Original article

The Neuroradiology Journal

0(00) 1–10

Arterial spin labeling perfusion: Prospective ! The Author(s) 2018

Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
MR imaging in differentiating neoplastic from DOI: 10.1177/1971400918783058
journals.sagepub.com/home/neu
non-neoplastic intra-axial brain lesions

Neetu Soni1 , Karthika Srindharan2, Sunil Kumar2, Prabhakar Mishra3,

Girish Bathla1, Jyantee Kalita4 and Sanjay Behari5

Abstract
Purpose: The purpose of this article is to assess the diagnostic performance of arterial spin-labeling (ASL) magnetic
resonance perfusion imaging to differentiate neoplastic from non-neoplastic brain lesions.
Material and methods: This prospective study included 60 consecutive, newly diagnosed, untreated patients with intra-axial
lesions with perilesional edema (PE) who underwent clinical magnetic resonance imaging including ASL sequences at 3T.
Region of interest analysis was performed to obtain mean cerebral blood flow (CBF) values from lesion (L), PE and normal
contralateral white matter (CWM). Normalized (n) CBF ratio was obtained by dividing the mean CBF value of L and PE by
mean CBF value of CWM. Discriminant analyses were performed to determine the best cutoff value of nCBFL and nCBFPE in
differentiating neoplastic from non-neoplastic lesions.
Results: Thirty patients were in the neoplastic group (15 high-grade gliomas (HGGs), 15 metastases) and 30 in the non-
neoplastic group (12 tuberculomas, 10 neurocysticercosis, four abscesses, two fungal granulomas and two tumefactive
demyelination) based on final histopathology and clincoradiological diagnosis. We found higher nCBFL (6.65  4.07 vs
1.68  0.80, p < 0.001) and nCBFPE (1.86  1.43 vs 0.74  0.21, p < 0.001) values in the neoplastic group than non-
neoplastic. For predicting neoplastic lesions, we found an nCBFL cutoff value of 1.89 (AUC 0.917; 95% CI 0.854 to 0.980;
sensitivity 90%; specificity 73%) and nCBFPE value of 0.76 (AUC 0.783; 95% CI 0.675 to 0.891; sensitivity 80%; specificity
58%). Mean nCBFL was higher in HGGs (8.70  4.16) compared to tuberculomas (1.98  0.87); and nCBFPE was higher in
HGGs (3.06  1.53) compared to metastases (0.86  0.34) and tuberculomas (0.73  0.22) (p < 0.001).
Conclusion: ASL perfusion may help in distinguishing neoplastic from non-neoplastic brain lesions.

Keywords
Three-dimensional pseudocontinuous arterial spin labeling, dynamic susceptibility contrast, normalized cerebral blood flow
lesion, normalized cerebral blood flow peri-lesional edema

