Expert Review of Neurotherapeutics

ISSN: 1473-7175 (Print) 1744-8360 (Online) Journal homepage: http://www.tandfonline.com/loi/iern20

Neuroimaging in amyotrophic lateral sclerosis:

current and emerging uses

Federica Agosta, Edoardo Gioele Spinelli & Massimo Filippi

To cite this article: Federica Agosta, Edoardo Gioele Spinelli & Massimo Filippi (2018):
Neuroimaging in amyotrophic lateral sclerosis: current and emerging uses, Expert Review of
Neurotherapeutics, DOI: 10.1080/14737175.2018.1463160

To link to this article: https://doi.org/10.1080/14737175.2018.1463160

Accepted author version posted online: 09

Apr 2018.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=iern20
Neuroimaging in amyotrophic lateral sclerosis: current and emerging
uses

Federica Agosta,1 Edoardo Gioele Spinelli,1,2 Massimo Filippi.1,2

1. Neuroimaging Research Unit, San Raffaele Scientific Institute, Vita-Salute San Raffaele

t
University, Milan, Italy.

ip
2. Department of Neurology, Institute of Experimental Neurology, Division of Neuroscience, San

cr
Raffaele Scientific Institute, Vita-Salute San Raffaele University, Milan, Italy.

us
Corresponding author:
an
Federica Agosta,
M
Neuroimaging Research Unit,

Institute of Experimental Neurology,

ed

Division of Neuroscience,
pt

San Raffaele Scientific Institute,

ce

Vita-Salute San Raffaele University,

Ac

Via Olgettina, 60,

20132 Milan,

Italy

E-mail: [email protected].
Abstract

Introduction: Several neuroimaging techniques have been used to define in vivo markers of

pathological alterations underlying amyotrophic lateral sclerosis (ALS). Growing evidence supports

the use of magnetic resonance imaging (MRI) and positron emission tomography (PET) for the non-

invasive detection of central nervous system involvement in patients with ALS.

t
ip
Areas covered: A comprehensive overview of structural and functional neuroimaging applications

in ALS is provided, focusing on motor and extra-motor involvement in the brain and the spinal

cr
cord. Implications for pathogenetic models, patient diagnosis, prognosis, monitoring and the design

us
of clinical trials are discussed.

Expert commentary: State-of-the-art neuroimaging techniques provide fundamental instruments

an
for the detection and quantification of upper motor neuron and extra-motor brain involvement in

ALS, with relevance for both pathophysiologic investigation and clinical practice. Network-based
M

analysis of structural and functional connectivity alterations and multimodal approaches combining

several neuroimaging measures are promising tools for the development of novel diagnostic and
ed

prognostic markers to be used at the individual patient level.

pt

Keywords: amyotrophic lateral sclerosis; magnetic resonance imaging; atrophy; white matter
ce

damage; resting state functional connectivity; positron emission tomography; spinal cord imaging;

network analysis.
Ac
 1. Introduction

Amyotrophic lateral sclerosis (ALS) is a heterogeneous neurodegenerative disorder characterized

by progressive damage to upper motor neurons (UMN) in the cortex and lower motor neurons

(LMN) in the brainstem and spinal cord. From a clinical standpoint, patients with ALS greatly

differ in terms of site of onset, differential UMN and LMN involvement, degree of extra-motor –

particularly, cognitive – impairment, and disease progression, configuring a wide spectrum of

t
ip
syndromes partially overlapping with frontotemporal dementia (FTD) with different prognostic

impact at the individual level [1]. Neuropathological studies identified cytoplasmic inclusions of

cr
ubiquitin-positive Transactive Response DNA Binding Protein 43 (TDP-43) as the molecular

us
hallmark in up to 98% of ALS cases [2], following a progressive spreading pattern that has been

classified into sequential stages [3]. Despite this unifying pathological feature, significant genetic
an
heterogeneity parallels the great variability in clinical presentations, both in familial and sporadic

cases [4]. Currently, no curative treatment is available for ALS. It has been suggested that the
M

difficulty to enroll patients in early phases of the disease using current diagnostic tools, together

with the well-known disease heterogeneity, might explain the high failure rate of pharmacological
ed

trials in ALS [5].

pt

In this framework, the development of reliable markers of ALS-related changes has a

multifold purpose: first, to allow an early diagnosis with high sensitivity and specificity, particularly
ce

when the so-called ‘ALS-mimic disorders’ (e.g., motor neuropathies, neuromuscular transmission

disorders, radiculopathies, spinal and bulbar muscular atrophy, etc.) are in the differential diagnosis;
Ac

second, to monitor disease progression and provide reproducible measures of pathological

impairment that support a correct stratification of patients in the design of clinical trials; and third,

to identify prognostic factors that may help to predict disease evolution.

In recent years, several neuroimaging techniques have proven useful in the search of

markers of central nervous system (CNS) degeneration in ALS. Neuroimaging provides an ideal

tool to explore objective measures of subclinical UMN impairment in a non-invasive way,

potentially reducing diagnostic delay and allowing an accurate monitoring of disease progression.

Furthermore, neuroimaging in ALS is sensitive to extra-motor brain involvement (e.g.,

frontotemporal cortical regions) that frequently underpins the development of executive and

behavioral deficits over the course of the disease, with significant prognostic implications.

Given the high number of recently published imaging papers in ALS, our aim is not to

provide an exhaustive systematic review but to discuss the current state of the art of brain and spinal

t
ip
cord neuroimaging in the study of structural and functional alterations in ALS, highlighting

potential benefits and shortcomings of the most widely used magnetic resonance imaging (MRI)

cr
and positron emission tomography (PET) techniques for both pathophysiologic investigation and

us
clinical practice. Moreover, we also highlight emerging developments in this field, like applications

of the graph theory in the context of network-based analyses and the ongoing effort to combine
an
multiple diagnostic/prognostic tools into multimodal approaches.
M

2. Clinical MRI

Historically, MRI has been used to exclude structural abnormalities that can mimic clinical UMN
ed

and LMN damage, both in the brain (e.g., vascular lesions, multiple sclerosis, tumors) and the
pt

spinal cord (e.g., radiculopathy, myelopathy, syringomyelia). Since the last decade of the 20th

century, some observations led to the definition of ALS-related MRI alterations, which are
ce

detectable on conventional imaging. Corticospinal tract (CST) hyperintensity on T2-weighted,

proton density or fluid-attenuated inversion recovery (FLAIR) sequences of brain MRI has been
Ac

described in a highly variable proportion of ALS cases, ranging from 15% to 76% across different

cohorts [6]. A hypointense rim in the precentral gyrus can also be observed on T2-weighted images

of patients with ALS [6]. However, these findings are neither sufficiently sensitive nor specific for

ALS. For example, increased T2-signal intensity in the CST has also been shown in healthy

individuals and patients with hepatic failure [6]. Therefore, the visual assessment of these

alterations on brain MRI is currently not recommended for making a firm diagnosis of ALS[6].
Only recently, the application of quantitative approaches to assess FLAIR hyperintensity in the CST

of ALS patients has been suggested to provide some benefit for the diagnosis of ALS, particularly

in those subjects with greater UMN damage [7].

CST involvement can also be detected using conventional MRI of the spinal cord, as T2 and

T1 hyperintensities have been shown in the anterolateral columns of the cervical cord of patients

with ALS [6]. Cord hyperintensities provided higher specificity compared with intensity alterations

t
ip
on brain MRI [6]. In this context, preliminary findings using ultra-high field (7 Tesla) MRI has

recently shown some potential to detect signal abnormalities in the spinal CST of ALS patients with

cr
high accuracy [8] (Figure 1).

us
3. Advanced brain neuroimaging techniques
an
3.1. Grey matter atrophy

Numerous studies assessed grey matter (GM) focal loss in patients with ALS using voxel-based
M

morphometry (VBM) on high-resolution 3D T1-weighted MRI scans. This reliable and reproducible

method provides subtle volume changes by performing a voxel-by-voxel comparison of the regional
ed

attenuation of brain GM intensity across groups of subjects. Initial VBM studies showed diverging
pt

results, with some reporting focal atrophy in motor/premotor regions [9, 10], others reporting

widespread frontotemporal GM atrophy sparing the motor cortex [11], and some reporting no
ce

significant atrophy[12]. Such variability likely results from differences in image processing

pipelines and statistical approaches. A recent meta-analysis pooled results from 29 VBM studies
Ac

comprising 638 subjects with ALS, showing significant GM volume reduction in the right

precentral gyrus and in bilateral inferior frontal cortex [13]. VBM studies have also demonstrated

different atrophy patterns for specific clinical presentations. For example, ALS patients with bulbar

and limb onset showed different GM involvement within the motor strip, consistent with their

clinical disability [14] (Figure 2). Patients with ALS and frontotemporal dementia (FTD) have

shown the most severe atrophy involving widespread frontotemporal cortical areas and the caudate
nucleus [15, 16], although significant GM loss in extensive cortical and subcortical regions -

including the caudate nucleus – has been observed also in patients with subtle cognitive and/or

behavioral impairment [17]. One longitudinal study showed a significant decrease of GM volume

over a 6-month follow-up in motor and extra-motor frontotemporal cortex, thalami, and caudate

nucleus, associated with clinical and cognitive decline [18].

