European Journal of Radiology 81 (2012) 508513

Contents lists available at ScienceDirect

European Journal of Radiology

journal homepage: www.elsevier.com/locate/ejrad

F-18 FDG PET-CT in patients with recurrent glioma: Comparison with contrast
enhanced MRI
Amburanjan Santra a,1 , Rakesh Kumar b, , Punit Sharma b , Chandrashekhar Bal b ,
Atin Kumar c , Pramod Kumar Julka d , Arun Malhotra b
a

Department of Nuclear Medicine, Medical College Kolkata, Kolkata, India

Department of Nuclear Medicine, All India Institute of Medical Sciences, New Delhi, India
c
Department of Radio-diagnosis, All India Institute of Medical Sciences, New Delhi, India
d
Department of Radiotherapy, All India Institute of Medical Sciences, New Delhi, India
b

a r t i c l e

i n f o

Article history:
Received 8 October 2010
Accepted 3 January 2011
Keywords:
Glioma
Recurrence
PET-CT
MRI
FDG

a b s t r a c t
Purpose: The purpose of the study was to compare the efcacies of FDG PET-CT and contrast enhanced
MRI in detection of recurrent gliomas.
Methods: Ninety histopathologically proven glioma patients with clinical suspicion of recurrence were
evaluated. All patients underwent FDG PET-CT scan and contrast enhanced MRI. Combination of clinical
follow up, repeat imaging and biopsy (when available) was taken as gold standard.
Results: Based on gold standard criteria, 59 patients were positive and 31 patients were negative for
recurrence. Overall sensitivity and specicity of FDG PET-CT were 70% and 97% respectively whereas that
for contrast enhanced MRI was 95% and 23%. FDG PET-CT also has higher accuracy (80%) as compared to
MRI (70%). FGD PET-CT has lower sensitivity than MRI in all grades, except for Grade II gliomas where
their sensitivities are comparable (95% and 90%). Very low specicity of MRI was observed in all grades
of tumour (1833%). In contrast the specicity of FDG PET-CT was high across all grades (83100%).
Conclusion: FDG PET-CT is a highly specic modality for detecting recurrence in patients with gliomas
and can effectively exclude post therapy changes.
2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction
Recurrence in glioma may occur after complete removal of
tumour or stabilization of tumour with treatment. In more than
90% cases, initial site of recurrence is at or within 2 cm of the
primary site [1]. Radiation injury and postoperative changes are
also usually found around the tumour bed. Moreover, radiation
necrosis and recurrent tumour can frequently coexist [2]. Differentiation between recurrence and radiation necrosis is crucial since
the two entities have completely different management and prognosis. Radiation necrosis can produce disruption in the bloodbrain
barrier (BBB) by vascular and astrocytic damage, and thus contrast
enhancement, edema, and cortical dysfunction that are indistinguishable from recurrent tumour on conventional computed
tomography (CT) or magnetic resonance imaging (MRI) [3,4].

Corresponding authorat: E-81, Ansari Nagar (East), AIIMS Campus, New Delhi
110029, India. Tel.: +91 11 26588017; fax: +91 11 26588663.
E-mail addresses: a ranjan santra@yahoo.co.in (A. Santra), rkphulia@yahoo.com
(R. Kumar).
1
Department of Nuclear Medicine, Medical College Kolkata, Kolkata 700073,
India. Tel.: +91 9433812043; fax: +91 11 26588663.
0720-048X/$ see front matter 2011 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ejrad.2011.01.080

Because of its high soft-tissue contrast, MRI has become the

rst-line method of choice for the assessment of brain tumours.
Though conventional contrast enhanced MRI allows for excellent
visualization of these tumours, there are still certain limitations.
This particularly includes dening tumour extension and grade, as
well as differentiating tumour recurrence from necrosis or scar.
Recurrent brain tumour usually demonstrates extensive edema
and enhances with intravenous paramagnetic contrast on MRI.
However, necrosis induced by therapy or occurring spontaneously
during tumour progression may also show contrast enhancement
and hence cannot be distinguished reliably from a solid tumour
after therapy on MRI [5,6].
Positron
emission
tomography
(PET)
with
F18uorodeoxyglucose (FDG) is widely used to provide information
regarding glucose metabolism in a wide variety of tumours,
including gliomas. FDG accumulation in gliomas is not dependent
on bloodbrain barrier (BBB) disruption [7]. FDG PET imaging
of brain tumours have been applied for dening the extent of
tumour, tumour grading, prognostication, treatment response
evaluation, as well as to differentiate between radiation necrosis
and recurrent tumour, and between benign and malignant lesions
[810]. The use of FDG PET to distinguish recurrent tumour from
radiation necrosis appears promising [11,12]. FDG PET however