gadolinium contrast agent and provides relative cere-

Introduction
Contrast magnetic resonance imaging (MRI) has a
well-established role in brain lesion evaluation by pro-
viding anatomical information; however, it has a lim-
ited role in the quantitative and functional assessment
of lesions thus making it less reliable in differentiating
tumors from tumor-like lesions. Contrast enhancement
depicts blood-brain barrier (BBB) impairment, so every
lesion that shows contrast enhancement is not a tumor
and every glioma that does not take contrast is not a
low-grade glioma and vice versa.1 Perfusion-weighted
imaging (PWI) serves as a useful adjunct to conven-
tional MRI and overcomes these limitations by estimat-
ing tumor neoangiogenesis, which helps in tumor
grading, guiding stereotactic biopsy and surgical plan-
ning. Dynamic susceptibility contrast (DSC), routinely
used with PWI in clinical practice, uses an exogenous
bral blood volume (rCBV) and relative cerebral blood
flow (rCBF) parameters that correlate well with tumor
grade and histology.2 Arterial spin labeling (ASL) is a
noninvasive PWI method that provides absolute
1
Neuroradiology Department, University of Iowa Hospitals and Clinics, Iowa
City, IA, USA
2
Department of Radiology, Sanjay Gandhi Postgraduate Institute of Medical
Sciences (SGPGIMS), Lucknow, Uttar Pradesh, India
3
Department of Biostatistics and Health Informatics, SGPGIMS, Lucknow,
Uttar Pradesh, India
4
Department of Neurology, SGPGIMS, Lucknow, Uttar Pradesh, India
5
Department of Neurosurgery, SGPGIMS, Lucknow, Uttar Pradesh, India
Corresponding author:
Neetu Soni, Neuroradiology Department, University of Iowa Hospitals and
Clinics, 200 Hawkins Drive, Iowa City, 52242-1009, IA, USA.
Emails: neetu_soni06@yahoo.co.in; drneetusoni98@gmail.com
2 The Neuroradiology Journal 0(00)
quantification of CBF in the absence of exogenous con-
trast, which makes ASL a promising technique for
studying perfusion in patients with allergies, renal fail-
ure, difficult intravenous access, repetitive follow-ups,
pregnant women and children.3 DSC is preferred over
ASL in brain lesion evaluation because of routine con-
trast administration and rapid data acquisition.
However, in contrast-enhanced perfusion techniques,
CBF quantification is affected by T1 and T2 effects of
BBB disruption and susceptibility artifacts, which are
less with ASL.4 Previous studies have shown a very
strong correlation between ASL and DSC-MRI derived
perfusion parameters for evaluation of brain tumors
and support the possibility that ASL can be used as
an alternative to DSC-MRI.5–7 Gliomas are the most
common primary brain tumors and sometimes differen-
tiation from solitary enhancing cerebral metastases can
be difficult on conventional MRI. Primary tumors have
shown statistically significant higher perfusion in peri-
lesional edema (PE) compared to metastases that have
been ascribed to tumor infiltration along white matter
tracts in glioma.8–12
Contrast enhancements of infective brain lesions
is due to disrupted BBB and lack of angiogenesis in
enhancing lesions on PWI favors non-neoplastic
origin, which helps to differentiate them from tumors.
In endemic countries, tuberculosis (TB) and neurocys-
ticercosis (NCC) are the most frequent infective brain
lesions and are sometimes difficult to differentiate from
primary brain tumors and metastases on conventional
MRI.13–15
In this study, we evaluated the practicability of ASL-
derived CBF parameters in differentiation of various
intra-axial lesions in our population and assessed the
feasibility of using ASL as an alternative to DSC when
needed.
Materials and methods
Participants
This prospective study protocol (July 2014 through
December 2016) was approved by our institute’s
ethics committee and all participants provided written
informed consent at the Sanjay Gandhi Postgraduate
Institute of Medical Sciences, Lucknow, India. We
included 60 consecutive, newly diagnosed, untreated
patients (mean age 39 years; age range 18–66 years;
male:female 34:26) with single or multiple enhancing
intra-axial brain lesions of size more than 1 cm with
perilesional edema (PE) detected either on computed
tomography or MRI. We excluded patients with any
form of prior specific medical or surgical treatment,
unknown disease, and lesions less than 1 cm.
MRI study
MRI examination was performed with a 3 T MR scan-
ner (Signa Hdxt, General Electric, Milwaukee, WI,
USA), using a 12-channel head coil. Routine imaging
protocols included axial T2, T1/T2 fluid-attenuated
inversion recovery (FLAIR), diffusion-weighted
imaging (DWI) (b value ¼ 0 and 1000 mm2/sec) and
post-contrast T1. Whole-brain three-dimensional pseu-
docontinuous ASL (3D PCASL) was acquired with the
following parameters: label duration ¼ 1.5 s, post-
labeling delay ¼ 1.5 s, repetition time ¼ 4.4 s, echo time-
¼ 9.2 ms, frequency ¼ 512, phase ¼ 8, number of
excitations ¼ 3, number of slices ¼ 32, field of view ¼ 24,
slice thickness ¼ 6 mm, band width ¼ 62.50, pla-
ne ¼ axial, scan time ¼ four minutes; pulse labeling
plane ¼ just below the volume of interest.
Post-processing and data analysis
ASL data were transferred to the workstation
(ADW4.4, GE, USA) for post-processing and analyzed
using commercial software (FuncTool BrainStat
Software, General Electric, Milwaukee, WI, USA)
with automated generation of quantitative perfusion
CBF maps. Conventional MRI and perfusion maps
were co-analyzed by two subspecialty neuroradiologists
(reader 1, with 10 years’ experience, and reader 2, with
35 years’ experience) in consensus who were blinded to
the final diagnosis. The conventional MR sequence
assessed the lesion localization, characterization and
PE on T2 FLAIR. CBF values were calculated by pos-
itioning three circular regions of interest (ROIs) (5–10
mm2) over the lesions showing highest perfusion signal
as visually determined from the CBF maps and from
the nonenhancing PE (within 1 cm from the lesion)
excluding necrotic, cystic and hemorrhagic areas
(Figure 1).
To minimize measurement errors of CBF analysis,
uniformly sized ROIs (5–10 mm2) were kept and tar-
geted to the maximal abnormalities within the lesion.
An average of three CBF measurements were taken and
in cases of multiple lesions, the larger lesion with max-
imum perfusion value was targeted for measurements.
For normalization to avoid individual patient vari-
ation, three similar ROIs were placed on the normal
contralateral white matter (CWM) and averaged to
obtain mean CWM CBF values. Normalized CBF
ratios were calculated by dividing the average CBF
from lesion and PE to average CWM CBF (nCBFL
and nCBFPE).
Statistical analysis
The normality of continuous variables (ASL param-
eters) was tested using the Kolmogorov-Smirnov test.
For normally distributed data, mean  standard devi-
ation (SD) was used as descriptive statistics. nCBFL
and nCBFPE values between the two groups (neoplas-
tic and non-neoplastic) were compared using the inde-
pendent t test. Receiver operating characteristic (ROC)
curve was performed and area under curve (AUC)
calculated to seek the best cutoff values of nCBFL
Soni et al. 3

Figure 1. (a) ROC curve representing the discriminatory capability of nCBFL and nCBFPE in distinguishing neoplastic lesions from
non-neoplastic lesions with area under the curve. (b) ROC curve representing the discriminatory capability of nCBFPE in distinguishing
high-grade gliomas from metastases with area under curve.
nCBFL: normalized cerebral blood flow from lesion; nCBFPE: normalized cerebral blood flow from lesion perilesional edema;
ROC: receiver operating characteristic.