Surface-based morphometry (SBM) is an alternative method to measure subtle cortical

t
ip
changes in human brain, based on the estimation of cortical thickness on T1-weighted images. This

method requires the identification of the inner and outer cortical boundaries or surfaces in order to

cr
measure the cortical thickness. One of the most popular approaches to perform this calculation is

us
provided by the Freesurfer software package (https://surfer.nmr.mgh.harvard.edu, Center for

Biomedical Imaging, Charlestown, MA), which reconstructs a representation of the GM and WM

an
surfaces and provides a measure of GM thickness using the distance between corresponding points

on the two surfaces. SBM studies have consistently shown significant cortical thinning in the
M

primary motor areas of patients with ALS [19, 20], with additional extra-motor involvement that
ed

was particularly severe in those with cognitive or behavioral impairment, although present even in

subjects with exclusive motor symptoms [21, 22]. Focal cortical alterations of the motor areas were
pt

found to mirror specific clinical presentation [23]. Significant associations were observed between

primary motor cortical thinning over one year and the progression of clinical disability [24].
ce

Cortical motor thinning is specific to UMN degeneration, as it was not observed in patients with
Ac

prevalent LMN involvement and ALS-mimic syndromes [25, 26]. Several studies suggested that

cortical thinning of the motor cortex can be used as a sensitive diagnostic marker of ALS at the

individual patient level [20, 27, 28]. One longitudinal SBM study demonstrated the progression of

cortical thinning over time from primary motor to frontal, temporal, and parietal regions [29]. Such

extra-motor cortical damage and its pattern of progression is consistent with postmortem studies

that have recently led to the development of a pathological staging system of ALS [3], supporting

this model in vivo.

 Recent studies have applied VBM or SBM approaches to the specific assessment of

subcortical structures, with some promising results that need further validation. Notably, the

caudate nucleus, nucleus accumbens [30], and thalamus [18] were found to be involved in patients

with ALS, consistent with neuropathological studies [3]. Volume loss in the basal ganglia has also

been shown to correlate with shorter survival in ALS patients [31].

Cerebellar involvement shown by neuroimaging studies [32] is also in line with findings

t
ip
from postmortem studies [3]. A VBM study showed that cerebellar atrophy affects predominantly

the inferior lobules and vermis in patients with motor symptoms and the superior lobules and crus

cr
in patients with cognitive and neuropsychiatric symptoms [32].

3.2. White matter damage

us
an
WM tract alterations have been extensively investigated in patients with ALS using diffusion tensor

(DT) MRI. This technique is sensitive to the random thermal movement of water molecules
M

(Brownian motion) in neural tissues. Structures like cell walls or myelin are barriers to water
ed

diffusion, and so, by measuring the magnitude and directionality of water diffusion, it is possible to

obtain information about the microstructural integrity of brain structures. The DT model allows
pt

calculating the absolute magnitude of water diffusion, measured by mean diffusivity (MD), and the

degree of anisotropy of such diffusion, represented by fractional anisotropy (FA). WM is

ce

characterized by an anisotropic diffusion of water molecules due to the linear structure of tracts,
Ac

which allows the water molecules to diffuse freely along the axons but impede their diffusion

perpendicularly to the axonal main direction because of myelin. Therefore, changes of FA and/or

MD values are used to infer damage to the microarchitectural structure of neuronal projections.

Several DT MRI studies performed in ALS patients consistently showed decreased FA and

increased MD in a “signature” WM region involving the CSTs and the middle and posterior parts of

the corpus callosum, as confirmed by a recent large multicenter study from eight sites[33]. Similar

to GM atrophy, additional microstructural damage to extra-motor frontotemporal tracts has also

been shown, especially in ALS patients with variable degree of cognitive or neuropsychiatric

impairment [15, 22, 26, 34] (Figure 3).

The greatest and most widespread DT MRI alterations were found in patients with primary

lateral sclerosis, encompassing both motor and extra-motor areas and correlating with the severity

of cognitive deficits [35]. On the contrary, the least diffuse WM damage was observed in patients

with predominant LMN involvement, with diverging results from literature as to the extent and

t
ip
significance of such damage [26, 36, 37]. A recent two-center study suggested that these differences

among independent cohorts might be influenced by intrinsic clinical heterogeneity, as only patients

cr
with predominant LMN involvement and a higher rate of disease progression showed significant

us
WM damage, with a pattern resembling the one found in classic ALS [38].

The topographical patterns of WM alterations extending from motor to extra-motor tracts

an
can also be used to classify patients according to stage of disease [39] in agreement with

pathological studies [3]. In this context, longitudinal studies are fundamental to validate current
M

views about pathological spreading in ALS in vivo and dynamically. Relatively few longitudinal DT
ed

MRI studies have been published so far, mirroring the difficulties in enrolling enough patients with

a rapidly evolving disease who could undergo an appropriate number of follow-up scans. Most of
pt

these studies focused on alterations within the motor tracts, reporting a significant progression of

CST damage in the brain after 6 to 8 months [40-42]. Although a minority of studies could not
ce

detect such longitudinal CST damage [24, 43], these inconsistencies likely derive from the
Ac

heterogeneous disease course and great variability across studies in patient sample sizes, follow-up

intervals and functional impairment at baseline. Even taking this into account, several studies

showed a significant association between degeneration of WM motor tracts and worsening

disability [41, 42]. In addition to this, progressive involvement of extra-motor frontotemporal and

cerebellar WM areas has been shown by one longitudinal study [41].

More recently, another study used DT MRI to classify ALS patients according to the

involvement of WM tracts which are damaged at each neuropathological stage, showing that
individual increase in disease stage demonstrated by MRI was consistent with clinical progression

over a 6-month interval [42]. All these data support the use of DT MRI as an objective biomarker of

disease progression in ALS, although further studies describing the topographical pattern of

pathological progression and its correlation with cognitive variables and survival are still needed.

Moreover, the pathological correlates underlying WM diffusion changes in ALS are yet to be

clearly established. Ultra-high-field (7 T) MRI studies are expected to clarify this point,

t
ip
investigating the degree to which implicated factors (i.e., loss of myelinated axons, astrogliosis and

microglial activation) relatively contribute to DT MRI alterations [44].

cr
us
3.3. MR spectroscopy

Magnetic resonance spectroscopy (MRS) is a non-invasive method that measures the concentration
an
of specific biochemical compounds in the brain. This technique provides a spectrum rather than an

image, and can quantify spectra of many biologically important metabolites. The MR spectrum
M

consists of resonances or peaks that represent signal intensities as a function of frequency. Spectra
ed

are obtained either from one selected brain region in the case of single-voxel spectroscopy or from

multiple brain regions in the case of MR spectroscopic imaging. Proton (1H)-MRS is the most used
pt

in clinical practice, and has been used to investigate the metabolite content of brain tissue in

patients with ALS, using both single-voxel and, more recently, whole-brain approaches.
ce

Concentrations of N-acetyl aspartate (NAA) identified by proton MRS can estimate neuronal
Ac

density, whereas choline is a marker of cellular membrane turnover, creatine provides information

about cellular energy metabolism, and myo-inositol is a marker of reactive gliosis.

