15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.

com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
CME
JOURNAL OF MAGNETIC RESONANCE IMAGING 37:351–358 (2013)

Original Research

Correlation of Apparent Diffusion Coefficient Values

Measured by Diffusion MRI and MGMT Promoter
Methylation Semiquantitatively Analyzed With
MS-MLPA in Patients With Glioblastoma Multiforme
Leonard Sunwoo, MD,1 Seung Hong Choi, MD,1* Chul-Kee Park, MD,2*
Jin Wook Kim, MD,2 Kyung Sik Yi, MD,1 Woong Jae Lee, MD,1 Tae Jin Yoon, MD,1
Sang Woo Song, MD,2 Ja Eun Kim, MS,2 Ji Young Kim, MS,2 Tae Min Kim, MD,3
Se-Hoon Lee, MD, PhD,3 Ji-Hoon Kim, MD,1 Chul-Ho Sohn, MD,1
Sung-Hye Park, MD,4 Il Han Kim, MD,5 and Kee-Hyun Chang, MD1

Purpose: To retrospectively determine whether the appa- survival (PFS) was also correlated with the ADC values
rent diffusion coefficient (ADC) values correlate with O6- using Kaplan–Meier survival analysis.
methylguanine DNA methyltransferase (MGMT) promoter
Results: The mean methylation ratio was 0.21 6 0.20. By
methylation semiquantitatively analyzed by methylation-
MSP, there were 5 methylated and 21 unmethylated
specific multiplex ligation-dependent probe amplification
tumors. The mean ADC revealed a positive relationship
(MS-MLPA) in patients with glioblastoma.
with MGMT promoter methylation ratio (P ¼ 0.015) and
Materials and Methods: The study was approved by the was also significantly different according to MSP-deter-
Institutional Review Board and was Health Insurance mined methylation status (P ¼ 0.011). Median PFS was
Portability and Accountability Act (HIPAA) compliant. significantly related with methylation ratio (P ¼ 0.017)
Newly diagnosed patients with glioblastoma (n ¼ 26) were and MSP-derived methylation status (P ¼ 0.025). A posi-
analyzed with an ADC histogram approach based on tive relationship was demonstrated between PFS and the
enhancing solid portion. The methylation status of MGMT mean ADC value (P ¼ 0.001). The 5th percentile ADC val-
promoter was assessed by methylation-specific polymer- ues showed a significant negative relationship with Ki-67
ase chain reaction (MSP) and by MS-MLPA. MS-MLPA is a labeling index (P ¼ 0.036).
semiquantitative method that determines the methylation
Conclusion: We found that ADC values were significantly
ratio. The Ki-67 labeling index was also analyzed. The
correlated with PFS as well as with MGMT promoter
mean and 5th percentile ADC values were correlated with
methylation status. We believe that ADC values may merit
MGMT promoter methylation status and Ki-67 labeling
further investigation as a noninvasive biomarker for pre-
index using a linear regression model. Progression-free
dicting treatment response.
Key Words: apparent diffusion coefficient (ADC); O6-
1
methylguanine DNA methyltransferase (MGMT); methyla-
Department of Radiology, Seoul National University College of tion-specific multiplex ligation-dependent probe amplifi-
Medicine, Seoul, Korea.
2 cation (MS-MLPA); methylation ratio; glioblastoma
Department of Neurosurgery, Seoul National University College of
Medicine, Seoul, Korea. J. Magn. Reson. Imaging 2013;37:351–358.
3 C 2012 Wiley Periodicals, Inc.
V
Department of Internal Medicine, Cancer Research Institute, Seoul
National University College of Medicine, Seoul, Korea.
4
Department of Pathology, Seoul National University College of
Medicine, Seoul, Korea.
5
Department of Radiation Oncology, Cancer Research Institute, Seoul FOR PATIENTS WITH GLIOBLASTOMA MULTIFORME
National University College of Medicine, Seoul, Korea.
(GBM), the most common brain tumor in adults,
Contract grant sponsor: National R&D Program for Cancer Control,
Ministry of Health & Welfare, Republic of Korea; Contract grant the prognosis remains dismal in spite of the multidis-
number: 1120300; Contract grant sponsor: Korea Healthcare ciplinary treatments now used (1–3). Currently, the
Technology R&D Projects, Ministry for Health, Welfare & Family
Affairs; Contract grant number: A112028.
standard treatment includes maximal safe tumor
*Address reprint requests to: S.H.C., Department of Radiology, Seoul resection, followed by a DNA alkylating agent temozo-
National University College of Medicine, 28, Yongon-dong, Jongno-gu, lomide (TMZ) in combination with radiotherapy.
Seoul, 110-744, Korea. E-mail: verocay@snuh.org or C.K.P, Recent studies have shown that the methylation
nsckpark@paran.com
Received February 15, 2012; Accepted August 24, 2012.
of the O6-methylguanine DNA methyltransferase
DOI 10.1002/jmri.23838 (MGMT) gene promoter has been associated with a
View this article online at wileyonlinelibrary.com. longer survival in patients treated with radiotherapy
C 2012 Wiley Periodicals, Inc.
V 351