Table 1. MeanSD values of nCBFL and nCBFPE of neoplastic and non-neoplastic lesions.

nCBFL nCBFPE
Group Diagnosis (N) MeanSDb MeanSDb
a
Neoplastic (n ¼ 30) HGG (N ¼ 15) 8.70  4.16 3.06  1.53
Metastases (N ¼ 15) 6.53  3.56 0.86  0.34
Non-neoplastic (n ¼ 30) Tuberculoma (N ¼ 12) 1.98  0.87 0.73  0.22
NCC (N ¼ 10) 1.66  0.70 0.95  0.14
Abscess (N ¼ 4) 1.47  0.64 0.83  0.14
Tumefactive demyelination (N ¼ 2) 0.62  0.38 0.65  0.34
Fungal granulomas (N ¼ 2) 1.22  0.54 0.60  0.12
HGG: high-grade glioma; nCBFL: normalized cerebral blood flow from lesion; NCC: neurocysticercosis; nCBFPE: normalized cerebral blood flow from lesion
perilesional edema.
a
Between HGG and metastases: p ¼ 0.136 for nCBFL, p < 0.001 for nCBFPE (independent samples t test).
b
Values are given as meanSD; p < 0.05 is significant.

and nCBFPE to differentiate neoplastic from non-neo-
plastic lesions. Sensitivity, specificity and 95% confi-
dence intervals (CIs) were calculated when employing
the cutoff CBF values. A p < 0.05 was considered stat-
istically significant and p < 0.001 highly significant.
Statistical analysis was performed using the Statistical
Package for Social Sciences, version 23 (SPSS-23, IBM,
and Chicago, IL, USA).
Results
All the patients were broadly classified into two major
groups of neoplastic (male:female 21:9; age range 38–66
years) and non-neoplastic (male:female 13:17; age
range 18–46 years) based on the histopathological and
clinicoradiological diagnoses. Histopathology was
documented in patients with high-grade gliomas
(HGGs) (n ¼ 15), fungal granuloma (n ¼ 2) and pres-
ence of primary carcinoma in metastases (n ¼ 15).
Clinicoradiologic diagnoses were determined by a com-
bination of provided clinical history, typical imaging
findings, biochemical, culture and post-treatment
follow-ups in patients with TB (n ¼ 12), NCC (n ¼ 10),
abscess (n ¼ 4) and tumefactive demyelination (n ¼ 2).
In the neoplastic group, HGGs and metastasis (eight
cases of primary carcinoma of the breast, seven cases of
primary carcinoma of the lung) comprised the bulk of
patients while tuberculoma and NCC predominated in
the non-neoplastic group (Table 1). The mean nCBFL
and nCBFPE values of the neoplastic group were com-
pared with those of the non-neoplastic group and are
displayed in Table 2. Statistically significant (p < 0.001)
4 The Neuroradiology Journal 0(00)