The most consistent MRS finding in ALS patients is a reduction in NAA relative to choline

and creatine in the motor cortex and along the entire length of the intracranial CST, down to the

brainstem [45-47]. Decreased NAA levels in the primary motor cortex were associated with greater

disability and faster disease progression [46, 47]. Whole-brain spectroscopy showed decreased

NAA concentrations in widespread cortical and subcortical areas, including the premotor and
sensory cortex, basal ganglia, thalamus, and extra-motor frontal, parietal, and occipital regions [48,

49]. NAA levels in the motor cortex have been shown to consistently decrease over 3 to 6 months in

longitudinal studies [50, 51]. However, NAA concentrations did not show a linear pattern of decline

in these studies, and their relationship with clinical progression is currently unclear. An increase in

myo-inositol levels within the motor cortex has also been demonstrated [52], consistent with gliotic

alterations that parallel neurodegeneration.

t
ip
In line with pathogenic theories of increased excitatory and decreased inhibitory signaling in

ALS, some studies found higher glutamate-glutamine [53, 54] and lower GABA concentrations [45]

cr
in the motor cortex of ALS patients compared with healthy controls. However, other studies found

us
decreased glutamate levels within the same areas [55, 56], including a recent study performed using

a 7-Tesla scanner [56], which overcomes the significant overlap between glutamate and glutamine
an
peaks observed at lower field strengths. Inconsistencies between earlier and more recent findings

might be partially due to the effects of riluzole, a glutamate release inhibitor that has been used for
M

ALS treatment since the late 1990s. Therefore, although the assessment of glutamate and GABA
ed

concentrations has provided interesting results, the use of such measures as diagnostic biomarkers

in ALS needs further evidence, particularly in early, drug-naïve disease stages.

pt

3.4. Resting-state functional MRI

ce

Functional MRI (fMRI) is based on the measurement of fluctuations in blood flow and blood-
Ac

oxygen-level dependent contrast as a consequence of neuronal activation, exploiting the inherent

paramagnetic properties of the blood. Brain regions that co-activate under resting conditions

delineate the so-called “resting-state” (RS) functional networks. The assessment of connectivity

alterations between RS networks have provided important insights into brain functional

reorganization in several neurodegenerative diseases, including ALS, in which motor and – when

present – cognitive impairment may undermine the use of task-based fMRI.

 Several studies showed decreased functional connectivity of the sensorimotor network in

ALS patients [57-59], whereas others found increased connectivity [60, 61], or complex regional

patterns of decreased and increased functional connectivity [62]. Altered functional connectivity has

also been shown in brain networks related to cognition and behavior (especially the default mode

and frontoparietal networks) [63, 64], consistent with the multisystem involvement of ALS

pathology. Based on RS fMRI findings, it has been suggested that a widespread increase in brain

t
ip
functional connectivity might prevail in earlier stages of the disease as a compensatory mechanism,

with subsequent decrease as pathological burden accumulates. Consistent with this hypothesis,

cr
increased functional connectivity was found to be higher in patients with less severe microstructural

us
damage to the CST [65], and associated with lower rate of disease progression, shorter disease

duration [64], and preserved motor function [65]. Moreover, a recent longitudinal study showed
an
consistent decreased RS functional connectivity with the sensorimotor and thalamic networks,

paralleling progression of structural alterations and clinical decline over a two-year period in ALS
M

patients [66] (Figure 4). The co-occurrent progressive increase of functional connectivity in extra-
ed

motor networks, such as the left fronto-parietal and the temporal RS networks [66], is also

consistent with a “disconnection” hypothesis due to the loss of compensation. However, some
pt

studies also showed increased functional connectivity within the regions of structural disruption in

ALS correlating with faster disease progression [60], and greater clinical and executive cognitive
ce

impairment [61]. Therefore, a more direct pathogenic involvement of increased functional

Ac

connectivity related to the loss of local inhibitory circuitry within the primary motor and frontal

cortex is also possible, as also suggested by MRS [67] and 11C-flumazenil PET studies [68].

Although fMRI studies have consistently shown functional reorganization in patients with

ALS, providing some hints about the underlying pathophysiological processes, results are often

contradictory and hard to interpret, and might be influenced by several factors (disease stage,

medication, etc.). Future studies enrolling larger and more homogenous cohorts are needed to

validate this technique as a tool to be applied in the clinical setting. Other emerging techniques
providing information about brain functional reorganization, including PET imaging and

quantitative neurophysiological measures (e.g., spectral EEG, magnetoencephalography, and

transcranial magnetic stimulation) are promising tools to complement the results of fMRI studies.

3.5. PET imaging

Tracers labelled by radioactive isotopes are employed by PET to study the metabolism in specific

t
ip
brain areas or the density of particular receptors. Since glucose is the main source of energy for the

brain tissue, PET mainly uses radiolabeled glucose (i.e., [18F]-2-fluoro-2-deoxy-D-glucose

cr
([18F]FDG) to detect the regions of the brain where glucose uptake is low (hypometabolism) or high

us
(hypermetabolism). As for the evaluation of specific receptor density, the current availability and

continuing development of several PET radiotracers, targeting different molecular substrates, makes
an
this technique an interesting tool to investigate pathogenetic underpinnings of ALS.

The earliest PET imaging studies of cerebral glucose metabolism showed widespread
M

decreased FDG binding in the cortex and basal ganglia of ALS patients, particularly – but not
ed

exclusively – in those with greater UMN impairment [69, 70]. Frontotemporal hypometabolism was

associated with motor functional decline [70] and executive deficits [69]. Recently, these findings
pt

have been replicated in larger cohorts, demonstrating high diagnostic accuracy in differentiating

ALS patients from healthy controls using FDG PET [71, 72]. In addition to motor cortical and
ce

frontal hypometabolism, recent studies also showed increased metabolic activity in extra-motor
Ac

regions, such as the brainstem, medial temporal structures, and the cerebellum [71, 72]. Since glial

cells are the preferential site of glycolytic metabolism during brain activation, increased 18F-FDG

uptake has been related to reactive proliferation of astrocytes and microglia surrounding the

affected motor neurons and degenerating descending WM tracts in ALS patients [71].

An active role of neuroinflammation in the pathogenesis of ALS has been supported by PET

studies using radioligands binding to the translocator protein (TSPO), which is highly expressed in

activated glial cells. Increased binding of TSPO radioligands, such as 11C(R)-PK11195 and 11C-
PBR28, has been demonstrated in the primary and supplementary motor cortex, extra-motor

frontotemporal areas, thalamus, midbrain, pons and subcortical WM of patients with ALS [73, 74].

These findings are consistent with increased myo-inositol concentrations shown by MRS [52], and

indicate microglial activation and astrocytosis co-localizing with structural damage [75].

Reduced uptake of 11C-flumazenil, a radiotracer binding to GABA-A receptor subunits, has

been shown in the motor/premotor cortex and extra-motor frontotemporal cortical areas of patients

t
ip
with ALS [68, 76], suggesting impaired GABAergic neurotransmission that has been correlated

with poor performance at executive tasks [77]. The recent development of PET ligands targeting

cr
glutamatergic transmission [78] might provide further insights with regards to hypothesized

us
excitotoxic alterations in ALS.
an
4. Spinal cord imaging

Volumetric studies showed cord atrophy in the cervical and upper thoracic segments of patients
M

with ALS, correlating with disease duration, measures of clinical disability and upper arm muscle

strength [79-81]. DT MRI consistently showed decreased FA and/or increased MD in the cervical
ed

cord of patients with ALS [79-82], demonstrating damage involving not only the lateral columns
pt

but also the posterior sensory pathways [79]. Diffusivity alterations were found to be more severe at

more distal cervical segments [82], and to correlate with measures of disease severity and
ce

respiratory impairment (i.e., forced vital capacity) [79-82]. Significant progression of cord atrophy

was detected over a 9- to 11-month follow-up period, closely related with clinical indicators of
Ac

functional decline [43, 80]. Another study showed relative stability of diffusivity alterations over

time [80]. The few studies assessing MRS of the cervical spinal cord in ALS showed decreased

concentrations of NAA, correlating with clinical decline and reduced forced vital capacity [83, 84].

One of these studies also showed increased myo-inositol levels, suggesting the presence of gliotic

alterations in the cervical cord of ALS patients [83].

