Content uploaded by Giorgio Conte
Author content
All content in this area was uploaded by Giorgio Conte on Sep 16, 2020
Content may be subject to copyright.
NEURO
A susceptibility-weighted imaging qualitative score of the motor
cortex may be a useful tool for distinguishing clinical phenotypes
in amyotrophic lateral sclerosis
Conte Giorgio
1
&Sbaraini Sara
2
&Morelli Claudia
3
&Casale Silvia
1
&Caschera Luca
1
&Contarino Valeria Elisa
1
&
Scola Elisa
1
&Cinnante Claudia
1
&Trogu Francesca
3,4
&Triulzi Fabio
1,4
&Silani Vincenzo
3,4
Received: 24 March 2020 /Revised: 21 July 2020 /Accepted: 27 August 2020
#European Society of Radiology 2020
Abstract
Objectives To distinguish amyotrophic lateral sclerosis (ALS) and its subtypes from ALS mimics and healthy controls based on
the assessment of iron-related hypointensity of the primary motor cortex in susceptibility-weighted imaging (SWI).
Methods We enrolled 64 patients who had undergone magnetic resonance imaging studies with clinical suspicions of ALS. The
ALS group included 48 patients; the ALS-mimicking disorder group had 16 patients. The ALS group was divided into three
subgroups according to the prevalence of upper motor neuron (UMN) or lower motor neuron (LMN) impairment, with 12
subjects in the UMN-predominant ALS group (UMN-ALS), 16 in the LMN-predominant ALS group (LMN-ALS), and 20 with
no prevalent impairment (C-ALS). The Motor Cortex Susceptibility (MCS) score was defined according to the hypointensity of
the primary motor cortex in the SWI sequence. Its diagnostic accuracy in differentiating groups was evaluated.
Results The MCS was higher in the ALS group than in the healthy control and ALS-mimicking disorder groups (p<0.001).
Among ALS subgroups, the MCS was significantly higher in the UMN-ALS group than in the healthy control (p<0.001),ALS-
mimicking disorder (p= 0.002), and LMN-ALS groups (p= 0.002) and higher in the C-ALS group than in the healthy control
group (p=0.019).AnMCSvalue≥2 showed specificity and a positive predictive value of 100% in the detection of both UMN-
ALS and C-ALS patients.
Conclusions The assessment of MCS in the SWI sequence could be a useful tool in supporting diagnosis in patients suspicious
for ALS with prevalent signs of UMN impairment or with no prevalence signs of UMN or LMN impairment.
Key Points
•The hypointensity of the primary motor cortex in susceptibility-weighted imaging could support the diagnosis of ALS.
•Our new qualitative score called MCS shows high specificity and positive predictive value in differentiating ALS patients with
upper motor neuron impairment from patients with ALS-mimicking disorders and healthy controls.
Keywords Amyotrophic lateral sclerosis .Magnetic resonance imaging .Motor neuron disease .Primary motor cortex
Abbreviations
ALS Amyotrophic lateral sclerosis
ALSFRS-R Amyotrophic Lateral Sclerosis
Functional Rating Scale
AUC Area under the curve
C-ALS Amyotrophic lateral sclerosis with
no prevalence of upper motor
neuron or lower motor neuron impairment
FLAIR Fluid-attenuated inversion recovery
sequence
FSE Fast spin-echo
ICC Interclass correlation
kCohen’skappa
*Sbaraini Sara
sara.sbaraini@gmail.com
1
Neuroradiology Unit, Fondazione IRCCS Ca’Granda Ospedale
Maggiore Policlinico, via Francesco Sforza 35, Milan, Italy
2
Neuroradiology Unit, Department of Radiology, ASST Santi Paolo e
Carlo, San Carlo Borromeo Hospital, via Pio II n. 3, Milan, Italy
3
Department of Neurology-Stroke Unit and Laboratory of
Neuroscience, Istituto Auxologico Italiano IRCCS, piazzale Brescia
20, Milan, Italy
4
Department of Pathophysiology and Transplantation, Università
degli Studi di Milano, via Festa del Perdono 7, Milan, Italy
European Radiology
https://doi.org/10.1007/s00330-020-07239-0
LMN Lower motor neuron
LMN-ALS Lower motor neuron - predominant
amyotrophic lateral sclerosis
MCS Motor Cortex Susceptibility score
MRI Magnetic resonance imaging
ROC Receiver operating characteristic curve
SWI Susceptibility-weighted imaging
UMN Upper motor neuron
UMN-ALS Upper motor neuron - predominant
amyotrophic lateral sclerosis
Introduction
Amyotrophic lateral sclerosis (ALS) is a progressive neurode-
generative disorder that is characterized by a variable combi-
nation of upper motor neuron (UMN) and lower motor neuron
(LMN) dysfunction. There is a wide degree of clinical in-
volvement of the UMN and LMN, and each of them leads to
specific features. Currently, the diagnosis of ALS is challeng-
ing and often delayed due to a lack of any biological markers.
