Personalised, image-guided, noninvasive brain stimulation in gliomas: Rationale, challenges and opportunities

Abstract

Malignant brain tumours are among the most aggressive human cancers, and despite intensive efforts made over the last decades, patients' survival has scarcely improved. Recently, high-grade gliomas (HGG) have been found to be electrically integrated with healthy brain tissue, a communication that facilitates tumour mitosis and invasion. This link to neuronal activity has provided new insights into HGG pathophysiology and opened prospects for therapeutic interventions based on electrical modulation of neural and synaptic activity in the proximity of tumour cells, which could potentially slow tumour growth. Noninvasive brain stimulation (NiBS), a group of techniques used in research and clinical settings to safely modulate brain activity and plasticity via electromagnetic or electrical stimulation, represents an appealing class of interventions to characterise and target the electrical properties of tumour-neuron interactions. Beyond neuronal activity, NiBS may also modulate function of a range of substrates and dynamics that locally interacts with HGG (e.g., vascular architecture, perfusion and blood-brain barrier permeability). Here we discuss emerging applications of NiBS in patients with brain tumours, covering potential mechanisms of action at both cellular, regional, network and whole-brain levels, also offering a conceptual roadmap for future research to prolong survival or promote wellbeing via personalised NiBS interventions.

Full text

EBioMedicine 70 (2021) 103514
Contents lists available at ScienceDirect
EBioMedicine
journal homepage: www.elsevier.com/locate/ebiom
Review
Personalise d, image-guide d, noninvasive brain stimulation in gliomas:
Rationale, challenges and opportunities
Giulia Sprugnoli
a , b , c , d
, Simone Rossi
d
, Alexander Rotenberg
e
, Alvaro Pascual-Leone
f , g
,
Georges El-Fakhri
h
, Alexandra J. Golby
c , 1
, Emiliano Santarnecchi
a , 1 , ∗
a
Berenson-Allen Center for Noninvasive Brain Stimulation, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA
b
Radiology Unit, Department of Medicine and Surgery, University of Parma, Parma, Italy
c
Image Guided Neurosurgery laboratory, Department of Neurosurgery and Radiology, Brigham and Wo men’s Hospital, Harvard Medical School, Boston, MA,
USA
d
Brain investigation and Neuromodulation Laboratory (Si-BIN Lab), Department of Medicine, Surgery and Neuroscience, Neurology and Clinical
Neurophysiology Unit, University of Siena, Siena, Italy
e
Department of Neurology and Division of Epilepsy and Clinical Neurophysiology, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA
f
Hinda and Arthur Marcus Institute for Aging Research and Center for Memory Health, Hebrew Senior Life, Boston, MA, USA
g
Guttmann Brain Health Institute, Institut Guttmann, Universitat Autonoma, Barcelona, Spain
h
Gordon Center for Medical Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
a r t i c l e i n f o
Article history:
Received 20 April 2021
Revised 12 July 2021
Accepted 19 July 2021
Keywo rds:
Brain tumours
Glioma
HGG
NiBS
Noninvasive brain stimulation
Neuromodulation
a b s t r a c t
Malignant brain tumours are among the most aggressive human cancers, and despite intensive effort s
made over the last decades, patients’ survival has scarcely improved. Recently, high-grade gliomas (HGG)
have been found to be electrically integrated with healthy brain tissue, a communication that facilitates
tumour mitosis and invasion. This link to neuronal activity has provided new insights into HGG patho-
physiology and opened prospects for therapeutic interventions based on electrical modulation of neural
and synaptic activity in the proximity of tumour cells, which could potentially slow tumour growth. Non-
invasive brain stimulation (NiBS), a group of techniques used in research and clinical settings to safely
modulate brain activity and plasticity via electromagnetic or electrical stimulation, represents an appeal-
ing class of interventions to characterise and target the electrical properties of tumour-neuron interac-
tions. Beyond neuronal activity, NiBS may also modulate function of a range of substrates and dynamics
that locally interacts with HGG (e.g., vascular architecture, perfusion and blood-brain barrier permeabil-
ity). Here we discuss emerging applications of NiBS in patients with brain tumours, covering potential
mechanisms of action at both cellular, regional, network and whole-brain levels, also offering a concep-
tual roadmap for future research to prolong survival or promote wellbeing via personalised NiBS inter-
ventions.
©2021 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license
( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
1. Introduction
Due to limited therapeutic options, incidence in relatively
young people, delayed diagnosis, and infiltration into the brain
parenchyma, malignant brain tumours rank fourth among all can-
cers in terms of number of years of life lost, despite representing
only 2% of all cancers [1] . Among primary brain cancers, glioblas-
toma (GBM) is the most frequent and aggressive high-grade glioma
(HGG, WHO glioma IV), with a mean survival of approximately 16–
∗Corresponding author.
E-mail address: esantarn@bidmc.harvard.edu (E. Santarnecchi).
1 These authors contributed equally to the work.
18 months from diagnosis [2] . Many novel nonsurgical and non-
pharmacologic therapies have been tested over the last decades as
adjuncts to the standard of care that includes surgery, radiation
and chemotherapy [3] . Among these, a tumour treating [electrical]
fields (TTFs) device (Optune- NovoTTF-100A System) that generates
transcranial electrical stimulation utilising alternating current at a
frequency of 200 kHz may work by disrupting mitosis in cancerous
cells. The device has been cleared by the FDA and recommended
in combination with chemotherapy (CHT) by the National Com-
prehensive Cancer Network as a potentially effective treatment for
patients with newly diagnosed GBM [4] . However, the benefit of
TTF in terms of patient survival has been modest (20.9 months in
the TTF-CHT group vs 16. 0 months in the CHT group [ 5 , 6 ]), and no
https://doi.org/10.1016/j.ebiom.2021.103514
2352-3964/© 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
( http://creativecommons.org/licenses/by-nc-nd/4.0/ )

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
benefits have been reported for patients with recurrent GBM [6] .
Thus, novel therapeutic strategies for HGG remain an unmet need.
The recent discovery of neuron-to-glioma synaptogenesis of-
fers the possibility for novel strategies to interfere with this new
pathophysiological behaviour [ 7 , 8 ], also delineating the new field
of cancer neuroscience [9] . While the concept of HGG promot-
ing neuronal excitability is not new [10] , the existence of synaptic
neuron-to-tumour connections, which lead HGG cells to depolar-
ize in response to neuronal spiking and proliferate, is completely
novel. Glioma cells reflect cellular subpopulations at various stages
of astrocytic and oligodendrocytic differentiation, such as oligoden-
droglial precursor cells (OPCs) [11] . Normal OPCs are able to form
synapses with neurons, a communication that regulates progenitor
cell proliferation, migration, differentiation and even synaptic plas-
ticity in the human brain [11] . Interestingly, only OPCs and stem
and progenitor cells (NPCs) are provided with a subtype of amino-
3-hydroxy-5-methyl-4-isoxazole propionic acid receptors (AMPA-R,
ionotropic glutamate receptors) lacking of a GluR2 subunit that
makes the receptors Ca
2 +
-permeable, in contrast to the AMPA-
R type of the differentiated neurons that are not permeable to
Ca
2 + [11] . Of note, Ca
2 +
-permeable AMPA-R are also physiologi-
cally expressed in interneurons and characterised by fast kinetics
thought to be crucial for neural development and synaptic plastic-
ity, in terms of triggering long-term potentiation (LTP) and defining
synaptic efficacy [12] . Surprisingly, glioma cells express the same
excitatory synaptic structures consisting of Ca
2 +
-permeable AMPA-
R observed in normal OPCs [ 7 , 8 ]. Also, similarly to the normal OPCs
forming the axoglial synapses, glioma cells have been observed
only in the postsynaptic part of these neuron-glioma synapses [ 7 , 8 ]
and the depolarising currents cause a calcium-ion influx into the
glioma cells that ultimately promote tumour proliferation and in-
vasiveness [ 7 , 8 ].
The above findings highlight a positive feedback mechanism
between increased neuronal excitation triggered by gliomas and
its impact on mitosis and migration, suggesting that controlling
neuronal excitability in patients with HGG may inhibit tumour
growth and proliferation, ultimately prolonging patient survival. To
this end, we aim to bring attention to novel therapeutic oppor-
tunities offered by noninvasive brain stimulations (NiBS; Fig. 1 ),
a group of neurostimulation techniques that include transcranial
magnetic stimulation (TMS), and transcranial electrical stimulation
(tES) [ 13 , 14 ]. The possibility to interfere/interact with ongoing neu-
ronal activity and impact behaviour/cognition supported by a spe-
cific brain area has led to the implementation of ad-hoc neurostim-
ulation protocols to affect motor and/or cognitive functions in the
healthy brain, as well as FDA-approved devices and protocols for
treatment of certain neurological and psychiatric conditions. NiBS
protocols are safe and well-tolerated and have demonstrable ca-
pacity to modulate cortical excitability [ 13 , 14 ], intracortical excita-
tion/inhibition (E/I) balance and cortical plasticity [15] , even in a
long-lasting manner via modifications of synaptic strength and ef-
ficacy mediated by mechanisms resembling LT P and long-term de-
pression (LTD) [16] , thus may offer an interesting opportunity for
modulation of neuron-to-glioma functional circuitry activity.
In light of recent findings of electrical HGG integration, NiBS
can be regarded as a potential tool to suppress the proposed
neuron-to-glioma communication by, for instance, inhibiting neu-
ronal populations surrounding HGG via LTD-like effects. On the
other hand, NiBS could also affect tumour growth via non-neuronal
effects. Indeed, NiBS has recently been found to modify cere-
bral and intratumoural perfusion [17] , the permeability of the
blood brain barrier (BBB) [18] , and to interact with microglia [19] ,
suggesting additional interventional -still unexplored- noninvasive
stimulation strategies for patients with brain tumours.
2. Transcranial stimulation
TMS leverages the principle of electro-magnetic induction by
which an intracranial electric field (E-field) is induced by a rapidly
fluctuating (i.e., 300–350 μs) magnetic field that penetrates into
the brain through the scalp and skull [15] , focally depolarising
neurons ( Fig. 1 A). TMS was introduced in 19 85 and several TMS
devices are currently approved by the FDA and other regulatory
agencies worldwide for the treatment of drug-resistant depres-
sion, obsessive compulsive disorder and migraine with aura, as
well as for presurgical mapping of eloquent areas including motor
and language areas [13] . When multiple stimuli are applied in a
repetitive fashion (repetitive TMS - rTMS), the stimulation induces
long-term plastic changes in the brain, modifying the efficacy of
synaptic communication by triggering LTP or long-term depres-
sion (LTD)-like mechanisms, depending on the specific frequency
applied. In particular, high-frequency rTMS ( > 1Hz) or intermittent
TBS –iTBS (short trains of impulses) usually increases cortical ex-
citability and causes LTP-l ike effects, while low frequency rTMS
( ≤1 Hz) or continuous TBS – cTBS (single train of pulses) more
frequently causes a decrease of cortical excitability and eventually
LTD-like effects ( Fig. 1 A) [15] . Physiological LTD induction is de-
pendant on N-methyl-D-aspartate (NMDA) receptors that are usu-
ally mildly stimulated via low frequency stimulation (LFS), leading
to a modest intracellular Ca
2 + elevation that in turn activates pro-
tein phosphatases responsible of downregulation of AMPA-R [20] .
LTD (and LTP)-like phenomena induced by TMS have been related
to glutamatergic NMDA-mediated transmission and relative influx
of Ca
2 +
into the post synaptic cells that downregulate the AMPA-R,
both in vitro and in vivo [21] . Apart from LFS, LTD can be physiolog-
ically also induced via baseline synaptic stimulation contempora-
neously with depolarisation (i.e., pairing), by administration of an
appropriate receptor agonist or via timed back-propagating action
potentials (i.e., spike-timing dependant plasticity - STDP) [20] . TMS
protocols have been adapted to match each of these LTD-inducing
mechanisms (for a comprehensive review see [21] ). Finally, long-
term effects of TMS seem to depend also on activity-dependant
brain-derived neurotrophic factor (BDNF) plasticity, gene induction
and modulation of multiple neurotransmitter levels [22] . Therefore,
TMS is thought to act via numerous mechanisms and pathways
to modify synaptic plasticity, potentially relevant to decrease the
neuronal-induced tumour growth.
In contrast to TMS that depolarises neurons thus generating
action potentials, tES involves almost imperceptible electrical cur-
rents (~2 mA) delivered by scalp electrodes, that reach the cortex
where they modulate the resting membrane potential, thus affect-
ing the excitability of pyramidal cortical neurons [23] without di-
rectly inducing neuronal firing ( Fig. 1 B). Optimised tES protocols
enable multielectrode solutions to target cortical regions with a
few centimeters resolution. Among tES methods, transcranial di-
rect current stimulation - tDCS, involves subthreshold depolarisa-
tion of neurons under the anode (generally resulting in an in-
crease of their excitability) and hyperpolarisation of those under
the cathode (decrease of excitability), respectively leading to en-
hancement or suppression of regional brain activity usually par-
alleled by cognitive/behavioural modifications linked with the role
of the targeted regions/networks [24] . Short-term effect of tDCS in-
volves voltage-dependant ion channels, while stimulation extend-
ing over a few minutes promotes LTP or LTD-like plasticity that can
also affect interconnected cortical and subcortical areas [23] , again
relying on NMDA receptors and Ca
2 + influx [25] . In addition, tDCS
acts on cortical excitation/inhibition balance via a modulations of
γ-aminobutyric acid (GABA), BDNF and glutamate/glutamine con-
centrations [23] . Other forms of tES that deliver different type of
current (i.e., alternating current or random noise) have been de-
veloped and tested during the last years among healthy subjects
2

