scispace - formally typeset
Search or ask a question
Journal ArticleDOI

Perfusion and diffusion MRI signatures in histologic and genetic subtypes of WHO grade II-III diffuse gliomas

TL;DR: Results suggest a combination of rCBV, ADC, T2 hyperintense volume, and presence of contrast enhancement together may aid in non-invasively identifying genetic subtypes of diffuse gliomas.
Abstract: The value of perfusion and diffusion-weighted MRI in differentiating histological subtypes according to the 2007 WHO glioma classification scheme (ie astrocytoma vs oligodendroglioma) and genetic subtypes according to the 2016 WHO reclassification (eg 1p/19q co-deletion and IDH1 mutation status) in WHO grade II and III diffuse gliomas remains controversial In the current study, we describe unique perfusion and diffusion MR signatures between histological and genetic glioma subtypes Sixty-five patients with 2007 histological designations (astrocytomas and oligodendrogliomas), 1p/19q status (+ = intact/- = co-deleted), and IDH1 mutation status (MUT/WT) were included in this study In all patients, median relative cerebral blood volume (rCBV) and apparent diffusion coefficient (ADC) were estimated within T2 hyperintense lesions Bootstrap hypothesis testing was used to compare subpopulations of gliomas, separated by WHO grade and 2007 or 2016 glioma classification schemes A multivariable logistic regression model was also used to differentiate between 1p19q+ and 1p19q- WHO II-III gliomas Neither rCBV nor ADC differed significantly between histological subtypes of pure astrocytomas and pure oligodendrogliomas ADC was significantly different between molecular subtypes (p = 00016), particularly between IDHWT and IDHMUT/1p19q+ (p = 00013) IDHMUT/1p19q+ grade III gliomas had higher median ADC; IDHWT grade III gliomas had higher rCBV with lower ADC; and IDHMUT/1p19q- had intermediate rCBV and ADC values, similar to their grade II counterparts A multivariable logistic regression model was able to differentiate between IDHWT and IDHMUT WHO II and III gliomas with an AUC of 084 (p < 00001, 74% sensitivity, 79% specificity) Within IDHMUT WHO II-III gliomas, a separate multivariable logistic regression model was able to differentiate between 1p19q+ and 1p19q- WHO II-III gliomas with an AUC of 080 (p = 00015, 64% sensitivity, 82% specificity) ADC better differentiated between genetic subtypes of gliomas according to the 2016 WHO guidelines compared to the classification scheme outlined in the 2007 WHO guidelines based on histological features of the tissue Results suggest a combination of rCBV, ADC, T2 hyperintense volume, and presence of contrast enhancement together may aid in non-invasively identifying genetic subtypes of diffuse gliomas

Summary (3 min read)

Introduction

  • Diffuse gliomas are a heterogeneous group of primary brain tumors with high morbidity and variable outcomes.
  • Survival, growth characteristics, and therapeutic sensitivity for lower grade diffuse gliomas (WHO II–III) are highly dependent on the dominant cell lineage represented within the tumor.
  • This ambiguity can often lead to a range of challenging diagnoses including mixed features of both astrocytomas and oligodendrogliomas (i.e. oligoastrocytomas).
  • Diffusion-weighted imaging (DWI) is a physiologic imaging modality that exploits the diffusion of water molecules to create contrast between tissues.
  • Furthermore, the authors hypothesize low grade gliomas (WHO II) will have similar diffusion and perfusion MR characteristics across subtypes, whereas anaplastic gliomas (WHO III) will demonstrate significantly different patterns of diffusion 1 3 and perfusion MR characteristics depending on the specific histologic or genetic subtype.

Patients

  • All adult patients with WHO grade II or grade III gliomas between 2010 and 2016 were retrospectively reviewed (231 total).
  • Patients with the rare combination of 1p/19q co-deletion positive and IDH1 wild-type were excluded from statistical analysis due to their small sample size (Grade II: n = 1; Grade III: n = 1).
  • All patients in this study signed institutional review boardapproved informed consent to have their data stored in their neuro-oncology database and used for research purposes.
  • The images and histology were taken from this database without new review.

MR imaging and post‑processing

  • All patients included in this study had a T2-weighted FLAIR, a diffusion-weighted MR, and a perfusionweighted DSC-MR that was performed on either a 1.5 or 3  T scanner prior to surgery, with the average length between MRI and surgery at 10.6 ± 11.8 days, range 0–67 days.
  • Apparent diffusion coefficients (ADC) were calculated using a loglinear fit to all available b-values in the diffusion-weighted/ diffusion-tensor images using custom in-house code in MATLAB (Natick, MA).
  • A 0.025  mmol/kg preload dose of gadolinium contrast agent was administered prior to DSC-MRI with a 0.075 mmol/kg preload dose of gadolinium contrast agent used for DSC acquisition.
  • Relative cerebral blood volume (rCBV) was calculated using a recently introduced leakage correction algorithm using in-house custom MATLAB code that corrects for bidirectional contrast agent exchange [28, 29] and normalizing to contra-lateral normal-appearing white matter (NAWM) tissue.

Regions of interest

  • Regions of interest (ROIs) of suspected tumor and/ or edema were defined by abnormal hyperintensity on T2-weighted or T2-weighted FLAIR using semi-automated segmentation techniques, followed by manual inspection and adjustment of the resulting contour as described previously [30].
  • The median ADC and median normalized rCBV were obtained from each tumor lesion.
  • All tissues obtained from surgical resections were stained with hematoxylin–eosin.
  • PCR products were sequenced by BigDye Terminator v1.1 (Applied Biosystems), and sequences were determined via a 3730 sequencer (Applied Biosystems).