The revised El Escorial Criteria [1] are the most widely
accepted criteria for the diagnosis and classification of ALS.
However, they were originally developed for research pur-
poses, so they are not always easily applicable in clinical prac-
tice. Therefore, for that reason, neurologists currently use
many informal classifications in the clinical setting to improve
patient management. Recently, new phenotype classifications
of ALS were proposed to combine a formal classification with
clinical utility based on the degree of UMN and LMN impair-
ment at clinical examination: UMN-predominant ALS
(UMN-ALS), classic ALS with no prevalence of UMN or
LMN impairment (C-ALS), and LMN-predominant ALS
(LMN-ALS) [2].
The signs of UMN impairment are increased tendon re-
flexes, spasticity, and an extensor plantar response. Early di-
agnosis is often difficult because these signs are masked by
limb weakness caused by LMN degeneration. The assessment
of LMN involvement is supported by electromyography and
muscle biopsy, and transcranial magnetic stimulation has been
largely utilized in clinical practice to evaluate UMN degener-
ation [3,4]. Nevertheless, no reliable test is recognized in
guidelines to assess UMN dysfunction objectively [5,6].
The revised El Escorial Criteria recommend the use of neu-
roimaging studies only to exclude other possible causes of
UMN or LMN impairment. However, magnetic resonance
imaging (MRI) studies have recently identified the
hypointensity of the motor cortex using T2*-weighted images
and susceptibility-weighted imaging (SWI) as a valuable but
inconstant sign of UMN disease [7–9].This hypointensity was
proven to be associated with abnormal iron deposition in the
precentral cortex of ALS patients, and it is believed to con-
tribute to UMN damage [7]. Current evidence shows that
motor cortex hypointensity could be a suitable marker of
UMN impairment.Thus, the aim of this study is to build a
diagnostic algorithm for ALS based on the clinical predomi-
nance of UMN or LMN impairment, as well as the assessment
of iron-related hypointensity of the primary motor cortex in
susceptibility magnitude images.
Materials and methods
The research protocol was approved by the Ethics Committee
of the IRCCS Istituto Auxologico Italiano and is in accor-
dance with the principles of the Declaration of Helsinki for
experiments involving humans. Written informed consent was
obtained from all enrolled participants.
Study participants
WeWe retrospectively enrolled all patients who had consecu-
tively undergone brain MRI studies under clinical suspicion of
ALS at the IRCCS Istituto Auxologico Italiano-San Luca
Hospital of Milan (Italy) between January 2016 and
December 2017. According to the clinical protocol, the pa-
tients underwent MRI within 1 day of clinical assessment.
Clinical evaluation was performed by two neurologists with
more than 10 years of experience in the management of motor
neuron disorders. The following clinical data were collected:
disease duration, score on the revised Amyotrophic Lateral
Sclerosis Functional Rating Scale (ALSFRS-R), the Penn
UMN score, and the predominance of signs of UMN or
LMN impairment. The exclusion criteria were as follows: (a)
concomitant psychiatric or other neurological diseases; (b) the
presence of MRI artifacts; and (c) brain lesions involving the
motor cortex and corticospinal tracts not related to motor neu-
ron disease at the MRI study. Forty-two ALS patients were
reported in a previous study that investigated the susceptibility
properties of the motor cortex in a cohort of ALS patientswith
quantitative susceptibility mapping [10].
At the end of the diagnostic pathway and after a follow-up
period, the enrolled patients were divided into two groups.
The first group comprised those who were finally diagnosed
with ALS (ALS group) according to the Revised El Escorial
Criteria, while those with different diagnoses were assigned to
the group of ALS-mimicking disorders [11]. The patients di-
agnosed with ALS were then divided into three subgroups:
UMN-predominant ALS (UMN-ALS), classic ALS with no
prevalent signs of UMN or LMN impairment (C-ALS), and
LMN-predominant ALS (LMN-ALS) [2].The patients en-
rolled in the group of the ALS-mimicking disorders were
not classified according to the predominance of motor neuron
signs because of the small size of the sample.
