Mol Imaging Biol (2019)

DOI: 10.1007/s11307-019-01349-y
* World Molecular Imaging Society, 2019

RESEARCH ARTICLE

Comparison of [18F]Fluoroethyltyrosine PET

and Sodium MRI in Cerebral Gliomas: a Pilot
Study
Aliaksandra Shymanskaya,1 Wieland A. Worthoff ,1 Gabriele Stoffels,1
Johannes Lindemeyer,1 Bernd Neumaier,1 Philipp Lohmann,1 Norbert Galldiks,1,2,3
Karl-Josef Langen,1,4,5 N. Jon Shah1,5,6
1
Institute of Neuroscience and Medicine (3, 4, 5, 11), Forschungszentrum Jülich GmbH, 52425, Jülich, Germany
2
Department of Neurology, University of Cologne, Cologne, Germany
3
Center of Integrated Oncology (CIO), Universities of Bonn and Cologne, Cologne, Germany
4
Department of Nuclear Medicine, RWTH Aachen University, Aachen, Germany
5
Jülich-Aachen Research Alliance (JARA) – Section JARA-Brain, Aachen, Germany
6
Department of Neurology, RWTH Aachen University, Aachen, Germany

Abstract
Purpose: Positron emission tomography (PET) using O-(2-[18F]fluoroethyl)-L-tyrosine ([18F]FET)
improves the diagnostics of cerebral gliomas compared with conventional magnetic resonance
imaging (MRI). Sodium MRI is an evolving method to assess tumor metabolism. In this pilot
study, we explored the relationship of [18F]FET-PET and sodium MRI in patients with cerebral
gliomas in relation to the mutational status of the enzyme isocitrate dehydrogenase (IDH).
Procedures: Ten patients with untreated cerebral gliomas and one patient with a recurrent
glioblastoma (GBM) were investigated by dynamic [18F]FET-PET and sodium MRI using an
enhanced simultaneous single-quantum- and triple-quantum-filtered imaging of 23Na (SISTINA)
sequence to estimate total (NaT), weighted non-restricted (NaNR, mainly extracellular), and
restricted (NaR, mainly intracellular) sodium in tumors and normal brain tissue. [18F]FET uptake
and sodium parameters in tumors with a different IDH mutational status were compared. After
biopsy or resection, histology and the IDH mutational status were determined
neuropathologically.
Results: NaT (p = 0.05), tumor-to-brain ratios (TBR) of NaT (p = 0.02), NaNR (p = 0.003), and the
ratio of NaT/NaR (p G 0.001) were significantly higher in IDH-mutated than in IDH-wild-type
gliomas (n = 5 patients each) while NaR was significantly lower in IDH-mutated gliomas (p =
0.01). [18F]FET parameters (TBR, time-to-peak) were not predictive of IDH status in this small
cohort of patients. There was no obvious relationship between sodium distribution and [18F]FET
uptake. The patient with a recurrent GBM exhibited an additional radiation injury with strong
abnormalities in sodium MRI.

Aliaksandra Shymanskaya and Wieland A. Worthoff contributed equally to

this work.
Electronic supplementary material The online version of this article (https://
doi.org/10.1007/s11307-019-01349-y) contains supplementary material,
which is available to authorized users.

Correspondence to: Wieland Worthoff; e-mail: [email protected]

 Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

Conclusions: Sodium MRI appears to be more strongly related to the IDH mutational status than
are [18F]FET-PET parameters. A further evaluation of the combination of the two methods in a
larger group of high- and low-grade gliomas seems promising.
Key words: MRI, FET-PET, Sodium imaging, Gliomas, IDH mutational status

