T2-FLAIR mismatch sign in dysembryoplastic neuroepithelial tumor

概要

T2-FLAIR mismatch sign in dysembryoplastic neuroepithelial tumor
Shumpei Onishi, MD1,5, Vishwa Jeet Amatya, MD, PhD2, Manish Kolakshyapati MD, PhD1,
Motoki Takano, MD1, Ushio Yonezawa, MD1, Akira Taguchi, MD1, Yoko Kaichi MD, PhD4,
Yukio Takeshima MD, PhD2, Kazuo Awai, MD, PhD4, Kazuhiko Sugiyama, MD, PhD3,
Kaoru Kurisu, MD, PhD1, Fumiyuki Yamasaki, MD, PhD1
(1) Department of Neurosurgery, Graduate School of Biomedical and Health Sciences, Hiroshima
University, Hiroshima, Japan
(2) Department of Pathology, Graduate School of Biomedical and Health Sciences, Hiroshima
University, Hiroshima, Japan
(3) Department of Clinical Oncology and Neuro-oncology Program, Hiroshima University
Hospital, Hiroshima, Japan
(4) Department of Diagnostic Radiology, Graduate School of Biomedical and Health Sciences,
Hiroshima University, Hiroshima, Japan
(5) Department of Neurosurgery, National Hospital Organization, Kure Medical Center and
Chugoku Cancer Center, Hiroshima, Japan

Address:
Department of Neurosurgery, Graduate School of Biomedical and Health Sciences,
Hiroshima University, 1-2-3 Kasumi, Minami-ku, Hiroshima-city, Hiroshima, 734-8551, Japan
Telephone: +81-82-257-5227
Fax: +81-82-257-5229

ABSTRUCT
Purpose: T2-FLAIR mismatch sign was reported as specific imaging marker in non-enhancing
diffuse astrocytoma, IDH-mutant & 1p/19q non-codeleted. However, most of the previous studies
for T2-FLAIR mismatch sign were confirmed only among lower grade glioma. The aim of this
study is to assess the T2-FLAIR mismatch sign in dysembryoplastic neuroepithelial tumor
(DNET) and unveil the exception rules of the sign.
Method: Eleven patients with histopathologically confirmed DNET were included in this study.
The MR images were evaluated by 2 independent reviewers to assess (i) the presence or absence
of T2-FLAIR mismatch sign and (ii) the presence or absence of gadolinium enhancement. CT
was also performed to evaluate calcification and localized thinning of the skull bone. Interreviewer agreement with Cohen’s kappa (ȡ) was calculated.
Results: The T2-FLAIR mismatch sign was present in 8 cases (72.7%) and absent in 3 cases
(27.3%). None of them showed contrast enhancement on initial MR images. The inter-reviewer
agreement for T2-FLAIR mismatch and CT characteristics was excellent (ȡ=1.00). All of the
DNET without T2-FLAIR mismatch presented with calcification on CT. All of the DNET adjacent
to skull vault (5 cases) presented with localized bone thinning overlying the tumor.
Conclusions: The T2-FLAIR mismatch sign was observed in more than half of the DNET and
the sign is not specific for diffuse astrocytoma, IDH-mutant & 1p19q non-codeleted. The
localized skull bone thinning overlying the tumor might help for diagnosis of DNET in some
cases.

Keywords: T2-FLAIR mismatch, dysembryoplastic neuroepithelial tumor, imaging marker

Abbreviation: DNET:dysembryoplastic neuroepithelial tumor

Introduction:
The World Health Organization (WHO) classification of central nervous system tumors
was updated in 2016. Molecular status, IDH mutation and 1p/19q codeletion, are mandatory to
classify the grade II-III diffuse glioma, so-called lower grade glioma[1]. Lower grade glioma is
genetically divided into three groups; 1) diffuse astrocytoma, IDH-mutant (&1p/19q noncodeleted), 2) oligodendroglioma, IDH-mutant & 1p/19q codeleted, 3) diffuse astrocytoma, IDHwildtype. The preoperative differentiation among these three groups is of high clinical relevance
because the prognosis and treatment strategy could be different.
The T2-FLAIR mismatch was reported as an imaging biomarker for diffuse astrocytoma,
IDH-mutant & 1p/19q non-codeleted, which may be a valuable method for preoperative
differentiation. The T2-FLAIR mismatch sign represented homogeneously hyper-intense signal
on T2WI and peripheral hyper-intense rim with central hypo intensity on FLAIR[2]. Several
studies demonstrated that T2-FLAIR mismatch sign has 100% positive predictive value (PPV)
for diagnosis of diffuse astrocytoma with IDH-mutant & 1p/19q non-codeleted[2-5]. However,
the previous studies for T2-FLAIR mismatch were confirmed only among lower grade glioma, in
other words, “diffuse astrocytic and oligodendroglial tumors”[2-6]. In real clinical practice, we
need to distinguish these from other supratentorial intra-parenchymal tumor, such as other
astrocytic tumors, neuronal and mixed neuronal-glial tumors.