352
as well as with concurrent and adjuvant TMZ therapy
(4–7). In addition, researchers have proposed that
the methylation status of the MGMT promoter can
help differentiate true progression from pseudoprog-
ression (8,9).
Methylation-specific multiplex ligation-dependent
probe amplification (MS-MLPA) is a semiquantitative
method that has recently been introduced for testing
the methylation status of multiple genes simultane-
ously. MS-MLPA has also been shown to be a reliable
method for the determination of the MGMT promoter
methylation in glioma patients (10). MS-MLPA has
also been shown to be more accurate than methyla-
tion-specific polymerase chain reaction (MSP) (80%
vs. 68%) in reference to determining radiological pro-
gression after chemoradiotherapy (11).
An accurate measurement of MGMT promoter meth-
ylation is often difficult to obtain because of the small
biopsy specimens and the regional heterogeneity of
GBM (12). Therefore, a noninvasive surrogate marker
for MGMT promoter methylation would be useful in
the overall analysis of tumors. Magnetic resonance
imaging (MRI) is routinely used to diagnose, charac-
terize, and clinically manage GBM (13). It is a power-
ful and noninvasive diagnostic imaging tool that
provides global assessment of GBMs and the interac-
tion of these tumors with their local environments. In
its ability to extract structural, compositional, physio-
logical, and functional information, MRI captures the
multidimensional, in vivo portraits of GBMs (14).
Diffusion-weighted (DW)-MRI is especially proficient
at providing information concerning extracellular-
space tortuosity, tissue cellularity, and the integrity of
cellular membranes. In addition, DW-MRI can deter-
mine lesion aggressiveness (ie, the assessment of
cellularity), predict suitable therapies and monitor the
response to conventional and novel therapies (ie, the
assessment of cell membrane integrity) (15). To our
knowledge, there has been no attempt to correlate the
semiquantitative methylation status with MRI fea-
tures. In our study we correlated the ADC values
measured by DW-MRI with the semiquantitative
methylation status of the MGMT promoter that was
obtained with MS-MLPA. We also evaluated whether
the ADC values could be correlated with the clinical
outcome of patients with GBM.
MATERIALS AND METHODS
This study was approved by our Institutional Review
Board. Informed consent was waived.
Study Population
Sixty-five adult patients with newly diagnosed GBM
who had undergone initial MRI in our institution
between March 2008 and December 2010 were en-
rolled in this retrospective study. Inclusion criteria
were as follows: 1) a histopathological diagnosis of
GBM according to the World Health Organization
(WHO) criteria and 2) performance of preoperative
MRI including DW imaging at the standard b values
15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Sunwoo et al.
Figure 1. The flow diagram of the patient selection with
inclusion and exclusion criteria for the study. GBM ¼ glio-
blastoma multiforme.
(b ¼ 1000 sec/mm2). We excluded 39 patients
because of the following conditions: i) MRI obtained at
3 T (n ¼ 19), ii) DW imaging obtained with inadequate
MRI (n ¼ 3) and prior treatment (n ¼ 7), and iii) an
insufficient amount of DNA obtained to perform anal-
ysis (n ¼ 10). The patients who had undergone MRI
with a 3 T scanner were excluded to eliminate any
bias associated with the combination of data from MR
scanners with different field strengths. As a result, a
total of 26 patients (15 males, 11 females; mean age,
52.8 years) were included (Fig. 1).
Imaging
All the participants underwent MRI within 2 weeks
before surgery. All MR images were obtained with a
1.5 T MR imager (Signa HDx or HDxt; GE Medical
Systems, Milwaukee, WI) with an eight-channel head
coil. The imaging protocol included axial T2-weighted
fast spin-echo (FSE) sequence with TR/TE of 5000/
131 (msec), 25 sections, a 5-mm section thickness, a
1-mm intersection gap, a field of view of 220
220
mm, a matrix of 448
256, one acquired signal, an
echo train length of 16, and a voxel resolution of 0.5