Introduction restricted and non-restricted sodium [19, 20]. While the so-
called single-quantum coherences (SQCs) develop in both
Cerebral gliomas are the most common primary brain tumors areas with restricted (mostly intracellular sodium) and non-
with an incidence of 5–6 per 100,000 persons per year [1]. restricted sodium (mostly extracellular sodium), triple-
Despite intensive efforts to improve therapy, the majority of quantum coherences (TQCs) evolve in a restricted environ-
gliomas are associated with a poor prognosis [2]. The ment only [21], and can be measured when an appropriate
molecular genetic profile of cerebral gliomas such as the pulse sequence is applied [22, 23]. SISTINA sequences
gene mutation encoding for the isocitrate dehydrogenase allow simultaneous acquisition of signal originating from
enzyme (IDH) is of increasing importance for treatment restricted and non-restricted sodium [19, 20].
strategy [3]. Since this requires invasive tissue sampling, Initial studies on human brain tumors indicated that the
there is also a great interest in obtaining such information measurement of total sodium content could provide infor-
using modern molecular imaging techniques [4]. Initial mation on tumor response during radiotherapy and chemo-
studies have shown that a combination of conventional therapy [16, 24]. Furthermore, a recent study using an ultra-
magnetic resonance imaging (MRI) and advanced MRI high field MRI (7 T) in brain tumor patients has reported
methods such as perfusion-weighted imaging (PWI) and that the ratio of intracellular and extracellular sodium in
diffusion-weighted imaging (DWI) may predict IDH muta- cerebral gliomas has a high predictive value for the gene
tion status non-invasively [5, 6]. Furthermore, recent studies mutation encoding for the IDH [25]. An IDH mutation is an
report that amino acid positron emission tomography (PET) important molecular marker in the World Health Organiza-
is also highly predictive of IDH mutational status and that tion (WHO) classification of brain tumors in 2016 [3].
dynamic PET using O-(2-[ 18 F]fluoroethyl)-L-tyrosine Patients with an IDH mutation have an improved outcome
([18F]FET), especially, is an independent predictor of than do patients with IDH-wild-type (wt) gliomas. Similarly,
survival and superior to molecular parameters in terms of a relationship between [18F]FET-PET parameters and IDH
patient prognosis in some subgroups [7–12]. mutation has been reported [8–10]. The aim of this pilot
Another evolving MR method that may substantially study was to explore the relationship of metabolic parame-
contribute to the characterization of brain tumors is sodium ters derived from [18F]FET-PET and disturbances of
(23Na) MRI. Sodium is the most important electrolyte restricted and unrestricted sodium in patients with cerebral
which plays an essential role in human cell physiology gliomas with a different IDH mutational status.
[13], and its distribution reflects the functional status of the
sodium-potassium pump and sodium channels in cells [14].
The sodium-potassium pump supports a concentration Materials and Methods
gradient between intra- and extracellular sodium via energy Patients
mainly supplied by ATP hydrolysis [15]. This process is
disrupted if the metabolic energy supply is insufficient or if Eleven patients, referred to the Department of Nuclear
the cell membrane is damaged and can potentially lead to a Medicine of the University Clinic of Aachen, were investi-
rise in intracellular sodium concentration, change in cell gated by [18F]FET-PET for brain tumor assessment and had
volume, or malfunction of cells [16–18]. The chemical additional morphological imaging in a 4 T MRI scanner for
environment of sodium ion’s inside and outside cells comparison with [18F]FET-PET. Sodium imaging was
determines relaxation behavior of sodium ions thus gener- offered in addition to conventional MR sequences. All
ating MR contrasts. A number of techniques are under subjects gave a written informed consent prior to
development to differentiate MR signals originating from [18F]FET-PET and MR imaging. All 11 patients had a
intra- and extracellular sodium ions. Earlier studies did not neuropathological diagnosis based on the 2016 WHO
allow this distinction, which made it difficult to detect classification [3] (4 female, age 43 ± 12, and 7 male, age
alterations of intracellular sodium due to a high contribution 53 ± 15). Five patients had an untreated IDH-mutated glioma
from the extracellular compartment. In this study, a (1 astrocytoma WHO Grade II; 3 anaplastic astrocytoma
technique using enhanced simultaneous single-quantum- WHO grade III; 1 glioblastoma WHO grade IV), five
and triple-quantum-filtered imaging of 23Na (enhanced patients an untreated IDH-wild-type glioma (3 anaplastic
SISTINA) developed at our institute has been applied, glioma WHO Grade III; 1 diffuse midline glioma WHO IV;
allowing simultaneous acquisition of signal originating from 1 glioblastoma WHO grade IV), and one patient a pretreated
Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

Table 1. Details about the patient cohort

23 23 23
Group Pat. no. Diagnosis Location Age, sex Gd T1 Na NaT Na NaNR Na NaR [18F]FET-PET [18F]FET-PET
early TTP

Untreated IDH 1 A-II L frontal 34, m Neg. ++ ++ − − − No

mut gliomas 2 A-III L temporal 35, f Neg. ++ ++ − − − No
3 A-III L fronto-temp. 29, f Neg. ++ ++ − − − No
4 A-III L frontal 54, f Neg. + − − ++ No
5 GBM-IV L frontal insular 38, m Pos. ++ ++ − − ++ No
Untreated IDH 6 A-III L insular 51, m Neg. − − − − No
wt gliomas 7 DMG-IV R postero-lateral 52, f Neg. + − − − No
8 A-III L temporal 66, m Neg. + − + ++ No
9 A-III R basal ganglia 78, m Pos. + − − ++ Yes
10 GBM-IV L temporal 56, m Neg. − − − ++ No
Pretreated 11 Rec-GBM R parietal 50, m Pos. + + − + No
IDH-wt GBM Radiation L precentral Pos. + ++ − − − No
injury

IDH mut gliomas with mutation of the isocytrate dehydrogenase gene, IDH wt wild-type, gliomas without mutation of the isocitrate dehydrogenase gene, A
astrocytoma, GBM glioblastoma, DMG diffuse midline glioma H3 K27M-mutant, I–IV WHO grades, rec recurrent, NaT total sodium concentration, NaNR
non-restricted sodium SNR (mainly extracellular), NaR restricted sodium SNR (mainly intracellular), early TTP time to peak of the time activity curve earlier
than 17.5 min p.I. (Yes/No), B++^ strongly increased signal, B+^ slightly increased signal, B−B equal to background, B−−^ decreased signal.