A. Santra et al. / European Journal of Radiology 81 (2012) 508513

509

has some limitations in characterizing brain tumours because of

the high basal glucose metabolic rate of normal brain. Except for
very high grade tumours, most of the tumours have equal or lower
FDG uptake than that of normal brain parenchyma [13]. Under
these circumstances, fused PET-CT improves the sensitivity to
certain extent. Some studies have compared the roles of FDG PET
and contrast enhanced MRI for differentiating recurrence from
radiation necrosis. But because of the lesser number of patients,
results of these studies were variable [14,15]. The present study
was aimed to compare the diagnostic capabilities of FDG PET-CT
and MRI in detection of glioma recurrence in a large series of
patients.

2.4. Interpretation of images

2. Materials and methods

2.5. Gold standard

This prospective study was conducted after obtaining prior

approval of institutional review board. Patients were recruited in
the study between August, 2006 and February, 2008. A written
informed consent was obtained from all patients.

Combination of clinical follow up, repeat imaging and biopsy

(when available) were taken as gold standard. Patients who had disease related adverse event, progressive disease on imaging and/or
positive biopsy were taken positive for recurrence.

2.1. Patients
A total of 90 patients with glioma were included in the study.
Inclusion criteria were histopathologically proven glioma, previous treatment with surgery and/or radiotherapy with or without
chemotherapy and clinical suspicion of recurrence. Exclusion criteria were primary brain tumour other than gliomas and proven
malignancy of other sites where from metastasis can occur to brain.
All patients underwent FDG PET-CT scan and conventional MRI with
contrast in our institute within one week span.
2.2. F-18 FDG PET-CT imaging
PET-CT scans were taken on a dedicated PET-CT scanner present
in our institute (BIOGRAPH 2, SEIMENNS, Germany). It has LSO
(Lutetium oxyorthosilicate, Lu2 SiO5 : Ce) detectors with attenuation coefcient 0.89 cm1 , photofraction 30%, decay constant 40 ns
and energy resolution at 511 KeV, %FWHM is 10 with spatial resolution of 6 mm. All patients fasted for at least 4 h before the test.
Blood glucose level was below 140 mg/dl in all patients. A dose
of 370 MBq (10 mCi) of FDG was injected intravenously and the
patients rested in a quiet room. After a 4560-min uptake period,
patients were taken for scanning. In the PET-CT system, CT acquisition was performed on spiral dual slice CT with a slice thickness
of 4 mm and a pitch of 1. Image was acquired using a matrix of
512 512 pixels and pixel size of about 1 mm. After transmission
scan, 3D PET acquisition was done for 35 min per bed position
for one/two bed position. PET data were acquired using matrix of
128 128 pixels with a slice thickness of 1.5 mm. CT based attenuation correction of the emission images was employed. PET images
were reconstructed by iterative method ordered subset expectation
maximization (2 iterations and 8 subsets). Reconstructed images
were displayed and analyzed in transverse, sagittal and coronal
views.
2.3. MRI imaging
MR imaging were performed on a 1.5-T clinical MR imaging unit (Sonata/Avanto, Siemens, Germany) and images were
acquired using a standard head coil. Transaxial T1-weighted and
T2 weighted images were obtained from the second cervical vertebral body to the vertex. Slice thickness was adjusted to 1 mm.
Contrast-enhanced images were also obtained after intravenous
administration of Gadopentetate dimeglumine (Gd-DTPA) at a dose
of 0.1 mmol/kg using standard procedures.

PET-CT images were evaluated independently by two experienced Nuclear Medicine physicians. They were blinded to
the clinical and structural imaging ndings. PET-CT images
were interpreted as positive for recurrent tumour if there
was a denite lesion on CT images which was hypermetabolic/isometabolic/hypometabolic on PET images or if there was
an increased focal FDG uptake without any clearly discernible
lesion on CT. MR Radiologist was also blinded to clinical and PETCT ndings. Gd-DTPA enhancing lesions were considered positive
for recurrence in MRI images.