The exception of T2-FLAIR mismatch sign was reported in pilomyxoid astrocytoma
and in H3 K27M mutant midline glioma [7]. However, these tumors can be differentiated based
on their location or gadolinium enhancement pattern. Although, dysembryoplastic neuroepithelial
tumor (DNET) can occur in any region of the supratentorial cortex[8], they usually develop at
location similar to the lower grade glioma. Differentiation between DNET and lower grade glioma
is clinically important because of their remarkably different treatment strategy. However, the T2FLAIR match/mismatch is not confirmed in DNET yet.
The aim of this study is to assess the presence of T2-FLAIR mismatch sign in DNET.
In this study, we showed that diagnosis based on T2-FLAIR mismatch sign alone could be the
pitfall for the differential diagnosis between DNET and diffuse astrocytoma, IDH-mutant and
1p/19q non-codeleted.

Methods
-

Patients selection

This retrospective study was approved by the institutional review board of our institution. We
reviewed images of 11 patients (age ranging 3-42, median 17 years) with histologically proven
DNET from April 2001 to December 2018. All patients underwent preoperative imaging study in
our institution.

-

MR acquisition and evaluation

MR scans were acquired at 1.5T or 3.0T scanners and assessed by two independent authors (S.O
and F.Y). Investigators evaluated the T1-weighted imaging (T1WI), T2-weighted imaging
(T2WI), fluid-attenuated inversion recovery (FLAIR), diffusion-weighted imaging (DWI) and
post-contrast T1WI sequencing of each patient. The T2-FLAIR mismatch was defined similarly
to the original description of the sign[2]. The sign represented homogeneously hyper-intense
signal on T2WI and peripheral hyper-intense rim with central hypo intensity on FLAIR. Both
investigators independently assessed (i) presence or absence of the T2-FLAIR mismatch sign and
(ii) presence or absence of gadolinium contrast enhancement.

-

CT acquisition

CT scans were performed on a 320-detector CT scanner (Aquilion ONE; Canon Medical Systems)
or a 16 slices helical CT scanner (LightSpeed Ultra, General Electric Medical System). CT images
were also assessed by two independent authors (S.O & F.Y). Investigators evaluated the skull
thinning, cranial deformation and intratumoral calcification.

-

Histopathological and molecular classification

Tumor specimens after surgical resection were fixed in 10% phosphate-buffered formalin and
embedded in paraffin blocks. Representative slides were then stained with hematoxylin-eosin
reagent for standard histological diagnosis as described previously[9]. DNET was diagnosed
based on WHO classification for central nervous system tumors updated in 2016 by consensus of
two authors (V.J.A, & Y.T.).

-

Statistical analysis

Statistical analyses were performed with JMP pro ver. 13.0 (SAS institute, Cary, NC, USA). Interreviewer agreement was evaluated using the Kappa statistic (κ=0–0.40, poor; κ=0.41–0.60,
moderate; κ=0.61–0.80, good; κ =0.81–1.00, excellent).

Results
Eleven DNET cases were included in this study. All tumors showed high intensity on T2WI. The
T2-FLAIR mismatch sign was present in 8 cases (72.7%) and absent in 3 cases (27.3%). None of
them showed contrast enhancement with gadolinium on initial MR imaging. The summary of all
these cases is shown in Table 1.
Eight cases of DNET with T2-FLAIR mismatch sign were located in frontal lobe (3
case), lateral temporal lobe (5 case). Three temporal tumors without T2-FLAIR mismatch were

located in medial temporal lobe (2 case) and lateral temporal lobe (1 case). The inter-reviewer
agreement for the T2-FLAIR mismatch sign was excellent (ȡ=1.00). All DNET were low intense
on diffusion-weighted images, and hyper intense on ADC map. On CT images, localized thinning
of skull bone adjacent to DNET was observed in all of the tumors adjacent to the skull vault (5
cases). The deformation of the overlying part of the skull was also observed in 5 cases. In addition,
CT images revealed that all of the tumors without T2-FLAIR mismatch (3cases) presented with
partial calcification.
Representative cases with T2-FLAIR mismatch sign are shown in Fig. 1. And a
representative case without T2-FLAIR mismatch sign is shown in Fig. 2. These cases showed
oligodendroglioma-like cells in a mucin rich background and glioneuronal element with floating
neuron in HE staining (Fig. 1D,1H and 2D).