multifocal glioblastoma WHO grade IV (IDH-wild-type). hemisphere was obtained from an appropriately placed ROI
The latter patient showed a recurrent tumor and a radiation [30]. The selection of an appropriate ROI based on thresh-
injury in different brain areas as confirmed by further olding was infeasible as some lesions show an [18F]FET
follow-up. Given its retrospective nature, the local ethics uptake similar to that of normal brain tissue; therefore, an
committee of the University of Aachen waived the require- irregular ROI was positioned by hand around peculiarities in
ment for additional approval. The study adheres to the the proton MR images. The co-registered [18F]FET-PET
standards established in the Declaration of Helsinki. Further data is assigned to this ROI. [18F]FET uptake was expressed
details about the patient cohort are provided in Table 1. as the standardized uptake value (SUV) by dividing the
radioactivity (Bq/ml) in the tissue by the radioactivity
injected per gram of body weight. Mean and maximal
tumor-to-brain ratios (TBRmean,TBRmax) were calculated on
[18F]FET-PET
summed [18F]FET-PET images from 20 to 40 min postin-
jection by dividing the mean and maximal ROI value (Bq/
The amino acid [18F]FET was produced as described previously
ml) of the lesion by the mean ROI value of the normal brain.
[26, 27]. According to the German guidelines for brain tumor
If the TBRmax was higher than the threshold of 1.6, the
imaging using radiolabeled amino acid analogues, all patients
lesion was considered [18F]FET positive [31]. Time activity
had been fasting for at least 12 h before PET scanning [28].
curves (TAC) for the whole data set were produced using the
Dynamic PET studies were acquired up to 50 min after
appropriate ROIs and the time from the beginning of the
intravenous injection of approximately 200 MBq [18F]FET on
acquisition to the maximum [18F]FET uptake in the lesion
an ECAT EXACT HR+ scanner (Siemens Medical Systems,
(time-to-peak/TTP) was obtained. TTP earlier than 17.5 min
Inc.) in 3-dimensional mode (32 rings; axial field of view,
after injection was rated as early TTP and later than 17.5 min
15.5 cm). The reconstructed dynamic dataset consisted of 16
after injection as late TTP [8].
time frames (5 × 1 min; 5 × 3 min; 6 × 5 min). Attenuation
correction was based on a 10 min transmission scan measured
with three rotating line sources (68Ge/68Ga). Data were
corrected for random and scattered coincidences, and dead time Sodium MR Imaging
prior to iterative reconstruction of 63 image planes using the
OSEM algorithm (16 subsets, 6 iterations). The reconstructed Sodium images were acquired on a home-built 4 T MRI
image resolution was approximately 5.5 mm [29]. scanner based around a Siemens console with a dual-tuned
[18F]FET-PET scans were co-registered with conven- Na/H birdcage coil (RAPID Biomed, Germany) using an
tional MRI (see below) and evaluated by region-of-interest enhanced SISTINA sequence [20]. Enhanced SISTINA
(ROI) using an MPI tool (Version 3.28, ATV, Kerpen, consists of a triple-quantum filter (TQF) with an ultrashort
Germany). For tumors with increased [18F]FET uptake, the echo time (UTE) readout train after the first 90 ° pulse,
transaxial slice showing the highest lesion intensity was which gives total sodium concentration weighting, and a
chosen and a circular ROI with a diameter of 1.6 cm was Cartesian, multiple gradient echo (MGRE) readout after the
drawn automatically around the lesion maximum, making third 90 ° pulse, which yields information about single-
tumor evaluation independent on the size of the lesion. As a quantum (SQ) and triple-quantum (TQ) coherences after
reference, normal-appearing matter from the contralateral application of appropriate phase cycle [32].
 Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

The parameters used for simultaneous acquisition of Statistical Analysis

UTE, SQ, and TQ images are provided in Electronic
Supplementary Material (ESM) Table A1; sequence timing Descriptive statistics for sodium and [18F]FET parameters
and echo spacing were the same as in [20]. The repetition are provided as mean and standard deviation for all
time was 150 ms, yielding a measurement time of patients as well as for IDH-mutated and IDH-wild-type
approximately 10 min. Multiple GREs were acquired for groups separately. To compare sodium MRI and [18F]FET-
relaxometry measurements. The data were regridded to PET parameters in patients with and without IDH muta-
2.5 mm isotropic resolution. Further details on k-space tion, the Student t test for independent samples was used.
acquisition and data post-processing, calculation of unbiased The Wilcoxon test was used when variables were not
signal-to-noise ratio (SNR) in the sodium images, and normally distributed. Pearson Product Correlation was
correction procedures have been reported previously [33]. performed for testing correlations between two variables
The conventional MR images acquired at 4 T were co- due to data normality (Shapiro-Wilk test delivered p
registered with [18F]FET-PET scans and subsequently with 9 0.05). P values of 0.05 or less were considered
sodium images, as well as between UTE and Cartesian significant. Statistical analyses were performed using
sodium images using the FSL library (FMRIB, Oxford, UK) SigmaPlot software (SigmaPlot Version 11.0, Systat
[34–36]. Three regions were segmented manually using the Software Inc., San Jose, CA) and PASW Statistics
normal proton images: the tumor area, which corresponds software (Release 22.0.0, SPSS Inc., Chicago, IL, USA).
either to increased [18F]FET uptake in PET uptake or to T1- No correction for multiple testing was applied.
or T2-signal abnormalities in case of indifferent [18F]FET
uptake, a reference area in the normal-appearing contralat-
eral brain tissue, and normally appearing white matter. Results
Sodium concentration, NaT, in the ROIs was determined using Sodium Imaging
NaTROI = SNRROI·SNRref−1·NaTref, with the SNR in the sodium
UTE images (SNRROI), a sodium signal from the vitreous humor Visual evaluation of the sodium images of the patients
of the eyes (SNRref, taken as a reference), and NaTref = 135 mmol/ revealed a prominent relative increase of NaT and NaNR
l, which is an approximation from the literature [37–39]. compared to that of the normal brain and a decrease of NaR
Total sodium concentration, NaT, signal from sodium in four of five patients with IDH mutation. A typical
ions with predominantly non-restricted mobility (NaNR = example is shown in Fig. 1a (patient 2). In contrast, this
SNR of SQ-weighted images) and signal from sodium ions pattern was not observed in any of the five IDH-wild-type
with predominantly restricted mobility (NaR = SNR of TQ- patients (Table 1). A representative example is shown in
weighted images) were estimated. Furthermore, the TBRs of Fig. 2a (patient 7). Quantitative evaluation showed that the
the signals relative to healthy contralateral tissue and the group of tumors with IDH mutation showed a significantly
ratio between NaT and NaR were determined. Complete data higher NaT than IDH-wild-type gliomas (58.7 ± 10.4mmol/
for the patients is given in Table 2 and in ESM. l versus 40.5 ± 24.0mmol/l; p = 0.05), a significantly higher