2.6. Statistical analysis

Various descriptive statistics such as mean, median, range and
standard deviation (SD) were used to describe the baseline clinical
and demographic proles of all the patients. Sensitivity, specicity,
positive predictive value (PPV), negative predictive value (NPV) and
accuracy with 95% condence interval were calculated for each
modality. McNemar test was used to see the difference between
diagnostic accuracies of the modalities. Statistical package STATA
8.0(Stata Corporation, College Station, TX, USA) and SPSS 16.0(SPSS
Inc., Chicago, IL, USA) were used for all the statistical analyses.
3. Results
3.1. Patient characteristics
A total of 90 patients with mean age of 36.79 (1.25) years and
age range of 1268 years were evaluated. Patient characteristics
including sex, primary location of tumour, histology, grade, and
primary therapy are summarized in Table 1.
3.2. Follow up outcome
On the basis of clinical follow up, repeat imaging and/or biopsy
59 patients were positive and 31 patients were negative for recurrence. All patients were followed up for a period of at least 6
months. Twenty eight out of the total 90 patients had a disease
related adverse event. Five patients were re-operated and all of
them were found positive at histopathology (GBM-2 and anaplastic astrocytoma-3). Repeat imaging at follow up was done in the
remaining 55 patients and showed persistent or progressive disease in 26 patients.
3.3. Results of MRI
MRI was positive for recurrence in 80 patients and negative in
10 patients. Table 2 shows the overall and grade wise sensitivity, specicity, positive predictive value (PPV), negative predictive
value (NPV) and accuracy of MRI. Mean lesions size obtained from
contrast enhanced area in positive cases was 3.41 cm (0.40 cm)
and median size was 3.3 cm (range 1.17.2 cm).
3.4. Results of FDG PET-CT
FDG PET-CT was positive for recurrence in 42 and negative in
48 patients. Table 2 shows the overall and grade wise sensitiv-

Contrast enhancement of brain tumours depends on BBB damage. Necrosis induced by therapy or occurring spontaneously
during tumour progression may also show contrast enhancement
and hence cannot be distinguished reliably from a recurrent solid
tumour after therapy [46]. In addition, uptake of contrast agent
may be substantially reduced by dexamethasone given for reducing cerebral edema [16]. The fact that morphological imaging often
does not adequately reect the underlying tumour biology and its
metabolic activity imposes considerable demand to develop alter-

40% (7.383)
100% (39.6100)
100% (19.8100)
57.1% (20.288.2)
67%

MRI
PET/CT
MRI

95% (73.199.7)
17.6% (4.744.2)
57.6% (39.474)
75% (21.998.7)
59%
90% (66.998.2)
100% (56.1100)
100% (78.1100)
77.8% (40.296.1)
95%

PET/CT

Grade II

MRI

100% (81.5100)
33.3% (675.9)
84.6% (64.395)
100% (19.8100)
86%
68.2% (45.185.3)
83.3% (36.599.1)
93.8% (67.799.7)
41.7% (16.571.4)
71%

PET/CT

CI: condence interval; PPV: positive predictive value; NPV: negative predictive value; GBM: glioblastoma multiforme.

4. Discussion

83.3% (50.997.1)
25% (1.378.1)
76.9% (4693.8)
33.3% (1.897.5)
67%

Overall, MRI was highly sensitive (95%) and with poor specicity
(23%) for detection of recurrence. In contrast, FDG PET-CT has lower
sensitivity (70%) and higher specicity (97%). Comparison of overall
and grade wise sensitivity, specicity, PPV, NPV and accuracy of
FDG PET-CT and MRI are given in Table 2.
FDG PET-CT and MRI ndings were concordant in 52 patients
(Fig. 1). There was discordance between the ndings of MRI and
FDG PET-CT in remaining 38 patients (Fig. 2). On McNemar analysis
the difference was statistically signicant (p < 0.001). FDG PET-CT
was negative in all of these 38 patients while MRI was positive
in all (Table 3). Twenty out of these 38 cases were true negative
based on Gold standard. MRI was false negative only in 3 patients
(GBM-2; Grade II-1). In contrast FDG PET-CT was false negative in
18 patients (GBM-6; Grade III-7; Grade II-2; Grade I-3). There were
24 false positive cases on MRI but only one false positive case in
FDG PET-CT (Table 3). Moreover, FDG PET-CT was able to correctly
delineate mixed lesion of recurrent tumour and radiation necrosis
in 11 patients (Figs. 1 and 3).