 Despite the small number of studies in this field, current evidence supports the use of

cervical cord MRI as a tool to monitor ALS progression, as this technique is sensitive to

degeneration of both UMN and LMN. The use of high magnetic fields, recent technological

improvements providing greater contrast between GM, WM and cerebrospinal fluid, and the study

of lower spinal cord segments (i.e., thoracic and lumbar) might also help clarifying some

pathophysiological aspects of ALS, such as the differential involvement of GM and WM and the

t
ip
‘dying-back’ versus ‘dying-forward’ hypotheses – i.e., whether neurodegeneration in ALS starts in

the brain or in distal cord segments.

cr
us
5. Future directions

5.1. Network-based analyses

an
The human brain is a highly integrated neural network, consisting of several cortical and subcortical

regions that are structurally and functionally interconnected, forming many different co-operating
M

sub-networks. Graph theoretical models have conceptualized such complex organization as the

brain ‘connectome’, consisting of anatomic regions defined as ‘nodes’, which are linked by ‘edges’
ed

(i.e., structural or functional connections). Recently, neuroimaging research has focused on the
pt

study of changes in structural and functional connectivity at a brain-system level, rather than on

alterations in individual brain regions. Such approach has been already demonstrated as a powerful
ce

tool to measure structural and functional reorganization in the FTD spectrum of disorders [85], and

might be particularly helpful in the study of ALS, particularly to clarify the relationship between
Ac

motor and extra-motor alterations observed in the brain of affected patients.

Two independent studies applied network-based statistics to DT MRI of patients with ALS,

both demonstrating the presence of an impaired sub-network including bilateral primary motor

regions, bilateral supplementary motor areas, the basal ganglia and associative areas such as the

right posterior cingulate and precuneus [86, 87]. One longitudinal study has shown that the sub-

network of impaired connectivity expands over time, involving widespread frontal, temporal, and
parietal regions [88] in a way which is consistent with the proposed model of TDP-43 pathological

spreading [3]. In line with such hypothesis, a recent study evaluated brain structural connectivity in

a consistent set of healthy controls, showing that regions involved in subsequent stages of ALS

pathology are highly interconnected by WM tracts, which may serve as anatomical “infrastructures”

facilitating TDP-43 spread [89].

Two studies applied network-based analyses to the assessment of functional alterations in

t
ip
ALS patients using RS fMRI, demonstrating complex connectivity alterations encompassing

frontal, temporal and occipital regions [90, 91], mirroring widespread multisystem functional

cr
reverberation of TDP-43 deposition.

us
Although the high complexity of network-based analyses complicates their application to

clinical practice, the assessment of affected sub-networks in individual patients might facilitate
an
early diagnosis, prognostic patient stratification, and provide evidence for disease mechanisms,

which ultimately should allow gaining novel targets for therapeutic interventions.
M
ed

5.2. Multimodal approaches

Although a number of studies showed that a good level of diagnostic accuracy can be obtained in
pt

distinguishing ALS patients from healthy controls using cortical thickness of the precentral gyrus

[20, 25] or DT MRI measures of the motor tracts [35], each neuroimaging marker provides
ce

information about only a part of the whole pathological process affecting the CNS in ALS.
Ac

Therefore, the use of various neuroimaging techniques to build a complex data-driven

neuroimaging marker, as successfully performed for several other neurodegenerative conditions

[92], might allow the highest accuracy in the correct discrimination of ALS patients, with obvious

relevance for the clinical practice. Combinations of several structural neuroimaging markers (i.e.,

DT MRI, MRS, high-resolution volumetric imaging assessing both cortical and subcortical

structures) have demonstrated diagnostic accuracies ranging from 80% to over 90% in

discriminating ALS patients and healthy controls [45, 93, 94]. A recent study evaluated the
combination of cortical thickness and DT MRI measurements in the separation of patients with ALS

and ALS-mimicking conditions, showing a 94% accuracy using such combined model [28] (Figure

5).

Multimodal sets of MRI data have also been tested as prognostic indicators in ALS,

especially for the prediction of survival in combination with clinical variables. Both the use of

multimodal brain (i.e., SBM and DT MRI of motor regions) [95] and spinal cord neuroimaging

t
ip
techniques (i.e., cord volume and diffusivity alterations) [96] provided enhanced survival prediction

in ALS, suggesting a possible relevance of this approach for clinical practice and the design of

cr
clinical trials.

5.3. Imaging genetic cases

us
an
The great heterogeneity of clinical presentations within the ALS spectrum is likely to mirror, at

least partially, the variability of genetic backgrounds. Therefore, the study of neuroimaging
M

alterations in ALS patients with specific underlying genetic mutations might help to characterize
ed

their structural and functional signatures.

The most common genetic mutation in both familial and sporadic forms of ALS is the
pt

hexanucleotide expansion in the C9orf72 region [97], associated with greater prevalence of

cognitive impairment and full-blown FTD, when compared with sporadic ALS. MRI studies of
ce

ALS patients carrying C9orf72 expansions revealed more severe brain GM atrophy than patients
Ac

with sporadic ALS, involving widespread frontal, temporal, parietal and occipital cortical areas, as

well as subcortical regions such as the thalami and the cerebellum [98, 99]. DT MRI studies also

showed more diffuse involvement of WM tracts [98-100], and a recent longitudinal study

demonstrated the spreading of diffusivity alterations from anterior to posterior WM regions over a

6-month follow-up period in C9orf72 expansion carriers with ALS [100]. A recent study assessing

RS functional alterations in these patients showed rearrangements within the visual networks, likely

related to the involvement of posterior regions in this particular form of ALS [99].
 The few structural MRI studies performed in SOD1 ALS patients showed a relative sparing

of brain motor networks both from a structural [101-103] and a functional [103] point of view,

when compared with sporadic ALS patients. By contrast, a recent study showed marked cervical

cord atrophy in SOD1 ALS patients [103], suggesting that pathological alterations in ALS due to

SOD1 mutations may follow different progression patterns compared with ALS cases with TDP-43

pathology.

t
ip
6. Conclusions

cr
The success of new therapies for ALS critically depends on the ability to provide new drugs at the

us
earliest possible stage with a high diagnostic certainty and capacity to stage patients in prognostic

classes. Neuroimaging has demonstrated to provide reliable markers of ALS pathological processes.
an
Several MRI and PET techniques complement each other in providing information about structural

and functional alterations in the disease course, mirroring the selective involvement of motor
M

networks in early stages and the progressive spread to extra-motor cortical and subcortical areas that

accompanies functional decline and the development of cognitive and behavioral deficits.
ed

Particularly, cross-sectional studies consistently demonstrated that structural and functional

pt

alterations in the CNS of ALS patients correlate with motor and cognitive impairment, and provide

high diagnostic accuracy in the distinction from both healthy controls and patients with ALS-mimic
ce

disorders, especially in disease phenotypes with greater UMN involvement. Moreover, longitudinal

studies showed patterns of disease progression which are consistent with the recently proposed
Ac

neuropathological staging model [3], as early involvement of the primary motor cortex and motor

WM tracts was found to be followed by widespread frontotemporal damage. Structural

neuroimaging alterations in the brain and cervical cord were also shown to correlate with faster

disease progression and shorter survival, supporting a role of these measures as powerful prognostic

tools. Therefore, current evidence suggests that neuroimaging can be used to correctly diagnose and

stratify ALS patients into different stages and prognostic categories in a reliable way, particularly
when multimodal approaches are applied. These findings have obvious relevance for the clinical

practice and the design of pharmacological trials.