Table 2. Complete data of the patients

Group Patient Sodium MRI [18F]FET-PET

No NaT (mmol/l) NaT (TBR) NaNR (TBR) NaR (TBR) NaT/NaR TBRmean TBRmax TTP (min)

IDH mut 1 51.4 2.00 2.22 0.66 3.03 1.0 1.4 40

2 66.5 2.17 3.73 0.46 4.72 1.4 1.6 40
3 66.2 2.33 1.98 0.55 4.24 0.4 1.1 40
4 43.9 2.17 2.81 0.53 4.09 2.3 3.2 40
5 65.4 2.14 2.86 0.49 4.37 2.6 3.4 35
Mean N = 5 58.7 ± 10.4 2.16 ± 0.12 2.72 ± 0.68 0.54 ± 0.08 4.09 ± 0.64 1.54 ± 0.91 2.14 ± 1.08 39 ± 2
IDH wt 6 57.1 1.62 1.46 0.83 1.95 1.0 1.0 40
7 44.1 2.27 1.43 0.78 2.91 1.3 2.0 25
8 48.9 1.52 1.55 1.05 1.45 2.0 2.8 40
9 23.9 0.98 1.16 0.56 1.75 3.3 5.6 7
10 28.3 1.01 0.77 0.86 1.17 2.2 3.1 25
Mean N = 5 40.5 ± 14.0 1.48 ± 0.53 1.27 ± 0.32 0.82 ± 0.18 1.85 ± 0.66 1.96 ± 0.90 2.90 ± 1.71 27 ± 13
P value 0.05 0.02 0.003 0.01 G 0.001 0.48 0.78 0.10
Recurrence radiation injury 11 44.4 2.20 1.91 0.71 3.10 2.0 3.1 25
57.1 2.82 2.71 0.60 4.70 1.3 1.9 40

IDH mut gliomas with mutation of the isocytrate dehydrogenase gene, IDH wt wild-type, gliomas without mutation of the isocitrate dehydrogenase gene, NaT
total sodium concentration, NaNR non-restricted sodium SNR (mainly extracellular), NaR restricted sodium SNR (mainly intracellular), TAC time activity
curve, TBRmean mean tumor-to-brain ratio of [18F]FET uptake, TBRmax maximum tumor-to-brain ratio of [18F]FET uptake, TTP time-to-peak in the time-
activity curve of [18F]FET uptake in the tumor.
Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

Fig. 1. a MRI and PET of a 35-year-old patient with an IDH-mutated anaplastic astrocytoma WHO grade III (white arrow). The
tumor shows no contrast enhancement in the T1-weighted MRI and is clearly delineated in the FLAIR image. [18F]FET-PET
shows no significant tracer uptake in the tumor and an increasing time activity curve. Sodium imaging shows increased NaT,
increased NaNR, and decreased NaR, which is the typical finding for IDH-mutated gliomas. b MRI and PET of a 38-year-old
male patient with an IDH-mutated glioblastoma WHO grade IV. The tumor (white arrow) shows focal contrast enhancement in
the T1-weighted MRI and is clearly depicted in the FLAIR image. [18F]FET-PET shows significant tracer uptake in the tumor and
an increasing time activity curve. Similar to the IDH-mutated glioma in Fig. 1a, sodium imaging shows increased NaT, increased
NaNR, and decreased NaR.

TBR of NaT (2.16 ± 0.12 versus 1.48 ± 0.53; p = 0.02) and a 42 %, NaR decreased by 15 %) than the recurrent
significantly higher TBR of NaNR (2.72 ± 0.68 versus 1.27 glioblastoma (Fig. 4).
± 0.32, p = 0.003). Furthermore, TBR of NaR was signifi-
cantly decreased in IDH-mutated gliomas compared with
that of IDH-wild-type (0.54 ± 0.08 versus 0.82 ± 0.18, p = [18F]FET PET and Comparison with Sodium MRI
0.01). Correspondingly, the ratio NaT/NaR was significantly
different between the two groups (4.09 ± 0.64 versus 1.85 ± In contrast to the results with sodium imaging, quantitative
0.66, p = 0.001). The radiation injury of the pretreated evaluation of [18F]FET uptake parameters showed no
patient showed considerable stronger abnormality in sodium significant difference between IDH-mutated and IDH-wild-
imaging (NaT increased by 28 % and NaNR increased by type gliomas, which is not in line with the results in the
 Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