0.9
5.0 mm. An axial T1-weighted spin-echo (SE)
sequence was performed as follows: TR/TE, 466/11;
section thickness, 5 mm; intersection gap, 1 mm; field
of view, 220
220 mm; matrix, 320
192; voxel
resolution, 0.7
1.1
5 mm.
Echo-planar DW imaging was performed in the axial
plane before the injection of the contrast material
with TR/TE of 6000/63 (at b ¼ 0 and 1000 sec/
mm2), 25 sections, a bandwidth of 1953 Hz/voxel, a
5-mm section thickness, a 1-mm intersection gap, a
field of view of 240
240 mm, a matrix of 160
160,
two acquired signals, and a voxel resolution of 1.5

1.5
5.0 mm. DW images were acquired in three or-
thogonal directions and combined into a trace image.
Using these data, ADC maps at standard b values
(b ¼ 1000 sec/mm2) were calculated on a voxel-by-voxel
basis with the software incorporated in the MRI unit.

Correlation of ADC, MGMT Methylation in GBM
Figure 2. MRI of a 73-year-old man with a GBM at the left temporal lobe (arrows) showing how the regions of interest (ROIs)
were drawn on the ADC maps with reference to the postcontrast T1-weighted images. a: Axial T2-weighted FSE image (5000/
131). b: Axial postcontrast T1-weighted image (667/21). c: The ADC map calculated from DWI (b ¼ 1000 sec/mm2). Areas
showing necrosis or peritumoral edema were excluded from the ROI.
Axial T1-weighted sequences were repeated after the
intravenous administration of a single dose of 0.1
mmol/kg gadopentetate dimeglumine (Magnevist;
Bayer Schering Pharma, Germany). A fat suppression
pulse was added to the axial T1-weighted SE
sequence after the administration of the contrast
agent. Depending on the location of the primary tu-
mor, coronal and/or sagittal T1-weighted SE sequen-
ces without fat suppression were performed with
identical imaging parameters after administration of
the contrast agent.
After completion of radiotherapy with concurrent TMZ
therapy, follow-up MRI was performed every 2 or 3
months. The imaging protocol included axial T2-weighted
FSE, fluid-attenuated inversion recovery (FLAIR), and
pre- and postcontrast T1-weighted SE sequences.
Follow-up and Response Assessment
Four patients were lost to follow-up immediately after
the surgery. The remaining 22 patients were followed,
with a median follow-up of 8.8 months (range, 2–31
months). The patients were assessed clinically and
received follow-up MRI. The treatment responses were
evaluated using the Response Assessment in Neuro-
Oncology (RANO) criteria (16). All but one patient,
who showed early clinical deterioration, underwent
TMZ and concurrent radiotherapy.
Volume Acquisition/ADC Histograms
The MR data for the ADC maps were digitally trans-
ferred from the PACS workstation to a personal com-
puter and processed with ImageJ (available at http://
rsb.info.nih.gov/ij/) (17) and a software program
developed in-house using Microsoft Visual Cþþ.
Enhancing tumor volumes were manually seg-
mented on postcontrast T1-weighted images on
presurgical scans by carefully defining the region of
interest at the tumor margin. Definite areas of cystic,
necrotic, or hemorrhagic areas were excluded. The
15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
353
data acquired from each slice were summated to
derive voxel-by-voxel ADC values for the entire tumor
using the software developed in-house (Fig. 2).
ADC histograms were plotted with ADC values on
the x-axis with a bin size of 1
106 mm2/sec, and
the percentage of the total lesion volume calculated
by dividing the frequency in each bin by the total
number of voxels analyzed on the y-axis. We also per-
formed a cumulative analysis with the ADC histo-
grams, in which the cumulative number of observa-
tions in all of the bins up to the specified bin was
mapped onto the y-axis expressed as a percentage.
For the cumulative ADC histograms, the mean ADC
and the 5th percentile values (the point at which 5%
of the voxel values that form the histogram are found
to the left in the histogram) were generated (18,19).
Assessment of MGMT Promoter Methylation Status
The methylation status of the MGMT promoter was
assessed using MS-MLPA. The procedure was performed
according to the manufacturer’s protocol and has been
described in the literature (11). Briefly, the methylation
status was tested using the MS-MLPA probe mix pre-
pared by MRC-Holland (Salsa MS-MLPA Kit ME011
MMR, Amsterdam, The Netherlands). The MS-MLPA
probe mix included three probes specific for the MGMT
promoter regions that contained HhaI recognition sites.
The resultant polymerase chain reaction (PCR) frag-
ments were separated by capillary gel electrophoresis.
The methylation status was quantified using Gene-
Marker software (v. 1.5, Soft Genetics). To compensate
for the differences in the efficiency of the PCR for the
individual samples, the peak value of each probe was
normalized by dividing it by the peak of the control
probes. To evaluate the methylation status, the meth-
ylation ratio was calculated by the average of dividing
each normalized peak value of the digested sample by
that of the corresponding undigested sample. This
value corresponded to the percentage of methylated
sequences.