We also included 28 healthy controls, who were recruited
from volunteers and non-blood relatives of the patients. These
Eur Radiol
subjects were enrolled as a control group in a previous study
[10]. The exclusion criteria were as follows: (a) a history of
psychiatric or neurological disorders; (b) history of substance
abuse; (c) the presence of image artifacts; and (d) brain MRI
showing abnormal findings other than sporadic small gliotic
lesions in the white matter.
Image acquisition
The MRI study was performed with a 3-T SIGNA General
Electric scanner (GE Healthcare Medical Systems). The MRI
protocol included the whole-brain three-dimensional sagittal
FSPGR BRAVO T1-weighted sequence, the whole-brain 3D
sagittal fluid-attenuated inversion recovery (FLAIR) se-
quence, the axial T2-weighted fast spin-echo (FSE) sequence,
and the SWI sequence. The SWI sequence consisted of the
three-dimensional gradient–recalled multi-echo sequence
(SWAN). The susceptibility magnitude and phase images
were collected from the SWAN sequence with the following
parameters: repetition time = 39 ms; 7 echoes with TE1 =
24 ms and ΔTE = 3.3 ms; pixel spacing = 0.468 mm; slice
thickness = 1.4 mm; spacing between slices = 0.7 mm; flip
angle = 20°; and a 416 × 320 matrix. By default, the echoes
were averages, and the phase images were saved after high-
pass filtering by the scanner for clinical purposes.
Image analysis
The brain MRI was independently evaluated by two residents
who had 2 years of experience in neuroradiology and were
blinded to clinical information. A consensus assessment was
then carried out. A curvilinear multi-planar reconstruction of
the cerebral hemispheres in magnitude and phase images of
the SWI sequence was obtained along the cortical surface. The
primary motor cortex of both hemispheres was segmented into
three sub-regions according to the classic cortical homunculus
map: the upper limb sub-region based on the hand knob as an
anatomical landmark [12], the medial-dorsal sub-region cor-
responding to the cortical representation of muscles of the
lower limb, and the lateral-ventral sub-region corresponding
to the bulbar musculature.
The primary motor cortex intensity was assessed separately
for the six sub-regions (three in each hemisphere) using an
ordinal score of 0 to 2, with 0 indicating normal intensity
(similar to post-central and superior frontal gyri), 1 indicating
mild hypointensity (at least one-third of the sub-region cortex
similar to corpus callosum intensity), and 2 indicating marked
hypointensity (at least one-third of the sub-region cortex sim-
ilar to veins intensity) (Fig. 1)[9,13]. The sum of each indi-
vidual sub-region score was used to generate the overall score
called the Motor Cortex Susceptibility (MCS) score, which
ranged from 0 to 12 (maximum score: 2 points × 6 sub-
regions = 12 points). The MCS was calculated on both mag-
nitude and phase images.
Statistical analysis
All statistical analyses were performed using SPSS Statistics
software (version 22; IBM). All results with pvalues < 0.05
were considered significant. The inter-observer agreement for
each individual sub-region and the MCS score was calculated
with Cohen’s kappa (k) and interclass correlation (ICC), re-
spectively. The Shapiro-Wilk test was used to assess the nor-
mality of continuous variables. Spearman’s test was used to
assess the correlation between continuous variables.
Kruskal-Wallis tests were used to compare non-parametric
continuous variables between groups, and post hoc pairwise
tests were carried out with Bonferroni’s correction for multi-
ple comparisons. We used the area under the curve (AUC)
value of the receiver operating characteristic (ROC) curve to
evaluate the accuracy of MCS in differentiating subjects’
groups. Sensitivity, specificity, positive predictive value, and
negative predictive value were then calculated using the value
that maximizes specificity and positive predictive value as a
cut-off.
Results
Subject characteristics
A cohort of 64 patients with suspected ALS was examined.
The median duration of follow-up was 22 months (interquar-
tile range (IQR): 10–28 months). Forty-eight of them (75%)
were diagnosed with ALS (ALS group). The remaining 16
patients (25%) were diagnosed as reported in Table 1and
classified as ALS-mimicking disorders. Then, 28 healthy con-
trols were enrolled (median age 57 years; range 40–81 years,
males: 39%).