Statistical analyses

  • Due to the small sample size, bootstrap hypothesis testing was used for all statistical tests.
  • First, using the real data, the f-score was calculated from the one-way ANOVA test if three groups were being compared or a t-score if only two groups were being compared to each other.
  • The f-score or t-test was then calculated from the resampled, mean-shifted data 10,000 times to generate the null hypothesis distribution.
  • All bootstrap routines were performed in MATLAB using in-house custom code, while the receiver operating characteristic statistics, i.e. area under the curve, were performed in GraphPad Prism (La Jolla, CA) using the sensitivity and specificity from the highest likelihood ratio.
  • Since eight statistical tests comparing the glioma subtypes were performed (four for the 2007 analysis, four for the 2016 analysis, two tumor volume analyses, two presence of contrast enhancement analyses, and two receiver operating characteristic curve analyses), for multiple comparisons correction, a p value of 0.015/14 = 0.0036 was considered to be significant.

Results

  • Using the 2007 WHO glioma classification scheme, the authors observed a substantial overlap in perfusion and diffusion MR measurements between WHO grade and histological subtype (Fig. 2).
  • No significant differences in rCBV or ADC were observed between pure astrocytomas and pure Fig. 2 Comparisons of rCBV and ADC between histologic subtypes of WHO grade II and III gliomas using the 2007 WHO classification.
  • In particular, IDHMUT gliomas appeared to have relatively small variance and formed well-defined clusters exhibiting higher ADC and moderate rCBV compared with IDHWT anaplastic gliomas.
  • Univariate results suggested ADC was significant between these subtypes (Fig.  5g, p = 0.0018).
  • The coefficients for these models are outlined Table 1.

Discussion

  • In the current study, the authors explored whether diffusion and perfusion MR signatures could identify histologic or genetic subtypes of WHO II or III gliomas that form the basis for the 2007 and 2016 WHO glioma classification schemes, respectively.
  • On the other hand, the joint rCBV and ADC characteristics have much smaller variation in both IDH1 mutation subtypes.
  • Using rCBV and ADC along with the more traditional biomarkers of contrast enhancement and volume 1 3 Fig. 3 Comparisons of rCBV and ADC between genetic subtypes of WHO grade II and III gliomas using the 2016 WHO classification.
  • 6 Classification of IDH status in WHO II and III gliomas using rCBV, ADC, T2 hyperintense lesion volume, and presence of contrast enhancement.

Conclusion

  • In summary, the current study suggests that ADC better correlates with genetic subtypes of gliomas according to the 2016 WHO guidelines and imaging measurements were varied less when tumors were stratified based on histological features using the 2007 criteria.
  • Furthermore, using ADC in combination with rCBV, T2 volume enhancement, and contrast enhancement allowed us to distinguish between IDHWT and IDHMUT gliomas as well as between IDHMUT/1p19q+ and IDHMUT/1p19q− gliomas.

Did you find this useful? Give us your feedback

Content maybe subject to copyright    Report

UCLA
UCLA Previously Published Works
Title
Perfusion and diffusion MRI signatures in histologic and genetic subtypes of WHO grade
II-III diffuse gliomas.
Permalink
https://escholarship.org/uc/item/0tp3p5xw
Journal
Journal of neuro-oncology, 134(1)
ISSN
0167-594X
Authors
Leu, Kevin
Ott, Garrett A
Lai, Albert
et al.
Publication Date
2017-08-01
DOI
10.1007/s11060-017-2506-9
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California

Vol.:(0123456789)
1 3
J Neurooncol
DOI 10.1007/s11060-017-2506-9
CLINICAL STUDY
Perfusion anddiffusion MRI signatures inhistologic andgenetic
subtypes ofWHO grade II–III diffuse gliomas
KevinLeu
1,2,3
· GarrettA.Ott
1,2
· AlbertLai
4,6
· PhioanhL.Nghiemphu
4,6
·
WhitneyB.Pope
2
· WilliamH.Yong
7
· LindaM.Liau
8
· TimothyF.Cloughesy
4,6
·
BenjaminM.Ellingson
1,2,3,4,5,9
Received: 21 October 2016 / Accepted: 21 May 2017
© Springer Science+Business Media New York 2017
multivariable logistic regression model was also used to
differentiate between 1p19q
+
and 1p19q
WHO II–III glio-
mas. Neither rCBV nor ADC differed significantly between
histological subtypes of pure astrocytomas and pure oligo-
dendrogliomas. ADC was significantly different between
molecular subtypes (p = 0.0016), particularly between
IDH
WT
and IDH
MUT
/1p19q
+
(p = 0.0013). IDH
MUT
/1p19q
+
grade III gliomas had higher median ADC; IDH
WT
grade
III gliomas had higher rCBV with lower ADC; and
IDH
MUT
/1p19q
had intermediate rCBV and ADC val-
ues, similar to their grade II counterparts. A multivariable
logistic regression model was able to differentiate between
IDH
WT
and IDH
MUT
WHO II and III gliomas with an AUC
of 0.84 (p < 0.0001, 74% sensitivity, 79% specificity).
Within IDH
MUT
WHO II–III gliomas, a separate multi-
variable logistic regression model was able to differentiate
between 1p19q
+
and 1p19q
WHO II–III gliomas with an
AUC of 0.80 (p = 0.0015, 64% sensitivity, 82% specificity).
Abstract The value of perfusion and diffusion-weighted
MRI in differentiating histological subtypes according to
the 2007 WHO glioma classification scheme (i.e. astro-
cytoma vs. oligodendroglioma) and genetic subtypes
according to the 2016 WHO reclassification (e.g. 1p/19q
co-deletion and IDH1 mutation status) in WHO grade II
and III diffuse gliomas remains controversial. In the cur-
rent study, we describe unique perfusion and diffusion
MR signatures between histological and genetic glioma
subtypes. Sixty-five patients with 2007 histological des-
ignations (astrocytomas and oligodendrogliomas), 1p/19q
status (+ = intact/− = co-deleted), and IDH1 mutation sta-
tus (MUT/WT) were included in this study. In all patients,
median relative cerebral blood volume (rCBV) and appar-
ent diffusion coefficient (ADC) were estimated within T2
hyperintense lesions. Bootstrap hypothesis testing was used
to compare subpopulations of gliomas, separated by WHO
grade and 2007 or 2016 glioma classification schemes. A
* Benjamin M. Ellingson
bellingson@mednet.ucla.edu
1
UCLA Brain Tumor Imaging Laboratory (BTIL), Center
forComputer Vision andImaging Biomarkers, University
ofCalifornia, Los Angeles, LosAngeles, CA, USA
2
Department ofRadiological Sciences, David Geffen
School ofMedicine, University ofCalifornia, Los Angeles,
LosAngeles, CA, USA
3
Department ofBioengineering, Henry Samueli School
ofEngineering andApplied Science, University
ofCalifornia, Los Angeles, LosAngeles, CA, USA
4
UCLA Neuro-Oncology Program, University ofCalifornia,
Los Angeles, LosAngeles, CA, USA
5
Department ofBiomedical Physics, David Geffen School
ofMedicine, University ofCalifornia, Los Angeles,
LosAngeles, CA, USA
6
Department ofNeurology, David Geffen School ofMedicine,
University ofCalifornia, Los Angeles, LosAngeles, CA,
USA
7
Department ofPathology andLaboratory Medicine, David
Geffen School ofMedicine, University ofCalifornia, Los
Angeles, LosAngeles, CA, USA
8
Department ofNeurosurgery, David Geffen School
ofMedicine, University ofCalifornia, Los Angeles,
LosAngeles, CA, USA
9
Departments ofRadiological Sciences andPsychiatry,
David Geffen School ofMedicine, University ofCalifornia,
Los Angeles, 924 Westwood Blvd, Suite 615, LosAngeles,
CA90024, USA

