Multiple Sclerosis Journal - October 2017 - 1456

Multiple Sclerosis Journal 23(11)

Multiple Sclerosis Journal
2017, Vol. 23(11) 1456-1458
DOI:
10.1177/
https://doi.org/10.1177/1352458517713344
1352458517713345
https://doi.org/10.1177/1352458517713344
© The Author(s), 2017.
Reprints and permissions:
http://www.sagepub.co.uk/
journalsPermissions.nav

Advanced MRI measures like DTI or fMRI
should be outcome measures in future clinical
trials - NO
Paul M Matthews

Correspondence to:
PM Matthews
Division of Brain Sciences,
Department of Medicine,
Imperial College London,
Hammersmith Hospital,
E515, Du Cane Road,
London WC12 0NN, UK.
p.matthews@imperial.ac.uk
Paul M Matthews
Division of Brain Sciences,
Department of Medicine,
Imperial College London,
Hammersmith Hospital,
London, UK/Centre for
Neurotechnology, Imperial
College London, London,
UK/UK Dementia Research
Institute, Imperial College
London, London, UK

The power of a clinical trial to test drug efficacy is
related directly to the effect size of its outcome
measures, their precision and their sensitivity to
drug-associated changes. The initial trials of diseasemodifying therapies rapidly demonstrated the power
of conventional magnetic resonance imaging (MRI)
measures (counts of gadolinium-enhancing and T2
lesions) to detect changes with anti-inflammatory
treatments. Their acceptance as meaningful was
enhanced by post mortem correlative studies directly
relating the imaging measures to neuropathological
'gold standard' criteria for white matter inflammatory lesions.1 Subsequent demonstrations that they
could be related directly to the clinically meaningful
outcomes of relapse frequency and time to sustained
disability progression transformed perceptions of the
utility of these measures for the clinical development
of anti-inflammatory treatments.2
This history is a reminder that a useful clinical trial
endpoint measure as one that is
1. Able to be related to the concept of interest (in this
case, active white matter inflammatory lesions);
2. Able to be measured on multiple platforms and
by different observers;
3. Sensitive to treatment-related change; and,
4. Clinically meaningful (i.e. in terms of the way
that a person with disease feels or functions).
Any measure introduced into a clinical trial comes at
a cost. The cost can be in terms of time, inconvenience, risk or direct expense. Cost also is incurred in a
clinical trial by the complication or delay of decisionmaking by introducing uncertainties of interpretation.
The latter introduces a necessary 'biomarker conservatism' into clinical trials: we should not make measurements in clinical trials without a plan for using the
data meaningfully.3 This does not mean that exploratory measures should not be used. However, when
this is done, they need to be recognised as such.

1456

What do we know about measures derived from the
two newer MRI methods in question here: blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) and diffusion tensor
imaging (DTI)? BOLD fMRI is based on generating a
statistical image describing significant increases in
the local susceptibility weighted MRI signal arising
from increases in the (diamagnetic) oxygenated blood
volume that accompany neuronal activation in the
healthy brain.4 This so-called 'haemodynamic
response' is mediated by signalling from neurons to
astrocytes and then from astrocytes to pre-capillary
sphincters controlling local capillary dilation. Thus,
while BOLD fMRI may reflect local increases in synaptic activity, the measure is several steps removed
from its direct assessment.
DTI is based on a simpler idea. Water molecules diffuse in all directions equally in bulk liquids (e.g. ventricular cerebrospinal fluid (CSF)), but with the
physical constraints to free diffusion provided by cell
membranes, water in tissue has preferred directions of
diffusion, for example, with 'fast' diffusion along the
membranes of myelinated axons, rather than across
them. DTI is sensitive to this directionality (anisotropy) of diffusion. Water diffusion along tightly
packed, well-oriented myelinated axonal tracts, such
as those in the spinal cord, is highly anisotropic with
preferential diffusion parallel to the principle axis of
the axons. Demyelination or axonal degeneration is
associated with locally increased diffusion orthogonal
to the original axis of the tract.5
Both of these methods are powerful research tools. In
conjunction with other data, they have provided major
new insights concerning the functional organisation
of the brain and its dynamics in health and disease.6
Early studies with fMRI in multiple sclerosis (MS)
provided compelling evidence (since supported by
many additional observations) that functional reorganisation of the brain is an important mechanism of

journals.sagepub.com/home/msj

https://uk.sagepub.com/en-gb/journals-permissions https://uk.sagepub.com/en-gb/journals-permissions https://journals.sagepub.com/home/msj

Table of Contents for the Digital Edition of Multiple Sclerosis Journal - October 2017