These three groups (ALS, ALS-mimicking disorders, and
healthy controls) did not differ significantly in terms of age
(p= 0.09) and sex (p= 0.96). Among ALS patients, there were
12 in the UMN-ALS group (42%), 20 in the C-ALS group
(25%), and 16 in the LMN-ALS group (33%). Table 2shows
a comparison of the demographic and clinical data among the
three ALS subgroups. None of these subgroups differed from
the ALS-mimicking disorders and healthy control groups in
terms of age (p= 0.12) and sex (p=0.91).
Inter-observer agreement
Table 3shows the data on inter-observer agreement (Cohen’s
Kvalues) for the visual score of motor cortex hypointensity
for each sub-region on magnitude and phase images. In sum-
mary, on magnitude images, the inter-observer agreement was
Eur Radiol
almost perfect or substantial in the assessment of every corti-
cal sub-region except for moderate agreement in the left bul-
bar region. The inter-observer agreement was excellent for the
MCS on magnitude images (ICC: 95.5 (95%CI: 93.6–96.8)),
while it was fair on phase images (ICC: 57.5 (95%CI: 52.8–
61.8)). For this reason, phase images were not further consid-
ered in the statistical analysis.
MCS comparison between ALS, healthy controls, and
ALS-mimicking disorders
There was a significant difference in MCS among groups
(Kruskall-Wallis test, chi-squared = 15.95, p< 0.001). In par-
ticular, MCS was higher in the ALS group (median: 0; IQR:
0–2) than in the healthy control group (median: 0; IQR = 0–0)
(p= 0.001, see box plot in Fig. 2). There was no significant
difference between the ALS-mimicking disorder group (me-
dian: 0; IQR: 0–0) and the other two groups, although there
was a tendency toward significance for the comparison be-
tween ALS-mimicking disorders and ALS (p=0.081).
As shown in Fig. 2, ALS had great variability in MCS. In
this group, there was no correlation of MCS with disease
Table 1 ALS-mimicking disorders diagnosed in our cohort of patients
ALS-mimicking disorders Number
Cervical polyradiculopathy 1
Hereditary spastic paraparesis 2
Metabolic myelopathies 1
Corticobasal degeneration 3
Cervical myeloradiculopathy 2
Parkinsonism 5
Frontotemporal dementia 2
TOT. 16
Fig. 1 MCS assessment on the
curvilinear multi-planar recon-
struction of SWI sequence (a,c,
e), the same images are
shown with the colored marked
sub-regions (b,d,f): lower limbs
marked in blue, upper limbs in
red, bulbar in yellow. In the cases
represented in a(MCS = 0) and e
(MCS=8,score0+2+2forboth
sides), the observers totally
agreed, while in the case
represented in c, the observers
disagreedinscoringtheright
upper limb region (for the first
observer the MCS was 4, 0 + 1 +
1 on both side; for the second
observer, the MCS was 3, 0 + 0 +
1 on the right, 0 + 1 + 1 on the
left)
Eur Radiol
duration (Rho = 0.10, p= 0.49) and ALSFRS-R score (Rho =
0.13, p= 0.46), while MCS positively correlated with the
UMN score (Rho = 0.48, p= 0.001). The AUC of MCS was
65.4 (95%CI: 48–81) in differentiating ALS from healthy
controls and ALS-mimicking disorders.
MCS comparison between healthy controls, ALS-
mimicking disorders, and ALS phenotypes
There was a significant difference in terms of MCS among
groups (Kruskall-Wallis test, chi-squared = 36.73, p<0.001,
see box plot in Fig. 3). In particular, MCS was higher
(p< 0.002 for all comparisons) in the UMN-ALS group (me-
dian: 3.75; IQR: 0.25–5.75) than in the healthy control group
(median: 0; IQR: 0–0), ALS-mimicking disorder group (me-
dian: 0; IQR: 0–0), and LMN-ALS group (median: 0; IQR:
0–0). It was also higher (p= 0.019) in the C-ALS group (me-
dian: 0; IQR: 0–1.75) than in the healthy control group. No
other differences were detected.