J Neurooncol
1 3
ADC better differentiated between genetic subtypes of glio-
mas according to the 2016 WHO guidelines compared to
the classification scheme outlined in the 2007 WHO guide-
lines based on histological features of the tissue. Results
suggest a combination of rCBV, ADC, T2 hyperintense vol-
ume, and presence of contrast enhancement together may
aid in non-invasively identifying genetic subtypes of dif-
fuse gliomas.
Keywords Perfusion MRI· Diffusion MRI· Glioma·
WHO classification· IDH mutant
Introduction
Diffuse gliomas are a heterogeneous group of primary
brain tumors with high morbidity and variable outcomes.
High-grade, malignant gliomas represent a majority of
primary malignant brain tumors (70–80%) [13] and have
dismal prognoses of only 12–15 months for glioblastoma
(WHO grade IV) and 2–5 years for anaplastic gliomas
(WHO grade III) [4]. Survival, growth characteristics,
and therapeutic sensitivity for lower grade diffuse gliomas
(WHO II–III) are highly dependent on the dominant cell
lineage represented within the tumor. In 2007, the World
Health Organization (WHO) glioma classification scheme
used histological appearance on light microscopy to dis-
tinguish gliomas of different glial lineages. Astrocytomas
make up a majority of gliomas [4] and WHO grade III
anaplastic astrocytomas have a median survival of around
2–3 years [1]. Astrocytic tumors are comprised of irregu-
lar, hyperchromatic nuclei with a glial fibrillary acidic pro-
tein (GFAP)—positive cytoplasm. On the other hand, oli-
godendrogliomas are characterized by round nuclei with a
branching network of capillaries with possible calcification
[5]. They are more sensitive to chemoradiation, and over-
all, WHO grade III anaplastic oligodendrogliomas have a
slightly better prognosis compared with anaplastic astrocy-
tomas [6].
Although classification of diffuse gliomas based solely
on histological features of the resected tissue may be ben-
eficial for staging disease and developing more targeted
therapeutics, this strategy can be prone to sampling error
during tumor resection and bias based on relatively sub-
jective criteria. This ambiguity can often lead to a range
of challenging diagnoses including mixed features of both
astrocytomas and oligodendrogliomas (i.e. oligoastrocy-
tomas). As more molecular and genetic information about
these types of tumors have become mature in the litera-
ture it has became clear that sub-stratification of tumor
types should likely be performed using genetic tests for
common deletions and mutations. In 2016, the WHO
glioma classification scheme was restructured to reflect
two common molecular alterations in gliomas: 1p/19q co-
deletion (1p19q
) and isocitrate dehydrogenase-1 mutation
(IDH
MUT
) [7]. Co-deletion of 1p and 19q is most com-
monly associated with oligodendroglial tumors [8, 9] and is
both predictive of therapeutic response and prognostic for
survival [10, 11]. IDH mutations, which are known to be a
driver mutation in low-grade gliomas [12, 13], are associ-
ated with more favorable outcomes as they are known to be
more sensitive to chemoradiation [14, 15].
Diffusion- and perfusion-weighted magnetic resonance
imaging (MRI), have been extensively studied as non-
invasive tools for identifying glioma subtypes, character-
izing the aggressiveness of gliomas, and identifying early
malignant transformation. Diffusion-weighted imaging
(DWI) is a physiologic imaging modality that exploits the
diffusion of water molecules to create contrast between tis-
sues. One common measurement obtained from DWI is
the apparent diffusion coefficient (ADC). ADC is an esti-
mate of the magnitude of the diffusion of water molecules
within the tissue, and there is a strong negative correlation
between the ADC and tumor cellularity in gliomas [16, 17].
Perfusion-weighted imaging (PWI) is an MRI modality
that gives insights into the delivery of blood to tissues by
monitoring a bolus of contrast agent as it passes through
the blood vasculature. A common biomarker derived from
PWI is the relative cerebral blood volume (rCBV). Meas-
urements of rCBV have been shown to be higher in high-
grade tumors than in low-grade tumors and may correlate
with glioma vascularity [18, 19].
Several studies have illustrated differences in diffusion
and perfusion MR measurements between oligodendroglio-
mas from astrocytomas [20, 21], 1p/19q co-deleted tumors
(1p19q
) from non-1p/19q co-deleted tumors (1p19q
+
)
[2225], and IDH mutant (IDH
MUT
) from IDH wild-type
(IDH
WT
) diffuse gliomas [26]. Presumably, these observed
differences in diffusion MR measurements of apparent
diffusion coefficient (ADC) and perfusion MR imaging
measurements of relative cerebral blood volume (rCBV)
between subtypes reflects known differences in tumor cell
morphometry and aspects of vascular biology [27]. Despite
these interesting observations, there is a critical gap in our
current understanding of how diffusion and perfusion MRI
might be used together to better understand differences
between the histologic and genetic subtypes of tumors and
between WHO II and III tumors of the same subtype. We
hypothesize a combination of diffusion and perfusion MR
measurements will better separate tumors based on their
genetic characteristics (1p19q co-deletion and IDH1 muta-
tion status) [7] than the more subjective histologic criteria.
Furthermore, we hypothesize low grade gliomas (WHO II)
will have similar diffusion and perfusion MR characteris-
tics across subtypes, whereas anaplastic gliomas (WHO III)
will demonstrate significantly different patterns of diffusion