7. Expert Commentary

The diagnosis of ALS is based on the demonstration of UMN and LMN degeneration. Whereas

neurophysiological assessment is commonly used to assist the clinician in the identification of LMN

t
ip
damage, currently no markers of occult or subtle UMN involvement have been translated to the

clinical practice. Moreover, clinical measures of functional impairment (e.g., the ALSFRS-R scale)

cr
are rater-dependent and may not be sensitive to subtle changes at short follow-up time intervals.

us
Another key issue in the evaluation of ALS is the great clinical heterogeneity of this condition, as

patients might present with a wide variety of clinical pictures ranging from prevalent UMN to
an
prevalent LMN involvement, and might show greatly variable rates of progression to severe

disability, tracheostomy or death. Genetic heterogeneity underlies and partially explains this clinical
M

diversity, and objective measures of neurodegeneration that can describe the signature pattern of

disease evolution in each ALS-related mutation are needed. Neuroimaging techniques have shown
ed

to provide reproducible, non-invasive measures of UMN damage in the brain (e.g., FA of the CST,
pt

cortical thickness of the primary motor cortex, etc.) that hold the promise to effectively allow an

early diagnosis and a reliable monitoring of disease evolution, and should be included in the clinical
ce

work-up of patients with ALS. Spinal cord imaging also indicated possible markers to be used for

the clinical management of ALS (such as measures of cord atrophy and diffusivity alterations in the
Ac

lateral columns), although findings from these studies warrant further validation due to technical

difficulties that have been overcome only recently. Moreover, MR and PET imaging are candidate

markers for the evaluation of extra-motor involvement in ALS, which is closely related to the

development of cognitive and behavioral alterations in this condition that overlaps with FTD. The

main limitations of current research in this field are the relatively small sample sizes – especially in

longitudinal studies – due to enrollment difficulties in a rare and rapidly progressive disease, the
scarcity of large multicenter studies, and the prevalent use of a single technique for each study.

Considering the complexity of ALS, the use of multimodal approaches and advanced statistical

models such as network-based analyses will be needed to elaborate highly accurate diagnostic and

prognostic markers in this condition. Moreover, an effort should be made to enroll sufficient sample

sizes in follow-up studies, as failure to detect significant longitudinal evolution in some studies

might essentially derive from this issue. Finally, a greater number of large multicenter trials will be

t
ip
required to definitely establish the applicability of any proposed neuroimaging marker to the

clinical practice.

cr
us
8. Five-year view

Available findings from neuroimaging research in ALS suggest that the use of quantitative
an
measures of CNS damage provided by MRI and PET is soon to be translated into the clinical

practice for diagnostic, prognostic and disease monitoring purposes. The inclusion of these
M

assessments into the design of pharmacological trials is likely to provide objective, reliable outcome

measures that may help identifying effective disease-modifying drugs for a fatal rapidly evolving
ed

disease such as ALS. Large-scale, longitudinal studies will also validate current views supporting
pt

the pathological spreading of TDP-43 inclusions from motor to extra-motor areas in a prion-like

way, as previously suggested for other neurodegenerative diseases.

ce
Ac

9. Key issues

• Visual assessment of T2/FLAIR hyperintensity in the CST on brain and spinal cord MRI is

not recommended for making a firm diagnosis of ALS. Only recently, the application of

quantitative approaches and high magnetic fields has provided some benefit in those

subjects with greater UMN damage.

• The most consistent neuroimaging diagnostic markers of ALS are cortical thinning of the

primary motor cortex, diffusivity alterations of the CST and motor callosal fibres, and
 cervical cord atrophy. These measures showed high diagnostic accuracy and good

correlation with functional decline, especially when combined in multimodal approaches.

• GM loss and WM microstructural damage, demonstrated by MRI, and FDG PET

hypometabolism have been shown to involve not only the motor system, but also

widespread frontotemporal areas. Such diffuse degeneration correlates with greater

functional and cognitive impairment.

t
ip
• fMRI, MRS and PET ligand studies showed complex functional and neurometabolic

alterations over the ALS disease course, supporting current views regarding a pathogenetic

cr
role of excitotoxicity and neuroinflammation in this disease.

us
• In this review, we provide an overview of the different neuroimaging techniques assessing

structural and functional alterations in ALS, highlighting potential benefits and

an
shortcomings of each approach for both pathophysiologic investigation and translation to

clinical practice.
M
ed

Funding

This paper was supported by the Italian Ministry of Health (RF-2011-02351193).

pt

Declaration of interest
ce

F Agosta is Section Editor of NeuroImage: Clinical, has received speaker honoraria from Biogen

Idec, Novartis, and ExceMED – Excellence in Medical Education and receives or has received
Ac

research support from the Italian Ministry of Health, Fondazione Italiana di Ricerca per la SLA

(AriSLA), and the European Research Council. M Filippi is Editor-in-Chief for the Journal of

Neurology; serves on a scientific advisory board for Teva Pharmaceutical Industries; has received

compensation for consulting and/or speaking activities from Biogen Idec, Merck-Serono, Novartis,

Teva Pharmaceuticals; and receives research support from Biogen Idec, Teva Pharmaceuticals,

Novartis, Italian Ministry of Health, Fondazione Italiana Sclerosi Multipla, Cure PSP, Alzheimer’s
Drug Discovery Foundation, the Jacques and Gloria Gossweiler Foundation, and AriSLA. The

authors have no other relevant affiliations or financial involvement with any organization or entity

with a financial interest in or financial conflict with the subject matter or materials discussed in the

manuscript apart from those disclosed. Peer reviewers on this manuscript have no relevant financial

or other relationships to disclose.

t
ip
cr
us
an
M
ed
pt
ce
Ac
References
Papers of special note have been highlighted as:
* of interest
** of considerable interest
1. Swinnen B and Robberecht W, The phenotypic variability of amyotrophic lateral sclerosis.
Nat Rev Neurol, 2014. 10(11): p. 661-70.
2. Neumann M, Sampathu DM, Kwong LK, et al., Ubiquitinated TDP-43 in frontotemporal
lobar degeneration and amyotrophic lateral sclerosis. Science, 2006. 314(5796): p. 130-3.
3. Brettschneider J, Del Tredici K, Toledo JB, et al., Stages of pTDP-43 pathology in

t
amyotrophic lateral sclerosis. Ann Neurol, 2013. 74(1): p. 20-38.

ip
** A groundbreaking neuropathological study assessing the regional pattern of TDP-43
deposition in ALS patients post-mortem, establishing progressive stages of disease
pathology

cr
4. Renton AE, Chio A, and Traynor BJ, State of play in amyotrophic lateral sclerosis genetics.
Nat Neurosci, 2014. 17(1): p. 17-23.
5. Katyal N and Govindarajan R, Shortcomings in the Current Amyotrophic Lateral Sclerosis

us
Trials and Potential Solutions for Improvement. Front Neurol, 2017. 8: p. 521.
6. Filippi M, Agosta F, Abrahams S, et al., EFNS guidelines on the use of neuroimaging in the
management of motor neuron diseases. Eur J Neurol, 2010. 17(4): p. 526-e20.
7.
an
Fabes J, Matthews L, Filippini N, et al., Quantitative FLAIR MRI in Amyotrophic Lateral
Sclerosis. Acad Radiol, 2017. 24(10): p. 1187-1194.
8. Cohen-Adad J, Zhao W, Keil B, et al., 7-T MRI of the spinal cord can detect lateral
corticospinal tract abnormality in amyotrophic lateral sclerosis. Muscle Nerve, 2013. 47(5):
M
p. 760-2.
9. Agosta F, Pagani E, Rocca MA, et al., Voxel-based morphometry study of brain volumetry
and diffusivity in amyotrophic lateral sclerosis patients with mild disability. Hum Brain
Mapp, 2007. 28(12): p. 1430-8.
ed

10. Turner MR, Hammers A, Allsop J, et al., Volumetric cortical loss in sporadic and familial
amyotrophic lateral sclerosis. Amyotroph Lateral Scler, 2007. 8(6): p. 343-7.
11. Mezzapesa DM, Ceccarelli A, Dicuonzo F, et al., Whole-brain and regional brain atrophy in
pt

amyotrophic lateral sclerosis. AJNR Am J Neuroradiol, 2007. 28(2): p. 255-9.