In differentiating UMN-ALS versus healthy controls and
ALS-mimicking disorders, the ROC analysis of MCS showed
an AUC of 87.0 (95% CI: 72.0–100) with a cutoff value of ≥2
having an sensitivity of 75% (95CI%: 42.8–94.5), specificity
of 100% (95%CI: 100–100), positive predictive value of
100% (95%CI: 100–100), and negative predictive value of
95.2% (88.2–98.2), as shown in Fig. 4. In differentiating C-
ALS versus healthy controls and ALS-mimicking disorders,
the ROC analysis of MCS showed an AUC of 65.0 (95% CI:
48.7–81.2) with a cutoff value of ≥2 having a sensitivity of
26.3% (95CI%: 9.1–51.5), specificity of 100% (95%CI: 91.9–
100), positive predictive value of 100% (95%CI: 100–100),
and negative predictive value of 75.9% (70.6–80.4).
Discussion
Our study shows that the MCS is significantly higher in cases of
ALS than in the healthy control and ALS-mimicking disorder
groups. MCS ≥2 yields a high specificity and positive predictive
value in the detection of both UMN-ALS and C-ALS patients
(specificity = 100%; negative predictive value = 100%). We pro-
pose a diagnostic flowchart based on MCS (Fig. 5), in which
MRI is used to rule out other possible causes of motor neuron
impairment and to support the diagnosis of UMN-ALS and C-
ALS. In particular, if a patient has a clinical suspicion of ALS
with prevalent signs of UMN impairment, the MCS is supportive
for confirming the diagnosis when MCS ≥2 (positive predictive
value = 100%) and excluding this clinical hypothesis when MCS
is < 2 (negative predictive value = 95.2%).
If a patient has a clinical suspicion of C-ALS, the MCS is
useful for confirming the diagnosis (MCS ≥2: positive predictive
value = 100%) but not for excluding the hypothesis of ALS when
MCS < 2 (negative predictive value = 75.9%). On the other
hand, the MCS loses its diagnostic relevance when a patient
has clinical suspicion of ALS with prevalent signs of LMN im-
pairment. We think that it is crucial to evaluate the clinical prev-
alence of UMN or LMN impairment and the MCS together in
order to optimize the diagnostic pathway of ALS.
More than 25 years ago, a marked MRI signal loss was
described for the first time in T2-/T2*-weighted images of
the precentral gyrus of ALS patients [14,15]. More recently,
signal changes in the motor cortex of an ALS population were
Table 2 Demographic and clinical characteristics of ALS patients divided into phenotype subgroups. Qualitative variables are reported in frequencies,
continuous variables in median and interquartile range (IQR)
Demographic and clinical variables C-ALS n= 20 UMN-ALS n= 12 LMN-ALS n=16 pvalue
Sex (male/female) 7/13 5/7 8/8 0.66
Age (years) 58 (50–65) 62.5 (58–68.75) 67 (57–73) 0.59
Disease duration (months) 18 (6.25) 29 (8–5-62.25) 17 (12–18.5) 0.06
ALSFRS-R scale 39 (31–42.5) 38.5 (30.25–43.5) 38 (34–42.5) 0.98
Penn UMN score 10 (6–13) 21.5 (19–24) 3 (0–6) <0.001
Abbreviations:c-ALS, classic amyotrophic lateralsclerosis; UMN-ALS, amyotrophic lateral sclerosis with prevalence of upper motor neuron impairment;
LMN-ALS, amyotrophic lateral sclerosis with prevalence of lower motor neuron impairment; IQR, interquartile range
Note: variable with a significant pvalue (p< 0.05) is in italicized
Table 3 Inter-observer agreement in scoring motor cortex
hypointensity for each sub-region expressed as Cohen’skvalues with
95% confidence interval (CI)
Side Sub-region Agreement κ(95%CI)
Right Lower limb 0.83 (0.83–0.83)
Upper limb 0.82 (0.82–0.82)
Bulbar 0.77 (0.77–0.77)
Left Lower limb 0.71 (0.71–0.71)
Upper limb 0.69(0.69–0.69)
Bulbar 0.50 (0.50–0.50)
Abbreviations:CI, confidence interval
Eur Radiol
also demonstrated with the SWI sequence [9,13,16], which is
the most sensitive qualitative MRI technique in the detection
of iron deposition in brain structures [8,9]. Quantitative MRI
techniques such as quantitative susceptibility mapping have
also been performed recently to quantify the magnetic suscep-
tibility changes in the precentral cortex of ALS patients
[17–19]. These studies showed an increased magnetic suscep-
tibility in the motor cortex of ALS patients, which is in line
with our results. Notably, the quantitative susceptibility map-
ping calculation needs post-processing and is not easily appli-
cable in clinical practice, in contrast to the qualitative evalua-
tion of the SWI sequence. Moreover, in line with our results,
the inter-observer agreement in the assessment of motor cor-
tex hypointensity in SWI sequences has already been reported
to be good to excellent [13,16], thus supporting its reliability
in a clinical setting.