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
Fig. 1. NiBS techniques. (a) Transcranial Magnetic Stimulation (TMS) can be applied as a single stimulation pulse (single pulse TMS, spTMS), pairs of stimuli separated
by variable intervals (paired pulse TMS, ppTMS) delivered to the same or different brain areas, or as trains of repetitive stimuli at various frequencies (repetitive TMS,
rTMS) to respectively measure cortical excitability, excitation/inhibition balance and induce long-lasting neuromodulation effects and changes in plasticity. Outside the motor
cortex, accurate targeting is guaranteed by a neuronavigation system that provides, on the basis of individual MRI data, the spatial coordinates of a target area allowing
the coil to be
held in the correct position during a stimulation session, as well as in subsequent ones. Long Term Depression (LTD) and Long Term Potentiation (LTP)-
like effects are mediated by multiple mechanisms, such as actions on GABA and NMDA receptors, gene induction and modulation of numerous neurotransmitters. When
spTMS is applied to the motor cortex, TMS elicits motor evoked potentials (MEPs) recorded via surface electromyography, whose amplitude reflects the excitability and
integrity of the corticospinal system, the conduction properties along peripheral motor pathways, and the degree of excitability of the motor cortex itself [27] . When ppTMS
is delivered, depending on the
specific stimulation parameters, intracortical Inhibition and Facili tati on can be assessed, respectively measuring the activity of GABAergic
and glutamatergic (inter)neurons as well as excitation/inhibition balance (E/I). (b) Transcranial electrical stimulation (tES) can be applied using 2 electrode or adopting
multifocal montages that allow for finer customisation of stimulation protocol and targeting accuracy over the region(s) of interest. tDCS: Direct Current Stimulation causes
a subthreshold depolarisation of neurons under the anode (increased excitability) and hyperpolarization of neurons under the cathode (decreased excitability). Transcranial
alternating current stimulation (tACS) is capable of modulating cortical rhythms, i.e. to entrain neuronal firing at a specific frequency and
causing enhancement of related
brain functions [ 28 , 29 ]. tACS depolarised and hyperepolarised neurons at a specific frequency (e.g., 10Hz) , increasing the probability of neuronal spiking in response to
other inputs during their depolarised phase via stochastic resonance mechanism [23] . As per the STDP law, synapses of neuronal network that have a resonance frequency
matching that of repetitive inputs are strengthened. Transcranial random noise stimulation (tRNS) delivers electrical noise in a wide frequency band (1–640 Hz) to modulate
cortical excitability. The injection of noise is thought to promote the excitability of pyramidal cells via stochastic resonance mechanism, but
activation of voltage-gated
sodium channels has been documented as well [23] . tRNS after-effects seems to be mediated by GABA receptors and voltage-gated sodium channels [23]
to increase their cognitive abilities as well on patient populations
to restore physiological neural activity [23] (see Fig. 1 B).
LTD-like changes in synaptic excitability could be relevant in
the new neuron-to-glioma context, where reduced probability of
neuronal firing after a presynaptic event would reduce the activa-
tion of Ca
2 +
AMPA-R in the post-synaptic glioma cell, leading to a
limited inflow of Ca
2 + signal mitosis-promoting and thus poten-
tially limiting the neuronal contribution to glioma growth [7–9] .
Notably, NiBS protocols differ conceptually from the currently-used
TTF paradigm in which alternating current is delivered at extra-
physiological frequencies (200 kHz) and is aimed at directly inter-
fering with cancer cell mitosis, rather than modulating neuronal
activity [26] . TTFs requires patients to shave their head and wear
the devices at least 18 h/day in order to be effective, whereas tES
offers a light and highly portable alternative that typically pro-
duces neurophysiological (i.e., modulation of cortical excitability)
and cognitive effects (i.e., performance increase) even with single,
relatively short (i.e., 30 min) sessions. Finally, a considerable per-
centage of patients receiving TTFs (up to 43%) reports dermatolog-
ical adverse effects such as dermatitis, erosions, ulcers, and infec-
tions due to the continuous wearing of the device [26] , while NiBS
has a good safety profile, with no adverse effects other than occa-
sional scalp itching and redness, especially for tES [14] . As for TMS,
the risk of a seizure induction is low, even in patients with pre-
existing epileptic conditions or who are taking medications which
potentially lower the seizure threshold [13] . Also, it must be no-
ticed that patients with brain tumour usually assume antiepileptic
treatment lowering the possibility of the occurrence of seizures.
Moreover, NiBS protocols that increase the risk of seizure are those
inducing hyperexcitability in the brain, that reasonably will not be
those applied in the case of patients with brain tumours since the
aim is to suppress the neuron-to-glioma communication via in-
hibitory protocols of stimulation, thus further lowering the possi-
bility of inducing seizures.
In conclusion, many of the neurophysiological peculiarities of-
fered by NiBS can potentially impact tumour pathophysiological
mechanisms that are currently not targeted by other therapeu-
tic approaches, including TTF that supposedly affects tumour mi-
tosis. Considering the standard of care provided to patients with
HGG, consisting in surgical excision, radiotherapy plus concomi-
tant chemotherapy and adjuvant chemotherapy, a specific tempo-
ral framework can be proposed for the implementation of NiBS
3

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
Fig. 2. Te mp ora l Framework for potential NiBS Applications. Potential applications of NiBS in patients with brain cancers are presented, considering the timeline of standard
of care therapy (e.g., surgical excision, radiotherapy plus concomitant chemotherapy, adjuvant chemotherapy). NiBS could be carried out via repetitive application in a long-
term perspective to cause LTD-li ke effects, while other applications could be limited to a defined temporal window matching standard therapies, whose efficacy could
be enhanced by concomitant application of NiBS, e.g. promoting drug delivery during CHT. Given the portability of tES in particular, extended home-use may be feasible.
Baseline assessment should include neuroimaging and/or
nuclear imaging data for image-guided personalised interventions to maximise effects towards relevant areas with
minimal side effects, e.g. targeting the solid mass to reduc e intra-tumoural perfusion, or the surrounding brain regions to inhibit tumour-promoting neural activity. Finally,
amelioration of neurological, cognitive and psychiatric symptoms –thereby avoiding or reducing additional pharmacological treatments —could also be considered in patients
with HGG, given the positive effects documented in other populations of patients. Note: CHT = chemotherapy, MRI = Magnetic Resonance Imaging, PET = Positron Emission
Tomography, TAMs = Tumour-Associated Macrophages.
( Fig. 2 ). We foresee a roadmap of the many opportunities through
which NiBS applications could be offered to patients at multiple
timepoints in their clinical course, possibly in a personalised man-
ner (e.g., considering tumour location, type, size, extension of per-
itumoural infiltrated tissue; Fig. 2 ). In the following sections, bio-
logical mechanisms for local and network-based NiBS applications
are presented, with suggestions for potential clinical trial protocols
in patients with HGG.
3. Local therapy
NiBS may be used to modulate/suppress synaptic signalling of
neurons surrounding an HGG tumour, therefore slowing down its
mitosis and migration rate. Moreover, additional, non-neuronal ef-
fects of NiBS have begun to emerge, such as modulation of perfu-
sion and permeability, as well as activation of microglia, each one
potentially relevant for brain tumour management.
3.1. Suppression of neuronal activity-regulated cancer growth
Recent pivotal work by two independant groups has demon-
strated that neuronal activity promotes the proliferation and in-
vasiveness of HGG in vivo [ 7 , 8 ]. One of the two groups, led by
Michelle Monje, has shown that neural activity promotes the mi-
tosis of cancer cells via a specific pathway involving synaptic pro-
tein neuroligin-3 (NLGN3) for adult and paediatric GBM, diffuse in-
trinsic pontine gliomas (DIPG), and anaplastic oligodendrogliomas
[30] . NLGN3, secreted in an activity-dependant manner by neu-
rons, inversely correlates with overall survival (OS) of adult GBM
patients [30] , and patient-derived orthotopic xenografts of paedi-
atric GBM, DIPG and adult GBM are unable to grow in NLGN3-
knockout mice [31] . These findings were recently complemented
by the observation of synaptic structures between HGG cells and
surrounding neurons in multiple models of HGG and DIPG, such
as in patient-derived xenografts, resected human tumour tissue,
and genetic mouse models [ 7 , 8 ]. As anticipated, these neuro-
gliomal synapses show the characteristics of glutamatergic chem-
ical synapses, specifically involving Ca
2 +
-permeable AMPA recep-
tors, with the glioma cells being exclusively postsynaptic. AMPA-
R have been observed along the tumour microtubes, represent-
ing long cellular processes crucial for tumour invasion and allow-
ing the connection of glioma cells into a functional communicat-
ing network, essential for transferring growth elements and factors
favouring treatment resistance [8] . The investigators observed fast
excitatory postsynaptic current propagating inside the HGG cells,
mediated by AMPA receptors and time-locked with the neuronal
spiking of the presynaptic neurons. This fast response is followed
by a long-lasting depolarizing current, probably depending on ex-
tracellular concentration of potassium ions rather than on synap-
tic activity. In total, approximately 31% of the observed GBM cells
showed at least one of these electrical responses [8] . Additionally,
electrical currents were found to spread in the network of inter-
connected glioma cells via gap junctions, strengthening also the
connectivity between the cells’ network. Finally, depolarisation cur-
rents cause a calcium transient inside the cells, ultimately driving
mitosis and invasiveness. Pharmacologic and genetic blockage of
AMPA receptors, as well as drug targeting gap junctions, reduced
the invasion and mitosis of glioma cells, in turn promoting mouse
survival.
Interestingly, such neuron-to-glioma communication is not uni-
lateral. In clinical settings, glioma patients have an increased risk
of seizures, and worsening seizure control correlates with recur-
rence [10] . Even if glioma cells are not able to spike, they have
been found to promote neuronal firing in order to create a posi-
tive feedback with neurons for their further activation via multiple
mechanisms, such as synaptogenic factor secretion, non-synaptic
glutamate release, and by reducing the activity of inhibitory in-
terneurons in the surrounding microenvironment [10] . Accordingly,
epileptiform activity has been found to arise in the infiltrated
parenchyma of glioma xenotransplants [32] , strongly dependant on
glutamate release [10] . Monje and colleagues further confirmed
and extended these data by showing increased high gamma ( γ,
70–110 Hz) activity–an index of neuronal activation— in infiltrated
tissue of patients with HGG via intraoperative electrocorticography
[7] . Overall, HGG is integrated into the brain and is even capable
of initiating a vicious circle to promote its mitosis and invasiveness
by amplifying the excitability of the surroundings neurons. Further
research is needed, including investigations directed at LGG, con-
4