differences were found between mean nCBFL
(6.65  4.07 vs 1.68  0.80) and mean nCBFPE
(1.86  1.43 vs 0.74  0.21) of the neoplastic compared
to the non-neoplastic group (Table 2). ROC analysis
results indicated that when mean nCBFL ¼ 1.89 was
set as the threshold value, the best diagnostic perform-
ance in predicting neoplastic lesions was achieved
with 90%, sensitivity (95% CI 0.723–0.973) 73%, speci-
ficity (95% CI 0.538–0.871) and AUC of 0.917 (95% CI
0.854–0.980) (Figure 1(a)). When mean nCBFPE ¼ 0.76
was set as the threshold value, the best diagnostic per-
formance in predicting neoplastic lesions was achieved
with 80% sensitivity (95% CI 0.608–0.916), 58% speci-
ficity (95% CI 0.377–0.741) and 0.783 AUC (95% CI
0.675 to 0.891); with a further increased cutoff value of
1.02, specificity increases to 90% (95% CI 0.723–0.974)
at the cost of decreased sensitivity of 60% (95% CI
0.408–0.768) (Figure 1).
After the main group comparison, we found statis-
tically higher (p < 0.001) mean nCBFPE in HGG
(Figure 2) compared to metastasis (Figure 3); however,
the mean difference in nCBFL was statistically insignifi-
cant (p ¼ 0.125) (Table 1). ROC analysis results
(Figure 1(b)) indicated that when nCBFPE ¼ 1.2 was
set as the threshold value, the best diagnostic
Table 2. MeanSD values of nCBFL and nCBFPE values in
neoplastic and non-neoplastic groups.
Neoplastic Non-neoplastic
b
ASL parameter group (N ¼ 30) group (N ¼ 30) p value
a
nCBFL 6.65  4.07 1.68  0.80 <0.001
nCBFPEa 1.86  1.43 0.74  0.21 <0.001
ASL: arterial spin-labeling; HGG: high-grade glioma; nCBFL: normalized
cerebral blood flow from lesion; NCC: neurocysticercosis; nCBFPE: normal-
ized cerebral blood flow from lesion perilesional edema.
a
Values are given as meanSD.
b
Independent samples t test used; p < 0.001 is highly significant
performance in predicting HGG lesions was achieved
with a sensitivity of 94% (95% CI 0.765–0.988), speci-
ficity of 83.3% (95% CI 0.645–0.937) and AUC of 0.94
(95% CI 0.86–1.00), while with a low cutoff value of
0.73, sensitivity increases to 100% (95% CI 0.859–1.00)
at the cost of decreased specificity of 38% (95% CI
0.205–0.561). When HGG was compared to tubercu-
loma (Figure 4), mean nCBFL and nCBFPE were
higher in HGG compared to tuberculoma (p < 0.001)
(Table 1). We also compared two closely resembled
lesions, metastases and tuberculoma, and observed
higher mean nCBFL in metastases compared to tuber-
culoma (p < 0.001) with no statistically significant
(p ¼ 0.263) difference in mean nCBPE (Table 1).
A cutoff nCBFL value of 2.8 differentiated metastasis
from tuberculoma with 87% sensitivity (95% CI 0.584–
0.977), 84% specificity (95% CI 0.509–0.971), 87%
positive predictive value (95% CI 0.584–0.977), 83%
negative predictive value (95% CI 0.509–0.971) and
0.89 AUC (95% CI 0.74–1.00, p ¼ 0.001).
We found almost similar nCBFL and nCBFPE
values in tuberculoma and NCC (Figure 5) with no
statistically significant difference (p ¼ 0.351 and
p ¼ 0.295) (Table 1). We were not able to specifically
compare between other non-neoplastic lesions such as
abscess (Figure 6), tumefactive demyelinating lesion
(TDL) (Figure 7), and fungal granulomas (Figure 8)
with neoplastic lesions because of the limited number
of patients in our sample.
Discussion
In this study, we found significantly higher (p < 0.001)
mean nCBFL and nCBFPE in the neoplastic group
with a good ability of ASL to define neoplastic lesions
when nCBFL of 1.89 was considered as the cutoff value
(sensitivity 90% and specificity 73%). Although many
brain perfusion studies have been reported in evalu-
ation of tumors and a few on infections, most of

Figure 2. Glioblastoma multiforme in a 50-year-old man. (a) Axial T2-weighted MR image showing a heterogeneous mass with
peritumoral white matter edema in the left high frontoparietal region. (b) Axial contrast-enhanced T1-weighted MR image showing
irregularly enhanced mass. (c) ASL color map demonstrating an area of hypervascularity (high CBF value, red area) in the tumor.
(d) ASL-derived CBF maps, overlaid on post-contrast axial three-dimensional T1-weighted images acquired with same planning.
ASL: arterial spin labeling; CBF: cerebral blood flow; MR: magnetic resonance.
Soni et al. 5

Figure 3. MR images of 56-year-old female with metastases of the breast. (a) T2-weighted imaging showing a hyperintense lesion with
peripheral hypointense rim and perilesional edema in the left temporal region. (b) T1-weighted post-contrast image shows predominant
central enhancement rather than peripheral rim enhancement. (c) ASL perfusion maps show red hue in high CBF solid enhancing part
and blue hue in non-enhancing perilesional edema (red arrow).
ASL: arterial spin labeling; CBF: cerebral blood flow; MR: magnetic resonance.

Figure 4. MR images of a 35-year-male with multiple tuberculomas. (a) T2-weighted imaging showing multiple hypointense lesions with
perilesional edema in the right temporal, thalamus and midbrain regions. (b) T1-weighted post-contrast image showing conglomerate
pattern of enhancement (c) ASL perfusion map shows blue hue indicating lower CBF in both lesions (black arrow) and immediate
perilesional edema.
ASL: arterial spin labeling; CBF: cerebral blood flow; MR: magnetic resonance.