J Neurooncol
1 3
and perfusion MR characteristics depending on the spe-
cific histologic or genetic subtype. Finally, we hypothesize
that using diffusion, perfusion, T2-enhancement volume,
and contrast enhancement together will allow us to clas-
sify IDH
WT
from IDH
MUT
as well as IDH
MUT
/1p19q
+
from
IDH
MUT
/1p19q
tumors.
Materials andmethods
Patients
All adult patients with WHO grade II or grade III gliomas
between 2010 and 2016 were retrospectively reviewed
(231 total). Patients were only included if they met all of
the following inclusion criteria: (1) histologic diagnosis
of WHO grade II or grade III gliomas [1]; (2) dynamic
susceptibility contrast (DSC) perfusion-weighted MRI,
diffusion-weighted MRI, T2-weighted, and post-contrast
T1-weighted anatomical scan performed at initial diagno-
sis and prior to any surgery; (3) diagnosis of astrocytoma,
mixed glioma, or oligodendroglioma via histology; and
(4) known IDH1 mutation and 1p/19q co-deletion status.
Patients with the rare combination of 1p/19q co-deletion
positive and IDH1 wild-type were excluded from statisti-
cal analysis due to their small sample size (Grade II: n = 1;
Grade III: n = 1). A total of 65 patients (38 men, 27 women;
average age 46.5 ± 15.8 years; age range 21–85), with 31
WHO II and 34 WHO III gliomas, fit the inclusion criteria.
All patients in this study signed institutional review board-
approved informed consent to have their data stored in our
neuro-oncology database and used for research purposes.
The images and histology were taken from this database
without new review.
MR imaging andpost‑processing
All patients included in this study had a T2-weighted
FLAIR, a diffusion-weighted MR, and a perfusion-
weighted DSC-MR that was performed on either a 1.5
or 3 T scanner prior to surgery, with the average length
between MRI and surgery at 10.6 ± 11.8 days, range 0–67
days.
Diffusion MR acquisition parameters echo times (TE)
varied from 67 to 100 ms, repetition times (TR) ranged
from 7 to 10 s, flip angle was 90°, b-values used were
0 s/mm
2
and either 700 or 1000 s/mm
2
with matrix size
128 × 128. In some cases, diffusion-weighted images were
diffusion tensor images with 6–12 directions. Apparent
diffusion coefficients (ADC) were calculated using a log-
linear fit to all available b-values in the diffusion-weighted/
diffusion-tensor images using custom in-house code in
MATLAB (Natick, MA).
A 0.025 mmol/kg preload dose of gadolinium con-
trast agent was administered prior to DSC-MRI with a
0.075mmol/kg preload dose of gadolinium contrast agent
used for DSC acquisition. For DSC-MRI, echo times (TE)
ranged from 23 to 35 ms, repetition times (TR) ranged
from 1250 to 2000, flip angles were 35, 60, or 90, with 40
to 120 temporal time points at a slice thickness of 4–6mm
with an interslice gap of 0–1mm. A total of 12–25 slices
were collected with matrix size ranging from 80 × 96 to
128 × 128. Relative cerebral blood volume (rCBV) was
calculated using a recently introduced leakage correction
algorithm using in-house custom MATLAB code that cor-
rects for bidirectional contrast agent exchange [28, 29] and
normalizing to contra-lateral normal-appearing white mat-
ter (NAWM) tissue.
All images were registered to the T2-weighted image
using a mutual information algorithm and a 12-degree free-
dom transformation using FSL (FMRIB; http://www.fmrib.
ox.ac.uk/fsl/) or tkregister2 (Freesurfer, surfer.nmr.mgh.
harvard.edu; Massachusetts General Hospital, Harvard
Medical School).
Regions ofinterest
Regions of interest (ROIs) of suspected tumor and/
or edema were defined by abnormal hyperintensity on
T2-weighted or T2-weighted FLAIR using semi-automated
segmentation techniques, followed by manual inspection
and adjustment of the resulting contour as described previ-
ously [30]. The median ADC and median normalized rCBV
were obtained from each tumor lesion.
Histologic characterization (2007 WHO glioma
guidelines)
All tissues obtained from surgical resections were stained
with hematoxylin–eosin. Tissues were classified and
graded according to the 2007 WHO criteria [1] with stand-
ard hematoxylin–eosin staining. Gliomas were classified
according to grade (II or III) and one of the following three
categories: astrocytoma, mixed glioma, and oligodendro-
glioma (Fig.1), though mixed gliomas were removed from
the statistical tests because of their vague label. (Grade
II: 15 astrocytomas, 10 mixed gliomas, 6 oligodendro-
gliomas; Grade III: 14 astrocytomas, 4 mixed gliomas, 16
oligodendrogliomas).
IDH1 mutation status and1p/19q co‑deletion (2016
WHO glioma guidelines)
IDH1 mutation status was determined by sequencing for
codon 132 in the catalytic domain of IDH1 via stand-
ard genomic sequencing practices (Sanger sequencing