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
sidering that LGG–and not HGG- are more frequently associated
with epilepsy in the clinical settings [33] .
Given the ability of rTMS and tDCS of causing long-lasting sup-
pression of neuronal spiking/excitability via LTD-like effects, these
techniques could represent safe and useful tools to modulate neu-
ronal firing and thus, hopefully, limit glioma mitosis and invasive-
ness. In support of this, it has already been shown that both high-
frequency rTMS (600 pulses over 30 min daily) and low-frequency
rTMS (1800 pulses over 30 min daily at 1 Hz) protocols applied
to the left or right dorsolateral prefrontal cortex for drug-resistant
depression are able to induce long-lasting changes in local corti-
cal activity, connectivity, perfusion, and even affect structural brain
properties [ 34 , 35 ]. Given the focality of TMS electric fields (E-field,
1-2 cm
3
), rTMS could be applied to several locations surround-
ing a small cortical lesion, consequently targeting a relatively lim-
ited bordering neuronal population; alternatively, a local or distant
single area could be targeted based on significant communication
with the HGG cells detected via neuroimaging modalities such as
functional MRI-based connectivity.
Compared to TMS, tDCS applies a broader E-field that can be
useful in the presence of larger lesions receiving inputs from mul-
tiple other brain regions. Despite not being able to directly in-
hibit the spiking of pyramidal neurons, the cathodal field can re-
duce the probability of the neuronal spiking in the targeted cor-
tical areas via LTD-like effects as well, representing an appeal-
ing approach to interrupt diffuse neuron-to-glioma communication
(see also Network-based therapeutic opportunities paragraph). Ad-
ditionally, both techniques can be safely combined in the same
protocol to induce a synergistic effect on brain excitability [13] . A
recent study simultaneously applying rTMS at inhibitory frequency
and cathodal tDCS over the motor cortex showed a stronger in-
hibitory modulation of motor evoked potential (MEPs amplitude,
an index of corticospinal excitability) compared to each technique
separately [36] . Finally, direct modulation of glioma cellular depo-
larisation by NiBS could not be a-priori excluded, i.e., making the
glioma cells less able to depolarise in response to neuronal spiking.
The potential decrease of tumour responsiveness to neuronal spik-
ing via NiBS need to be assessed, possibly starting from in vitro
and preclinical models.
3.2. Perfusion and permeability
Recently, extra-neuronal effects of NiBS have begun to receive
attention, with animal and human studies exploring the effect of
tES on vascular brain components. The first study exploring perfu-
sion effects of tDCS in humans was conducted on healthy subjects
receiving direct current stimulation (tDCS) during the acquisition
of a perfusion-sensitive Magnetic Resonance Imaging scan (Arterial
Spin Labelling, ASL [37] ). The authors showed modulation of Cere-
bral Blood Flow (CBF) in the targeted cortical regions, with a pos-
itive correlation to stimulation intensity. In particular, both anodal
and cathodal stimulation induced a significant CBF increase with
respect to baseline level, with anodal stimulation having a stronger
and more reliable vascular modulatory effect across repeated stim-
ulation blocks [37] . In gliomas, tumour perfusion is positively cor-
related with WHO grade and negatively with survival [38–40] . Par-
ticularly for GBM, neoangiogenesis and high perfusion are the most
distinctive histopathological features, related to extreme invasive-
ness and aggressive growth, as well as with markers of cell pro-
liferation (e.g., Ki67 index [ 38 , 39 ]). CBF and CBV (Cerebral Blood
Volume) represent validated and reliable imaging markers of tu-
mour progression, with increased CBF on perfusion-sensitive MRI
sequences predicting shorter PFS and OS [40] . Therefore, consider-
ing the potential importance of inhibiting tumour perfusion, an-
tibodies targeting neoangiogenesis pathways (e.g., Bevacizumab)
have been developed and tested, without finding however signifi-
cant benefit in OS in newly diagnosed nor in recurrent GBM, prob-
ably due to the single pathway targeting that causes a compen-
satory increase in other neoangiogenesis strategies [41] .
Electrical current has been shown to reduce intratumoural per-
fusion in extracranial tumours (i.e., breast and lung cancer, liver
metastases) when applied via two or more platinum electrodes lo-
cated directly inside of the tumour or in the surrounding tissue. In
the last two decades, the application of electrical stimulation (Elec-
trical Treatment, ET) to malignant visceral tumours has been found
to reduce tumour perfusion via a vasoconstriction phenomenon,
and even cause tumour necrosis when applied over multiple days
[42] . In addition, ET (usually delivering 10 0 0 V/cm at 5 kHz) seems
to potentiate the effect of chemotherapy (CHT), by decreasing tu-
mour blood flow, which in turn prolongs the contact with and con-
sequently the action of CHT agents on tumour cells. ET also mod-
ulates tumour trans-membrane permeability favoring drug inter-
nalization [42] , aligning with tDCS evidence in rat and endothelial
monolayer models [ 18 , 43 ]. For these reasons, many trials applying
electrochemotherapy, as it is now known, are being conducted on
colorectal tumours, basal cell carcinoma and even spine metastases
[44] . Targete d delivery of electrical stimulation to the solid tumour
mass could, in theory, reduce intratumoural perfusion similarly to
that observed in extracranial tumours and eventually even induce
tumour cell necrosis.
A recent pilot study by our group tested this possibility in pa-
tients with GBM ( n = 6) and lung metastasis in the brain ( n = 2)
[17] . Multifocal tDCS was delivered for 20 min with an MRI-
compatible device while the patient was inside the MRI scanner
allowing contemporaneous assessment of perfusion variation via
a CBF-sensitive sequence (ASL) in a single experimental session.
Direct current stimulation was applied according to individualised
biophysical models based on manually traced tumour masks of the
necrotic core, solid tumour and T2-hyperintense region, in order to
maximise the E-field over the solid tumour mass. All the patients
completed the study without reporting any adverse effects, and a
decrease in intratumoural perfusion (-36% of CBF respect to base-
line) was observed, while no significant changes were detected in
the surrounding edema, necrotic core nor other control regions of
the brain. Importantly, 3 patients also underwent another single
session of tDCS inside the MRI scanner in the post-surgical setting,
demonstrating the safety of applying tDCS in patients with skull
breaches and skull fixation hardware [17] . Despite being a prelim-
inary investigation, carried out with a single stimulation session
instead of repeated daily applications, this study supports the pos-
sible role of tDCS as a novel therapeutic approach for brain cancers,
especially if combined with standard CHT regimens and applied for
repeated sessions, as currently performed with electrochemother-
apy in extracranial tumours where intratumoural perfusion reduc-
tion seems to prolong drug persistence in the tumour vessels, pos-
sibly boosting their actions [42] . Given the perfusion reduction ob-
served following a single session of tDCS (20 minutes), it is rea-
sonable to assume a greater –and clinically meaningful - decrease
following repeated sessions of tDCS, as observed in extracranial
tumours with electrochemotherapy that can even lead to necro-
sis [45] . However, potential negative effects need to be carefully
taken into consideration before scaling up this type of investiga-
tion in brain tumours patients, such as the theoretical reduction
of cancer sensitivity to radiotherapy treatment (due to the lower
oxygen concentration consequent to the perfusion reduction [46] ),
potential perfusion decrease in the healthy brain tissue surround-
ing the lesion, or the promotion of the recently discovered neuron-
to-glioma communications. For these reasons, tDCS targeting the
tumour perfusion could be tested at recurrence when radiother-
apy usually has been already delivered and therapeutic options are
very limited [47] , possibly in combination with antiepileptic treat-
ments (e.g., Perampanel) aiming to decrease neuronal excitability.
5