354
The MGMT promoter methylation status was also
analyzed by MSP using standard protocols. The
prepared DNA was modified with a sodium bisulfite
treatment. The annealing temperature was 64 C. The
obtained PCR products were electrophoresed in 2%
agarose gels and visualized under ultraviolet illumina-
tion after staining with ethidium bromide.
Histopathological Analysis
Immunohistochemistry was used to measure the Ki-
67 labeling index. The routinely used formalin-fixed,
paraffin-embedded tissue blocks were sectioned at
4-mm thickness and then used for immunohistochem-
istry. The areas with the highest cellularity on inspec-
tion were selected and the Ki-67 labeling index was
evaluated by the avidin-biotin complex immunohisto-
chemical technique (20).
Statistical Methods
With a Pearson linear regression model, the mean
ADC and the 5th percentile ADC values described
above were correlated with the methylation status of
the MGMT promoter and the Ki-67 labeling index. To
exclude possible interference from the age and gender
factors, age and gender were correlated with the mean
ADC values and the methylation ratio, respectively.
The age factor was analyzed using linear regression
model, and to evaluate the gender factor the patients
were divided into male and female groups and inde-
pendent Student’s t-test was used to compare the
mean values of the mean ADC and the methylation ra-
tio for each group. Progression-free survival (PFS) was
measured from the time of the operation to disease
progression or to the date of the last follow-up and was
analyzed with either a linear regression model or the
Kaplan–Meier method. The 95% confidence intervals
were calculated using the associated estimated stand-
ard errors. Student’s t-test was used to compare the
mean ADC values associated with the MSP-derived
methylation status. Probability values of less than
0.05 were considered statistically significant. These
statistical analyses were performed with the aid of
MedCalc (v. 12.1.0.0 for Microsoft Windows 2000/XP/
Vista/7, MedCalc Software, Mariakerke, Belgium).
RESULTS
The overall mean ADC and the 5th percentile values
were 1130.5
106 mm2/sec 6 191 (standard devia-
tion) and 740.2
106 mm2/sec 6 141, respectively.
The calculated methylation ratio ranged from 0–0.92
(mean, 0.21). The findings from MSP showed that
there were 5 methylated tumors and 21 unmethylated
tumors. The median PFS was 4.0 months (range, 0.5–
31 months). Patients’ age was not significantly related
to the mean ADC values or the methylation ratio (P ¼
0.294 and P ¼ 0.989, respectively). The mean of the
mean ADC values and that of the methylation ratio
between the male and female groups were also not
significantly different (P ¼ 0.544 and P ¼ 0.124,
respectively). The baseline demographic data are sum-
marized in Table 1.
15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Sunwoo et al.
Table 1
Baseline Demographic Data of the Patients
Total (n¼26; male¼15,
female¼11)
Age (years) 52.8 6 11.7
Age range (years) 32-73
Median PFS (months) 4.0 6 8.3
Mean ADC (10-6 mm2/sec) 1130.5 6 191.1
5th percentile (10-6 mm2/sec) 740.3 6 140.7
Methylation ratio 0.21 6 0.20
MSP Methylated, 5;
unmethylated, 21
Ki-67 (%) 17.4 6 9.9
The values represent the mean 6 standard deviation. The appa-
rent diffusion coefficient (ADC) measurements are expressed in

10-6 square millimeters per second. PFS ¼ progression-free sur-
vival, MSP ¼ methylation-specific polymerase chain reaction.
The mean ADC values of the enhancing tumor por-
tion revealed a positive relationship with the MGMT
promoter methylation ratio (Figs. 3 and 4, R2 ¼ 0.22,
P ¼ 0.015). However, the 5th percentile ADC values
failed to show a statistically significant relationship
with methylation ratio (R2 ¼ 0.04, P ¼ 0.32). Accord-
ing to the MSP results, methylated tumors had lower
mean ADC values: 1319.4 vs. 1085.5
106 mm2/
sec (Fig. 5, P ¼ 0.011). The mean of the 5th percentile
ADC values for methylated and unmethylated tumors
was not quite significantly different: 844.8 vs. 715.4