these studies utilized the DSC perfusion technique and
very few used ASL-MRI.16,17 Studies using DSC-MRI
have shown statistically significant higher rCBV values
in neoplastic lesions than infectious lesions.18–20
Floriano et al.18 found higher (p < 0.001) rCBV in neo-
plastic (4.28  2.11) than non-neoplastic lesions
(0.63  0.49) with a cutoff rCBV value of 1.3 (sensitivity
97.8% and specificity 92.6%) for predicting malig-
nancy. Hourani et al.20 found higher rCBV (4.9  3.1)
in high-grade tumors than non-neoplastic lesions
(1.0  0.4; p < 0.001) with a cutoff rCBV value of 1.5
(sensitivity 77.8% and specificity 91.7%) for predicting
malignancy. The slightly higher nCBFL cutoff value in
our study than in the other published studies might be
because of patient age, immunological status, hetero-
geneity and variability of included lesions, degree of
angiogenesis, scanning protocol, perfusion techniques
and post-processing software used.
The next comparison between closely resembled
HGGs and metastases showed statistically significant
(p < 0.001) higher mean nCBFPE values in HGG
while an insignificant difference in nCBFL (Table 1).
Glioblastoma multiforme (GBM) is the most common
glial brain tumor, often having a similar imaging
appearance as solitary metastases with heterogeneous
enhancement and extensive perilesional edema.
6 The Neuroradiology Journal 0(00)

Figure 5. Multiple neurocysticercosis. (a) Axial T2 image shows multiple hyperintense lesions with perilesional vasogenic edema.
Intralesional eccentric hypointense scolex is also seen. (b) Axial post-contrast-enhanced T1-weighted MR image shows thin rim
enhancement. (c) Hypoperfusion is seen within the lesion on ASL perfusion map (red arrow).
ASL: arterial spin-labeling; MR: magnetic resonance.

Figure 6. Pyogenic brain abscess. (a) Axial T2 image shows the lesion with a high-signal intensity center and low-signal intensity
capsule. On DWI (b) high signal was detected in the abscess cavity with significant low ADC values (c). On the perfusion MRI (d) lesion
shows low CBF (arrow).
ADC: apparent diffusion coefficient; CBF: cerebral blood flow; DWI: diffusion-weighted imaging; MRI: magnetic resonance imaging.

Gliomas are infiltrative in nature and show statistically
significant higher peritumoral blood flow values
secondary to neo-angiogenesis and plenty of tumor-
specific extracellular matrix components. Brain metas-
tases capillaries lack BBB resulting in extensive
vasogenic edema in the absence of tumor cells, which
further decreases microcirculation due to local pressure
and accounts for a decrease in peritumoral rCBV and
CBF.8,9,21–24 A recent prospective study by Lin et al.25
using 3D PCASL documented significantly (p < 0.038)
higher CBF values in the peritumoral regions of glio-
blastomas than those in metastases with a cutoff CBF
of 1.92 (sensitivity 92.86% and specificity 100%).
A retrospective ASL study by Sunwoo et al.26 in
128 patients with GBM and metastases found signifi-
cantly higher peritumoral blood flow in GBM. Our
results are also consistent with two other ASL studies
conducted by Weber et al.10 and Bai et al.11 that
demonstrated significantly higher perilesional CBF in
glioblastomas than metastases. Weber et al.10 also
demonstrated significantly higher microvessel density
in GBM than brain metastases and suggested that
ASL-derived CBF may estimate brain tumors’ micro-
vascular densities.
Mainly two factors contribute to decreased peritu-
moral perfusion in metastases. First, brain metastases
capillaries do not possess the unique BBB that makes
them highly leaky resulting in vasogenic edema in the
Soni et al. 7

Figure 7. Tumefactive demyelination. (a) Axial T2-FLAIR image shows a lesion with perilesional edema in the right high frontal region.
(b) Sagittal post-contrast image shows characteristic ‘‘open ring’’ of contrast enhancement. (c) Hypoperfusion is seen within the lesion on
ASL perfusion map (red arrow).
ASL: arterial spin labeling; FLAIR: fluid-attenuated inversion recovery.