J Neurooncol
1 3
method), as previously described [31]. In brief, tumor
DNA was isolated from the frozen or formalin-fixed tissue
using DNeasy Blood and Tissue Kit (Qiagen). A 236-bp
fragment that included codon 132 was amplified using the
primers 5-GCG TCA AAT GTG CCA CTA TC-3 and 5-GCA
AAA TCA CAT TAT TGC CAAC-3 to generate a 236 bp
fragment. PCR products were sequenced by BigDye Ter-
minator v1.1 (Applied Biosystems), and sequences were
determined via a 3730 sequencer (Applied Biosystems).
1p/19q co-deletion status was determined by fluores-
cence insitu hybridization specific probes for the 1p36 and
19q13 loci. A deletion of >50% of the nuclei examined for
both 1p and 19q constituted a co-deletion.
The gliomas were separated according to grade and
one of the following categories (Fig. 1): IDH1 wild-
type (IDH
WT
); IDH1 mutant with intact 1p or 19q
(IDH
MUT
/1p19q
+
); or IDH1 mutant with 1p/19q co-
deleted (IDH
MUT
/1p19q
). (Grade II: 8 IDH
WT
, 16
IDH
MUT
/1p19q
+
, 7 IDH
MUT
/1p19q
; Grade III: 14 IDH
WT
,
12 IDH
MUT
/1p19q
+
, 8 IDH
MUT
/1p19q
).
Statistical analyses
Due to the small sample size, bootstrap hypothesis testing
was used for all statistical tests. First, using the real data,
the f-score was calculated from the one-way ANOVA test
if three groups were being compared or a t-score if only
two groups were being compared to each other. In order to
generate the null hypothesis distribution, the mean of the
data for each group was subtracted from their respective
groups because the null hypothesis was that the means of
all groups are equal. Next, the mean-shifted data was ran-
domly resampled with replacement such that the newly res-
ampled data had the same number of data points as the real
data for each group. The f-score or t-test was then calcu-
lated from the resampled, mean-shifted data 10,000 times
to generate the null hypothesis distribution. The p value
was then computed by counting the number of f-scores
greater than the f-score generated from the real data or the
number of t-scores that were more extreme than the t-score
generated from the real data.
For analyses pertaining to both the 2007 and the 2016
WHO glioma classification, all three categories were used
in the statistical comparison. All bootstrap routines were
performed in MATLAB using in-house custom code,
while the receiver operating characteristic statistics, i.e.
area under the curve, were performed in GraphPad Prism
(La Jolla, CA) using the sensitivity and specificity from the
highest likelihood ratio.
A logistic regression was performed to classify IDH
MUT
and IDH
WT
gliomas (both grade II and grade III tumors)
using median rCBV, median ADC, presence or absence
of contrast enhancement, and volume of T2-enhancement
using in-house custom MATLAB code. Sensitivity and
specificity were picked based on the point on the ROC
where the product was maximized. To generate the p val-
ues for the presence or absence of contrast enhancement,
a Fisher’s exact test was employed. Since eight statistical
tests comparing the glioma subtypes were performed (four
for the 2007 analysis, four for the 2016 analysis, two tumor
volume analyses, two presence of contrast enhancement
analyses, and two receiver operating characteristic curve
analyses), for multiple comparisons correction, a p value of
0.015/14 = 0.0036 was considered to be significant.
The elliptical error bounds in Figs.3 and 4 were cal-
culated using custom code in MATLAB. The major and
minor axes were determined by the eigenvectors, and the
lengths of the axes represent one standard deviation in
Fig. 1 Diagram illustrating the 2007 and 2016 WHO classification
criteria for gliomas. (Top) Categories of grade II and III gliomas
under the 2007 WHO criteria based on histological features using
light microscopy and hematoxylin and eosin staining. (Bottom) Cate-
gories of grade II and III gliomas under the 2016 WHO criteria based
on molecular genotype using IDH1 mutation status and 1p/19q co-
deletion