106 mm2/sec (P ¼ 0.063).
We found a positive relationship between PFS and
the mean ADC values (Fig. 6a, R2 ¼ 0.44, P ¼ 0.001).
However, the 5th percentile ADC values were not stat-
istically significant (R2 ¼ 0.14, P ¼ 0.095). There was
also a significant relationship between PFS and the
methylation ratio (Fig. 6b, R2 ¼ 0.27, P ¼ 0.017). The
Kaplan–Meier survival curves associated with the
MGMT promoter methylation status determined by
the MSP are shown in Fig. 6c. The median PFS was
significantly influenced, being 14.5 months in the
MGMT promoter methylated patients and 4.0 months
(95% confidence interval, 3.0–9.0 months) in the
MGMT promoter unmethylated patients (P ¼ 0.025).
The mean value of the Ki-67 labeling index was
17.4% 6 9.9. The 5th percentile ADC values showed a
statistically significant inverse relationship with the
Ki-67 labeling index (Fig. 7, R2 ¼ 0.17, P ¼ 0.036).
However, the mean ADC values were not statistically
significant (R2 ¼ 0.02, P ¼ 0.51).
DISCUSSION
In the present study the methylation ratio, which was
generated using MS-MLPA, showed a positive relation-
ship with the mean ADC values of the enhancing por-
tion of the tumor, but there was no significant correla-
tion with the 5th percentile ADC values from the ADC
histograms. The MSP-derived methylation status
yielded a similar pattern. In addition, our results
showed a significant correlation between the mean
ADC values and PFS (P ¼ 0.001). In terms of the
methylation status of the MGMT promoter in GBM,
 15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Correlation of ADC, MGMT Methylation in GBM 355

Figure 3. An example of a high-

ADC tumor and a low-ADC tu-
mor. A 41-year-old man with
GBM involving the right frontal
lobe and the corpus callosum. a:
Axial postcontrast T1-weighted
image (450/8). b: On the ADC
map, the mass shows relatively
high ADC values with a mean
value of 1656.8
106 mm2/sec.
The methylation ratio of this tu-
mor was 0.49, and the tumor was
methylated according to the MSP.
A GBM at the left thalamus in a
67-year-old female. c: Axial post-
contrast T1-weighted image
(650/20). d: On the ADC map,
the mass shows relatively low
ADC values with a mean value of
776.9
106 mm2/sec. The
methylation ratio of this tumor
was 0.08, and the tumor was
unmethylated according to the
MSP.

the patients with methylated MGMT promoter showed The MGMT gene codes for a DNA repair enzyme
a longer median PFS than those with unmethylated that removes alkyl-groups from the O6-position of
MGMT promoter (14.5 vs. 4.0 months, P ¼ 0.025). guanine, which is one of the targets of alkylating

Figure 4. The linear regression plot of the Pearson correla-
tion analysis, which evaluated the significance of the associ-
ation between the methylation ratio measured by MS-MLPA
and the mean ADC values (R2 ¼ 0.22, P ¼ 0.015). [Color fig-
ure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
Figure 5. A box-and-whisker plot of the mean ADC values
in the MGMT promoter methylated and unmethylated
tumors according to the MSP results. The mean values are
significantly different between the two groups (P ¼ 0.011, in-
dependent t-test). [Color figure can be viewed in the online
issue, which is available at wileyonlinelibrary.com.]
 15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
356 Sunwoo et al.

agents such as TMZ or nitrosourea (7,21). Epigenetic

silencing of the MGMT gene by promoter methylation
has recently been highlighted because the methyla-
tion status of its promoter could be used as a bio-
marker for stratifying GBM patients and predicting
treatment response (5). However, the difficulty obtain-
ing an accurate MGMT promoter methylation status
is a major obstacle for the clinical application of this
marker. Alterations in the MGMT promoter methyla-
tion status following tumor resection and chemother-
apy and loss of methylation associated with resistance
to TMZ therapy have also been reported (22,23).
Because of such clinical difficulties, the search for a
reliable noninvasive surrogate marker for the MGMT
promoter methylation status is important.
Previous work has reported certain MRI features,
such as ring enhancement, to be associated with the
MGMT promoter methylation status (24). Recently,
there have been some attempts to correlate the ADC
values with MGMT promoter methylation status, but
the results have been inconsistent. Pope et al (25)
found an association between the ADC values and the
MGMT promoter methylation status. They applied
ADC histograms processed by using a two-mixture
normal distribution to provide optimal curve-fitting
and calculated mean values for the lower peak ADC
(ADCL) in GBMs; finally, they found that tumors with
the MGMT promoter methylation had significantly
lower mean ADCL. In addition, our results are sup-
ported by another recent work; Moon et al (26)
showed a positive relationship in terms of the ADC
values and the methylation status.
In the current work, we used the recently intro-
duced method of MS-MLPA along with MSP, which is
currently regarded as the standard method for assess-
ing the MGMT promoter methylation status. Correla-
tion analyses can be performed with MS-MLPA, which
establishes a semiquantitative methylation ratio with
the use of only small quantities of DNA (10,27). In our
study, both methods uniformly showed a correlation
between higher ADC values and methylated tumors.