Discussion
In this study, we showed that DNET is a typical cerebral tumor with imaging
characteristics of T2-FLAIR mismatch. More than half (8 out of 11 cases (72.7%) of DNETs
presented with the T2-FLAIR mismatch sign without contrast enhancement on preoperative MRI.
While interpreting the T2-FLAIR mismatch sign in cerebral glioma population, physicians need
to be aware that it is not specific for diffuse astrocytoma, IDH-mutant &1p/19q non-codeleted in

cerebral hemispheric tumors.
T2-FLAIR mismatch sign has been originally described as having 100% PPV for diffuse
astrocytoma, IDH-mutant &1p/19q non-codeleted in 2017[2]. Subsequent studies also presented
the T2-FLAIR mismatch sign as good surrogate imaging marker for diffuse astrocytoma, IDHmutant &1p/19q non-codeleted with high PPV and high specificity [3-5]. However, the
pathophysiological characteristics of T2-FLAIR mismatch sign remain unclear. The T2-FLAIR
mismatch sign was previously described as “T2 FLAIR suppression” in protoplasmic astrocytoma,
presenting with thin FLAIR rim [10]. The tumor demonstrated that microcystic change was more
evident in central lesion than peripheral lesion of the tumor. Therefore, T2-FLAIR mismatch sign
may represent the difference in the histological background of the tumor cells.
Although the T2-FLAIR mismatch sign was of high clinical relevance, there were some
exceptional cases in gliomas. The exceptional cases were reported in pilomyxoid astrocytoma,
H3 K27M-mutant midline glioma, oligodendroglioma with IDH mutant &1p/19q codeleted, and
low-grade astrocytoma harboring MYB rearrangement[7]. Pilomyxoid astrocytoma usually show
the characteristics homogeneous enhancement, which would help for differentiation from diffuse
astrocytoma, IDH-mutant &1p/19q non-codeleted. H3 K27M-mutant midline glioma develops at
pons, thalamus, medulla oblongata, and spinal cord, and it is not so difficult to distinguish them
from diffuse astrocytoma, IDH-mutant &1p/19q non-codeleted. While, differentiation between

diffuse astrocytoma with IDH-mutant &1p/19q non-codeleted from oligodendroglioma with IDH
mutant &1p/19q codeleted is difficult. Juratli, et al. reported false positive mismatch sign in
28.5% of oligodendroglioma with IDH mutant &1p/19q codeleted including contrast-enhancing
tumor among 316 lower grade glioma cases. They also mentioned a criteria specific for diffuse
astrocytoma, IDH mutant &1p/19q non-codeleted, with 3 combining parameters: younger than
40 years old, larger than 6 cm and presence of T2-FLAIR mismatch sign. The criteria increased
the preoperative diagnostic accuracy for diffuse astrocytoma, IDH mutant &1p/19q non-codeleted
(PPV 88.6% and specificity 96%) [6].
Unfortunately, previous studies on T2-FLAIR mismatch sign focused only on grade II
and III diffuse glioma[2-6], and other supratentorial tumors such as DNET were excluded from
these studies. DNET and lower grade glioma mainly occur in supratentorial region. Moreover,
both diffuse astrocytoma, IDH-mutant &1p/19q non-codeleted and DNET have the characteristics
of “non-enhancement” with gadolinium. In addition, the largest DNET case series reported that
the mean age of seizure onset was 14.6 years (3 months – 54 years) and the mean age at surgery
was 30.5 years (6 – 65 years)[11]. In view of possibility of overlap between the patients with
DNET and lower grade glioma, we need to differentiate these tumors in clinical practice.
In our study, the T2-FLAIR mismatch sign was observed in 72.7% of DNET cases. The
MRI characteristics of DNET was previously reported as a “FLAIR ring sign” [12]. The FLAIR