MRI

3.5. Comparison between MRI and FDG PET-CT

PET/CT

ity, specicity, positive predictive value (PPV), negative predictive

value (NPV) and accuracy of FDG PET-CT.

MRI

GBM: glioblastoma multiforme; Sx: surgery; RT: radiotherapy.

PET/CT

51
31
3
5

Grade III

48
34
8 (2 + 6)

GBM

9
37
28
16

Overall

7
24
33
18
3
5

Grade I/others

16
35
24
10
5

50% (22.377.7)
100% (39.6100)
100% (51.7100)
40% (13.772.6)
62.5%

66
24

94.9% (84.998.7)
22.9% (10.341.5)
70% (58.679.5)
70% (35.491.9)
70%

No. of Patients

Parameter

Sex
Male
Female
Type of glioma
GBM
Astrocytoma
Oligodendroglioma
Mixed glioma
Others
Primary site of glioma
Corpus callosal
Temporal
Frontal
Parietal
Multilobar
Posterior fossa
Histopathologic grade of primary tumour
Grade I
Grade II
Grade III
Grade IV
Primary treatment received
Sx + RT
Sx + RT + CT
Sx or RT
Treatment given after evaluation of recurrence
No treatment
Temozolamide
RT + CT
Re-Sx

Table 2
Sensitivity, specicity, PPV, NPV and accuracy of FDG PET-CT and MRI in detection of recurrence glioma; overall and in patients with different histopathological grades (with 95% CI).

Parameter

69.5% (5680.5)
96.8% (81.599.8)
97.6% (85.999.9)
62.5% (47.375.5)
80%

Table 1
Patients characteristics.

100% (46.3100)
25% (1.378.1)
62.5% (25.989.8)
100% (5.5100)
67%

A. Santra et al. / European Journal of Radiology 81 (2012) 508513

Sensitivity
Specicity
PPV
NPV
Accuracy

510

A. Santra et al. / European Journal of Radiology 81 (2012) 508513

511

Fig. 1. A 35 year female with right temporal Grade II astrocytoma, primarily treated with Surgery and Radiotherapy presented with severe headache and seizures. (A)
Transaxial T1W MRI showing large hypointense temporoparietal lesion, (B) Transaxial T1W contrast enhanced MRI showing intense contrast enhancement; ndings are
suggestive of recurrent tumour,(C) FDG PET and (D) FDG PET-CT images showing temporoparietal lesion with intense FDG uptake; ndings are positive for recurrence. The
lesion seen in MRI includes both viable tumour and necrotic area. FDG PET-CT effectively delineate viable tumour from necrotic area. High FDG uptake in recurrent tumour
suggested higher grade transformation. Patient died within 3 weeks of PET-CT scan.

Fig. 2. A 66 years old male patient of right frontal GBM primarily treated with complete surgical removal, radiotherapy and temozolamide. (A)Transaxial T1W contrast
enhanced MRI showing contrast enhancing lesion in the right frontoparietal area, suggestive of recurrent tumour. (B) FDG PET-CT was negative, suggestive of radiation
necrosis. This patient was followed up for >1.5 years and was ne without any further treatment.

512

A. Santra et al. / European Journal of Radiology 81 (2012) 508513

Table 3
FDG PET-CT and MRI ndings in all patients with different primary grades.
Parameter

TP
FP
TN
FN

Overall

GBM

Grade III

Grade II

Grade I/others

PET/CT

MRI

PET/CT

MRI

PET/CT

MRI

PET/CT

MRI

PET/CT

MRI

41
1
30
18

56
24
7
3

6
0
4
6

10
3
1
2

15
1
5
7

22
4
2
0

18
0
17
2

19
14
3
1

2
0
4
3

5
3
1
0

TP: true positive; TN: true negative; FP: false positive; FN: false negative; GBM: glioblastoma multiforme.