Figure 8. Aspergillus granuloma. Axial T2-weighted image (a) showing an irregular hypointense lesion with perilesional edema in the
right temporal region; The lesion is hyperintense on DWI (b) with low ADC (c) and showing intense solid enhancement on post-contrast
T1-weighted image (d). The lesion is isointense (red arrow) to gray matter on ASL-derived gray (f) and colored CBF maps (f).
ADC: apparent diffusion coefficient; ASL: arterial spin labeling; CBF: cerebral blood flow; DWI: diffusion-weighted imaging.
8 The Neuroradiology Journal 0(00)
absence of tumor cells, and second, the presence of
local microcirculation compression by extravasated
edema fluid. In HGG, peritumoral edema demonstrates
impervious peritumoral neural vasculature with BBB
breakdown, tumor infiltration and tumor-specific extra-
cellular matrix components, possibly explaining the
increased rCBV and CBF.24,27,28 Bulakbasi et al.29
reported perilesional rCBV cutoff ratios of 1.2 in differ-
entiating metastases from high-grade glial tumors,
respectively. In our study, the proposed reason for the
increased nCBFPE in neoplastic lesions is due to
increased vascularity following tumor infiltration in
HGG as these lesions comprised the majority of the
neoplastic lesions included.
Similarly, perilesional blood flow values can be used
to differentiate glial tumors from enhancing tuberculo-
mas. Intra-axial tuberculomas are difficult to differenti-
ate from primary brain tumors and metastases on
conventional MRI. Tuberculomas were the most fre-
quent non-neoplastic lesion in our study and showed
decreased (p < 0.001) nCBFL and nCBFPE compared
to HGG. Batra and Tripathi30 reported high intrale-
sional rCBV in tuberculomas like cerebral gliomas;
however, perilesional rCBV was reported to be low in
tuberculomas. Gupta et al.31 reported raised rCBV in
tuberculomas, which correlated with reactive neovascu-
larization and increased vascular endothelial growth
factor expression.
The comparison between two closely resembled
lesions, metastases and tuberculoma, showed statistic-
ally significant (p < 0.001) higher nCBFL in metastases
compared to tuberculoma (Table 1) while no statistic-
ally significant difference (p ¼ 0.263) in nCBFPE.
Similar results shown by another retrospective study
by Chatterjee et al.15 and Shankhe et al.14 in distin-
guishing metastases from tuberculomas using DSC
found statistically significant high rCBV from the
enhancing portion of metastases than in tuberculomas.
However, the rCBV values from PE were not different
between tuberculomas and metastases. The authors
also reported that multidrug-resistant TB may demon-
strate a drastic increase in rCBV value, and MR perfu-
sion can help predict therapeutic response in patients
with cerebral tuberculomas.14,32
NCC shows variable MRI appearances depending on
the evolving stage of infection and closely resembles
tuberculoma where both exist as endemic disease.
Increased perfusion generally favors diagnosis of
neoplastic lesion rather than NCC. Gupta et al.33
demonstrated significant differences using dynamic con-
trast-enhanced MRI-derived perfusion parameters
among all the clinically active stages of NCC and
found the highest in the colloidal stage. We found no
statistically significant difference in lesion and perile-
sional perfusion values between NCC and tuberculomas.
To compare our results, we did not find any ASL study
in the literature differentiating tuberculomas and NCC.
Sometimes, rim-enhancing brain abscesses mimic
cystic brain neoplasms and metastases, making
differentiation difficult on conventional MRI. PWI
along with DWI is helpful and few studies have demon-
strated higher rCBV values from cystic HGG and
metastases over cerebral abscesses.34 Erdogan et al.35
reported low rCBV ratios from the capsular portion
of abscesses as compared to HGG and metastases. In
our study, we have reported lower nCBFL and
nCBFPE values in cerebral abscess than HGG and
metastases.
The clinical presentation of patients with TDL is
variable and poses a diagnostic challenge to distinguish
it from HGG on conventional MRI, which often fol-
lows biopsy. TDLs usually show decreased rCBV
values in comparison to HGG; however, they can
also present with elevated rCBV values and mimic
HGG, making differentiation difficult.36,37 A recent
study by Hiremath et al.38 has shown lower mean
rCBV in TDL (2.11  1.12) compared to mean rCBV
in HGGs (3.77  1.65).
Although the results from our current study are
encouraging, there are a few limitations that should
be noted. First, a relatively small number of patients
were included in each group and their subgroups.
Henceforth, a large sample size would be necessary to
further establish the role of ASL in differentiation of
various intracranial lesions. Second, we did not have
histopathological confirmation in most of the non-neo-
plastic lesions. However, histopathological confirm-
