Glioblastoma broadband power as a longitudinal biomarker for tumor progression

Glioblastoma (grade IV glioma) is a devastating disease with dismal prognosis,
for which individualized treatment through biomarkers is often aimed for, but
rarely achieved. This lack of success necessitates a radically different way of
thinking, which our group has been exploring in the past years. Recent work
shows that glioma growth may not only depend on characteristics of the tumor
itself, but is also determined by activity of the healthy tissue around it.
Higher activity of neurons surrounding glioma in an animal model has been shown
to cause an acceleration of glioma growth, through glutamate-dependent
*neurogliomal synapses*. Development of these synapses is promoted through
increased neuroligin-3 (NLGN3) expression, which is also caused by higher
neuronal activity. This causal relationship indicates that higher neuronal
activity leads to faster tumor growth.
In patients, neuronal activity is difficult to measure directly, but the
broadband power of electroencephalography (EEG) and magnetoencephalography
(MEG) may most straightforwardly relate to neuronal activity. EEG and MEG allow
for easy, non-invasive and low-burden measurement of brain activity: a five
minute resting recording usually suffices.
We already translated some of these preclinical results, first showing that
lower broadband power (indicating less neuronal activity) at the time of
diagnosis relates to lower NLGN3 expression in resected tumor tissue in glioma
patients.16 Moreover, lower broadband power relates to longer progression-free
survival (PFS), with a hazard ratio (HR) of 2.1, adjusted for known predictors
of PFS. In a second study, we replicated these results in the postoperative
phase of the disease, showing again that glioma patients with greater broadband
power have shorter PFS (HR 2.5, adjusted for known predictors).
These promising results raise the question whether broadband power might be
useful as a prognostic biomarker during the disease course of glioblastoma, and
particularly during the standard first line treatment period comprised of
chemoradiation and adjuvant chemotherapy. In this timeframe, distinguishing
between pseudoprogression and actual tumor progression is an important clinical
challenge, which limits optimization of treatment.

We will investigate (1) whether broadband power can be measured repeatedly and
accurately within this patient population with limited variation in case of
stable disease, (2) whether meaningful changes in broadband power in the case
of real progression exceed normal variations, and (3) whether it is possible to
establish a cut-off score for broadband power that may provide clinically
relevant diagnostic accuracy (taking the REMARK guidelines into account).

Longitudinal, observational study.

MEG/EEG does not bear any risks: this is a completely non-invasive measurement
of brain activity, that does not make use of radiation. The additional MRI
sequences do not present risks either: CEST is a safe sequence, which is
already a clinical/commercial product of some vendors inside the Netherlands
(e.g. Philips Healthcare). Task and rsfMRI are routinely performed in these
patients and also do not bear any risk. Patients will be aware that they can
refuse to undergo the extra scan times anytime including during the ongoing
scan.
The burden of participation consists of undergoing these extra measurements at
maximum five times. These measurements will be combined with necessary visits
to the hospital as much as possible, to limit effort and time investment of the
patients. There is no benefit of participation for patients.