Fig. 2. a MRI and PET of a 52-year-old patient with a diffuse midline glioma WHO grade IV showing IDH-wild-type. The tumor
(white arrow) shows no contrast enhancement in the T1-weighted MRI and is clearly depicted in the T2-weighted MRI. [18F]FET-
PET shows no significant tracer uptake in the tumor and an increasing time activity curve. Sodium imaging shows a minor
increase of NaT, but no abnormality in NaNR and NaR. b MRI and PET of a 78-year-old male patient with an IDH-wild-type
anaplastic astrocytoma WHO grade III in the right basal ganglia. The tumor (white arrows) shows focal contrast enhancement in
the T1-weighted MRI and diffuse abnormalities in the FLAIR image. [18F]FET-PET shows significant tracer uptake in the tumor
and a time activity curve with an early peak. Sodium imaging shows no abnormality in NaT, NaNR, or NaR. The signal decrease
in the midfrontal area in NaR is due to an artifact caused by air in the ethmoid sinuses.

literature obtained in larger groups of patients (Table 2). In contrast, high [18F]FET uptake was found in both tumors
the group of IDH-mutated gliomas, two patients showed an with abnormal and inconspicuous sodium distribution (Figs.
increased [18F]FET uptake and all tumors in this group 1b and 2b). In the latter patient with an IDH-wild-type,
showed a late TTP. In the group of IDH-wild-type tumors, however, the [18F]FET TAC showed an early TTP peak
two patients showed low [18F]FET uptake and three patients followed by a constant descent, which is indicative of a
showed a late TTP, which is usually associated with an IDH highly malignant tumor. Correlation analysis between
mutation (Table 1). Furthermore, there was no obvious sodium MRI and [18F]FET-PET parameters in the tumors
relationship between [18F]FET uptake and sodium imaging. yielded no significant correlation except a weak negative
Cases with low [18F]FET uptake showed both abnormal and relationship between the TBRmax of [18F]FET uptake and
inconspicuous sodium distribution (Figs. 1a and 2a). In NaT (Fig. 3).
Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

Fig. 3. Comparison of the TBRmax of [18F]FET uptake in brain tumors and total sodium content. There is a weak inverse
relationship, i.e., tumors with increased NaT appear to have lower [18F]FET uptake.

Discussion significantly lower in gliomas with IDH mutation than in

IDH-wild-type gliomas. In that study, the NaR/NaT ratio
Sodium MRI is an evolving method in this field and thus far,
was an independent predictor of progression-free survival
only little data are available with this method in relation to
[25]. We here used the enhanced SISTINA method, which
molecular markers of gliomas. A recent study by Biller et al.
provides further insights into the intratumoral distribution of
[25] reported that the ratio of 23Na ions with restricted 23
Na ions with restricted and unrestricted mobility in relation
mobility and total sodium concentration (NaR/NaT) was

Fig. 4. MRI and PET of a 50-year-old male patient with a recurrent IDH-wild-type glioblastoma WHO grade IV in the right
parietal lobe and a radiation injury on the contralateral side (white arrows), both confirmed by follow-up. Both lesions exhibit
contrast enhancement in the T1-weighted MRI and diffuse abnormalities in the T2-weigthed MRI. [18F]FET-PET is positive in the
recurrent tumor and negative in the necrosis. Sodium imaging shows more prominent abnormalities in the radionecrosis.
 Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