12. Abrahams S, Goldstein LH, Suckling J, et al., Frontotemporal white matter changes in
amyotrophic lateral sclerosis. J Neurol, 2005. 252(3): p. 321-31.
ce

13. Shen D, Cui L, Fang J, et al., Voxel-Wise Meta-Analysis of Gray Matter Changes in
Amyotrophic Lateral Sclerosis. Front Aging Neurosci, 2016. 8: p. 64.
14. Bede P, Bokde A, Elamin M, et al., Grey matter correlates of clinical variables in
amyotrophic lateral sclerosis (ALS): a neuroimaging study of ALS motor phenotype
Ac

heterogeneity and cortical focality. J Neurol Neurosurg Psychiatry, 2013. 84(7): p. 766-73.
15. Lillo P, Mioshi E, Burrell JR, et al., Grey and white matter changes across the amyotrophic
lateral sclerosis-frontotemporal dementia continuum. PLoS One, 2012. 7(8): p. e43993.
16. Masuda M, Senda J, Watanabe H, et al., Involvement of the caudate nucleus head and its
networks in sporadic amyotrophic lateral sclerosis-frontotemporal dementia continuum.
Amyotroph Lateral Scler Frontotemporal Degener, 2016. 17(7-8): p. 571-579.
17. Alruwaili AR, Pannek K, Coulthard A, et al., A combined tract-based spatial statistics and
voxel-based morphometry study of the first MRI scan after diagnosis of amyotrophic lateral
sclerosis with subgroup analysis. J Neuroradiol, 2017.
18. Menke RA, Korner S, Filippini N, et al., Widespread grey matter pathology dominates the
longitudinal cerebral MRI and clinical landscape of amyotrophic lateral sclerosis. Brain,
2014. 137(Pt 9): p. 2546-55.
** A large longitudinal MRI study comparing patterns of grey matter and white matter
damage over the disease course of ALS.
19. Verstraete E, van den Heuvel MP, Veldink JH, et al., Motor network degeneration in
amyotrophic lateral sclerosis: a structural and functional connectivity study. PLoS One,
2010. 5(10): p. e13664.
20. Agosta F, Valsasina P, Riva N, et al., The cortical signature of amyotrophic lateral sclerosis.
PLoS One, 2012. 7(8): p. e42816.
21. Schuster C, Kasper E, Dyrba M, et al., Cortical thinning and its relation to cognition in
amyotrophic lateral sclerosis. Neurobiol Aging, 2014. 35(1): p. 240-6.
22. Agosta F, Ferraro PM, Riva N, et al., Structural brain correlates of cognitive and behavioral
impairment in MND. Hum Brain Mapp, 2016. 37(4): p. 1614-26.

t
* A large study comparing structural alterations between ALS patients with and without

ip
cognitive impairment and their relationship with neuropsychological deficits.
23. Schuster C, Kasper E, Machts J, et al., Focal thinning of the motor cortex mirrors clinical
features of amyotrophic lateral sclerosis and their phenotypes: a neuroimaging study. J

cr
Neurol, 2013. 260(11): p. 2856-64.
24. Kwan JY, Meoded A, Danielian LE, et al., Structural imaging differences and longitudinal
changes in primary lateral sclerosis and amyotrophic lateral sclerosis. Neuroimage Clin,

us
2012. 2: p. 151-60.
25. Walhout R, Westeneng HJ, Verstraete E, et al., Cortical thickness in ALS: towards a marker
for upper motor neuron involvement. J Neurol Neurosurg Psychiatry, 2015. 86(3): p. 288-
94.
an
26. Spinelli EG, Agosta F, Ferraro PM, et al., Brain MR Imaging in Patients with Lower Motor
Neuron-Predominant Disease. Radiology, 2016. 280(2): p. 545-56.
27. Verstraete E, Veldink JH, Hendrikse J, et al., Structural MRI reveals cortical thinning in
M
amyotrophic lateral sclerosis. J Neurol Neurosurg Psychiatry, 2012. 83(4): p. 383-8.
28. Ferraro PM, Agosta F, Riva N, et al., Multimodal structural MRI in the diagnosis of motor
neuron diseases. Neuroimage Clin, 2017. 16: p. 240-247.
ed

* A multimodal structural MRI study supporting the use of automated, combined approaches
for the diagnosis of ALS, also using a disease control group.
29. Schuster C, Kasper E, Machts J, et al., Longitudinal course of cortical thickness decline in
amyotrophic lateral sclerosis. J Neurol, 2014. 261(10): p. 1871-80.
pt

30. Bede P, Elamin M, Byrne S, et al., Basal ganglia involvement in amyotrophic lateral
sclerosis. Neurology, 2013. 81(24): p. 2107-15.
31. Westeneng HJ, Verstraete E, Walhout R, et al., Subcortical structures in amyotrophic lateral
ce

sclerosis. Neurobiol Aging, 2015. 36(2): p. 1075-82.

32. Tan RH, Devenney E, Dobson-Stone C, et al., Cerebellar integrity in the amyotrophic lateral
sclerosis-frontotemporal dementia continuum. PLoS One, 2014. 9(8): p. e105632.
33. Muller HP, Turner MR, Grosskreutz J, et al., A large-scale multicentre cerebral diffusion
Ac

tensor imaging study in amyotrophic lateral sclerosis. J Neurol Neurosurg Psychiatry, 2016.
87(6): p. 570-9.
34. Trojsi F, Corbo D, Caiazzo G, et al., Motor and extramotor neurodegeneration in
amyotrophic lateral sclerosis: a 3T high angular resolution diffusion imaging (HARDI)
study. Amyotroph Lateral Scler Frontotemporal Degener, 2013. 14(7-8): p. 553-61.
35. Agosta F, Galantucci S, Riva N, et al., Intrahemispheric and interhemispheric structural
network abnormalities in PLS and ALS. Hum Brain Mapp, 2014. 35(4): p. 1710-22.
36. Prudlo J, Bissbort C, Glass A, et al., White matter pathology in ALS and lower motor
neuron ALS variants: a diffusion tensor imaging study using tract-based spatial statistics. J
Neurol, 2012. 259(9): p. 1848-59.
37. Rosenbohm A, Muller HP, Hubers A, et al., Corticoefferent pathways in pure lower motor
neuron disease: a diffusion tensor imaging study. J Neurol, 2016. 263(12): p. 2430-2437.
38. Muller HP, Agosta F, Riva N, et al., Fast progressive lower motor neuron disease is an ALS
variant: A two-centre tract of interest-based MRI data analysis. Neuroimage Clin, 2018. 17:
p. 145-152.
39. Kassubek J, Muller HP, Del Tredici K, et al., Diffusion tensor imaging analysis of
sequential spreading of disease in amyotrophic lateral sclerosis confirms patterns of TDP-43
pathology. Brain, 2014. 137(Pt 6): p. 1733-40.
40. van der Graaff MM, Sage CA, Caan MW, et al., Upper and extra-motoneuron involvement
in early motoneuron disease: a diffusion tensor imaging study. Brain, 2011. 134(Pt 4): p.
1211-28.
41. Keil C, Prell T, Peschel T, et al., Longitudinal diffusion tensor imaging in amyotrophic
lateral sclerosis. BMC Neurosci, 2012. 13: p. 141.

t
42. Kassubek J, Muller HP, Del Tredici K, et al., Imaging the pathoanatomy of amyotrophic

ip
lateral sclerosis in vivo: targeting a propagation-based biological marker. J Neurol
Neurosurg Psychiatry, 2017.
** A novel tract-based approach to stage ALS patients in vivo using diffusion tensor MRI.

cr
43. Agosta F, Rocca MA, Valsasina P, et al., A longitudinal diffusion tensor MRI study of the
cervical cord and brain in amyotrophic lateral sclerosis patients. J Neurol Neurosurg
Psychiatry, 2009. 80(1): p. 53-5.

us
44. Cardenas AM, Sarlls JE, Kwan JY, et al., Pathology of callosal damage in ALS: An ex-vivo,
7 T diffusion tensor MRI study. Neuroimage Clin, 2017. 15: p. 200-208.
45. Foerster BR, Carlos RC, Dwamena BA, et al., Multimodal MRI as a diagnostic biomarker
an
for amyotrophic lateral sclerosis. Ann Clin Transl Neurol, 2014. 1(2): p. 107-14.
46. Sivak S, Bittsansky M, Kurca E, et al., Proton magnetic resonance spectroscopy in patients
with early stages of amyotrophic lateral sclerosis. Neuroradiology, 2010. 52(12): p. 1079-85.
47. Pyra T, Hui B, Hanstock C, et al., Combined structural and neurochemical evaluation of the
M
corticospinal tract in amyotrophic lateral sclerosis. Amyotroph Lateral Scler, 2010. 11(1-2):
p. 157-65.
48. Verma G, Woo JH, Chawla S, et al., Whole-brain analysis of amyotrophic lateral sclerosis
by using echo-planar spectroscopic imaging. Radiology, 2013. 267(3): p. 851-7.
ed