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
As for the theoretical reduction of radiotherapy sensitivity, no data
are available to date, but hypoxia could be assessed and monitored
via
18
F-Fluoromisonidazole (
18
F-FMISO) PET scan. F-MISO corre-
lates with VEGF expression, can distinguish between LGG and HGG,
is able to predict poor prognosis and to estimate chronic tissue
hypoxia [48] . GBM is characterised by vascular proliferation and
necrosis, with the latter representing a consequence of the extreme
hypoxia in the tumour core due to high proliferation and lack of
adequate metabolic supply [39] . A possible solution to avoid the
hypoxia induced mechanisms with NiBS intervention, could be rep-
resented by the application of repetitive and prolonged sessions of
stimulations that do not allow the tumour to sufficiently activate
the neoangiogenesis response and rapidly prompt its necrosis, as
observed in extracranial tumours [45] .
With regard to the generalised perfusion decrease, accurate bio-
physical modelling to enable precise targeting of the solid mass
would ensure personalised stimulation to the tumour mass with-
out significantly affecting the surrounding brain, especially with
the adoption of multifocal montages controlling E-field distribu-
tion. Assuring physiological blood flow in peri-tumoural regions
could be fundamental not only to maintain basic physiology and
metabolism via the delivery of adequate amount of oxygen and
nutrients, but also to ensure the delivery of drugs/CHT in the infil-
trated tissues that, in case of an impaired perfusion, could not suf-
ficiently reach that region. The absence of significant CBF variation
in the rest of the brain -when adopting multifocal montages selec-
tively targeting the solid mass– needs to be confirmed in larger pa-
tient samples via non-invasive MRI assessment (e.g., non-contrast
enhanced perfusion acquisition such as ASL, and contrast-enhanced
perfusion scan such as Dynamic Susceptibility Contrast–DSC). tDCS
could even exert a spontaneous dissociable effect on brain tissue
regarding perfusion modulation. In particular, tDCS could increase
CBF in the healthy brain characterised by normal, intact vessels,
and at the same time induce a perfusion decrease in the tumour
tissue region, due to its abnormal vessels’ micro- and macrostruc-
ture (i.e., impaired shunts, lack of smooth muscle cells in the walls,
irregular macrovascular architecture [ 42 , 49 ]). Overall, exploring the
exact boundaries of perfusion reduction obtained with tDCS in
brain tumours is necessary to assure safety and maintain the ef-
ficacy of other concurrent therapy such as CHT. In particular, it is
crucial to accurately define the limits of perfusion reduction in the
case of GBM, whose recurrence often takes place within 2 cm of
the resection margins [50] . Thus, a controlled reduction of perfu-
sion in the immediate vicinity of the macroscopic tumour borders
could be useful to constrain the progression of the disease. In that
case, instead of restricting the targeting to the macroscopic tumour
area, NiBS could be targeted to the surrounding areas following an
approach analogous to radiotherapy, for instance including 2 cm
3
of radiographically healthy tissue [ 17 , 51 ].
Beyond effects on perfusion, tDCS has been shown in rats to
transiently increase the permeability of the BBB to small and large
molecules after only 20’ of stimulation [43] , with the increase
in water and solute flux being confirmed by subsequent experi-
ments in endothelial monolayers [18] . The magnitude of the in-
duced flow was found to linearly correlate with the intensity of
applied current and to be caused by an electroosmotic mechanism.
The changes in permeability were observed at electrical intensities
that are usually reached on the cortex via transcranial stimulation
in humans (0.5 mV) [18] . Therefore, tDCS seems to affect vascular
brain components in –at least— two ways: (i) by varying perfusion
at the macroscale level (i.e. CBF), and (ii) by changing BBB per-
meability at the microscale ( Fig. 3 A). Both mechanisms could be
extremely relevant in the context of brain tumour management,
the former for decreasing intratumoural perfusion and the latter
for enhancing the delivery of drugs through the BBB [52] . Interest-
ingly, in vitro as well as animal models of tDCS have demonstrated
an increase in permeability and concentration of molecules up to
70 kDa [ 18 , 43 ] while the standard chemotherapy agent for GBM
(i.e. Temozolomide) involves a small molecule of only 192 Daltons.
3.3. Microenvironment
Malignant brain cancers, such as GBM, are capable of interact-
ing with the entire tumour microenvironment, ranging from pre-
existing vessels to the immune system, in order to promote tu-
mour growth and resistance to treatments [53] . tES techniques
seem to have the potential to interfere with at least some of the
cellular components of the tumour microenvironment, raising the
possibility to counteract the self-promoting action of cancer cells
and restore more physiological cellular function of the tumour
environment. GBM is characterised by the ability to form a vir-
tual cellular continuum to transfer inorganic and genetic molecules
into surrounding healthy tissue via tumour microtubes, thereby
changing the phenotype of the microenvironment cells towards
one promoting tumour resistance and survival [53] . Microglia, the
innate immune cells of the brain, are involved in the defense
against pathogens and toxins. By disrupting the BBB, GBM al-
lows monocytes and macrophages -recruited by the tumour cy-
tokine and chemokine gradient— to infiltrate the lesion, form-
ing, together with the resident microglia, the so called “tumour-
associated macrophages” or “myeloid cells”–TAMs [54] . TAMs sup-
ports tumour growth by (i) producing matrix metalloproteinases
involved in extracellular matrix degradation, which is essential
to GBM migration and invasion, and (ii) promoting angiogenesis
via the secretion of VEGF-A and other molecules [53] . Therefore,
the suppression of pathological TAM activation could represent a
therapeutic option, with some antibodies targeting their recruiting
pathway currently under investigation [54] . Microglia modulation
via tES has been observed in healthy preclinical models [19] , and
promising evidence is emerging for an effect in Alzheimer’s Dis-
ease (AD) [ 55 , 56 ]. A recent preclinical study in an AD mouse model
showed that exogenously-induced increases in brain γoscillations
(i.e. gamma band) via optogenetic stimulation modulates the activ-
ity of microglia, modifies inflammatory brain processes, and leads
to clearance of β-amyloid and p-tau deposition [ 55 , 56 ]. Now, pilot
studies from our group (NCT03880240, NCT04425148) are investi-
gating translation of these findings in patients with neurodegener-
ative disease such as AD and Frontotemporal Dementia by deliver-
ing tACS capable of entraining neurons at the provided frequency,
such as in the gamma band (i.e. 40 Hz) with the goal of restoring
microglial function [57] . Considering the promising results in AD
models and preliminary evidence in tumor animal models, human
translational research is needed to explore eventual modulation of
microglia and/or of specific neuron type (i.e., parvalbumin-positive
interneurons) via one or more types of tES in brain tumour pa-
tients [58] .
3.4. Galvanotaxis
The discovery that cells can be oriented and guided in their mi-
gration when exposed to electric fields –or galvanotaxis— dates to
the nineteenth century [59] . This phenomenon has been explored
and confirmed in a variety of physiological and pathological con-
texts, such as embryonic development, nerve cell growth, angio-
genesis, wound healing, and cancer cell migration [59] . The mag-
nitude of migration, orientation and reactivity to electrical charges
varies across cell types and a major role for asymmetric ionic
flux and redistribution of charged membrane particles is seen at
the core of this phenomenon [59] . Generally, both healthy and
cancerous cells align and migrate towards the cathode, although
the opposite (e.g. anodal migration) is also possible, as observed
in metastatic lung or breast cancer cells [59] . GBM cells follow
6

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
Fig. 3. Local and network-based NiBS application. (a) Local (tumoural-peritumoural) approaches of different NiBS protocols for brain cancer management can include: (i)
reduction of tumour perfusion to slow down cancer growth via tDCS, (ii) increase of BBB permeability to enhance drugs delivery via tDCS, (iii) migration control of further
spreading leveraging galvanotaxis via tDCS, (iv) stimulation of TAMs tumour suppressive phenotype via tACS, and (v) suppression of neuronal tumour-promoting activity
via inhibitory NiBS protocols. (b) Additionally, network-based approached based on advanced MRI and PET imaging could be developed to optimise interventions aimed at
restoring cognition and/or controlling neuronal tumour-promoting activity, optimising
growth control, and slowing spread of cancer cells. On top of using the aminoacidic
tracer to delineate tumour extent, PET imaging optimize treatment planning and assessment of response to therapy [65] , and could also be tested as predictor of migration
and/or sustainment of tumour growth by looking at the classic
18
F-2-fluoro-2-deoxy-D-glucose (
18
F-FDG)-PET, whose regional standard uptake volume might unveil the brain
regions characterised by the highest glucose uptake [66] , possibly having the strongest influence on neuronal-promoted tumour growth. Perfusion-weighted imaging (PWI)
could refine targeting by identifying recently discovered resting-state perfusion networks, beyond their utility as measures of target engagement by
NiBS [ 67 , 68 ]. Note: red
dots = tumour; orange dots = site of migration and thus of neuronal suppression; blue dots = most impaired brain regions leading to cognitive deficits; BBB = blood
brain barrier; dMRI = diffusion MRI; fALFF = fractional Amplitude of Low-Frequency Fluctuations; fMRI = functional Magnetic Resonance Imaging; PET = Positron Emission
Tomography; rs-fMRI = resting-state fMRI; TAMs = Tumour-Associated Macrophages.
the same migration routes of immature neurons and stem cells,
that is along brain vessels and white matter tracts (e.g. the so-
called Scherer’s structures [60] ). Additionally, GBM cells tend to
reside in perivascular niches to easily extract nutrients from the
bloodstream, particularly the subpopulation of glioma cells capa-
ble of self-renewing (Brain Tumour Initiating cells – BTICs) that ul-
timately seems to be responsible for the constant self-renewal ca-
pacity of GBM [61] . A recent study tested the migration pattern of
BTICs from three different human GBM types, in both 2D and 3D
environments under the influence of an E-field (EF = 0,5–1 V/cm
[62] ). The investigators showed that the migration pattern is af-
fected by the environment, namely that glioma cells migrate to-
wards the anode in a 2D environment and towards the cathode
when posed in a 3D extracellular matrix [62] . Even if the inten-
sity of the electric field typically used to induce the galvanotaxis
phenomenon is notably higher than the one reached in the hu-
man brain with tDCS (around 0,008 V/cm in tDCS compared to
0,5–1 V/cm in galvanotaxis), a similar phenomenon during tDCS
cannot be excluded, especially in the case of repetitive tES ses-
sions to obtain long-lasting changes of excitability/perfusion. Ac-
cordingly, neuronal stem cells were found to increase their migra-
tory activity under the influence of tDCS delivered within typical
human parameters in a rat model, even if a direct migration trend
towards the cathode or electrode was not detected [63] .
In this framework, the cathodal field generated by tDCS in the
region surrounding the tumour aiming to control neuronal ex-
citability could also benefit patients by restricting migration of
cancer cells localised in the cathodal field (i.e., in the immediate
vicinity of the tumour borders) and potentially interfering with mi-
gration and infiltration across the brain. Also, human peripheral
blood mononuclear cells (i.e., lymphocytes and monocytes) have
been found to migrate towards the cathode as well. T cells are typ-
ically downregulated in GBM by the tumour-promoting microglia
phenotype [64] . A potentially beneficial colocalisation of T cells in
the tumour regions could be obtained in the regions targeted by
the cathode, with the aim of promoting the immunological anti-
tumoural response [64] . This could promote the immunotherapy
treatments which currently have not shown positive results, due to
the “cold” aspect of GBM (insensitive to immunotherapy) related to
multiple mechanisms, such as strong immunosuppressive tumour
action, impaired tumour antigen presentation, and the highly hy-
poxic and necrotic environment [64] .
In conclusion, the potentials of NiBS techniques need to be ex-
tensively explored for their local therapeutic applications at mul-
tiple levels and timing in patients with brain cancers ( Fig. 3 A).
These approaches could be investigated in combination with exist-
ing treatments such as CHT and TTF to act on tumour management
at multiple scales. NiBS applications in pilot human studies, in an-
imal and cellular models support the possibility of (i) suppressing
neuronal activity-regulated cancer growth and slowing glioma mi-
tosis/invasiveness, (ii) reducing tumour perfusion and potentially
decreasing cancer growth, (iii) increasing vascular permeability and
promote drug delivery, (iv) restoring microglia function, and (v)
controlling cancer cell migration, with a favorable safety profile
and minimal side effects for patients.
7