to the IDH mutation status of gliomas. We observed that population and, in particular, the inclusion of a larger
80 % of IDH-mutated gliomas exhibited a significant fraction of patients with WHO grade II gliomas.
increase of total 23Na concentration NaT, a relative increase In our study, there was also a case of a recurrent
of 23Na ions with unrestricted mobility (mainly extracellular, glioblastoma patient with radiation injury, as confirmed by
NaNR) and a relative decrease of 23Na ions with restricted repeated PET/MR imaging after 5 months (patient 11).
mobility (mainly intracellular, NaR). In contrast, none of the Interestingly, the radiation injury showed a considerable
IDH-wild-type gliomas showed this pattern and all param- stronger abnormality in sodium imaging (increased NaT and
eters of sodium imaging were significantly different com- NaNR, decreased NaR) than the recurrent glioblastoma (Fig.
pared with those of IDH-mutated gliomas although the 4). Our study has several limitations. The number of patients
sample size was relatively small. These findings are in line in this pilot study is rather small, and the distribution of
with the observations by Biller et al. mentioned above [25]. WHO grades of the tumors in the different groups with and
Similar results have also been reported in a recent study without IDH mutation is not completely identical. Therefore,
which, however, included only one single case of a IDH- a confirmation of the results in a larger collective with
wild-type glioma [40]. Thus, the results of this pilot study homogeneous and more representative collectives is needed.
support the hypothesis that abnormalities in sodium imaging Some caution is advised when comparing the results of
are related to the molecular profile of gliomas such as the sodium imaging with the literature, because the methods of
IDH mutation status, which is related to the aggressiveness MR imaging are different [13]. A relative decrease in the
of the tumor. In contrast to sodium imaging, this small series concentration of sodium ions with restricted mobility may
of patient’s [18F]FET-PET parameters did not show differ- also be caused by an increase of the intracellular space, or
ences between IDH-mutated and IDH-wild-type gliomas, the increase of extracellular space, e.g., by edema. Further-
which is contrary to the results observed in larger series of more, sodium MRI suffers from low SNR and low
patients [8–10]. The lack of statistical significance may be resolution, which will be improved in the near future by
explained by the relatively small group of patients, but the the increasing availability of ultra-high field scanners at 7 T
apparent discrepancy with sodium imaging gives the [32, 46] and higher [47, 48]. The possible PVE from the
impression that advanced sodium imaging is a very sensitive high SNR regions such as ventricles and choroid plexus are
method with regard to this issue and is possibly superior to however balanced by the appropriate choice of the reference
[18F]FET-PET. Furthermore, our data give rise to the ROIs. Nevertheless, the better image resolution offered by
assumption that the sodium signal and amino acid transport ultra-high field systems, in particular for the acquisition of
abnormalities are not closely linked to each other but reflect images with restricted sodium weighting, will significantly
different metabolic properties of gliomas. enhance the quality of sodium quantification.
The importance of sodium metabolism in the develop-
ment of gliomas has gained increasing interest in recent
years. It has been shown that glioblastoma patients with Conclusions
mutations in sodium channels show a significantly shorter
survival compared to patients with no sodium channel In summary, the results of this pilot study support the
mutations [41, 42]. Increased intracellular sodium concen- hypothesis that sodium imaging adds information to predict
tration and increased cellular pH is a manifestation of Na+/ the IDH mutational status. Sodium MRI appears to be more
H+ exchanger isoform 1 (NHE1) upregulation, which has strongly related to IDH mutational status than are [18F]FET-
been linked to glioma proliferation, invasion, and resistance PET parameters. However, with respect to the excellent
to temozoleomide therapy [43, 44]. Increased intracellular performance of [18F]FET-PET in previous reports, a further
sodium concentration, which is commonly exhibited by evaluation of the combination of the two methods in gliomas
rapidly dividing cells [45], may indicate a breakdown of the is promising. Sodium imaging suffers from low SNR and
Na+/K+-ATPase and/or Na+ cotransporters. Nevertheless, the resolution but, as already noted, increasing availability of 7
IDH-mutated gliomas showed a higher total sodium con- T MRI scanners will allow these methods to be combined
centration than in the normal brain (Fig. 1 a, b), which is and open new horizons for the non-invasive assessment of
probably caused by an increased extracellular volume. These the molecular profile of gliomas.
relationships in gliomas with different IDH genotypes are
elucidated for the first time by the SISTINA method used Acknowledgements. The authors thank Petra Engels, Elke Bechholz, Anita
here, which allows the simultaneous detection of the Köth, Suzanne Schaden, Elisabeth Theelen, Silke Frensch, Kornelia Frey,
Stefan Schwan, and Lutz Tellmann for assistance in the patient studies;
restricted and non-restricted sodium. It is tempting to Johannes Ermert, Silke Grafmüller, Erika Wabbals and Sascha Rehbein for
speculate that IDH-mutated gliomas exhibit either a lower radiosynthesis of [18F]FET.
amount of restricted, mainly intracellular sodium, or a lower
level cellularity than IDH-wild-type gliomas, suggesting that Compliance with Ethical Standards
the extra/intracellular sodium concentration gradient is better
maintained in these tumors (Figs. 1 and 2). Confirmation of Conflict of Interest
this hypothesis, however, requires a significantly larger The authors declare that they have no conflict of interest.
Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