ring sign is defined as an area of sharp well-defined hyperintensity at the border of tumors
separating the tumor from surrounding normal brain tissue on FLAIR. DNET usually showed
hypointense on T1WI and hyperintense on T2WI. Hence, the characteristic of FLAIR ring sign
resembles T2-FLAIR mismatch sign. The peripheral loose glioneuronal elements was speculated
as a cause of the FLAIR ring sign based on pathological findings [12]. Four out of 8 T2-FLAIR
mismatch positive cases demonstrated FLAIR ring sign in our series. The other cases with T2FLAIR mismatch sign showed peripheral hyper-intensity but not complete sharp ring on FLAIR.
There has been no report focusing on the incidence of FLAIR ring sign in diffuse glioma, and the
pathological/radiological difference between T2-FLAIR mismatch and FLAIR ring sign should
be confirmed with future larger studies.
In our series, some DNET (27.3%) did not present T2-FLAIR mismatch sign. All DNET
without T2-FLAIR mismatch sign showed calcification on CT, the calcification might have
influence on MR signal. However, among DNETs with/without T2-FLAIR mismatch sign, they
did not show significant difference on HE staining(Fig.1D, 1H and 2D). In our study, only central
portions of the tumors were submitted to the pathological examination, peripheral portions of the
tumor were not collected. Also in diffuse glioma, not all diffuse astrocytoma with IDH-mutant &
1p/19q non-codeleted presented with T2-FLAIR mismatch sign[2-5], and the previous reports did
not mentioned about the pathological difference between peripheral and central portions of the

tumors[2-5]. Future studies are necessary to confirm the pathological characteristics in central
and peripheral portion of the tumors and unveil the etiology of T2-FLAIR mismatch sign.
Accordingly, the presence of T2-FLAIR mismatch sign cannot be relied on to
distinguish diffuse astrocytoma, IDH-mutant &1p/19q non-codeleted from DNET. Rather,
physicians should focus on the deformations of the overlying part of the skull observed in 44%
of DNET[13]. Characteristics of highest apparent diffusion coefficient in DNET obtained from
diffusion-weighted imaging may also help the diagnosis of this tumor[14]. Further larger study is
necessary to confirm the predictive values of this imaging characteristics for DNET or diffuse
astrocytoma, IDH-mutant &1p19q non-codeleted.
Conclusions
The T2-FLAIR mismatch sign was observed in DNET and the sign is not specific for
diffuse astrocytoma, IDH-mutant &1p19q non-codeleted. The localized skull bone thinning might
help for the diagnosis of DNET in some cases.

Conflict of interest
The authors report no conflict of interest.

Figure legends:

Figure1:Representative cases with T2-FLAIR mismatch sign in DNETs (case No. 1 and 8).
Case No. 1: Axial T2WI (A) and FLAIR (B) image of a patient with right frontal DNET. T2WI
demonstrate homogeneous hyperintense signal throughout the lesion (A). FLAIR display
relatively hypointense signal in majority of the lesion with peripheral hyperintense signal rim (B).
CT image showed the localized skull bone thinning of the overlying part of the skull(C).
Hematoxylin-eosin staining showed oligodendroglioma-like cells in a mucin rich background and
glioneuronal element with floating neuron (D). Case No. 8: Axial T2WI (E) and FLAIR (F) image
of a patient with left lateral temporal DNET. T2WI demonstrate homogeneous hyperintense signal
throughout the lesion (E). FLAIR display relatively hypointense signal in majority of the lesion
with peripheral hyperintense signal rim (F). CT image showed the localized skull bone thinning
overlying the DNET (G). Hematoxylin-eosin staining showed oligodendroglioma-like cells in a
mucin rich background and glioneuronal element with floating neuron (H).

Figure2:Representative case without the T2-FLAIR mismatch sign in DNET at mesial temporal
lobe (case No. 5). Axial T2WI (A) and FLAIR (B) image of a patient with right mesial temporal
DNET. T2WI demonstrate mixed signal intensity with multicystic appearance (A). FLAIR also

display mixed signal intensity with hyperintense signal (B). CT showed the partial calcification
of low-density tumor (C). Hematoxylin-eosin staining showed oligodendroglioma-like cells in a
mucin rich background and glioneuronal element with floating neuron (D).

References:
[1]

D.N. Louis, A. Perry, G. Reifenberger, A. Deimling, D. Figarella-Branger, W.K.
Cavenee, et al., The 2016 World Health Organization Classification of Tumors of the
Central Nervous System: a summary, Acta Neuropathol. 131 (2016) 803–820.

[2]

S.H. Patel, L.M. Poisson, D.J. Brat, Y. Zhou, L. Cooper, M. Snuderl, et al., T2-FLAIR
Mismatch, an Imaging Biomarker for IDH and 1p/19q Status in Lower-grade Gliomas:
A TCGA/TCIA Project, Clinical Cancer Research. 23 (2017) 6078–6085.