49. Usman U, Choi C, Camicioli R, et al., Mesial prefrontal cortex degeneration in amyotrophic
lateral sclerosis: a high-field proton MR spectroscopy study. AJNR Am J Neuroradiol, 2011.
32(9): p. 1677-80.
pt

50. Unrath A, Ludolph AC, and Kassubek J, Brain metabolites in definite amyotrophic lateral
sclerosis. A longitudinal proton magnetic resonance spectroscopy study. J Neurol, 2007.
254(8): p. 1099-106.
ce

51. Rule RR, Suhy J, Schuff N, et al., Reduced NAA in motor and non-motor brain regions in
amyotrophic lateral sclerosis: a cross-sectional and longitudinal study. Amyotroph Lateral
Scler Other Motor Neuron Disord, 2004. 5(3): p. 141-9.
52. Kalra S, Hanstock CC, Martin WR, et al., Detection of cerebral degeneration in amyotrophic
Ac

lateral sclerosis using high-field magnetic resonance spectroscopy. Arch Neurol, 2006.
63(8): p. 1144-8.
53. Pioro EP, Majors AW, Mitsumoto H, et al., 1H-MRS evidence of neurodegeneration and
excess glutamate + glutamine in ALS medulla. Neurology, 1999. 53(1): p. 71-9.
54. Han J and Ma L, Study of the features of proton MR spectroscopy ((1)H-MRS) on
amyotrophic lateral sclerosis. J Magn Reson Imaging, 2010. 31(2): p. 305-8.
55. Bowen BC, Pattany PM, Bradley WG, et al., MR imaging and localized proton spectroscopy
of the precentral gyrus in amyotrophic lateral sclerosis. AJNR Am J Neuroradiol, 2000.
21(4): p. 647-58.
56. Atassi N, Xu M, Triantafyllou C, et al., Ultra high-field (7tesla) magnetic resonance
spectroscopy in Amyotrophic Lateral Sclerosis. PLoS One, 2017. 12(5): p. e0177680.
57. Mohammadi B, Kollewe K, Samii A, et al., Changes of resting state brain networks in
amyotrophic lateral sclerosis. Exp Neurol, 2009. 217(1): p. 147-53.
58. Fekete T, Zach N, Mujica-Parodi LR, et al., Multiple kernel learning captures a systems-
level functional connectivity biomarker signature in amyotrophic lateral sclerosis. PLoS
One, 2013. 8(12): p. e85190.
59. Trojsi F, Esposito F, de Stefano M, et al., Functional overlap and divergence between ALS
and bvFTD. Neurobiol Aging, 2015. 36(1): p. 413-23.
60. Douaud G, Filippini N, Knight S, et al., Integration of structural and functional magnetic
resonance imaging in amyotrophic lateral sclerosis. Brain, 2011. 134(Pt 12): p. 3470-9.
61. Agosta F, Canu E, Inuggi A, et al., Resting state functional connectivity alterations in
primary lateral sclerosis. Neurobiol Aging, 2014. 35(4): p. 916-25.

t
62. Zhou F, Xu R, Dowd E, et al., Alterations in regional functional coherence within the

ip
sensory-motor network in amyotrophic lateral sclerosis. Neurosci Lett, 2014. 558: p. 192-6.
63. Agosta F, Canu E, Valsasina P, et al., Divergent brain network connectivity in amyotrophic
lateral sclerosis. Neurobiol Aging, 2013. 34(2): p. 419-27.

cr
64. Luo C, Chen Q, Huang R, et al., Patterns of spontaneous brain activity in amyotrophic
lateral sclerosis: a resting-state FMRI study. PLoS One, 2012. 7(9): p. e45470.
65. Agosta F, Valsasina P, Absinta M, et al., Sensorimotor functional connectivity changes in

us
amyotrophic lateral sclerosis. Cereb Cortex, 2011. 21(10): p. 2291-8.
66. Menke RAL, Proudfoot M, Talbot K, et al., The two-year progression of structural and
functional cerebral MRI in amyotrophic lateral sclerosis. Neuroimage Clin, 2018. 17: p.
an
953-961.
67. Foerster BR, Callaghan BC, Petrou M, et al., Decreased motor cortex gamma-aminobutyric
acid in amyotrophic lateral sclerosis. Neurology, 2012. 78(20): p. 1596-600.
68. Lloyd CM, Richardson MP, Brooks DJ, et al., Extramotor involvement in ALS: PET studies
M
with the GABA(A) ligand [(11)C]flumazenil. Brain, 2000. 123 ( Pt 11): p. 2289-96.
69. Kew JJ, Goldstein LH, Leigh PN, et al., The relationship between abnormalities of cognitive
function and cerebral activation in amyotrophic lateral sclerosis. A neuropsychological and
ed

positron emission tomography study. Brain, 1993. 116 ( Pt 6): p. 1399-423.

70. Hoffman JM, Mazziotta JC, Hawk TC, et al., Cerebral glucose utilization in motor neuron
disease. Arch Neurol, 1992. 49(8): p. 849-54.
71. Pagani M, Chio A, Valentini MC, et al., Functional pattern of brain FDG-PET in
pt

amyotrophic lateral sclerosis. Neurology, 2014. 83(12): p. 1067-74.

72. Van Laere K, Vanhee A, Verschueren J, et al., Value of 18fluorodeoxyglucose-positron-
emission tomography in amyotrophic lateral sclerosis: a prospective study. JAMA Neurol,
ce

2014. 71(5): p. 553-61.

73. Turner MR, Cagnin A, Turkheimer FE, et al., Evidence of widespread cerebral microglial
activation in amyotrophic lateral sclerosis: an [11C](R)-PK11195 positron emission
Ac

tomography study. Neurobiol Dis, 2004. 15(3): p. 601-9.

74. Zurcher NR, Loggia ML, Lawson R, et al., Increased in vivo glial activation in patients with
amyotrophic lateral sclerosis: assessed with [(11)C]-PBR28. Neuroimage Clin, 2015. 7: p.
409-14.
75. Alshikho MJ, Zurcher NR, Loggia ML, et al., Glial activation colocalizes with structural
abnormalities in amyotrophic lateral sclerosis. Neurology, 2016. 87(24): p. 2554-2561.
76. Turner MR, Hammers A, Al-Chalabi A, et al., Distinct cerebral lesions in sporadic and
'D90A' SOD1 ALS: studies with [11C]flumazenil PET. Brain, 2005. 128(Pt 6): p. 1323-9.
77. Wicks P, Turner MR, Abrahams S, et al., Neuronal loss associated with cognitive
performance in amyotrophic lateral sclerosis: an (11C)-flumazenil PET study. Amyotroph
Lateral Scler, 2008. 9(1): p. 43-9.
78. Warnock G, Sommerauer M, Mu L, et al., A first-in-man PET study of [(18)F]PSS232, a
fluorinated ABP688 derivative for imaging metabotropic glutamate receptor subtype 5. Eur
J Nucl Med Mol Imaging, 2017.
79. Cohen-Adad J, El Mendili MM, Morizot-Koutlidis R, et al., Involvement of spinal sensory
pathway in ALS and specificity of cord atrophy to lower motor neuron degeneration.
Amyotroph Lateral Scler Frontotemporal Degener, 2013. 14(1): p. 30-8.
80. El Mendili MM, Cohen-Adad J, Pelegrini-Issac M, et al., Multi-parametric spinal cord MRI
as potential progression marker in amyotrophic lateral sclerosis. PLoS One, 2014. 9(4): p.
e95516.
* A study suggesting the potential use of combined spinal cord imaging techniques as a
multimodal marker of clinical progression in ALS

t
81. Valsasina P, Agosta F, Benedetti B, et al., Diffusion anisotropy of the cervical cord is

ip
strictly associated with disability in amyotrophic lateral sclerosis. J Neurol Neurosurg
Psychiatry, 2007. 78(5): p. 480-4.
82. Nair G, Carew JD, Usher S, et al., Diffusion tensor imaging reveals regional differences in

cr
the cervical spinal cord in amyotrophic lateral sclerosis. Neuroimage, 2010. 53(2): p. 576-
83.
83. Ikeda K, Murata K, Kawase Y, et al., Relationship between cervical cord 1H-magnetic

us
resonance spectroscopy and clinoco-electromyographic profile in amyotrophic lateral
sclerosis. Muscle Nerve, 2013. 47(1): p. 61-7.
84. Carew JD, Nair G, Pineda-Alonso N, et al., Magnetic resonance spectroscopy of the cervical
an
cord in amyotrophic lateral sclerosis. Amyotroph Lateral Scler, 2011. 12(3): p. 185-91.
85. Agosta F, Sala S, Valsasina P, et al., Brain network connectivity assessed using graph theory
in frontotemporal dementia. Neurology, 2013. 81(2): p. 134-43.
86. Verstraete E, Veldink JH, Mandl RC, et al., Impaired structural motor connectome in
M
amyotrophic lateral sclerosis. PLoS One, 2011. 6(9): p. e24239.
87. Buchanan CR, Pettit LD, Storkey AJ, et al., Reduced structural connectivity within a
prefrontal-motor-subcortical network in amyotrophic lateral sclerosis. J Magn Reson
Imaging, 2015. 41(5): p. 1342-52.
ed