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
4. Network-based therapeutic opportunities
Anatomical and physiological studies have demonstrated that
cognitive and motor processes arise from the interaction between
multiple and distant brain regions, forming so-called “large scale
brain networks” [69] . The first type of brain network to be stud-
ied has been the anatomical/structural “connectome”, that defines
the structural connections – white matter fiber tracts- between
brain regions. In GBM patients, the structural connectome iden-
tified via diffusion MRI (dMRI) has gained particular attention
since GBM cells are known to spread along white matter fiber
tracts [70] . Multiple models have been proposed to describe the
growth/migration pathways of GBM cells, however without appli-
cability to clinical practice, beyond the important application to
guiding surgical resection (i.e. sparing eloquent fiber tracts) [71] .
In addition to structural connectomics, “functional networks”
have been identified on the basis of the synchronous interactions
between brain areas, independently of their anatomical connec-
tions [72] . Typically, functional networks are extracted from func-
tional MRI (fMRI), by analyzing correlations in the fluctuations of
the blood oxygen level dependant (BOLD) signal, and thus obtain-
ing the functional connectivity –FC (i.e. the temporal synchroni-
sation between the activity of brain areas [72] ). fMRI can be ac-
quired with subjects performing a specific task or, as emerged in
the last decade, during a resting-state condition (i.e., with a sub-
ject lying in the MRI scanner without performing any task), al-
lowing the identification of several so-called resting-state networks
(RSNs) [73] . Each RSN is composed of regions involved in spe-
cific functions, ranging from the sensory domain (i.e., visual, mo-
tor, and auditory) to high-order cognitive processes (i.e., reason-
ing and attention). Importantly, neurological and psychiatric con-
ditions have been linked to specific functional network alterations
including Schizophrenia, Depression and Alzheimer’s Disease [74] .
Considering that inter- and intra-network activity is responsible for
human cognition and action, modulating the specific interplay be-
tween two or more brain regions could promote (or inhibit) the
function controlled by the targeted brain regions. In this context,
NiBS has been successfully used in healthy subjects, to promote
intra- and inter-network connectivity [75] –and thus enhancing
subject performance during cognitive [76] and motor tasks [77] .
This approach has also been applied in patients, in an attempt
to restore proper network dynamics, as in the case of Alzheimer’s
Disease and Depression [78] .
The possibility of interacting with an entire network rather than
with a single brain area could help in suppressing the neural ac-
tivity regulating cancer growth and tumour spread ( Fig. 3 B). In-
hibition of an entire network could be more effective in slow-
ing neuronal-related cancer growth, with tools derived from Net-
work Control Theory potentially representing a valuable approach
to select the most relevant stimulation targets [79] . Additionally,
network-based approaches mapping network topology and evo-
lution could help predict tumour spread ( Fig. 3 B). As previously
mentioned, dMRI based white matter mapping is essential to visu-
alise the white matter fibers near the tumour and guide the surgi-
cal resection [71] . Therefore, dMRI could also be crucial to identify
the anatomic pathways along which tumour cells can spread, guid-
ing the placement of stimulation electrodes, especially if combined
with functional imaging [ 80 , 81 ] ( Fig. 3 B). Of course, possible cogni-
tive/motor/physiological side effects of network inhibition need to
be carefully monitored and patient wellbeing should be considered
a primary aim in addition to the promotion of survival.
On the other hand, the dynamics of regional brain activity rely
on a complex and balanced interplay between many directly or in-
directly connected areas and networks. Therefore, as in any other
physiological process, a homoeostatic response is frequently ob-
served after perturbation of a limited neuronal population, lead-
ing to the restoration of the initial network state after cessation
of stimulation [82] . However, the cortical excitation/inhibition (E/I)
ratio is typically altered in patients with gliomas, as well as in
other neurological and psychiatric conditions, often associated with
cognitive deficits and symptoms [82] . In patients with gliomas, the
E/I imbalance is involved with the emergence of epileptiform ac-
tivity, and with subsequent neuronal death, paralleled by tumour
progression [83] . Due to NiBS action on membrane channels, it has
been suggested that tDCS can also be useful in restoring the E/I
balance of brain network(s) [84] . In particular, daily application of
tDCS with a large cathodal field to inhibit tumour-promoting neu-
ronal activity could also exert beneficial effects on cortical E/I bal-
ance, hopefully leading to an interruption or even prevention of
epileptogenesis [83] , a common and debilitating symptom of the
disease that may also induce a positive feedback on neurons to
promote tumour growth and invasiveness, as previously discussed
[ 7 , 8 ]. Related encouraging results are emerging from the applica-
tion of tDCS in drug-resistant childhood epilepsy, with data show-
ing a decrease of at least 42% in seizure frequency after 10 sessions
of tDCS (1h/day) and pilot data showing a similar trend for epilep-
tic patients with previous surgical interventions [85] .
Finally, TMS could provide markers of tumour progression
and/or diagnosis by assessing the neuronal response to pertur-
bation of the surrounding apparently healthy regions, i.e., assess
cortical excitability and excitation/inhibition balance in the peritu-
moural areas ( Fig. 1 ). To this end, TMS could be combined with si-
multaneous electroencephalography (EEG) for measuring local and
distant brain responses to stimulation, given that a focal TMS pulse
typically evokes activation in secondary interconnected cortical ar-
eas (e.g. TMS-evoked potentials). In this context, TMS-EEG has
emerged as a method to study not only local cortical reactivity,
but also the causal communication between distant brain regions
at high temporal resolution lacking in the context of fMRI stud-
ies. TMS-EEG could provide insights into the mechanisms of effec-
tive connectivity (i.e. the influence of one neural system over an-
other), and into higher-order cognitive processes at the individual
level [86] . If integrated with imaging data and prediction theories,
TMS-EEG could provide temporally and spatially-optimised mark-
ers of neuron-to-glioma communications and eventually of glioma
spreading trajectory, promoting the development of strategies to
control this new pathophysiological mechanism.
5. Network-based approaches for symptom management
Resting-state connectivity has also recently been used to map
neurological and psychiatric symptoms, providing a novel frame-
work to understand the basis of altered behaviour, especially in
relation to classic anatomical lesion studies [87] . In this approach,
the connectivity of the lesions located in different parts of the
brain causing the same symptomatology is mapped to find out
the common region/network whose connectivity is altered. This
approach, called “Connectome lesion-based mapping”, has helped
disentangle the pathophysiological correlates of many neurological
and psychiatric conditions, such as amnesia, Parkinsonism, cervical
dystonia and delusional misidentifications [87] .
The connectome lesion-based mapping approach could be also
applied in patients with brain tumours to possibly identify core
regions whose functionality is altered, causing the core symptoma-
tology presents in patients. As recently suggested, symptoms of pa-
tients with gliomas are an integral part of the disease, but their
characterisation has been largely ignored in favor of molecular
classifications, consequently limiting the therapeutic opportunities
for symptoms relief [88] . Developing a symptom-sensitive thera-
peutic strategy could be valuable to improve patients’ quality of
life in addition to prolonging survival (considering that cognitive
impairment has been recognised as a significant prognostic fac-
8

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
tor for patient survival [89] ), and for those who cannot receive or
refuse aggressive oncologic treatments. NiBS could offer a safe tool
to help patients mitigate cognitive impairment present in approxi-
mately 80% of adult glioma patients as a consequence of radiother-
apy, CHT, surgery, and the tumour itself [90] . Gliomas can directly
alter brain connectivity and cause cognitive deficits via: (i) mass
effect, im pairing the activity of neighboring brain regions via com-
pression, (ii) infiltration and spreading along white matter tracts,
(iii) reduction of perfusion in nearby areas due to increased tu-
mour vascularisation [90] . In general, gliomas – especially GBM-
seem to induce local and widespread, network functional changes
as visualised by FC fMRI analysis [91] . Decreased FC has been ob-
served both in close proximity to the tumour as well as in distant
regions, including the contralateral hemisphere [91] . A longitudi-
nal study in three GBM patients revealed that lesions located in
close proximity to a functional hub (i.e. a region among those with
highest density of functional connections with other brain regions)
strongly affect long-range connectivity, while the same effect has
not been seen in the case of lesions involving a peripheral node
of the same network [92] . Surgery and radiotherapy interventions
were found to restore physiological connectivity to a certain ex-
tent in the affected regions, with the effect potentially due to the
release from compression as well as to reduction of edema and in-
flammation. However, after the initial treatment period, FC usually
decreases again, possibly indicating the progression of the disease
and further spreading of the tumour, along with the deleterious
effects of treatments themselves [92] .
In a group of patients with HGG and LGG followed for 1 year,
memory and attention deficits were reported in patients with le-
sions overlapping with the Default Mode and Attention Network,
respectively [93] . Additionally, neurite density (a marker of axon
and dendrite concentration derived from dMRI) was also found to
correlate with memory and attention recovery after surgery [93] .
As for patients with metastatic disease, there is a specific cognitive
decline starting 4–6 months after whole brain radiation therapy
and continuing indefinitely, mainly impacting memory and learn-
ing performance [94] . It can be speculated whether NiBS could
counteract these functional declines especially at the beginning of
treatment or in those who cannot undergo RT/surgery. Indeed, en-
hancement of behavioural performance has been demonstrated in
many domains by NiBS [24] , ranging from motor to cognitive, and
also including complex processes such as decision making and so-
cial behaviour, in patients as well as healthy subjects [23] . Recently
promoted network-based targeting approaches for tES can refine
the effects obtained with classical neuromodulation applications,
which have been shown to improve performance in healthy sub-
jects and patients in domains such as abstract reasoning, mem-
ory, attention and motor function [95] ( Fig. 3 B). Importantly, ex-
citatory protocols (i.e. tDCS, tRNS, high-frequency rTMS) are usu-
ally applied to enhance cognitive functions. Therefore, the stim-
ulation site(s) should be carefully selected to avoid those con-
nected with the glioma cell population to avoid a potential pro-
motion of cancer growth as a collateral effect. Accurate selection
of patients on the basis of the localisation of the lesion respect
to the target brain area (i.e., lesion not in the immediate proxim-
ity of the target region) would be of extreme importance along
with the adoption of personalised stimulation protocols. Indeed,
current flow across the brain can be accurately modelled and pre-
dicted when considering all the cranial compartments, the tumour
mass and/or surgical cavity as well as metallic clips to obtain a
personalised image-guided stimulation ( [96] , Fig. 4 A). Neurostimu-
lation techniques have gained more reliability and efficacy thanks
to their integration with imaging techniques (anatomical MRI, CT,
fMRI, PET) to personalise the stimulation solution based on each
patient’s anatomy and physiology, rather than using standard ap-
proaches based on general templates, thus allowing to reach un-
precedented fine and predictable electric field ( Fig. 4 B).
Finally, NiBS has the potential to induce cortical plasticity and
reorganisation into the healthy brain potentially enabling greater
surgical resection in eloquent cortex affected by the tumour, as
demonstrated in a few seminal studies [ 97 , 98 ]. The extent of resec-
tion is a strong predictor of prognosis and the proximity or direct
presence of the tumour in eloquent cortices (e.g., motor cortex,
language areas) profoundly limits the tumour resection, even when
the most advanced techniques are used intraoperatively. LGGs can
induce plasticity and reorganisation in cortical areas distant to the
lesion enabling the maintenance of a function, due to their slow
growth, but may also harbor critical structures within the glioma
itself [99] . Suppression of the eloquent cortex near/infiltrated by
the glioma (WHO grade II and III) paired to an intensive training of
the targeted function (median training: 16 days) allowed a greater
resection associated with cortical reorganisation with new, distant
brain activation and no significant performance deficits detected at
follow-up assessments [98] . NiBS could be implemented to induce
such cortical plasticity in LGG and HGG in order to maximise the
resection and prolong the survival, while also limiting functional
deficits.
Given the extensive and successful research made in the last
decade on different clinical and non-clinical populations via NiBS,
network approaches integrating multiple stimulation modalities
could be leveraged to tackle cognitive symptomatology and func-
tional deficits, which negatively impacts patients’ quality of life.
6. Conclusion
Cellular, animal, and human models suggest realistic NiBS
roles in targeting a range of glioma pathophysiological substrates,
ranging from neuron-to-glioma synaptic communication, tumour
microenvironment, perfusion to galvanotaxis. Combined with its
safety, noninvasiveness and tolerability, NiBS may offer novel ther-
apeutic approaches to control cancer growth and ameliorate pa-
tients’ disability. In particular, as demonstrated in multiple trials
[100] , tES is also suitable for home-based intervention, a signifi-
cant benefit in the present neurooncological landscape.
In this promising scenario, there are some challenges that need
to be addressed in order to optimise the application of NiBS in
brain tumour patients. For instance, neuromodulation effects could
be reduced by the antiepileptic therapy usually received by pa-
tients for seizure prevention. Indeed, antiepileptic drugs act on var-
ious ion channels, the same target of NiBS [101] . However, while
the blockage of voltage dependant Na + and Ca
2 + channels respec-
tively eliminates or decreases the hyperexcitability caused by an-
odal stimulation, the same effect is not observed during cathodal
stimulation [101] . In addition, the most frequently administered
antiepileptic drug in glioma patients is levetiracetam which does
not inhibit voltage-dependant Na + channels or GABAergic trans-
mission, but binds a synaptic vesicle glycoprotein (SV2A) and in-
hibits presynaptic Ca
2 + channels [102] . Therefore, the modulatory
action of levetiracetam could be of no -or minimal- significance
in brain tumour patients undergoing neuromodulation protocols,
given the limited reduction of NiBS effects observed with Ca
2 +
channels blockers. Moreover, the protocols aimed at inhibiting the
neuron-to-glioma communications would see the application of
cathodal stimulation, reducing the implications of the assumption
of antiepileptic drug.
On the other hand, in the case of application of excitatory pro-
tocol, for instance to restore cognitive and/or motor ability, the ad-
ministration of such antiepileptic drugs could partially limit the
expected excitatory effects induced by NiBS, but, at the same time,
control for potential collateral effects (i.e., seizure) and tumour-
promoting neuronal activity. This scenario is probably the most dif-
9