Citations
More filters
Journal ArticleDOI
TL;DR: Clinical best practice recommendations for glioma imaging assessment are proposed and the current role of advanced MRI modalities in routine use is addressed.
Abstract: Objectives: At a European Society of Neuroradiology (ESNR) Annual Meeting 2015 workshop, commonalities in practice, current controversies and technical hurdles in glioma MRI were discussed. We aimed to formulate guidance on MRI of glioma and determine its feasibility, by seeking information on glioma imaging practices from the European Neuroradiology community. Methods: Invitations to a structured survey were emailed to ESNR members (n=1,662) and associates (n=6,400), European national radiologists’ societies and distributed via social media. Results: Responses were received from 220 institutions (59% academic). Conventional imaging protocols generally include T2w, T2-FLAIR, DWI, and pre- and post-contrast T1w. Perfusion MRI is used widely (85.5%), while spectroscopy seems reserved for specific indications. Reasons for omitting advanced imaging modalities include lack of facility/software, time constraints and no requests. Early postoperative MRI is routinely carried out by 74% within 24–72 h, but only 17% report a percent measure of resection. For follow-up, most sites (60%) issue qualitative reports, while 27% report an assessment according to the RANO criteria. A minority of sites use a reporting template (23%). Conclusion: Clinical best practice recommendations for glioma imaging assessment are proposed and the current role of advanced MRI modalities in routine use is addressed. Key Points: • We recommend the EORTC-NBTS protocol as the clinical standard glioma protocol.• Perfusion MRI is recommended for diagnosis and follow-up of glioma.• Use of advanced imaging could be promoted with increased education activities.• Most response assessment is currently performed qualitatively.• Reporting templates are not widely used, and could facilitate standardisation.

128 citations


Cites background from "Perfusion and diffusion MRI signatu..."

  • ...could provide greater information on tissue microstructure for the distinction of glioma molecular subgroups [25, 26, 83] and to support early response assessment, e....

    [...]

Journal ArticleDOI
TL;DR: Relaxation-compensated multipool CEST MRI, particularly dns-APT imaging, enabled prediction of IDH mutation status and differentiation of LGG versus HGG and should therefore be considered as a non-invasive MR biomarker in the diagnostic workup.
Abstract: Background Early identification of prognostic superior characteristics in glioma patients such as isocitrate dehydrogenase (IDH) mutation and O6-methylguanine-DNA-methyltransferase (MGMT) promoter methylation status is of great clinical importance. The study purpose was to investigate the non-invasive predictability of IDH mutation status, MGMT promoter methylation, and differentiation of low-grade versus high-grade glioma (LGG vs HGG) in newly diagnosed patients employing relaxation-compensated multipool chemical exchange saturation transfer (CEST) MRI at 7.0 Tesla. Methods Thirty-one patients with newly diagnosed glioma were included in this prospective study. CEST MRI was performed at a 7T whole-body scanner. Nuclear Overhauser effect (NOE) and isolated amide proton transfer (APT; downfield NOE-suppressed APT = dns-APT) CEST signals (mean value and 90th signal percentile) were quantitatively investigated in the whole tumor area with regard to predictability of IDH mutation, MGMT promoter methylation status, and differentiation of LGG versus HGG. Statistics were performed using receiver operating characteristic (ROC) and area under the curve (AUC) analysis. Results were compared with advanced MRI methods (apparent diffusion coefficient and relative cerebral blood volume ROC/AUC analysis) obtained at 3T. Results dns-APT CEST yielded highest AUCs in IDH mutation status prediction (dns-APTmean = 91.84%, P 0.05). Conclusions Relaxation-compensated multipool CEST MRI, particularly dns-APT imaging, enabled prediction of IDH mutation status and differentiation of LGG versus HGG and should therefore be considered as a non-invasive MR biomarker in the diagnostic workup.

97 citations

Journal ArticleDOI
TL;DR: A feasible radiopharmacodynamics approach to support the rapid clinical translation of rationally designed drugs targeting IDH1/2 mutations for personalized and precision medicine of glioma patients is demonstrated.
Abstract: Inhibitors of the mutant isocitrate dehydrogenase 1 (IDH1) entered recently in clinical trials for glioma treatment. Mutant IDH1 produces high levels of 2-hydroxyglurate (2HG), thought to initiate oncogenesis through epigenetic modifications of gene expression. In this study, we show the initial evidence of the pharmacodynamics of a new mutant IDH1 inhibitor in glioma patients, using non-invasive 3D MR spectroscopic imaging of 2HG. Our results from a Phase 1 clinical trial indicate a rapid decrease of 2HG levels by 70% (CI 13%, P = 0.019) after 1 week of treatment. Importantly, inhibition of mutant IDH1 may lead to the reprogramming of tumor metabolism, suggested by simultaneous changes in glutathione, glutamine, glutamate, and lactate. An inverse correlation between metabolic changes and diffusion MRI indicates an effect on the tumor-cell density. We demonstrate a feasible radiopharmacodynamics approach to support the rapid clinical translation of rationally designed drugs targeting IDH1/2 mutations for personalized and precision medicine of glioma patients.

89 citations

Journal ArticleDOI
TL;DR: MRI showed the potential to non-invasively predict IDH mutation in patients with glioma and 2-Hydroxyglutarate MRS shows higher pooled sensitivity than other imaging modalities.
Abstract: To evaluate the imaging features of isocitrate dehydrogenase (IDH) mutant glioma and to assess the diagnostic performance of magnetic resonance imaging (MRI) for prediction of IDH mutation in patients with glioma. A systematic search of Ovid-MEDLINE and EMBASE up to 10 October 2017 was conducted to find relevant studies. The search terms combined synonyms for ‘glioma’, ‘IDH mutation’ and ‘MRI’. Studies evaluating the imaging features of IDH mutant glioma and the diagnostic performance of MRI for prediction of IDH mutation in patients with glioma were selected. The pooled summary estimates of sensitivity and specificity and their 95% confidence intervals (CIs) were calculated using a bivariate random-effects model. The results of multiple subgroup analyses are reported. Twenty-eight original articles in a total of 2,146 patients with glioma were included. IDH mutant glioma showed frontal lobe predominance, less contrast enhancement, well-defined border, high apparent diffusion coefficient (ADC) value and low relative cerebral blood volume (rCBV) value. For the meta-analysis that included 18 original articles, the summary sensitivity was 86% (95% CI, 79%–91%) and the summary specificity was 87% (95% CI, 78–92%). In a subgroup analysis, the summary sensitivity of 2-hydroxyglutarate magnetic resonance spectroscopy (MRS) [96% (95% CI, 91–100%)] was higher than the summary sensitivities of other imaging modalities. IDH mutant glioma consistently demonstrated less aggressive imaging features than IDH wild-type glioma. Despite the variety of different MRI techniques used, MRI showed the potential to non-invasively predict IDH mutation in patients with glioma. 2-Hydroxyglutarate MRS shows higher pooled sensitivity than other imaging modalities. • IDH mutant glioma showed frontal lobe predominance, less contrast enhancement, well-defined border, high ADC value, and low rCBV value. • The diagnostic performance of MRI for prediction of IDH mutation in patients with glioma is within a clinically acceptable range, the summary sensitivity was 86% (95% CI, 79–91%) and the summary specificity was 87% (95% CI, 78–92%). • In a subgroup analysis, the summary sensitivity of 2-hydroxyglutarate MRS [96% (95% CI, 91–100%)] was higher than the summary sensitivities of other imaging modalities.