Fig. 3. A 40 years male with right frontal Oligoastrocytoma, primarily treated with Surgery and Radiotherapy. The patient was symptomatic at the time of evaluation. (A)
Transaxial T1W contrast enhanced MRI showing contrast enhancing lesion in the posterior margin of right posterior frontal gliotic cavity, (B) FDG PET and (C) FDG PET-CT
images reveal two foci of intense FDG uptake at the posterior margin of the posterior frontal gliotic cavity suggestive of recurrent tumour and an area of radiation necrosis
in between.

native imaging modalities to correctly characterize the lesions. PET

with F18-FDG can identify the metabolic activity of the lesion, so
it can potentially differentiate recurrent tumour from post radiation or post surgical changes [11,12]. Steroids, when used to
control cerebral edema have no effect on the FDG uptake though
it affects grey and white matter differentiation [17]. However,
FDG PET has certain limitations. Firstly, grey matter of the brain
also show increased FDG uptake. Secondly, except for very highgrade tumours (Grade IV-Glioblastoma) most of the tumours are
isometabolic or hypometabolic compared to adjacent grey matter; hence it is very difcult to differentiate them from normal
brain tissue. Fusion imaging with CT can overcome some of these
limitations. As per published literature overall sensitivity of FDG
PET is about 8090% and specicity varied from 50 to 90% [11,12].
Belohlavek et al. evaluated 29 patients to compare the role of FDG
PET and MRI in the diagnosis of recurrent glioma and concluded that
MRI has higher sensitivity (95.8%) and diagnostic accuracy (86%),
whereas FDG PET has better specicity (83.3%) [14]. Another study
by Chen et al. in 23 patients of suspected recurrence showed that
overall diagnostic accuracy of FDG PET is better than that of MRI
[15]. In our study, we found that MRI is highly sensitive (95%) for
detection of recurrence but it has very low specicity and a large
number of false positive results which impacts its overall diagnostic accuracy. Being a metabolic marker, FDG shows no uptake in
the necrotic area and hence it can clearly exclude the possibilities
of viable tumour in that area. This is reected in very high specicity of FDG PET-CT (97%). The diagnostic accuracy was also better
than that of MRI (80% vs. 70%) in our study. On subgroup analysis
specicity of MRI remained very poor in patients with tumours of
all grades.
High-grade gliomas have very poor survival rates even with
optimal treatment [18]. Grade I tumours are very slow growing and
usually cured with primary treatment. On the contrary, recurrent
Grade II gliomas invariably transform to anaplastic grade. These

groups of patients need more attention as early detection and treatment of recurrence can improve the survival [19,20]. Interestingly,
FDG PET-CT had 90% sensitivity and 95% accuracy in detecting
recurrence in the patients with grade II gliomas along with very
high specicity (100%). Another advantage of FDG PET-CT is in
cases where there is mixed lesion of recurrent tumour and radiation necrosis. In these cases FDG PET-CT can clearly delineate
the metabolically active area and necrotic area so that the actual
tumour volume could be determined. This can accurately guide
radiotherapy planning or surgery.
Although FDG PET-CT fares well in overall diagnosis, its high
false negative rate is of a major concern. Moreover due to high normal brain uptake of FDG and lack of denite lesion in CT, some
lesions were missed. Because of lower sensitivity FDG PET-CT it
may not be wise to use it as a primary screening modality for detection of recurrence in glioma patients. It might be prudent to use
it to characterize any abnormal lesion found on MRI. The present
study has certain limitations. Although total number of patients
in this study was enough for statistical analysis, the numbers of
patients in each subgroup based on histological grade were relatively less. Hence any inference derived from this study needs to be
revalidated with a larger study with sufcient number of patients
in each subgroup. Also, histopathological conrmation of recurrent
tumour was available only in 5 patients.

5. Conclusion
FDG PET-CT is a highly specic modality for detecting recurrence in patients with gliomas and can effectively exclude radiation
necrosis and other therapy related changes. However, MRI has
higher sensitivity, therefore, a combination of these two modalities or better still hybrid imaging in the form of PET-MRI might be
more useful in these group of patients.