ation in non-neoplastic lesions may be practically
difficult and we suggest the utility of more noninvasive
ASL imaging methods to improve the reliability of our
current findings. Third, our study results might be
biased because of manual ROI placement, which usu-
ally affects all ROI-based studies, although previous
studies have shown that this method for the measure-
ment of maximal abnormality provides the highest
intra- and interobserver reproducibility in perfusion
measurements.39,40 On the other hand, Emblem et al.
reported higher interobserver agreement and diagnostic
accuracy of histogram analysis compared with the hot-
spot method.41 Fourth, ASL is an echo planar imaging-
based technique so is inherently associated with low
signal-to-noise ratio (SNR), lower temporal and spatial
resolution and greater susceptibility artifacts. We
used the 3D PCASL sequence, which is a fast spin
echo-based readout and gives a better SNR, fewer sus-
ceptibility artifacts and better labeling efficiency. Our
experience with ASL suggests that being non-contrast,
it can be useful in different pathologies at the cost of
only a few minutes.42,43
Conflict of interest
The authors declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
This research received no specific grant from any funding
agency in the public, commercial, or not-for-profit sectors.
Soni et al.
ORCID iD
Neetu Soni http://orcid.org/0000-0003-2082-1634
References
1. Omuro AM, Leite CC, Mokhtari K, et al. Pitfalls in the
diagnosis of brain tumours. Lancet Neurol 2006; 5:
937–948.
2. Law M, Yang S, Babb JS, et al. Comparison of cerebral
blood volume and vascular permeability from dynamic
susceptibility contrast-enhanced perfusion MR imaging
with glioma grade. AJNR Am J Neuroradiol 2004; 25:
746–755.
3. Wolf RL and Detre JA. Clinical neuroimaging using
arterial spin-labeled perfusion magnetic resonance ima-
ging. Neurotherapeutics 2007; 4: 346–359.
4. Grade M, Hernandez Tamames JA, Pizzini FB, et al.
A neuroradiologist’s guide to arterial spin labeling MRI
in clinical practice. Neuroradiology 2015; 57: 1181–1202.
5. Soni N, Dhanota DPS, Kumar S, et al. Perfusion MR
imaging of enhancing brain tumors: Comparison of arter-
ial spin labeling technique with dynamic susceptibility
contrast technique. Neurol India 2017; 65: 1046–1052.
6. Jiang J, Zhao L, Zhang Y, et al. Comparative analysis of
arterial spin labeling and dynamic susceptibility contrast
perfusion imaging for quantitative perfusion measure-
ments of brain tumors. Int J Clin Exp Pathol 2014; 7:
2790–2799.
7. Ata ES, Turgut M, Eraslan C, et al. Comparison between
dynamic susceptibility contrast magnetic resonance ima-
ging and arterial spin labeling techniques in distinguish-
ing malignant from benign brain tumors. Eur J Radiol
2016; 85: 1545–1553.
8. Cha S, Lupo JM, Chen MH, et al. Differentiation of
glioblastoma multiforme and single brain metastasis by
peak height and percentage of signal intensity recovery
derived from dynamic susceptibility-weighted contrast-
enhanced perfusion MR imaging. AJNR Am J
Neuroradiol 2007; 28: 1078–1084.
9. Law M, Cha S, Knopp EA, et al. High-grade gliomas and
solitary metastases: Differentiation by using perfusion
and proton spectroscopic MR imaging. Radiology 2002;
222: 715–721.
10. Weber MA, Zoubaa S, Schlieter M, et al. Diagnostic per-
formance of spectroscopic and perfusion MRI for distinc-
tion of brain tumors. Neurology 2006; 66: 1899–1906.
11. Bai Y, Shi D, Wang M, et al. Differentiation between
glioblastomas and solitary brain metastases using
pseudo-continuous arterial spin labeling perfusion MRI
in peritumoral edema. European Congress of Radiology
2013, poster number C-2569.
12. Tourdias T, Rodrigo S, Oppenheim C, et al. Pulsed arter-
ial spin labeling applications in brain tumors: Practical
review. J Neuroradiol 2008; 35: 79–89.
13. Gupta RK, Haris M, Husain N, et al. Relative cerebral
blood volume is a measure of angiogenesis in brain tuber-
culoma. J Comput Assist Tomogr 2007; 31: 335–341.
14. Sankhe S, Baheti A, Ihare A, et al. Perfusion magnetic
resonance imaging characteristics of intracerebral tuber-
culomas and its role in differentiating tuberculomas from
metastases. Acta Radiol 2013; 54: 307–312.
15. Chatterjee S, Saini J, Kesavadas C, et al. Differentiation
of tubercular infection and metastases presenting as ring
9
enhancing lesion by diffusion and perfusion magnetic res-
onance imaging. J Neuroradiol 2010; 37: 167–171.
16. Haller S, Zaharchuk G, Thomas DL, et al. Arterial spin
labeling perfusion of the brain: Emerging clinical appli-
cations. Radiology 2016; 281: 337–356.
17. Noguchi T, Yakushiji Y, Nishihara M, et al. Arterial
spin-labeling in central nervous system infection. Magn
Reson Med Sci 2016; 15: 386–394.
18. Floriano VH, Torres US, Spotti AR, et al. The role of
dynamic susceptibility contrast-enhanced perfusion MR