79 citations

Journal ArticleDOI
TL;DR: 2HG MRS demonstrated excellent specificity for prediction of IDH mutant glioma, with TE being associated with heterogeneity in the sensitivity, and Echo time was associated with studyogeneity in the meta-regression.
Abstract: Background Noninvasive and accurate modality to predict isocitrate dehydrogenase (IDH) mutant glioma may have great potential in routine clinical practice. We aimed to investigate the diagnostic performance of 2-hydroxyglutarate (2HG) magnetic resonance spectroscopy (MRS) for prediction of IDH mutant glioma and provide an optimal cutoff value for 2HG. Methods A systematic literature search of Ovid-MEDLINE and EMBASE was performed to identify original articles investigating the diagnostic performance of 2HG MRS up to March 20, 2018. Pooled sensitivity and specificity were calculated using a bivariate random-effects model. Subgroup analysis and meta-regression were performed to explain heterogeneity effects. An optimal cutoff value for 2HG was calculated from studies providing individual patient data. Results Fourteen original articles with 460 patients were included. The pooled sensitivity and specificity for the diagnostic performance of 2HG MRS for prediction of IDH mutant glioma were 95% (95% CI, 85-98%) and 91% (95% CI, 83-96%), respectively. The Higgins I2 statistic demonstrated that heterogeneity was present in the sensitivity (I2 = 50.69%), but not in the specificity (I2 = 30.37%). In the meta-regression, echo time (TE) was associated with study heterogeneity. Among the studies using point-resolved spectroscopy (PRESS), a long TE (97 ms) resulted in higher sensitivity (92%) and specificity (97%) than a short TE (30-35 ms; sensitivity of 90%, specificity of 88%; P < 0.01). The optimal 2HG cutoff value of 2HG using individual patient data was 1.76 mM. Conclusion 2HG MRS demonstrated excellent specificity for prediction of IDH mutant glioma, with TE being associated with heterogeneity in the sensitivity. Key Points 1. HG MRS has excellent diagnostic performance in the prediction of IDH mutant glioma. 2. The pooled sensitivity was 95% and the pooled specificity was 91%. 3. Echo time was associated with study heterogeneity in the meta-regression.

78 citations

References
More filters
Journal ArticleDOI
TL;DR: The fourth edition of the World Health Organization (WHO) classification of tumours of the central nervous system, published in 2007, lists several new entities, including angiocentric glioma, papillary glioneuronal tumour, rosette-forming glioneurs tumour of the fourth ventricle, Papillary tumourof the pineal region, pituicytoma and spindle cell oncocytoma of the adenohypophysis.
Abstract: The fourth edition of the World Health Organization (WHO) classification of tumours of the central nervous system, published in 2007, lists several new entities, including angiocentric glioma, papillary glioneuronal tumour, rosette-forming glioneuronal tumour of the fourth ventricle, papillary tumour of the pineal region, pituicytoma and spindle cell oncocytoma of the adenohypophysis. Histological variants were added if there was evidence of a different age distribution, location, genetic profile or clinical behaviour; these included pilomyxoid astrocytoma, anaplastic medulloblastoma and medulloblastoma with extensive nodularity. The WHO grading scheme and the sections on genetic profiles were updated and the rhabdoid tumour predisposition syndrome was added to the list of familial tumour syndromes typically involving the nervous system. As in the previous, 2000 edition of the WHO ‘Blue Book’, the classification is accompanied by a concise commentary on clinico-pathological characteristics of each tumour type. The 2007 WHO classification is based on the consensus of an international Working Group of 25 pathologists and geneticists, as well as contributions from more than 70 international experts overall, and is presented as the standard for the definition of brain tumours to the clinical oncology and cancer research communities world-wide.

13,134 citations

Journal ArticleDOI
TL;DR: The 2016 World Health Organization Classification of Tumors of the Central Nervous System is both a conceptual and practical advance over its 2007 predecessor and is hoped that it will facilitate clinical, experimental and epidemiological studies that will lead to improvements in the lives of patients with brain tumors.
Abstract: The 2016 World Health Organization Classification of Tumors of the Central Nervous System is both a conceptual and practical advance over its 2007 predecessor. For the first time, the WHO classification of CNS tumors uses molecular parameters in addition to histology to define many tumor entities, thus formulating a concept for how CNS tumor diagnoses should be structured in the molecular era. As such, the 2016 CNS WHO presents major restructuring of the diffuse gliomas, medulloblastomas and other embryonal tumors, and incorporates new entities that are defined by both histology and molecular features, including glioblastoma, IDH-wildtype and glioblastoma, IDH-mutant; diffuse midline glioma, H3 K27M-mutant; RELA fusion-positive ependymoma; medulloblastoma, WNT-activated and medulloblastoma, SHH-activated; and embryonal tumour with multilayered rosettes, C19MC-altered. The 2016 edition has added newly recognized neoplasms, and has deleted some entities, variants and patterns that no longer have diagnostic and/or biological relevance. Other notable changes include the addition of brain invasion as a criterion for atypical meningioma and the introduction of a soft tissue-type grading system for the now combined entity of solitary fibrous tumor / hemangiopericytoma-a departure from the manner by which other CNS tumors are graded. Overall, it is hoped that the 2016 CNS WHO will facilitate clinical, experimental and epidemiological studies that will lead to improvements in the lives of patients with brain tumors.