The iron-related hypointensity is known to be very heteroge-
neous in ALS populations, with more conspicuous changes ob-
served in patients with higher UMN impairment [9,13,20].
These data are in agreement with our results indicating both a
wide range of MCS in the ALS group and a strong correlation
between UMN score and MCS. Until now, only Vazquez-Costa
et al [13] had investigated the susceptibility changes in the motor
cortex by grouping ALS patients according to the predominance
of UMN or LMN impairment signs (i.e., UMN-ALS, C-ALS,
and LMN-ALS). In agreement with Vazquez-Costa et al, our
study showed a similar MCS between C-ALS and UMN-ALS
groups and higher MCS in UMN-ALS patients than LMN-ALS
patients. We can speculate that the difference in MCS between
UMN-ALS and LMN-ALS may reflect a difference in the path-
ogenetic mechanism.
Histological studies have shown that the hypointensity of
the precentral gyrus in ALS is due to pathological iron accu-
mulation in the form of ferritin in the middle and deep layers
of the motor cortex. This iron probably causes oxidative
stress, microglial activation, and motor neuron degeneration
[7,9,14].We think that iron overload in the motor cortex
might play a primary role in the pathogenesis of UMN-ALS.
In LMN-ALS, the damage of UMNs could be due to an-
other mechanism resulting in primary LMN degeneration and
Fig. 2 Box plot representing the
MCS in healthy controls (HC), in
ALS and in ALS-mimicking
disorder (MIMIC) groups
Fig. 3 Box plot representing the
values of MCS in ALS
phenotypes (i.e., C-ALS, UMN-
ALS, and LMN-ALS), healthy
controls (HC), and ALS-
mimicking disorders (MIMIC)
Eur Radiol
unrelated to iron accumulation in the motor cortex.
Interestingly, iron deposition into the primary motor cortex
was histologically proven in patients with UMN impairment
[7,9], but not in patients with isolated LMN impairment (pro-
gressive muscular atrophy) at clinical examination. Further
studies are required to deepen the understanding of the role
of iron deposition in both UMN-ALS and LMN-ALS.
The C-ALS group did not statistically differ from the
UMN-ALS group in terms of MCS, but it showed a wide
range of MCS values (see box plot in Fig. 3), which could
reflect heterogeneity of the patients enrolled in the C-ALS
group. Our results are also interesting in light of recent phar-
macological trials investigating the possible therapeutic role
of iron chelators in ALS patients. A reduction of oxidative
stress and lower disease progression have already been de-
scribed in murine models of ALS [21,22], and preliminary
results of a pilot clinical trial reported lower disease progres-
sion in humans as well [23]. A reliable systematic evaluation
of the MCS in the SWI sequence might be a useful supportive
tool in the selection of ALS patients who would most benefit
from therapy with iron chelators.
Some advantages of assessing the MR images with a cur-
vilinear multi-planar reconstruction in comparison with the
original axial MR images should be acknowledged: (1) the
possibility to visualize simultaneously the six primary motor
cortex sub-regions of each side, corresponding to the upper
limbs, lower limbs, and bulb; (2) an easier assessment of the
ventral-lateral region (bulb) whose the orientation is almost
perpendicular to the axial plane of the original acquisition.
Our study has some limitations that need to be discussed.
First, we have to take into consideration the small sample size
of the groups, which also prevented us from categorizing pa-
tients with ALS-mimicking disorders in the UMN-ALS or
LMN-ALS mimickers. In addition, the predominance of
Fig. 5 Proposal of a diagnostic
flowchart based on MCS and
ALS phenotypes. Abbreviations:
ALSFRS-R, Amyotrophic Lateral
Sclerosis Functional Rating
Scale; LMN, lower motor neuron;
MCS, Motor Cortex
Susceptibility score; MR,
magnetic resonance; NPV,
negative predictive value; PPV,
positive predictive value; SWI,
susceptibility-weighted imaging;
UMN, upper motor neuron
Fig. 4 ROC curve showing the accuracy of MCS in differencing UMN-
ALS from healthy controls (HC) and ALS-mimicking disorders
(MIMIC). Area under the curve (AUC) = 87. The red dot shows the cutoff
(MCS = 2)
Eur Radiol
UMN or LMN impairment (i.e., UMN-ALS, C-ALS, and
LMN-ALS) can be defined only clinically in the absence of
objective scales. Since iron deposition in the cerebral cortex
has been described in other neurologic conditions which have
not been included in this study (such as neurodegenerative and
inflammatory diseases [24,25]), further studies are needed for
the validation of the proposed diagnostic algorithm. Finally,
the follow-up duration did not allow us to confirm the diag-
nosis in patients with suspected ALS.
