Contributions
Abstract: EP1512
Type: Poster Sessions
Abstract Category: Pathology and pathogenesis of MS - MRI and PET
Purpose: Explore differences in lesion formation and growth in primary progressive and relapsing remitting MS patients using MRI segmentation with icobrain ms.
Materials and methods: Longitudinal MRI follow-up of 30 relapsing remitting (RR) and 20 primary progressive (PP) MS patients was conducted using a 1.5T MR system (Sonata Siemens). For each patient, MRI sessions from 2-10 time points were available, 0.5-1 year apart. Automated image analysis was conducted with icobrain ms. This software employs T1w and FLAIR images from two time points simultaneously to segment lesions and indicate their changes over time. Distinction is made between FLAIR lesions and T1 hypointesities (blackholes).Longitudinal changes in FLAIR lesions and T1 blackholes were first assessed in terms of total volume change by linearly regressing each patient's volumes in function of time. The slopes of all linear models were compared between RRMS and PPMS using Wilcoxon rank-sum test at 5% significance level. Secondly, a distinction was made between volumes of new lesions and volumes of enlarging regions around existing lesions. Potential differences in these volumes between the two groups were evaluated with a bivariate rank sum test.
Results: FLAIR and T1 lesion volume evolution was well fitted by a linear model in most patients, with less than 15% of patients showing a nonlinear volume evolution in time (R2< 0.25). Both groups had a broad intra-class variability, including relatively stable patients (< 0.01ml lesion volume change per year) to highly lesion-active patients (>1ml per year). There was no significant difference between PPMS and RRMS slope coefficients, with median [interquartile range] values of 0.1 [-0.07; 0.75] ml/year in RRMS and 0.2 [-0.08; 0.34] ml/year in PPMS for FLAIR lesions, and 0.2 [0.06; 0.51] ml/year in RRMS and 0.2 [-0.01; 0.42] ml/year in PPMS for T1 blackholes.
Bivariate distributions of new and enlarging lesion volumes were significantly different (rank sum test, p< 0.001) between PPMS and RRMS groups. For T1 blackholes, there was a clear pattern of more new lesions in RRMS patients and more lesion enlargement in PPMS patients.
Conclusion: Inter-individual variability is too large in RRMS and PPMS to reveal significant differences in total lesion volume change. Nevertheless, lesion segmentation with icobrain ms confirms that lesion activity in PPMS manifests more as enlargement of existing lesions rather than formation of new lesions, as opposed to RRMS.
Disclosure: This research received financial support from Roche.
Diana M. Sima, Dirk Smeets, Wim Van Hecke are employed by icometrix.
Gabriel Kocevar, Dominique Sappey-Marinier: nothing to declare.
Françoise Durand-Dubief has received consultancy fees, speaker fees, research grants (non-personal) or honoraria from Biogen, Genzyme-Sanofi, Medday, Merck-Serono, Novartis, Roche and Teva.
Abstract: EP1512
Type: Poster Sessions
Abstract Category: Pathology and pathogenesis of MS - MRI and PET
Purpose: Explore differences in lesion formation and growth in primary progressive and relapsing remitting MS patients using MRI segmentation with icobrain ms.
Materials and methods: Longitudinal MRI follow-up of 30 relapsing remitting (RR) and 20 primary progressive (PP) MS patients was conducted using a 1.5T MR system (Sonata Siemens). For each patient, MRI sessions from 2-10 time points were available, 0.5-1 year apart. Automated image analysis was conducted with icobrain ms. This software employs T1w and FLAIR images from two time points simultaneously to segment lesions and indicate their changes over time. Distinction is made between FLAIR lesions and T1 hypointesities (blackholes).Longitudinal changes in FLAIR lesions and T1 blackholes were first assessed in terms of total volume change by linearly regressing each patient's volumes in function of time. The slopes of all linear models were compared between RRMS and PPMS using Wilcoxon rank-sum test at 5% significance level. Secondly, a distinction was made between volumes of new lesions and volumes of enlarging regions around existing lesions. Potential differences in these volumes between the two groups were evaluated with a bivariate rank sum test.
Results: FLAIR and T1 lesion volume evolution was well fitted by a linear model in most patients, with less than 15% of patients showing a nonlinear volume evolution in time (R2< 0.25). Both groups had a broad intra-class variability, including relatively stable patients (< 0.01ml lesion volume change per year) to highly lesion-active patients (>1ml per year). There was no significant difference between PPMS and RRMS slope coefficients, with median [interquartile range] values of 0.1 [-0.07; 0.75] ml/year in RRMS and 0.2 [-0.08; 0.34] ml/year in PPMS for FLAIR lesions, and 0.2 [0.06; 0.51] ml/year in RRMS and 0.2 [-0.01; 0.42] ml/year in PPMS for T1 blackholes.
Bivariate distributions of new and enlarging lesion volumes were significantly different (rank sum test, p< 0.001) between PPMS and RRMS groups. For T1 blackholes, there was a clear pattern of more new lesions in RRMS patients and more lesion enlargement in PPMS patients.
Conclusion: Inter-individual variability is too large in RRMS and PPMS to reveal significant differences in total lesion volume change. Nevertheless, lesion segmentation with icobrain ms confirms that lesion activity in PPMS manifests more as enlargement of existing lesions rather than formation of new lesions, as opposed to RRMS.
Disclosure: This research received financial support from Roche.
Diana M. Sima, Dirk Smeets, Wim Van Hecke are employed by icometrix.
Gabriel Kocevar, Dominique Sappey-Marinier: nothing to declare.
Françoise Durand-Dubief has received consultancy fees, speaker fees, research grants (non-personal) or honoraria from Biogen, Genzyme-Sanofi, Medday, Merck-Serono, Novartis, Roche and Teva.