[3]

M.P.G. Broen, M. Smits, M.M.J. Wijnenga, H.J. Dubbink, M.H.M.E. Anten, O.E.M.G.
Schijns, et al., The T2-FLAIR mismatch sign as an imaging marker for non-enhancing
IDH-mutant, 1p/19q-intact lower-grade glioma: a validation study, Neuro-Oncology.
20 (2018) 1393–1399.

[4]

A. Lasocki, F. Gaillard, A. Gorelik, M. Gonzales, MRI Features Can Predict 1p/19q
Status in Intracranial Gliomas, AJNR Am J Neuroradiol. 39 (2018) 687–692.

[5]

P.P. Batchala, T.J.E. Muttikkal, J.H. Donahue, J.T. Patrie, D. Schiff, C.E. Fadul, et al.,
Neuroimaging-Based Classification Algorithm for Predicting 1p/19q-Codeletion Status
in IDH-Mutant Lower Grade Gliomas, AJNR Am J Neuroradiol. 61 (2019) 6713.

[6]

T.A. Juratli, S.S. Tummala, A. Riedl, D. Daubner, S. Hennig, T. Penson, et al.,
Radiographic assessment of contrast enhancement and T2/FLAIR mismatch sign in
lower grade gliomas: correlation with molecular groups, J. Neurooncol. 141 (2019)
327–335.

[7]

D.R. Johnson, T.J. Kaufmann, S.H. Patel, A.S. Chi, M. Snuderl, R. Jain, There is an
exception to every rule-T2-FLAIR mismatch sign in gliomas, Neuroradiology. 61
(2019) 225–227.

[8]

A.R. Campos, H. Clusmann, M. von Lehe, P. Niehusmann, A.J. Becker, J. Schramm,
et al., Simple and complex dysembryoplastic neuroepithelial tumors (DNT) variants:
clinical profile, MRI, and histopathology, Neuroradiology. 51 (2009) 433–443.

[9]

V. Karlowee, V.J. Amatya, H. Hirano, T. Takayasu, R. Nosaka, M. Kolakshyapati, et
al., Multicentric Glioma Develops via a Mutant IDH1-Independent Pathway:
Immunohistochemical Study of Multicentric Glioma, Pathobiology. 84 (2017) 99–107.

[10]

K.L. Tay, A. Tsui, P.M. Phal, K.J. Drummond, B.M. Tress, MR imaging
characteristics of protoplasmic astrocytomas, Neuroradiology. 53 (2011) 405–411.

[11]

M. Thom, A. Toma, S. An, L. Martinian, G. Hadjivassiliou, B. Ratilal, et al., One
hundred and one dysembryoplastic neuroepithelial tumors: an adult epilepsy series
with immunohistochemical, molecular genetic, and clinical correlations and a review
of the literature, J. Neuropathol. Exp. Neurol. 70 (2011) 859–878.

[12]

H.A. Parmar, C. Hawkins, R. Ozelame, S. Chuang, J. Rutka, S. Blaser, Fluidattenuated inversion recovery ring sign as a marker of dysembryoplastic
neuroepithelial tumors, J Comput Assist Tomogr. 31 (2007) 348–353.

[13]

R. Stanescu Cosson, P. Varlet, F. Beuvon, C. Daumas Duport, B. Devaux, F.
Chassoux, et al., Dysembryoplastic neuroepithelial tumors: CT, MR findings and
imaging follow-up: a study of 53 cases, Journal of Neuroradiology. 28 (2001) 230–
240.

[14]

F. Yamasaki, K. Kurisu, K. Satoh, K. Arita, K. Sugiyama, M. Ohtaki, et al., Apparent
Diffusion Coefficient of Human Brain Tumors at MR Imaging, Radiology. 235 (2005)
985–991.

Table 1.
Summary of clinical and radiological features of dysembryoplastic neuroepithelial tumor cases
Age at

Gender

Case surgery,

Location

y

T2-FLAIR
mismatch sign

Gadolinium

Localized

contrast

skull bone

enhancement

thinning

Calcification

1

7

F

frontal

+

-

+

-

2

20

M

frontal

+

-

-

-

3

42

F

frontal

+

-

-

-

4

14

M

lateral temporal

+

-

+

-

5

41

M

mesial temporal

-

-

-

+

6

32

M

lateral temporal

-

-

-

+

7

16

M

mesial temporal

-

-

-

+

8

3

M

lateral temporal

+

-

+

-

9

12

M

lateral temporal

+

-

+

-

10

11

M

lateral temporal

+

-

+

-

11

18

M

mesial temporal

+

-

-

-