imaging in differentiating between infectious and neo-
plastic focal brain lesions: Results from a cohort of 100
consecutive patients. PLoS One 2013; 8: e81509.
19. Haris M, Gupta RK, Singh A, et al. Differentiation of
infective from neoplastic brain lesions by dynamic con-
trast-enhanced MRI. Neuroradiology 2008; 50: 531–540.
20. Hourani R, Brant LJ, Rizk T, et al. Can proton MR
spectroscopic and perfusion imaging differentiate
between neoplastic and nonneoplastic brain lesions in
adults? AJNR Am J Neuroradiol 2008; 292: 366–372.
21. Kong L, Chen H, Yang Y, et al. A meta-analysis of arter-
ial spin labelling perfusion values for the prediction of
glioma grade. Clin Radiol 2017; 72: 255–261.
22. Liang R, Wang X, Li M, et al. Meta-analysis of peritu-
moural rCBV values derived from dynamic susceptibility
contrast imaging in differentiating high-grade gliomas
from intracranial metastases. Int J Clin Exp Med 2014;
7: 2724–2729.
23. Bauer AH, Erly W, Moser FG, et al. Differentiation of
solitary brain metastasis from glioblastoma multiforme:
A predictive multiparametric approach using combined
MR diffusion and perfusion. Neuroradiology 2015; 57:
697–703.
24. Zhang M and Olsson Y. Hematogenous metastases of the
human brain—Characteristics of peritumoral brain
changes: A review. J Neurooncol 1997; 35: 81–89.
25. Lin L, Xue Y, Duan Q, et al. The role of cerebral blood
flow gradient in peritumoral edema for differentiation of
glioblastomas from solitary metastatic lesions.
Oncotarget 2016; 7: 69051–69059.
26. Sunwoo L, Yun TJ, You SH, et al. Differentiation of
glioblastoma from brain metastasis: Qualitative and
quantitative analysis using arterial spin labeling MR ima-
ging. PloS One 2016; 11: e0166662.
27. Long DM. Capillary ultrastructure in human metastatic
brain tumors. J Neurosurg 1979; 51: 53–58.
28. Hossman KA and Blöink M. Blood flow and regulation
of blood flow in experimental peritumoral edema. Stroke
1981; 12: 211–217.
29. Bulakbasi N, Kocaoglu M, Farzaliyev A, et al.
Assessment of diagnostic accuracy of perfusion MR ima-
ging in primary and metastatic solitary malignant brain
tumors. AJNR Am J Neuroradiol 2005; 26: 2187–2199.
30. Batra A and Tripathi RP. Perfusion magnetic resonance
imaging in intracerebral parenchymal tuberculosis:
Preliminary findings. J Comput Assist Tomogr 2003; 27:
882–888.
31. Gupta RK, Haris M, Husain N, et al. Relative cerebral
blood volume is a measure of angiogenesis in brain tuber-
culoma. J Comput Assist Tomogr 2007; 31: 335–341.
32. Haris M, Gupta RK, Husain M, et al. Assessment of
therapeutic response in brain tuberculomas using serial
dynamic contrast-enhanced MRI. Clin Radiol 2008; 63:
562–574.
10
33. Gupta RK, Awasthi R, Garg RK, et al. T1-weighted
dynamic contrast-enhanced MR evaluation of different
stages of neurocysticercosis and its relationship with
serum MMP-9 expression. AJNR Am J Neuroradiol
2013; 34: 997–1003.
34. Holmes TM, Petrella JR and Provenzale JM. Distinction
between cerebral abscesses and high-grade neoplasms by
dynamic susceptibility contrast perfusion MRI. AJR Am
J Roentgenol 2004; 183: 1247–1252.
35. Erdogan C, Hakyemez B, Yildirim N, et al. Brain abscess
and cystic brain tumor: Discrimination with dynamic sus-
ceptibility contrast perfusion-weighted MRI. J Comput
Assist Tomogr 2005; 29: 663–667.
36. Cha S, Pierce S, Knopp EA, et al. Dynamic contrast-
enhanced T2*-weighted MR imaging of tumefactive
demyelinating lesions. AJNR Am J Neuroradiol 2001;
22: 1109–1116.
37. Blasel S, Pfeilschifter W, Jansen V, et al. Metabolism and
regional cerebral blood volume in autoimmune inflam-
matory demyelinating lesions mimicking malignant gli-
omas. J Neurol 2011; 258: 113–122.
38. Hiremath SB, Muraleedharan A, Kumar S, et al.
Combining diffusion tensor metrics and DSC perfusion
The Neuroradiology Journal 0(00)
imaging: Can it improve the diagnostic accuracy in dif-
ferentiating tumefactive demyelination from high-grade
glioma? AJNR Am J Neuroradiol 2017; 38: 685–690.
39. Young R, Babb J, Law M, et al. Comparison of region-
of-interest analysis with three different histogram analysis
methods in the determination of perfusion metrics in
patients with brain gliomas. J Magn Reson Imaging
2007; 26: 1053–1063.
40. Wetzel SG, Cha S, Johnson G, et al. Relative cerebral
blood volume measurements in intracranial mass lesions:
Interobserver and intraobserver reproducibility study.
Radiology 2002; 224: 797–803.
41. Emblem KE, Nedregaard B, Nome T, et al. Glioma grad-
ing by using histogram analysis of blood volume hetero-
geneity from MR-derived cerebral blood volume maps.
Radiology 2008; 247: 808–817.
42. Soni N, Jain A, Kumar S, et al. Arterial spin labeling
magnetic resonance perfusion study to evaluate the
effects of age and gender on normal cerebral blood
flow. Neurol India 2016; 64 (Suppl): S32–S38.
43. Soni N. Author’s reply: Arterial spin labeling. Neurol
India 2018; 66: 285.