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
Fig. 4. Personalised image-guided biophysical modeling. (a) After MRI/CT acquisition, images can be segmented into different tissues (skin, air, bone, cerebrospinal fluid-
CSF, gray matter - GM, white matter - WM), along with the tumour masks. All the tissues (manually or semi-automatically created), as well as metallic device and skull
holes/breaches, are imported into the 3D volume rendering of the single subject brain anatomy and conductivity values are attributed to each tissue [17] . (b) Examples of the
best multielectrode montage maximizing the normal E-field (En) on the tumour target selected among all the 10/ 20 EEG combinations available for electrode’s positioning
in a presurgical (left) and a post-operative (right) patient.
ficult to optimise, and must include a multidisciplinary effort in-
volving, for instance, expertise related to biophysical modelling in
order to maximize targeting, as well as knowledge from the anato-
mopathologist to estimate the percentage of glial cell in the surgi-
cal/biopsy tissue samples actually forming synapses with the sur-
roundings neuronal tissue, and therefore estimate risk of possible
tumour proliferation induced via NiBS along with the differential
contribution to proliferiation by the different neurons.
Multiple complementary possibilities in terms of targeting ap-
proaches, stimulation parameters, and timing of the intervention
should be tested in animal models, whereas some interventions
might be ready for first-in-human trials. A collaborative effort be-
tween neurooncologists, neurosurgeons, neurologists and neuro-
physiologists, neuroradiologists, physicists, engineers, neurobiolo-
gists and neuroscientists is needed to rapidly evaluate the potential
of NiBS in brain tumour patients.
Outstanding questions
In the light of recent findings of electrical glioma integration,
NiBS can be regarded as a potential tool to suppress neuron-to-
glioma communication, in addition to other potential effects on
tumour features and microenvironment. Scientists and physicians
should cooperate to operationalise the above formulated priorities
in order to potentially provide a new therapeutic and safe option
to glioma patients:
(1) To what extent can NiBS affect the newly discovered neuron-
to-glioma communication and the related tumour progression?
(2) Can the NiBS-induced tumour perfusion reduction and its ef-
fects on BBB permeability lead to tumour shrinkage if applied
in a daily protocol basis?
(3) How NiBS affects the glioma microenvironment?
(4) Can NiBS be applied to improve the quality of life of glioma
patients, especially to control cognitive and motor deficits?
(5) Can NiBS in combination with neuroimaging and electrophysi-
ological data improve the understanding of glioma pathophysi-
ology?
Search strategy and selection criteria
Data for this review were identified by searches of PubMed
and references from relevant English articles using the search
terms “glioma”, “brain cancer”, “fMRI”, “NIBS”, “brain stimulation”,
“tDCS”, “tACS”, “tRNS” without any type of restriction. Selection of
the most appropriate references was made by the authors due to
the journal constraints.
Declaration of Competing Interest
APL is listed as an inventor on several issued and pending
patents on the real-time integration of noninvasive brain stimu-
lation with electroencephalography and magnetic resonance imag-
ing and he is a member of the scientific advisory board of Starlab,
Magstim and MedRhythm; APL is co-inventor of Linus Health and
TI Solutions. The other authors declare no conflicts of interest.
Acknowledgements
E.S. is supported by the NIH ( R01 MH117063-01 , R01
AG060981-01 ) and by the Alzheimer’s Drug Discovery Founda-
tion (ADDF) & Association for Frontotemporal Dementia (AFTD) via
GA 201902-2017902 . A.J.G. is supported by NIH P41EB015898 , the
Haley Distinguished Chair in the Neurosciences, the Jennifer Op-
penheimer Cancer Research Initiative, the NIH via R21CA198740 .
A.P.L. is supported by NIH via R24AG06142 , R01AG059089 and P01
AG031720 . G.E-L. is supported by NIH via P41EB022544 .
References
[1] Gould J. Breaking down the epidemiology of brain cancer. Nature
2018;561:S40–1. doi: 10.1038/d41586- 018- 06704- 7 .
10