11,197 citations


"Perfusion and diffusion MRI signatu..." refers background or methods in this paper

  • ...We hypothesize a combination of diffusion and perfusion MR measurements will better separate tumors based on their genetic characteristics (1p19q co-deletion and IDH1 mutation status) [7] than the more subjective histologic criteria....

    [...]

  • ...In 2016, the WHO glioma classification scheme was restructured to reflect two common molecular alterations in gliomas: 1p/19q codeletion (1p19q−) and isocitrate dehydrogenase-1 mutation (IDHMUT) [7]....

    [...]

Journal ArticleDOI
TL;DR: The Central Brain Tumor Registry of the United States (CBTRUS), in collaboration with the Centers for Disease Control and Prevention and National Cancer Institute, is the largest population-based registry focused exclusively on primary brain and other central nervous system (CNS) tumors in the US.
Abstract: The Central Brain Tumor Registry of the United States (CBTRUS), in collaboration with the Centers for Disease Control (CDC) and National Cancer Institute (NCI), is the largest population-based registry focused exclusively on primary brain and other central nervous system (CNS) tumors in the United States (US) and represents the entire US population. This report contains the most up-to-date population-based data on primary brain tumors (malignant and non-malignant) and supersedes all previous CBTRUS reports in terms of completeness and accuracy. All rates (incidence and mortality) are age-adjusted using the 2000 US standard population and presented per 100,000 population. The average annual age-adjusted incidence rate (AAAIR) of all malignant and non-malignant brain and other CNS tumors was 23.79 (Malignant AAAIR=7.08, non-Malignant AAAIR=16.71). This rate was higher in females compared to males (26.31 versus 21.09), Blacks compared to Whites (23.88 versus 23.83), and non-Hispanics compared to Hispanics (24.23 versus 21.48). The most commonly occurring malignant brain and other CNS tumor was glioblastoma (14.5% of all tumors), and the most common non-malignant tumor was meningioma (38.3% of all tumors). Glioblastoma was more common in males, and meningioma was more common in females. In children and adolescents (age 0-19 years), the incidence rate of all primary brain and other CNS tumors was 6.14. An estimated 83,830 new cases of malignant and non-malignant brain and other CNS tumors are expected to be diagnosed in the US in 2020 (24,970 malignant and 58,860 non-malignant). There were 81,246 deaths attributed to malignant brain and other CNS tumors between 2013 and 2017. This represents an average annual mortality rate of 4.42. The 5-year relative survival rate following diagnosis of a malignant brain and other CNS tumor was 23.5% and for a non-malignant brain and other CNS tumor was 82.4%.

9,802 citations

Journal ArticleDOI
TL;DR: Mutations of NADP(+)-dependent isocitrate dehydrogenases encoded by IDH1 and IDH2 occur in a majority of several types of malignant gliomas.
Abstract: Background A recent genomewide mutational analysis of glioblastomas (World Health Organization [WHO] grade IV glioma) revealed somatic mutations of the isocitrate dehydrogenase 1 gene (IDH1) in a fraction of such tumors, most frequently in tumors that were known to have evolved from lower-grade gliomas (secondary glioblastomas). Methods We determined the sequence of the IDH1 gene and the related IDH2 gene in 445 central nervous system (CNS) tumors and 494 non-CNS tumors. The enzymatic activity of the proteins that were produced from normal and mutant IDH1 and IDH2 genes was determined in cultured glioma cells that were transfected with these genes. Results We identified mutations that affected amino acid 132 of IDH1 in more than 70% of WHO grade II and III astrocytomas and oligodendrogliomas and in glioblastomas that developed from these lower-grade lesions. Tumors without mutations in IDH1 often had mutations affecting the analogous amino acid (R172) of the IDH2 gene. Tumors with IDH1 or IDH2 mutations h...

4,853 citations

Journal ArticleDOI
TL;DR: The authors found that approximately 5% of patients with malignant gliomas have a family history of glioma and most of these familial cases are associated with rare genetic syndromes, such as neurofibromatosis types 1 and 2, the Li−Fraumeni syndrome (germ-line p53 mutations associated with an increased risk of several cancers), and Turcot's syndrome (intestinal polyposis and brain tumors).
Abstract: Approximately 5% of patients with malignant gliomas have a family history of gliomas. Some of these familial cases are associated with rare genetic syndromes, such as neurofibromatosis types 1 and 2, the Li−Fraumeni syndrome (germ-line p53 mutations associated with an increased risk of several cancers), and Turcot’s syndrome (intestinal polyposis and brain tumors). 10 However, most familial cases have

3,823 citations

Related Papers (5)
Frequently Asked Questions (2)
Q1. What contributions have the authors mentioned in the paper "Perfusion and diffusion mri signatures in histologic and genetic subtypes of who grade ii–iii diffuse gliomas" ?

In this paper, the authors used histological appearance on light microscopy to distinguish gliomas of different glial lineages. 

Future studies aimed at testing these specific hypotheses through serial imaging in tumors are warranted to further test how shifts in rCBV and ADC may reflect increasing tumor malignancy within a certain genotype at a time in Fig.