In conclusion, the SWI sequence allows for a reliable evalu-
ation of hypointensity in the motor cortex of ALS patients. We
have proposed MCS as a new qualitative score, which showed
high specificity and positive predictive value in differentiating
UMN-ALS and C-ALS from ALS-mimicking disorders and
healthy controls. The results suggest that it could have clinical
usefulness in the confirmation of these diagnoses.
Funding The authors state that this work has not received any funding.
Compliance with ethical standards
Guarantor The scientific guarantor of this publication is Dr. Giorgio
Conte.
Conflict of interest The authors of this manuscript declare no relation-
ships with any companies whose products or services may be related to
the subject matter of the article.
Statistics and biometry One of the authors has significant statistical
expertise.
Informed consent Written informed consent was obtained from all sub-
jects in this study.
Ethical approval Institutional Review Board approval was obtained.
Study subjects or cohorts overlap Some study subjects or cohorts have
been previously reported in Contarino VE, Conte G, Morelli C, Trogu F,
Scola E, Calloni SF, Sanmiguel Serpa LC, Liu C, Silani V, Triulzi F
(2020) Toward a marker of upper motor neuron impairment in amyotro-
phic lateral sclerosis: a fully automatic investigation of the magnetic sus-
ceptibility in the precentral cortex. Eur J Radiol 124:108815.
Methodology
•Retrospective
•observational
•performed at one institution
References
1. Lodoloh A, Drory V, Hardiman O et al (2015) A revision of the El
Escorial criteria –2015. Amyotroph Lateral Scler frontotemporal
Degener 16(5–6):291–292
2. Al-Chalabi A, Hardiman O, Kiernan MC, Chiò A, Rix-Brooks B,
Van Den Berg LH (2016) Amyotrophic lateral sclerosis: moving
towards a new classification system. Lancet Neurol 15:1182–11943
3. Pouget J, Trefouret S, Attarian S (2000) Transcranial magnetic
stimulation (TMS): compared sensitivity of different motor re-
sponse parameters in ALS. Amyotroph Lateral Scler Other Motor
Neuron Disord 1(Suppl. 2):S45–S49
4. Van den Bos MAJ, Geevasinga N, Higashihara M, Menon P, Vucic
S (2019) Pathophysiology and diagnosis of ALS: insights from
advances in neurophysiological techniques. Int J Mol Sci 20(11):
2818
5. Huynh W, Simon NG, Grosskreutz J, Turner MR, Vucic S, Kiernan
MC (2016) Assessment of the upper motor neuron in amyotrophic
lateral sclerosis. Clin Neurophysiol 127:2643–2660
6. Swash M (2012) Why are upper motor neuron signs difficult to
elicit in amyotrophic lateral sclerosis?: figure 1. J Neurol
Neurosurg Psychiatry 83:659–662
7. Kwan JY, Jeong SY, Van Gelderen P et al (2012) Iron accumula-
tion in deep cortical layers accounts for MRI signal abnormalities in
ALS: correlating 7 tesla MRI and pathology. PLoS One 7:e35241
8. Roeben B, Wilke C, Bender B, Ziemann U, Synofzik M (2019) The
motor band sign in ALS: presentations and frequencies in a consec-
utive series of ALS patients. J Neurol Sci 406:116440
9. Adachi Y, Sato N, Saito Y et al (2015) Usefulness of SWI for the
detection of Iron in the motor cortex in amyotrophic lateral sclero-
sis. J Neuroimaging 25:443–451
10. Contarino VE, Conte G, Morelli C et al (2020) Toward a marker of
upper motor neuron impairment in amyotrophic lateral sclerosis: a
fully automatic investigation of the magnetic susceptibility in the
precentral cortex. Eur J Radiol 124:108815. https://doi.org/10.