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
[2] Nayak L, Reardon DA. High-grade Gliomas. Contin Minneap Minn
2017;23:1548–63. doi: 10 .1212/CON .0 0 0 0 0 0 0 0 0 0 0 0 0554 .
[3] Bunevicius A, McDannold NJ, Golby AJ. Focused ultrasound strategies for
brain tumor therapy. Oper Neurosurg 2019 Hagerstown Md. doi: 10.1093/ons/
opz374 .
[4] Fabia n D, Eibl MD PGP, Alnahhas I, Sebastian N, Giglio P, Puduvalli V, et al.
Treatment of glioblastoma (GBM) with the Addition of tumor-treating fields
(TTF): a review. Cancers 2019;11. doi: 10.3390/cancers11020174 .
[5] Stupp R, Taillibert S, Kanner A, Read W, Steinberg D, Lhermitte B, et al. Ef-
fect of
tumor-treating fields plus maintenance temozolomide vs maintenance
temozolomide alone on survival in patients with glioblastoma: a randomised
clinical trial. JAMA 2017;318:2306–16. doi: 10.1001/jama.2017.18718 .
[6] Stupp R, Wong ET, Kanner AA, Steinberg D, Engelhard H, Heidecke V, et al.
NovoTTF-10 0A versus physician’s choice chemotherapy in recurrent glioblas-
toma: a rand omised phase III trial of a novel treatment modality. Eur J Cancer
Oxf Engl 2012;48:2192–202 1990. doi: 10.1016/j.ejca.2012.04.011 .
[7] Venkatesh HS, Morishita W, Geraghty AC, Silverbush D, Gillespie SM, Arzt M,
et al. Electrical and synaptic integration of glioma into neural circuits. Nature
2019;573:539–45. doi: 10.1038/s41586- 019- 1563- y .
[8] Venkataramani V, Tane v DI, Strah le C, Studier-Fischer A, Fankhauser L,
Kessler T, et al. Glutamatergic synaptic input to glioma cells drives brain tu-
mour progression. Natu re 2019;573:532–8. doi: 10.1038/s41586- 019- 156 4- x .
[9] Monje M, Borniger JC, D’Silva NJ, Deneen B, Dirks PB, Fattahi F, et al. Roadmap
for the emerging field of cancer neuroscience. Cell 2020;181:219–22. doi: 10.
1016/j.cell.2020.03.034 .
[10] Buckingham SC, Campbell SL, Haas BR, Montana V, Robel S, Ogunrinu T, et al.
Glutamate release by primary brain tumors induces epileptic activity. Nat
Med 2011;17:1269–74. doi: 10.1038/nm.2453 .
[11] Jung E, Alfonso J, Osswald M, Monyer H, Wick W, Winkler F. Emerg-
ing intersections between neuroscience and glioma biology. Nat Neurosci
2019;22:1951–60. doi: 10.1038/s41593- 019- 0540- y .
[12] Liu SJ, Zukin RS. Ca
2 +
-permeable AMPA receptors in synaptic plasticity and
neuronal death. Trends Neurosc i 2007;30:126–34. doi: 10.1016/j.tins.2007.01.
006 .
[13] Rossi S, Antal A, Bestmann S, Bikson M, Brewer C, Brockmöller J, et al. Safety
and recommendations for TMS use in healthy subjects and patient popula-
tions, with updates on training, ethical and
regulatory issues: expert guide-
lines. Clin Neurophysiol 2020. doi: 10.1016/j.clinph.2020.10.0 03 .
[14] Antal A, Alekseichuk I, Bikson M, Brockmöller J, Brunoni AR, Chen R, et al.
Low intensity transcranial electric stimulation: safety, ethical, legal regulatory
and application guidelines. Clin Neurophysiol 2017;128:1774–809 Off J Int Fed
Clin Neurophysiol. doi: 10.1016/j.clinph.2017.06.001 .
[15] Klomjai W, Katz R, Lackmy-Vallée A. Basic principles of transcranial mag-
netic stimulation (TMS) and repetitive TMS (rTMS). Ann Phys Rehabil Med
2015;58:208–13. doi: 10.1016/j.rehab.2015.05.005 .
[16] Cirillo G, Di Pino G, Capone F, Ranieri F, Florio L, Todisco V, et al. Neurobiolog-
ical after-effects of non-invasive brain stimulation. Brain Stimulat
2017;10:1–
18. doi: 10.1016/j.brs.2016.11.009 .
[17] Sprugnoli G, Monti L, Lippa L, Neri F, Mencarelli L, Ruffini G, et al. Reduction
of intratumoral brain perfusion by noninvasive transcranial electrical stimu-
lation. Sci Ad v 2019;5:eaau9309. doi: 10.1126/sciadv.aau9309 .
[18] Cancel LM, Arias K, Bikson M, Tarbe ll JM. Direct current stimulation of en-
dothelial monolayers induces a transient and reversible increase in trans-
port due to the electroosmotic effect. Sci Rep 2018;8:9265. doi: 10.1038/
s41598- 018- 27524- 9 .
[19] Gellner AK, Reis J, Fritsch B. Glia: a neglected player in non-invasive direct
current brain stimulation. Front Cell Neurosci 2016;10:188. doi: 10.3389/fncel.
2016.00188 .
[20] Collingridge GL, Peineau S, Howland JG, Wang YT. Long-term depression in
the CNS. Nat Rev Neurosci 2010;11:459–73. doi: 10.1038/nrn2867 .
[21] Huerta PT, Volpe BT. Transcranial magnetic stimulation, synaptic plastic-
ity and network oscillations. J Neuroengin Rehabil 2009;6:7. doi: 10.1186/
1743- 0 0 03-6-7 .
[22] Bolognini N, Pascual-Leone A,
Fregni F. Using non-invasive brain stimulation
to augment motor training-induced plasticity. J Neuroengin Rehabil 2009;6:8.
doi: 10.1186/1743- 0 0 03-6-8 .
[23] Reed T, Cohen Kadosh R. Transcranial electrical stimulation (tES) mechanisms
and its effects on cortical excitabil ity and connectivity. J Inherit Metab Dis
2018. doi: 10.1007/s10545- 018- 0181- 4 .
[24] Polanía R, Nitsche MA, Ruff CC. Stu dying and modifying brain function with
non-invasive brain stimulation. Nat Neurosci 2018;21:174–87. doi: 10.1038/
s41593- 017- 0054- 4 .
[25] Huang YZ, Lu MK, Antal A, Classen J, Nitsche M, Ziemann U, et al. Plasticity
induced by non-invasive transcranial brain stimulation: a position paper. Clin
Neurophysiol 2017;128:2318–29 Off J Int Fed Clin Neurophysiol. doi: 10 .1016 /
j.clinph.2017.09.007 .
[26] Mrugala MM, Ruzevick J, Zlomanczuk P, Lukas RV. Tumor treating fields
in neuro-oncological practice. Curr Oncol Rep 2017;19:53. doi: 10. 100 7/
s11912- 017- 0611- 8 .
[27] Rossi S, Hallett M, Rossini PM, Pascual-Leone A. Safety of TMS consensus
group. safety, ethical considerations, and application guidelines for the use
of transcranial magnetic stimulation in clinical practice and research. Clin
Neurophysiol 20 09;120:20 08–39 Off J Int Fed Clin Neurophysiol. doi: 10. 1016/
j.clinph.2009.08.016 .
[28] Santarnecchi E, Polizzotto NR, Godone M, Giovannelli F, Feurra M, Matzen L,
et al. Frequency-dependant enhancement of fluid intelligence induced by
transcranial oscillatory potentials. Curr Biol 2013;23:1449–53 CB. doi: 10. 1016/
j.cub.2013.06.022 .
[29] Santarnecchi E, Sprugnoli G, Bricolo E, Costantini G, Liew SL, Musaeus CS,
et al. Gamma tACS over the temporal lobe increases the occurrence of Eu-
reka! moments. Sci Rep 2019;9:5778. doi: 10 .10 38 /s 4159 8- 019- 42192- z .
[30] Venkatesh HS, Johung TB, Caretti V, Noll A, Tang Y, Nagaraja S, et al. Neu-
ronal activity promotes glioma growth through neuroligin-3 secretion. Cell
2015;161:803–16. doi: 10.1016/j.cell.2015.04.012 .
[31] Venkatesh HS, Tam LT, Woo PJ, Lennon J, Nagaraja S, Gillespie SM, et al.
Targ eti ng neuronal activity-regulated neuroligin-3 dependency in high-grade
glioma. Nature 2017;549:533–7. doi: 10.1038/nature24014 .
[32] Köhling R, Senner V, Paulus W, Speckmann EJ. Epileptiform activity preferen-
tially arises outside tumor invasion zone in glioma xenotransplants. Neurobiol
Dis 2006;22:64–75. doi: 10.1016/j.nbd.20 05.10.001 .
[33] Venkataramani V, Tanev DI, Kuner T, Wick W, Winkler F. Synaptic input to
brain tumors: clinical implications. Neuro Oncol 2020. doi: 10.1093/neuonc/
noaa158 .
[34] Notzon S, Steinberg C, Zwanzger P, Junghöfer M. Modulating emotion per-
ception: opposing effects of inhibitory and excitatory prefrontal cortex stim-
ulation. Biol Psychiatry Cogn Neurosci Neuroimaging 2018;3:329–36. doi: 10 .
1016/j.bpsc.2017.12.007 .
[35] Singh A, Erwin-Grabner T, Goya-Maldonado R, Antal A. Transcranial mag-
netic and direct current stimulation in the treatment of depression: basic
mechanisms and challenges of two commonly used brain stimulation meth-
ods in interventional psychiatry. Neuropsychobiology 2019:1–11. doi: 10. 1159/
0 0 0502149 .
[36] Han T, Xu Z, Liu C, Li S, Song P, Huang Q, et al. Simultaneously applying catho-
dal tDCS with low frequency rTMS at the motor cortex boosts inhibitory af-
tereffects. J Neurosci Methods 2019;324:108308. doi: 10.1016/j.jneumeth.2019.
05.017 .
[37] Zheng X, Alsop DC, Schlaug G. Effects of transcranial direct current stimula-
tion (tDCS) on human regional cerebral blood flow. NeuroImage 2011;58:26–
33. doi: 10.1016/j.neuroimage.2011.06.018 .
[38] Hu LS, Eschbacher JM, Dueck AC, Heiserman JE, Liu S, Karis JP, et al. Cor-
relations between perfusion MR imaging cerebral blood volume, microvessel
quantification, and clinical outcome using stereotactic analysis in recurrent
high-grade glioma. AJNR Am J Neuroradiol 2012;33:69–76. doi: 10 .3174 /aj nr.
A2743 .
[39] Louis DN, Perry A, Reifenberger G, Von
Deimling A, Figarella-Branger D, Cave-
nee WK, et al. The 2016 world health organization classification of tumors of
the central nervous system: a summary. Acta Neuropathol 2016;131:803–20
(Berl). doi: 10 .10 07/s0 0401-016-1545-1 .
[40] Qiao XJ, Ellingson BM, Kim HJ, Wan g DJJ, Salamon N, Linetsky M, et al. Arte-
rial spin-labeling perfusion MRI stratifies progression-free survival and corre-
lates with epidermal growth factor receptor status in glioblastoma. AJNR Am
J Neuroradiol 2015;36:672–7. doi: 10.3174/ajnr.A4196 .
[41] Chen Z, Xu N, Zhao C, Xue T, Wu X, Wa ng Z. Bevacizumab combined
with chemotherapy vs single-agent therapy in recurre nt glioblastoma: evi-
dence from
randomize d controlled trials. Cancer Manag Res 2018;10:2193–
205. doi: 10.2147/CMAR.S173323 .
[42] Ciria HMC, González MM, Zamora LO, Cabrales LEB, Sierra González GV, De
Oliveira LO, et al. Antitumor effects of electrochemical treatment. Chin J Can-
cer Res 2013;25:223–34 Chung-Kuo Yen Cheng Yen Chiu. doi: 10.3978/j.issn.
10 0 0-9604.2013.03.03 .
[43] D.W. Shin, N. Khadka, J. Fan, M. Bikson, B.M. Fu Transcranial direct current
stimulation transiently increases the blood-brain barrier solute permeabil-
ity in vivo . Proc. SPIE 9788, Medical Imaging 2016: Biomedical Applications
in Molecular, Structural, and Functional Imaging, 97881X (29 March 2016).
https://doi.org/10.1117/12.2218197 .
[44] Esmaeili N, Friebe M. Electrochemotherapy: a review of current sta-
tus, alternative igp approaches, and future perspectives. J Healthc Eng
2019;2019:2784516. doi: 10.1155/2019/2784516 .
[45] Jarm T , Cemazar M , Steinberg F , Stre ffer C , Sersa G , Miklavcic D . Perturba-
tion of blood flow as a mechanism of anti-tumour action of
direct current
electrotherapy. Physiol Meas 2003;24:75–90 .
[46] Kelley K, Knisely J, Symons M, Ruggieri R. Radioresistance of brain tumors.
Cancers 2016;8. doi: 10.3390/cancers8040042 .
[47] Weller M, Cloughesy T, Perry JR, Wick W. Standards of care for treatment of
recurrent glioblastoma–are we there yet? Neuro Oncol 2013;15:4–27. doi: 10.
1093/neuonc/nos273 .
[48] Muzi M, Peterson LM, O’Sullivan JN, Fink JR, Rajendran JG, McLaughlin LJ,
et al. 18F-fluoromisonidazole quantification of hypoxia in human cancer pa-
tients using image-derived blood surrogate tissue reference regions. J Nucl
Med 2015;56:1223–8 Off Publ Soc Nucl Med. doi: 10.2967/jnumed.115.158717 .
[49] Jarm T, Cemazar M, Miklavcic D, Sersa G. Antivascular effects of elec-
trochemotherapy: implications in treatment of bleeding metastases. Expert
Rev Anticancer Ther 2010;10:729–46. doi: 10.1586/era.10.43 .
[50] De Bonis P, Anile C, Pompucci A, Fiorentino A, Balducci M, Chiesa S, et al.
The influence of surgery on recurrence pattern of glioblastoma. Clin Neurol
Neurosurg 2013;115:37–43. doi: 10.1016/j.clineuro.2012.04.005 .
[51] Burnet NG, Thomas SJ, Burton KE, Jefferies SJ. Defining the tumour and target
volumes for radiotherapy. Cancer Imaging 2004;4:153–61 Off Publ Int Cancer
Imaging Soc. doi: 10 .110 2/14 70- 7 330 .2 0 04.0 054 .
[52] Bender E. Getting cancer drugs into the brain. Nature 2018;561:S46–7. doi: 10.
1038/d41586- 018- 06707- 4 .
[53] Broekman ML, Maas SLN, Abels ER, Mempel TR, Krichevsky AM, Breake-
11

