as independent predictors of phenotype (P=.038, P=.002 respectively).
Conclusion: CGMr's are increased (>1) in total cohort (70.3%), independent of active therapy. CGMr's are higher in RRMS compared to SPMS, conceivably due to inflammatory and remodelling effects. In contrast, DGMr's are lower (< 1) in total cohort (92,8%), without difference across subgroups. logCGMr and disease duration emerged as independent predictors of phenotype. These findings need to be replicated and validated in a prospective cohort.
Disclosure:
Jeroen Fleer*: nothing to disclose
Sébastien Vermeulen*: nothing to disclose
Paul Parizel: Board member of icometrix
Patrick Cras: nothing to disclose
Barbara Willekens: The institution (Antwerp University Hospital) received fees for speaking, consultancy and travel support from Biogen, Teva, Sanofi, Roche, Novartis, Merck. All fees are used for research purposes.
Source of funding: Antwerp University Hospital (see disclosure above).
*These authors contributed equally to this work