Ethical Approval 18. Nagel AM, Bock M, Hartmann C, Gerigk L, Neumann JO, Weber
MA, Bendszus M, Radbruch A, Wick W, Schlemmer HP, Semmler
All procedures performed in studies involving human participants were in W, Biller A (2011) The potential of relaxation-weighted sodium
accordance with the ethical standards of the institutional and/or national magnetic resonance imaging as demonstrated on brain tumors.
research committee and with the 1964 Helsinki declaration and its later Investig Radiol 46:539–547
amendments or comparable ethical standards. 19. Fiege DP, Romanzetti S, Mirkes CC, Brenner D, Shah NJ (2013)
Simultaneous single-quantum and triple-quantum-filtered MRI of
Informed Consent 23
Na (SISTINA). Magn Reson Med 69:1691–1696
Informed written consent was obtained from all individual participants 20. Worthoff WA, Shymanskaya A, Shah NJ (2019) Relaxometry and
included in the study. quantification in simultaneously acquired single and triple quantum
filtered sodium MRI. Magn Reson Med 81:303–315
21. Jaccard G, Wimperis S, Bodenhausen G (1986) Multiple-quantum
References Nmr-spectroscopy of S=3/2 spins in isotropic-phase - a new probe for
multiexponential relaxation. J Chem Phys 85:6282–6293
22. Keltner JR, Wong ST, Roos MS (1994) Three-dimensional triple-
1. Ostrom QT, Gittleman H, Liao P et al (2014) CBTRUS statistical quantum-filtered imaging of 0.012 and 0.024 M sodium-23 using
report: primary brain and central nervous system tumors diagnosed in short repetition times. J Magn Reson B 104:219–229
the United States in 2007-2011. Neuro Oncol 16(Suppl 4):iv1–i63 23. Woessner DE (2001) NMR relaxation of spin-(3)/(2) nuclei: effects of
2. Ohgaki H, Kleihues P (2005) Population-based studies on incidence, structure, order, and dynamics in aqueous heterogeneous systems.
survival rates, and genetic alterations in astrocytic and oligodendrog- Concept Magn Res 13:294–325
lial gliomas. J Neuropathol Exp Neurol 64:479–489 24. Babsky AM, Zhang H, Hekmatyar SK, Hutchins GD, Bansal N
3. Louis DN, Perry A, Reifenberger G, von Deimling A, Figarella- (2007) Monitoring chemotherapeutic response in RIF-1 tumors by
Branger D, Cavenee WK, Ohgaki H, Wiestler OD, Kleihues P, Ellison single-quantum and triple-quantum-filtered 23Na MRI, (1)H diffusion-
DW (2016) The 2016 World Health Organization Classification of weighted MRI and PET imaging. Magn Reson Imaging 25:1015–1023
Tumors of the Central Nervous System: a summary. Acta Neuropathol 25. Biller A, Badde S, Nagel A, Neumann JO, Wick W, Hertenstein A,
131:803–820 Bendszus M, Sahm F, Benkhedah N, Kleesiek J (2016) Improved
4. Langen KJ, Galldiks N, Hattingen E, Shah NJ (2017) Advances in brain tumor classification by sodium MR imaging: prediction of IDH
neuro-oncology imaging. Nat Rev Neurol 13:279–289 mutation status and tumor progression. AJNR Am J Neuroradiol
5. Leu K, Ott GA, Lai A, Nghiemphu PL, Pope WB, Yong WH, Liau 37:66–73
LM, Cloughesy TF, Ellingson BM (2017) Perfusion and diffusion 26. Hamacher K, Coenen HH (2002) Efficient routine production of the
MRI signatures in histologic and genetic subtypes of WHO grade II- 18F-labelled amino acid O-2-18F fluoroethyl-L-tyrosine. Appl Radiat
III diffuse gliomas. J Neuro-Oncol 134:177–188 Isot 57:853–856
6. Xing Z, Yang X, She D, Lin Y, Zhang Y, Cao D (2017) Noninvasive 27. Wester HJ, Herz M, Weber W, Heiss P, Senekowitsch-Schmidtke R,
assessment of IDH mutational status in World Health Organization Schwaiger M, Stöcklin G (1999) Synthesis and radiopharmacology of
grade II and III astrocytomas using DWI and DSC-PWI combined O-(2-[18F]fluoroethyl)-L-tyrosine for tumor imaging. J Nucl Med
with conventional MR imaging. AJNR Am J Neuroradiol 38:1138– 40:205–212
1144 28. Langen KJ, Bartenstein P, Boecker H, Brust P, Coenen HH, Drzezga
7. Kunz M, Albert NL, Unterrainer M et al (2018) Dynamic 18F-FET A, Grünwald F, Krause BJ, Kuwert T, Sabri O, Tatsch K, Weber WA,
PET is a powerful imaging biomarker in gadolinium-negative gliomas. Schreckenberger M (2011) German guidelines for brain tumour
Neuro-Oncology imaging by PET and SPECT using labelled amino acids. Nuklearme-
8. Suchorska B, Giese A, Biczok A, Unterrainer M, Weller M, Drexler dizin 50:167–173
M, Bartenstein P, Schüller U, Tonn JC, Albert NL (2018) Identifica- 29. Herzog H, Tellmann L, Hocke C, Pietrzyk U, Casey ME, Kuwert T
tion of time-to-peak on dynamic 18F-FET-PET as a prognostic marker (2004) NEMA NU2-2001 guided performance evaluation of four
specifically in IDH1/2 mutant diffuse astrocytoma. Neuro-Oncology siemens ECAT PET scanners. IEEE Trans Nucl Sci 51:2662–2669
20:279–288 30. Unterrainer M, Vettermann F, Brendel M, Holzgreve A, Lifschitz M,
9. Verger A, Stoffels G, Bauer EK, Lohmann P, Blau T, Fink GR, Zähringer M, Suchorska B, Wenter V, Illigens BM, Bartenstein P,
Neumaier B, Shah NJ, Langen KJ, Galldiks N (2018) Static and Albert NL (2017) Towards standardization of 18F-FET PET imaging:
dynamic 18F-FET PET for the characterization of gliomas defined by do we need a consistent method of background activity assessment?
IDH and 1p/19q status. Eur J Nucl Med Mol Imaging 45:443–451 EJNMMI Res 7:48
10. Unterrainer M, Winkelmann I, Suchorska B, Giese A, Wenter V, 31. Pauleit D, Floeth F, Hamacher K et al (2005) O-(2-[18F]fluoroethyl)-
Kreth FW, Herms J, Bartenstein P, Tonn JC, Albert NL (2018) L-tyrosine PET combined with MRI improves the diagnostic
Biological tumour volumes of gliomas in early and standard 20-40 assessment of cerebral gliomas. Brain 128:678–687
min 18F-FET PET images differ according to IDH mutation status. Eur 32. Fiege DP, Romanzetti S, Tse DH et al (2013) B0 insensitive multiple-
J Nucl Med Mol Imaging 45:1242–1249 quantum resolved sodium imaging using a phase-rotation scheme. J
11. Rohrich M, Huang K, Schrimpf D et al (2018) Integrated analysis of Magn Reson 228:32–36
dynamic FET PET/CT parameters, histology, and methylation 33. Worthoff WA, Shymanskaya A, Shah NJ (2018) Relaxometry and
profiling of 44 gliomas. Eur J Nucl Med Mol Imaging 45:1573–1584 quantification in simultaneously acquired single and triple quantum
12. Lopci E, Riva M, Olivari L, Raneri F, Soffietti R, Piccardo A, Bizzi filtered sodium MRI. Magn Reson Med Sci in press
A, Navarria P, Ascolese AM, Rudà R, Fernandes B, Pessina F, 34. Woolrich MW, Jbabdi S, Patenaude B, Chappell M, Makni S, Behrens
Grimaldi M, Simonelli M, Rossi M, Alfieri T, Zucali PA, Scorsetti M, T, Beckmann C, Jenkinson M, Smith SM (2009) Bayesian analysis of
Bello L, Chiti A (2017) Prognostic value of molecular and imaging neuroimaging data in FSL. Neuroimage 45:S173–S186
biomarkers in patients with supratentorial glioma. Eur J Nucl Med 35. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ,
Mol Imaging 44:1155–1164 Johansen-Berg H, Bannister PR, de Luca M, Drobnjak I, Flitney DE,
13. Madelin G, Regatte RR (2013) Biomedical applications of sodium Niazy RK, Saunders J, Vickers J, Zhang Y, de Stefano N, Brady JM,
MRI in vivo. J Magn Reson Imaging 38:511–529 Matthews PM (2004) Advances in functional and structural MR image
14. Shah NJ, Worthoff WA, Langen KJ (2016) Imaging of sodium in the analysis and implementation as FSL. Neuroimage 23(Suppl 1):S208–
brain: a brief review. NMR Biomed 29:162–174 S219
15. Murphy E, Eisner DA (2009) Regulation of intracellular and 36. Jenkinson M, Beckmann CF, Behrens TE et al (2012) Fsl. Neuro-
mitochondrial sodium in health and disease. Circ Res 104:292–303 image 62:782–790
16. Thulborn KR, Lu A, Atkinson IC, Damen F, Villano JL (2009) 37. Kaufman PL, Adler FH, Levin LA, Alm A (2011) Adler’s physiology
Quantitative sodium MR imaging and sodium bioscales for the of the eye. Elsevier Health Sci
management of brain tumors. Neuroimaging Clin N Am 19:615–624 38. Coe JI (1969) Postmortem chemistries on human vitreous humor. Am
17. Ouwerkerk R, Bleich KB, Gillen JS, Pomper MG, Bottomley PA J Clin Pathol 51:741–750
(2003) Tissue sodium concentration in human brain tumors as
measured with 23Na MR imaging. Radiology 227:529–537
 Shymanskaya A. et al.: FET PET and Sodium MRI in Gliomas