A. Santra et al. / European Journal of Radiology 81 (2012) 508513

References
[1] Wallner KF, Galicich JH, Krol G, Arbit E, Malkin MG. Patterns of failure following
treatment for glioblastoma and anaplastic astrocytoma. Int J Radiol Oncol Biol
Phys 1989;16:1409.
[2] Burger PC. Malignant astrocytic neoplasms: classication, pathologic anatomy,
and response to treatment. Semin Oncol 1986;13:1626.
[3] Mikhael MA. Radiation necrosis of the brain: correlation between, computed tomography, pathology and dose distribution. J Comput Assist Tomogr
1978;2:7180.
[4] Van Dellen JR, Danziger A. Failure of computerized tomography to differentiate
between radiation necrosis and cerebral tumour. S Afr Med J 1978;53:1712.
[5] Price SJ. The role of advanced MR imaging in understanding brain tumour
pathology. Br J Neurosurg 2007;21:56275.
[6] Dooms GC, Hecht S, Brant-Zawadzki M, Berthiaume Y, Norman D, Newton TH.
Brain radiation lesions: MR imaging. Radiology 1986;158:14955.
[7] Delbeke D, Meyerowitz C, Lapidus RL, et al. Optimal cutoff levels of F-18 uorodeoxyglucose uptake in the differentiation of low grade from high-grade
brain tumors with PET. Radiology 1995;195:4752.
[8] Roelcke U, Rad EW, von Ammon K, Hausmann O, Maguire RP, Leenders
KL. Alteration of bloodbrain barrier in human brain tumors: comparison of
[18 F]uorodeoxyglucose, [11 C]methionine and rubidium-82 using PET. J Neurol
Sci 1995;132:207.
[9] Kaschten B, Stevenaert A, Sadzot B, et al. Preoperative evaluation of 54 gliomas
by PET with uorine-18-uorodeoxyglucose and/or carbon-11-methionine. J
Nucl Med 1998;39:77885.
[10] Wong TZ, van der Westhuizen GJ, Coleman RE. Positron emission tomography
imaging of brain tumors. Neuroimaging Clin N Am 2002;12:61526.

513

[11] Ricci PE, Karis JP, Heiserman JE, Fram EK, Bice AN, Drayer BP. Differentiating
recurrent tumor from radiation necrosis: time for re-evaluation of positron
emission tomography. Am J Neuroradiol 1998;9:40713.
[12] Asensio C, Perez-Castejon MJ, Maldonado A, et al. The role of PET-FDG in questionable diagnosis of relapse in the presence of radionecrosis of brain tumors.
Rev Neurol 1998;27:44752.
[13] Di Chiro G, DeLaPaz RL, Brooks RA, et al. Glucose utilization of cerebral gliomas
measured by (18F)uorodeoxyglucose and positron emission tomography.
Neurology 1982;32:13239.
[14] Belohlavek KJ, Vymazal J, Dbaly V, Tovarys F. The diagnostics of recurrent
gliomas using FDG-PET: still questionable? Nucl Med Rev Cent East Eur
2002;5:12730.
[15] Chen YR, Li WX, Lin YR, Chen LH. Value of 18F-FDG PET imaging in diagnosing
tumor residue of intracranial glioma after surgery and radiotherapy. Ai Zheng
2004;23(10):12102.
[16] Ostergaard L, Hochberg FH, Rabinov JD, et al. Early changes measured by magnetic resonance imaging in cerebral blood ow, blood volume, and bloodbrain
barrier permeability following dexamethasone treatment in patients with
brain tumors. J Neurosurg 1999;90:3005.
[17] Kaye A, Laws E. Brain tumors: an encyclopedic approach. 2nd ed. London:
Harcourt Publishers Ltd.; 2001.
[18] Wang SX, Boethius J, Ericson K. FDG-PET on irradiated brain tumor: ten years
summary. Acta Radiol 2006;47:8590.
[19] Devaux BC, Ofallon JR, Kelly PJ. Resection, biopsy and survival in malignant
glioma, a retrospective study of clinical parameters, therapy and outcomes. J
Neurosurg 1993;78:76775.
[20] Okamoto Y, Di Patre PL, Burkhard C, et al. Population-based study on incidence,
survival rates, and genetic alterations of low-grade diffuse astrocytomas and
oligodendrogliomas. Acta Neuropathol 2004;108:4956.