Figure 6. Graphs showing the correlation analyses between

PFS and several other parameters. a,b: The linear regression
plots of PFS against (a) the mean ADC (R2 ¼ 0.44, P ¼ 0.001)
and (b) the methylation ratio (R2 ¼ 0.27, P ¼ 0.017). c: The
Kaplan–Meier analysis according to the MSP-derived methyl-
ation status. Patients with methylated tumors had signifi-
cantly longer survival rates than those with unmethylated
tumors (P ¼ 0.025). M ¼ Methylated, U ¼ Unmethylated. Figure 7. The correlation study between the 5th percentile
[Color figure can be viewed in the online issue, which is ADC values and the Ki-67 labeling index using a linear
available at wileyonlinelibrary.com.] regression model. A significant negative relationship was
noted (R2 ¼ 0.17, P ¼ 0.036). [Color figure can be viewed in
the online issue, which is available at wileyonlinelibrary.com.]

Correlation of ADC, MGMT Methylation in GBM
We believe that the ADC analysis method of the pres-
ent study can be ascribed to the different result from
that of the previous study (25). Considering that more
highly methylated tumors have a better prognosis, we
inferred that these tumors may have less cellularity or
contain a larger proportion of necrotic tissue, which
would increase the ADC values. In terms of PFS, our
result is similar to that of the study by Pope et al (25),
in which GBMs of lower ADC values had lower PFS
that those with higher ADC values.
Several studies have shown a significant correlation
between the ADC values and patient outcome (28,29).
We have also shown that the mean ADC values are
significantly related to the median PFS. We have also
demonstrated a longer median PFS in the methylated
group, which is in agreement with previous studies
(7,30). Our work provides further evidence of the con-
nection between ADC values and MGMT promoter
methylation as well as the connection between MGMT
promoter methylation and PFS. This finding may pro-
vide the opportunity to identify the causal relation-
ships between ADC, methylation status, and PFS.
However, there may be many other genetic or molecu-
lar alterations contributing to the ADC values and
overall prognosis.
Ki-67, a nuclear antigen specific for proliferating
cells (31), is used for the evaluation of tumor prolifer-
ation and its positive relationship with higher cell
density has been established (32). In this study, we
measured the Ki-67 labeling index to validate the
ADC values because ADC values are known as a bio-
marker of cell density in tumors. The 5th percentile
values showed a significant negative relationship to
the Ki-67 labeling index, while the mean ADC did not.
Because an elevated Ki-67 labeling index reflects
increased cellularity, it may be reasonable to expect
low ADC values for high Ki-67-tumors, which explains
the inverse relationship between the 5th percentile
ADC and the Ki-67 labeling index. Pathologists usu-
ally select the fields where Ki-67-labeled cells are
most closely packed (ie, areas with the highest cellu-
larity) for measuring the Ki-67 labeling index. This
may explain the insignificant results of the mean ADC
value, which is representative of the entire tumor.
Higano et al (29) have shown that minimum ADC val-
ues correlate well with the Ki-67 labeling index in
high-grade gliomas. In our study, we used the 5th
percentile value instead of the minimum because of
the risk that a single erroneous voxel may have an
extremely low value and skew the data. With the use
of a histogram analysis, there is little risk of data con-
tamination by a few erroneous voxels when assessing
the large number of whole tumor voxels. A previous
study (19) showed that the 5th percentile ADC value