G. Sprugnoli, S. Rossi, A. Rotenberg et al. EBioMedicine 70 (2021) 103514
field XO. Multidimensional communication in the microenvirons of glioblas-
toma. Nat Rev Neurol 2018;14:482–95. doi: 10.1038/s41582- 018- 0025- 8 .
[54] Poon CC, Sarkar S, Yong VW, Kelly JJP. Glioblastoma-associated microglia
and macrophages: targets for therapies to improve prognosis. Brain J Neurol
2017;140:1548–60. doi: 10.1093/brain/aww355 .
[55] Iaccarino HF, Singer AC, Martorell AJ, Rudenko A, Gao F, Gillingham TZ, et al.
Gamma frequency entrainment attenuates amyloid load and modifies mi-
croglia. Nature 2016;540:230–5. doi: 10.1038/nature20587 .
[56] Adaikkan C, Middleton SJ, Marco A, Pao PC, Mathys H, Kim DNW, et al.
Gamma entrainment binds higher-order brain regions and offers neuropro-
tection. Neuron 2019. doi: 10.1016/j.neuron.2019.04.011 .
[57] Thomson H. How flashing lights and pink noise might banish
Alzheimer’s, improve memory and more. Nature 2018;555:20–2.
doi: 10.1038/d41586- 018- 02391- 6 .
[58] Tantillo E, Vannini E, Cerri C, Spalletti C, Colistra A, Mazzanti CM, et al. Differ-
ential roles of pyramidal and fast-spiking, GABAergic neurons in the control
of glioma cell proliferation. Neurobiol Dis 2020;141:104942. doi: 10 .1016 /j .n bd .
2020.104942 .
[59] Cortese B, Palamà IE, D’Amone S, Gigli G. Influence of electrotaxis on cell
behaviour. Integr Biol 2014;6:817–30 Quant Biosci Nano Macro. doi: 10.1039/
c4ib00142g .
[60] Cuddapah VA , Robel S , Watkins S
, Sontheimer H . A neurocentric perspective
on glioma invasion. Nat Rev Neurosci 2014;15:455–65 10.1038/nrn3765 .
[61] Calabrese C, Poppleton H, Kocak M, Hogg TL, Fuller C, Hamner B, et al. A
perivascular niche for brain tumor stem cells. Cancer Cell 2007;11:69–82.
doi: 10.1016/j.ccr.2006.11.020 .
[62] Huang YJ, Hoffmann G, Wheeler B, Schiapparelli P, Quinones-Hinojosa A,
Searson P. Cellular microenvironment modulates the galvanotaxis of brain tu-
mor initiating cells. Sci Rep 2016;6:21583. doi: 10.1038/srep21583 .
[63] Keuter s MH, Aswendt M, Tennstaedt A, Wiedermann D, Pikhovych A, Rot-
thues S, et al. Transcranial direct current stimulation promotes the mobility
of engrafted NSCs
in the rat brain. NMR Biomed 2015;28:231–9. doi: 10. 100 2/
nbm.3244 .
[64] Lim M, Xia Y, Bettegowda C, Weller M. Current state of immunother-
apy for glioblastoma. Nat Rev Clin Oncol 2018;15:422–42. doi: 10.1038/
s41571- 018- 0 0 03- 5 .
[65] Wern er JM, Lohmann P, Fink GR, Langen KJ, Galldiks N. Current landscape
and emerging fields of PET imaging in patients with brain tumors. Molecules
2020;25 Basel Switz. doi: 10.3390/molecules25061471 .
[66] Kim MM, Parolia A, Dunphy MP, Venneti S. Non-invasive metabolic imag-
ing of brain tumours in the era of precision medicine. Nat Rev Clin Oncol
2016;13:725–39. doi: 10.1038/nrclinonc.2016.108 .
[67] Patel P, Baradaran H, Delgado D, Askin G, Christos P, John Tsiouris A, et al.
MR perfusion-weighted imaging in the evaluation of high-grade gliomas after
treatment: a systematic review and meta-analysis. Neuro Oncol 2017;19:118–
27. doi: 10.1093/neuonc/now148 .
[68] Li Z, Vidorreta M, Katchmar N, Alsop DC, Wolf DH, Detre
JA. Effects of rest-
ing state condition on reliability, trait specificity, and network connectivity of
brain function measured with arterial spin labeled perfusion MRI. NeuroIm-
age 2018;173:165–75. doi: 10.1016/j.neuroimage.2018.02.028 .
[69] Bassett DS, Sporns O. Network neuroscience. Nat Neurosci 2017;20:353–64.
doi: 10.1038/nn.4502 .
[70] Colombo MC, Giverso C, Faggiano E, Boffano C, Acer bi F, Ciarletta P. Towa rd s
the personalised treatment of glioblastoma: integrating patient-specific clin-
ical data in a continuous mechanical model. PloS One 2015;10:e0132887.
doi: 10.1371/journal.pone.0132887 .
[71] Golby AJ, Kindlmann G, Norton I, Yarmarkovich A, Pieper S, Kikinis R. Inter-
active diffusion tens or tractography visualisation for neurosurgical planning.
Neurosurgery
2011;6 8:4 96–505. doi: 10.1227/NEU.0b013e3182061ebb .
[72] Biswal B , Ye tk in FZ , Haughton VM , Hyde JS . Functional connectivity in the
motor cortex of resting human brain using echo-planar MRI. Magn Reson
Med 1995;34:537–41 .
[73] Zhang D, Raichle ME. Disease and the brain’s dark energy. Nat Rev Neurol
2010;6:15–28. doi: 10.1038/nrneurol.2009.198 .
[74] Van den Heuvel MP, Scholtens LH, Kahn RS. Multiscale neuroscience of psy-
chiatric disorders. Biol Psychiatry 2019. doi: 10.1016/j.biopsych.2019.05.015 .
[75] Santarnecchi E, Momi D, Sprugnoli G, Neri F, Pascual-Leone A, Rossi A, et al.
Modulation of network-to-network connectivity via spike-timing-dependant
noninvasive brain stimulation. Hum Brain Mapp 2018;39:4870–83. doi: 10 .
1002/hbm.24329 .
[76] Momi D, Neri F, Coiro G, Smeralda C, Veniero D, G S, et al. Cognitive en-
hancement via network-targeted cortico-cortical associative brain stimulation.
Cereb Cortex 2019;1991 N Y N. doi: 10.1093/cercor/bhz182 .
[77] Fiori F, Chiappini E, Avenanti A. Enhanced action performance following TMS
manipulation of associative plasticity in ventral premotor-motor pathway.
NeuroImage 2018;183:847–58. doi: 10.1016/j.neuroimage.2018.09.002 .
[78] Fox MD, Buckner RL, Liu H, Chakravarty MM, Lozano AM, Pascual-Leone A.
Resting-state networks link invasive and noninvasive brain stimulation across
diverse psychiatric and neurological diseases. Proc Natl Acad Sci U S A
2014;111:E4367–75. doi:
10.1073/pnas.1405003111 .
[79] Liu YY, Slotine JJ, Barabási AL. Controllability of complex networks. Nature
2011;473:167–73. doi: 10.1038/nature10011 .
[80] O’Donnell LJ, Rigolo L, Norton I, Wells WM, Westi n CF, Golby AJ. fMRI-DTI
modeling via landmark distance atlases for prediction and detection of fiber
tracts. NeuroImage 2012;60:456–70. doi: 10.1016/j.neuroimage.2011.11.014 .
[81] Propper RE, O’Donnell LJ, Whalen S, Tie Y, Norton IH, Suarez RO, et al. A com-
bined fMRI and DTI examination of functional language lateralisation and ar-
cuate fasciculus structure: Effects of degree versus direction of hand prefer-
ence. Brain Cogn 2010;73:85–92. doi: 10.1016/j.bandc.2010.03.004 .
[82] Tatti R, Haley MS, Swanson OK, Ts elh a T, Maffei A. Neurophysiology and regu-
lation of the balance between excitati on and inhibition in neocortical circuits.
Biol Psychiatry 2017;81:821–31. doi: 10.1016/j.biopsych.2016.09.017 .
[83] Pallud J, Le Van Quyen M, Bielle F, Pellegrino C, Varlet P, Cresto N, et al. Cor-
tical GABAergic excitation contributes to epileptic activities around human
glioma. Sci Transl Med 2014;6 244ra89. doi: 10.1126/scitranslmed.3008065 .
[84] Krause B, Márquez-Ruiz J,
Cohen Kadosh R. The effect of transcranial direct
current stimulation: a role for cortical excitation/inhibition balance? Front
Hum Neurosci 2013;7:602. doi: 10.3389/fnhum.2013.0 0602 .
[85] Targ ete d Non-Invasive Neurostimulation Meaningfully Reduces Seizure Fre-
quency in Refractory Epilepsy Patients with Prior Surgical Interventions. Py-
zowski P, Ruffini G, Salvador R, Rotenberg A, Kaye HL, Shafi M, San J.D. Ab-
stract presented at the annual meeting of the American Epilepsy Society
(AES) on December 09, 2019.
[86] Ozdemir RA, Tadayo n E, Boucher P, Momi D, Karakhanyan KA, Fox MD, et al.
Individualised perturbation of the human connectome reveals reproducible
biomarkers of network dynamics relevan t to cognition.
Proc Natl Acad Sci U
S A 2020;117:8115–25. doi: 10.1073/pnas.1911240117 .
[87] Fox MD. Mapping symptoms to brain networks with the human connectome.
N Engl J Med 2018;379:2237–45. doi: 10. 105 6/ NE JM ra170615 8 .
[88] Armstrong T. See brain cancer as more than just the sum of biology. Nature
2018;561:S45. doi: 10.1038/d41586- 018- 06706-5 .
[89] Klein M , Postma TJ , Taphoorn MJB , Aaronson NK , Vandertop WP ,
Muller M , et al. The prognostic value of cognitive functioning in the
survival of patients with high-grade glioma. Neurology 2003;61:1796–8
10.1212/01.wnl.0 0 0 0 098892.33018.4c .
[90] Lee J, Chaloner Winton Hall
R. The impact of gliomas on cognition and capac-
ity. J Am Acad Psychiatry Law 2019;47:350–9. doi: 10.29158/JAAPL.003841-19 .
[91] Ghinda DC, Wu JS, Duncan NW, Northoff G . How much is enough-can rest-
ing state fMRI provide a demarcation for neurosurgical resection in glioma?
Neurosci Biobehav Rev 2018;84:245–61. doi: 10.1016/j.neubiorev.2017.11.019 .
[92] Tuovinen N, de Pasquale F, Caulo M, Caravasso CF, Giudice E, Miceli R, et al.
Transient effects of tumor location on the functional architecture at rest in
glioblastoma patients: three longitudinal case studies. Radiat Oncol Lond Engl
2016;11:107. doi: 10.1186/s13014-016-0683- x .
[93] R. Romero-Garcia, J. Suckling, M. Owen, M. Assem,
R. Sinha, P. Coelho, et al.
Memory recovery is related to default mode network impairment and neu-
rite density during brain tumours treatment 2019. https://doi.org/10.1101/
19008581 .
[94] Valiente M, Ahluwalia MS, Boire A, Brastianos PK, Goldberg SB, Lee EQ, et al.
The evolving landscape of brain metastasis. Trends Cancer 2018;4:176–96.
doi: 10.1016/j.t recan .2018.01.0 03 .
[95] Santarnecchi E, Brem AK, Levenbaum E, Thompson T, Kadosh RC, Pascual-
Leone A. Enhancing cognition using transcranial electrical stimulation. Curr
Opin Behav Sci 2015;4:171–8. doi: 10.1016/j.cobeha.2015.06.003 .
[96] Song B, Wen P, Ahfock T, Li Y. Numeric investigation of brain tumor influence
on the current distributions during transcranial direct current stimulation.
IEEE Trans Biomed Eng 2016;63:176–87. doi: 10 .1109 /T BM E. 2015 .2 46 86 72 .
[97] Barcia JA, Sanz A, González-Hidalgo M, de Las Heras C, Alonso-Lera P, Díaz P,
et al. rTMS stimulation to induce plastic changes at the language motor area
in a patient with a left recidivant brain tumor affecting Broca’s area. Neuro-
case 2012;18:132–8. doi: 10.1080/13554794.2011.568500 .
[98] Rivera-Rivera PA, Rios-Lago M, Sanchez-Casarrubios S, Salazar O, Yus M,
González-Hidalgo M, et al. Cortical plasticity catalyzed by prehabilitation en-
ables extensive resection of brain tumors in eloquent areas. J Neurosurg
2017;126:1323–33. doi: 10.3171/2016.2.JNS152485 .
[99] Robles SG, Gatignol P, Lehéricy S, Duffau H. Long-term brain plasticity al-
lowing a multistage surgical approach to world health organization grade
II gliomas in eloquent areas. J Neurosurg 2008;109:615–24. doi: 10. 3171/ JN S/
2008/109/10/0615 .
[100] Sandran N, Hillier S, Hordacre B. Strategies to implement and moni-
tor in-home transcranial electrical stimulation in neurological and psychi-
atric patient populations: a systematic review. J Neuroengineering Rehabil
2019;16:58. doi: 10.1186/s12984- 019- 0529- 5 .
[101] Nitsche MA, Fr icke K, Henschke U, Schlitterlau A, Liebetanz D, Lang N, et al.
Pharmacological modulation of cortical excitabil ity shifts induced by tran-
scranial direct current stimulation in humans. J Physiol 2003;553:293–301.
doi: 10.1113/jphysiol.2003.049916 .
[102] Sills GJ, Rogawski MA. Mechanisms of action of currently used antiseizure
drugs. Neuropharmacology 2020;168:107966. doi: 10.1016/j.neuropharm.2020.
107966 .
12