39. Ouwerkerk R, Bleich KB, Gillen JS, Pomper MG, Bottomley PA 45. Cameron IL, Smith NK, Pool TB, Sparks RL (1980) Intracellular
(2003) Tissue sodium concentration in human brain tumors as concentration of sodium and other elements as related to mitogenesis
measured with Na-23 MR imaging. Radiology 227:529–537 and oncogenesis in vivo. Cancer Res 40:1493–1500
40. Nunes Neto LP, Madelin G, Sood TP, Wu CC, Kondziolka D, 46. Romanzetti S, Mirkes CC, Fiege DP, Celik A, Felder J, Shah NJ
Placantonakis D, Golfinos JG, Chi A, Jain R (2018) Quantitative (2014) Mapping tissue sodium concentration in the human brain: a
sodium imaging and gliomas: a feasibility study. Neuroradiology comparison of MR sequences at 9.4 Tesla. Neuroimage 96:44–53
60:795–802 47. Shah NJ (2015) Multimodal neuroimaging in humans at 9.4 T: a
41. Joshi AD, Parsons DW, Velculescu VE, Riggins GJ (2011) Sodium technological breakthrough towards an advanced metabolic imaging
ion channel mutations in glioblastoma patients correlate with shorter scanner. Brain Struct Funct 220:1867–1884
survival. Mol Cancer 10:17 48. Shah NJ, Oros-Peusquens AM, Arrubla J, Zhang K, Warbrick T,
42. Wang R, Gurguis CI, Gu W, Ko EA, Lim I, Bang H, Zhou T, Ko JH Mauler J, Vahedipour K, Romanzetti S, Felder J, Celik A, Rota-Kops
(2015) Ion channel gene expression predicts survival in glioma E, Iida H, Langen KJ, Herzog H, Neuner I (2013) Advances in
patients. Sci Rep 5:11593 multimodal neuroimaging: hybrid MR-PET and MR-PET-EEG at 3 T
43. Cong D, Zhu W, Shi Y, Pointer KB, Clark PA, Shen H, Kuo JS, Hu S, and 9.4 T. J Magn Reson 229:101–115
Sun D (2014) Upregulation of NHE1 protein expression enables
glioblastoma cells to escape TMZ-mediated toxicity via increased
H(+) extrusion, cell migration and survival. Carcinogenesis 35:2014–
2024
44. Zhu W, Carney KE, Pigott VM, Falgoust LM, Clark PA, Kuo JS, Sun Publisher’s Note Springer Nature remains neutral with regard to jurisdic-
D (2016) Glioma-mediated microglial activation promotes glioma tional claims in published maps and institutional affiliations.
proliferation and migration: roles of Na+/H+ exchanger isoform 1.
Carcinogenesis 37:839–851