aided in differentiating high-grade from low-grade
gliomas, which indicates that the 5th percentile ADC
could also be used as a reliable parameter in assess-
ing the highest cellular portion in gliomas.
It is interesting to note that the 5th percentile ADC
value has a weaker correlation with PFS than the
mean ADC value (P ¼ 0.010 vs. P ¼ 0.075). It seems
that the mean ADC values predict the overall progno-
sis more accurately for GBM patients than the ADC
15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
357
values of the most cellular portion of the tumor.
Although the exact mechanism remains unclear, we
believe that this is probably because the 5th percen-
tile ADC values, despite its ability to reflect the
aggressiveness of the tumor, cannot assess tumor
heterogeneity.
Our study has several limitations. First, the study
population was relatively small. With regard to correla-
tion between the PFS and the mean ADC values, the
statistical significance might have affected by the two
outliers to some extent, considering Fig. 6a. It would
have been preferable to include more patients to
strengthen the statistical power. Second, our methods
require the coregistration of postcontrast T1-weighted
images with the ADC map, which could be a source of
error, particularly if the section thickness between the
pulse sequences is not the same. Third, although we
excluded lesions with grossly cystic or hemorrhagic
areas to avoid the effect of volume averaging and
ADC changes, there was no direct spatial correlation
between the ADC values and the pathologic specimens.
In conclusion, we have demonstrated a positive
relationship between the mean ADC value and the
semiquantitative methylation status of the MGMT pro-
moter. In addition, we have shown a significant rela-
tionship between a higher mean ADC value and clini-
cal outcome. These results indicate that mean ADC
values could serve as an independent biomarker for
predicting treatment response, following further vali-
dation studies.
ACKNOWLEDGMENT
The authors thank So Young Yun for continuous sup-
port with updating and organizing the clinical data.
REFERENCES
1. Davis FG, Freels S, Grutsch J, Barlas S, Brem S. Survival rates
in patients with primary malignant brain tumors stratified by
patient age and tumor histological type: an analysis based on
Surveillance, Epidemiology, and End Results (SEER) data, 1973–
1991. J Neurosurg 1998;88:1–10.
2. deAngelis L. Brain tumors. N Engl J Med 2001;344:114–122.
3. Stark AM, Nabavi A, Mehdorn HM, Bl€ omer U. Glioblastoma mul-
tiforme—report of 267 cases treated at a single institution. Surg
Neurol 2005;63:162–169.
4. Stupp R, Dietrich PY, Ostermann Kraljevic S, et al. Promising
survival for patients with newly diagnosed glioblastoma multi-
forme treated with concomitant radiation plus temozolomide fol-
lowed by adjuvant temozolomide. J Clin Oncol 2002;20:
1375–1382.
5. Hau P, Stupp R, Hegi ME. MGMT methylation status: the advent
of stratified therapy in glioblastoma? Dis Markers 2007;23:
97–104.
6. Hegi ME, Liu L, Herman JG, et al. Correlation of O6-methylgua-
nine methyltransferase (MGMT) promoter methylation with clini-
cal outcomes in glioblastoma and clinical strategies to modulate
MGMT activity. J Clin Oncol 2008;26:4189–4199.
7. Hegi ME, Diserens AC, Gorlia T, et al. MGMT gene silencing and
benefit from temozolomide in glioblastoma. N Engl J Med 2005;
352:997–1003.
8. Brandes AA, Franceschi E, Tosoni A, et al. MGMT promoter
methylation status can predict the incidence and outcome of
pseudoprogression after concomitant radiochemotherapy in
newly diagnosed glioblastoma patients. J Clin Oncol 2008;26:
2192–2197.

