Image Processing Reference
In-Depth Information
4
Discussion and Conclusion
We proposed a method to determine whether monophasic demyelination differs from
the chronic demyelination that characterizes MS, with respect to impact on age-
expected brain growth. To achieve this, we fitted for each subject the brain and thal-
amus growth curves using the available follow-up data and we used these growth
curves to predict the brain and thalamus volumes and growth rates at the time of the
first attack.
A predictive modeling of disease in children is challenging, as we have to take into
account not only the continual development of the brain but also the normal inter-
subject variability in this development [19]. We proposed a method that compares the
brain changes due to the disease with the normal changes due to brain development. As
the volume and growth rate are dependent on age and sex, we computed z-scores to
compare the volume and growth rate of each patient with age and sex-matched healthy
subjects. The z-score metric takes into account the inter-subject normal variability.
We worked on relevant structures known to be implicated in MS disease (the brain
and the thalamus [20]). The normalized thalamus is a good metric to indicate if the thal-
amus is more impacted by MS compared to the brain. We studied not only the volume
of these structures but also dynamic information, i.e. the growth rate at the time of the
first attack. We have shown that neither brain or thalamus volume are significantly im-
pacted at the time of the first attack. However, the brain growth rates in the patient
groups were significantly reduced compared to the normal control group, indicating that
even after a single demyelinating event maturation of the brain seems to be affected.
The patient normalized thalamus growth rates were even more impacted than over-
all brain growth. Thalamic growth differed between MS and MONO groups, and
between ADEM groups and MONO group but not between MS and ADEM.
One of the roles of the thalamus is that of a relay center in the brain, regulating
numerous of sensory-motor and cognitive pathways. Different processes may explain
the thalamic volume loss [20]: it may be the direct consequence of local inflammation
or secondary affected by retrograde degenerating axons reaching the thalamus from
inflammatory areas. ADEM patients often show involvement of the thalamus (even
bilateral), and already the single attack seems to influence the age-expected growth in
our cohort. Monophasic non-ADEM demyelination is less likely to involve the thala-
mus so the effect of retrograde axonal damage might be negligible. That is in agree-
ment with our results: the MONO group is less impacted compared to MS and
ADEM. As the normalized thalamus z-scores of the monophasic group were homoge-
neous, the monophasic subjects were well classified in the discriminant analysis. Con-
trary to the monophasic group, the MS group z-scores were heterogeneous, resulting
in a bad discrimination. It could be explained by the fact that the first demyelinating
episode is not necessary the beginning of the disease, the MS disease could have
started earlier for some subjects. The heterogeneity could be explained by different
disease durations among subjects if we hypothesize that the thalamus growth rate is
more reduced as the disease progresses. We will apply the method at different periods
Search WWH ::




Custom Search