1016/j.ejrad.2020.108815
11. Turner MR, Talbot K (2013) Mimics and chameleons in motor
neurone disease. Pract Neurol 13:153–164
12. Yousry TA, Schmid UD, Alkadhi H et al (1997) Localization of the
motor hand area to a knob on the precentral gyrus. A new landmark.
Brain 120(Pt 1):141–157
13. Vázquez-Costa JF, Mazón M, Carreres-Polo J et al (2018) Brain
signal intensity changes as biomarkers in amyotrophic lateral scle-
rosis. Acta Neurol Scand 137:262–271
14. Oba H, Araki T, Ohtomo K et al (1993) Amyotrophic lateral scle-
rosis: T2 shortening in motor cortex at MR imaging. Radiology
189:843–846
15. Ishikawa K, Nagura H, Yokota T, Yamanouchi H (1993) Signal
loss in the motor cortex on magnetic resonance images in amyotro-
phic lateral sclerosis. Ann Neurol 33:218–222
16. Sheelakumari R, Madhusoodanan M, Radhakrishnan A, Ranjith G,
Thomas B (2016) A potential biomarker in amyotrophic lateral
sclerosis: can assessment of brain iron deposition with SWI and
corticospinal tract degeneration with DTI help? AJNR Am J
Neuroradiol 37:252–258
17. Costagli M, Donatelli G, Biagi L et al (2016) Magnetic susceptibil-
ity in the deep layers of the primary motor cortex in amyotrophic
lateral sclerosis. NeuroImage Clin 12:965–969
18. Schweitzer AD, Liu T, Gupta A et al (2015) Quantitative suscepti-
bility mapping of the motor cortex in amyotrophic lateral sclerosis
and primary lateral sclerosis. AJR Am J Roentgenol 204:1086–
1092
19. Callaghan MF, Freund P, Draganski B et al (2014) Widespread age-
related differences in the human brain microstructure revealed by
quantitative magnetic resonance imaging. Neurobiol Aging 35:
1862–1872
20. Bowen BC, Pattany PM, Bradley WG et al (2000) MR imaging and
localized proton spectroscopy of the precentral gyrus in amyotro-
phic lateral sclerosis. AJNR Am J Neuroradiol 21:647–658
21. Golko-Perez S, Amit T, Bar-Am O et al (2017) A novel Iron
Chelator-radical scavenger ameliorates motor dysfunction and im-
proves life span and mitochondrial biogenesis in SOD1G93A ALS
mice. Neurotox Res 31:230–244
Eur Radiol
22. Jeong SY, Rathore KI, Schulz K, Ponka P, Arosio P, David S
(2009) Dysregulation of iron homeostasis in the CNS contributes
to disease progression in a mouse model of amyotrophic lateral
sclerosis. J Neurosci 29:610–619
23. Moreau C, Danel V, Devedjian JC et al (2018) Could conservative
Iron chelation lead to neuroprotection in amyotrophic lateral scle-
rosis? Antioxid Redox Signal 29:742–748
24. Park M, Moon Y, Han SH, Moon WJ (2019) Motor cortex
hypointensity on susceptibility-weighted imaging: a potential im-
aging marker of iron accumulation in patients with cognitive im-
pairment. Neuroradiology 61(6):675–683. https://doi.org/10.1007/
s00234-019-02159-3
25. Al-Radaideh A, Athamneh I, Alabadi H, Hbahbih M (2019)
Cortical and subcortical morphometric and iron changes in
relapsing-remitting multiple sclerosis and their association with
white matter T2 lesion load : a 3-tesla magnetic resonance imaging
study. Clin Neuroradiol 29(1):51–64. https://doi.org/10.1007/
s00062-017-0654-0
26. Petri S, Körner S, Kiaei M (2012) Nrf2/ARE signaling pathway:
key mediator in oxidative stress and potential therapeutic target in
ALS. Neurol Res Int 2012:1–7
27. Cosottini M, Cecchi P, Piazza S et al (2013) Mapping cortical
degeneration in ALS with magnetization transfer ratio and voxel-
based Morphometry. PLoS One 8:e68279
28. Carrì MT, Ferri A, Cozzolino M, Calabrese L, Rotilio G (2003)
Neurodegeneration in amyotrophic lateral sclerosis: the role of ox-
idative stress and altered homeostasis of metals. Brain Res Bull 61:
365–374
Publisher’snoteSpringer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.
Eur Radiol