88. Verstraete E, Veldink JH, van den Berg LH, et al., Structural brain network imaging shows
expanding disconnection of the motor system in amyotrophic lateral sclerosis. Hum Brain
Mapp, 2014. 35(4): p. 1351-61.
pt

** A network-based analysis of structural MRI alterations in ALS at two timepoints, showing

progressive disconnection consistent with pathological spreading from motor to extra-motor
networks.
ce

89. Schmidt R, de Reus MA, Scholtens LH, et al., Simulating disease propagation across white
matter connectome reveals anatomical substrate for neuropathology staging in amyotrophic
lateral sclerosis. Neuroimage, 2016. 124(Pt A): p. 762-769.
90. Zhou C, Hu X, Hu J, et al., Altered Brain Network in Amyotrophic Lateral Sclerosis: A
Ac

Resting Graph Theory-Based Network Study at Voxel-Wise Level. Front Neurosci, 2016.
10: p. 204.
91. Geevasinga N, Korgaonkar MS, Menon P, et al., Brain functional connectome abnormalities
in amyotrophic lateral sclerosis are associated with disability and cortical hyperexcitability.
Eur J Neurol, 2017. 24(12): p. 1507-1517.
92. Oxtoby NP, Alexander DC, and Euro Pc, Imaging plus X: multimodal models of
neurodegenerative disease. Curr Opin Neurol, 2017. 30(4): p. 371-379.
93. Schuster C, Hardiman O, and Bede P, Development of an Automated MRI-Based
Diagnostic Protocol for Amyotrophic Lateral Sclerosis Using Disease-Specific
Pathognomonic Features: A Quantitative Disease-State Classification Study. PLoS One,
2016. 11(12): p. e0167331.
94. Bede P, Iyer PM, Finegan E, et al., Virtual brain biopsies in amyotrophic lateral sclerosis:
Diagnostic classification based on in vivo pathological patterns. Neuroimage Clin, 2017. 15:
p. 653-658.
95. Schuster C, Hardiman O, and Bede P, Survival prediction in Amyotrophic lateral sclerosis
based on MRI measures and clinical characteristics. BMC Neurol, 2017. 17(1): p. 73.
96. Querin G, El Mendili MM, Lenglet T, et al., Spinal cord multi-parametric magnetic
resonance imaging for survival prediction in amyotrophic lateral sclerosis. Eur J Neurol,
2017. 24(8): p. 1040-1046.
97. Rohrer JD, Isaacs AM, Mizielinska S, et al., C9orf72 expansions in frontotemporal dementia
and amyotrophic lateral sclerosis. Lancet Neurol, 2015. 14(3): p. 291-301.
98. Bede P, Bokde AL, Byrne S, et al., Multiparametric MRI study of ALS stratified for the

t
C9orf72 genotype. Neurology, 2013. 81(4): p. 361-9.

ip
99. Agosta F, Ferraro PM, Riva N, et al., Structural and functional brain signatures of C9orf72
in motor neuron disease. Neurobiol Aging, 2017. 57: p. 206-219.
100. Floeter MK, Danielian LE, Braun LE, et al., Longitudinal diffusion imaging across the

cr
C9orf72 clinical spectrum. J Neurol Neurosurg Psychiatry, 2018. 89(1): p. 53-60.
101. Blain CR, Brunton S, Williams VC, et al., Differential corticospinal tract degeneration in
homozygous 'D90A' SOD-1 ALS and sporadic ALS. J Neurol Neurosurg Psychiatry, 2011.

us
82(8): p. 843-9.
102. Stanton BR, Shinhmar D, Turner MR, et al., Diffusion tensor imaging in sporadic and
familial (D90A SOD1) forms of amyotrophic lateral sclerosis. Arch Neurol, 2009. 66(1): p.
an
109-15.
103. Agosta F, Spinelli EG, Marjanovic IV, et al., Unraveling ALS due to SOD1 mutation
through the combination of brain and cervical cord MRI. Neurology, 2018.
M
ed
pt
ce
Ac
Figure legends

Figure 1. (A) Axial view of 7-T T*2-weighted images in a control subject and in a patient with

ALS. Slices are centered at the C2 (top row) and C4 (bottom row) vertebral levels. Hyperintensity is

clearly visible in both lateral segments of the spinal cord in the ALS patient, corresponding to the

corticospinal tract (CST) (arrows). Orientation: V = ventral, R = right. (B) Quantification of the

t
ip
T*2-weighted signal in both lateral segments of the spinal cord, between C2 and mid-C5. Signal in

the CST was normalized by the average signal in the dorsal segment of the cord (see regions of

cr
interest definition in right panel). The t-test compares normalized signal in the CST between control

us
and patient. *P < 10-7 (corrected). Reproduced with permission from [8].
an
Figure 2. Focal atrophy of the motor homunculus, when comparing patients with bulbar onset ALS

and those with limb onset. The blue colour indicates focal atrophy in patients with bulbar onset
M

compared with those with limb onset, while the orange colour indicates focal atrophy in patients

with limb onset compared with those with bulbar onset (p < 0.05 family-wise error). The relevant
ed

clusters are overlaid onto the study-specific averaged grey matter map. Reproduced with permission
pt

from [14].
ce

Figure 3. Axial and coronal T1-weighted images of the Montreal Neurologic Institute standard

brain show white matter (WM) tract damage in patients with classic ALS compared with healthy
Ac

control subjects. Voxel-wise group differences in fractional anisotropy (red), mean diffusivity

(blue), and radial diffusivity (purple) are shown. Results were overlaid on WM skeleton images

(light green; p < 0.05, corrected for multiple comparisons). L = left. R = right. Reproduced with

permission from [26].

Figure 4. Progressive decreases in resting state functional connectivity within the sensorimotor,

thalamic and visual resting state networks (RSN, shown thresholded at z > 3 in red-yellow in the

left panel, shown in green otherwise) over time and in relation to ALS functional rating scale

(ALSFRS-R) decline. Clusters of significant change (encircled in red for improved visibility) are

overlaid onto the respective RSN (green), as well as the mean filtered functional image in Montreal

Neurologic Institute space in the middle and right panel. Blue-lightblue – results without voxel-

t
ip
based morphometry (VBM) regressor. Pink – results with VBM regressor. Reproduced from [66]

(doi: 10.1016/j.nicl.2017.12.025; http://creativecommons.org/licenses/by/4.0/).

cr
us
Figure 5. Diagnostic ability of MRI models. Figure shows the ability of the combined MRI model

(i.e., cortical thickness + DT MRI) and the DT MRI measures in differentiating ALS and PUMN
an
patients from healthy controls and ALS-mimic disorders, respectively. Scaling coordinates of the

proximity matrix from random forest were used to represent the distance – in terms of MRI values –
M

between subjects with different diagnoses. Abbreviations: DT = diffusion tensor; HC = healthy

ed

controls; PUMN = predominantly upper motor neuron. Reproduced from [28].

pt
ce
Ac
 Figure 1
Ac
ce
pt
ed
M
an
us
cr
ip
t
Figure 3
Figure 2
Ac
ce
pt
ed
M
an
us
cr
ip
t
 Figure 4
Ac
ce
pt
ed
M
an
us
cr
ip
t
 Figure 5
Ac
ce
pt
ed
M
an
us
cr
ip
t