358
9. Kong DS, Kim ST, Kim EH, et al. Diagnostic dilemma of pseudo-
progression in the treatment of newly diagnosed glioblastomas:
the role of assessing relative cerebral blood flow volume and oxy-
gen-6-methylguanine-DNA methyltransferase promoter methyla-
tion status. AJNR Am J Neuroradiol 2011;32:382–387.
10. Jeuken JW, Cornelissen SJ, Vriezen M, et al. MS-MLPA: an
attractive alternative laboratory assay for robust, reliable, and
semiquantitative detection of MGMT promoter hypermethylation
in gliomas. Lab Invest 2007;87:1055–1065.
11. Park CK, Kim J, Yim SY, et al. Usefulness of MS-MLPA for detection
of MGMT promoter methylation in the evaluation of pseudoprogres-
sion in glioblastoma patients. Neurooncology 2010;13:195–202.
12. Preusser M, Hassler, M., Elandt, K., et al. Analysis of MGMT pro-
moter methylation status in high grade glioma patients with long
term and conventional survival times: a retrospective study. J
Clin Oncol 2007;25:2084.
13. Henson JW, Gaviani P, Gonzalez RG. MRI in treatment of adult
gliomas. Lancet Oncol 2005;6:167–175.
14. Diehn M, Nardini C, Wang DS, et al. Identification of noninvasive
imaging surrogates for brain tumor gene-expression modules.
Proc Natl Acad Sci U S A 2008;105:5213–5218.
15. Patterson DM, Padhani AR, Collins DJ. Technology insight: water
diffusion MRI—a potential new biomarker of response to cancer
therapy. Nat Clin Pract Oncol 2008;5:220–233.
16. Wen PY, Macdonald DR, Reardon DA, et al. Updated response
assessment criteria for high-grade gliomas: response assessment in
neuro-oncology working group. J Clin Oncol 2010;28:1963–1972.
17. Abramoff MD, Magalhães PJ, Ram SJ. Image processing with
ImageJ. Biophotonics Int 2004;11:36–42.
18. Tozer DJ, J€ ager HR, Danchaivijitr N, et al. Apparent diffusion
coefficient histograms may predict low-grade glioma subtype.
NMR Biomed 2007;20:49–57.
19. Kang Y, Choi SH, Kim YJ, et al. Gliomas: histogram analysis of
apparent diffusion coefficient maps with standard- or high-b-
value diffusion-weighted MR imaging—correlation with tumor
grade. Radiology 2011;261:882–890.
20. Hsu SM, Raine L, Fanger H. A comparative study of the peroxi-
dase-antiperoxidase method and an avidin-biotin complex
method for studying polypeptide hormones with radioimmunoas-
say antibodies. Am J Clin Pathol 1981;75:734–738.
21. Friedman HS, McLendon RE, Kerby T, et al. DNA mismatch
repair and O6-alkylguanine-DNA alkyltransferase analysis and
15222586, 2013, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.23838, Wiley Online Library on [04/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Sunwoo et al.
response to Temodal in newly diagnosed malignant glioma. J Clin
Oncol 1998;16:3851–3857.
22. Jung TY, Jung S, Moon KS, et al. Changes of the O6-methylgua-
nine-DNA methyltransferase promoter methylation and MGMT
protein expression after adjuvant treatment in glioblastoma.
Oncol Rep 2010;23:1269–1276.
23. Zhang J, Stevens MF, Laughton CA, Madhusudan S, Bradshaw
TD. Acquired resistance to temozolomide in glioma cell lines: mo-
lecular mechanisms and potential translational applications. On-
cology 2010;78:103–114.
24. Drabycz S, Roldan G, de Robles P, et al. An analysis of image tex-
ture, tumor location, and MGMT promoter methylation in glio-
blastoma using magnetic resonance imaging. NeuroImage 2010;
49:1398–1405.
25. Pope WB, Lai A, Mehta R, et al. Apparent diffusion coefficient his-
togram analysis stratifies progression-free survival in newly diag-
nosed bevacizumab-treated glioblastoma. Am J Neuroradiol
2011;32:882–889.
26. Moon WJ, Choi JW, Roh HG, Lim SD, Koh YC. Imaging parame-
ters of high grade gliomas in relation to the MGMT promoter
methylation status: the CT, diffusion tensor imaging, and perfu-
sion MR imaging. Neuroradiology 2012;54:555–563.
27. Nygren AO, Ameziane N, Duarte HM, et al. Methylation-specific
MLPA (MS-MLPA): simultaneous detection of CpG methylation
and copy number changes of up to 40 sequences. Nucleic Acids
Res 2005;33:e128.
28. Oh J, Henry RG, Pirzkall A, et al. Survival analysis in patients
with glioblastoma multiforme: predictive value of choline-to-N-
acetylaspartate index, apparent diffusion coefficient, and relative
cerebral blood volume. J Magn Reson Imaging 2004;19:546–554.
29. Higano S, Yun X, Kumabe T, et al. Malignant astrocytic tumors:
clinical importance of apparent diffusion coefficient in prediction
of grade and prognosis. Radiology 2006;241:839–846.
30. Esteller M, Garcia-Foncillas J, Andion E, et al. Inactivation of the
DNA-repair gene MGMT and the clinical response of gliomas to
alkylating agents. N Engl J Med 2000;343:1350–1354.
31. Scholzen T, Gerdes J. The Ki-67 protein: from the known and the
unknown. J Cell Physiol 2000;182:311–322.
32. Kiss R, Dewitte O, Decaestecker C, et al. The combined determi-
nation of proliferative activity and cell density in the prognosis of
adult patients with supratentorial high-grade astrocytic tumors.
Am J Clin Pathol 1997;107:321–331.