Neuroimaging of pediatric infratentorial tumors and the value of diffusion-weighted imaging (DWI) in determining tumor grade

Abstract

Background

Diffusion-weighted imaging (DWI) provides information about the cellular density of tumors. This feature is useful in grading and identifying different tumor types.

Purpose

To assess the value of diffusion restriction and apparent diffusion coefficient (ADC) values in differentiating pediatric infratentorial tumors.

Material and Methods

This was a retrospective review of the magnetic resonance imaging (MRI) of 82 children (age range 1–16 years) with infratentorial tumors. Histopathological grading after surgical excision/biopsy was categorized as low grade (WHO grades I and II) (n = 31; 29 pilocytic astrocytomas, 2 ependymomas) and high grade (WHO grade III and IV) (n = 51; 40 medulloblastomas, 8 anaplastic ependymomas, 1 anaplastic astrocytoma, 2 atypical rhabdoid teratoid tumors [ATRT]). MRI features and ADC values were compared among tumor types and grades using a two-tailed t test, Mann–Whitney U test for continuous data and Chi-square test for categorical variables.

Results

Diffusion restriction and low ADC value was a feature of high-grade tumors (P<0.001). The mean ADC values of the low-grade and high-grade tumors were 1.567 × 10⁻³mm²/s and 0.661 × 10⁻³mm²/s, respectively. Using 0.9 × 10⁻³mm²/s as the cut-off value, the sensitivity, specificity, positive and negative predictive values for differentiating the grades was 87%, 100%, 100%, and 81.8%, respectively. Significant differences were found between the mean ADC values of the individual tumor types (P<0.05), except between medulloblastoma and ATRT.

Conclusion

ADC values and visual assessment of diffusion restriction are useful in tumor grading. The individual tumor types can be identified by an algorithmic approach, using DWI in conjunction with other described MRI features.

Keywords

Magnetic resonance imaging diffusion-weighted imaging infratentorial tumors grading children

Introduction

Infratentorial tumors account for 45–60% of brain tumors in children (1), the common tumors being juvenile pilocytic astrocytoma (JPA), medulloblastoma and ependymoma. The other less common tumors include brainstem glioma, atypical rhabdoid teratoid tumor (ATRT), hemangioblastoma, metastases, and extra-axial lesions. The peak incidence is in the age range of 5–13 years, 3–4 years, 3–5 years, and <2 years for pilocytic astrocytoma, medulloblastoma, ependymoma, and ATRT, respectively (2).

In view of different imaging protocols, treatment options and prediction of outcome required for these tumors, preoperative differentiation becomes essential (3).

A conventional magnetic resonance imaging (MRI) study allows evaluation of tumor location and its extent accurately. However, the commonly used parameters such as signal intensity, homogeneity, and contrast enhancement can be atypical and is often limiting in determining tumor type and grade (4,5).

Diffusion-weighted imaging (DWI) is now widely used and forms a part of routine MRI protocol. Hyperintensity on DWI and the corresponding apparent diffusion coefficient (ADC) value provides information about the cellular density of tumors (6). As the grade of tumor is directly related to its cellular content, studies have tried to find a correlation between tumor grade and ADC values, but with conflicting results (1,5,7 –12). The aim of the present study was to establish the ADC values of common pediatric infratentorial tumors and to assess the significance of diffusion restriction and ADC value as a quantitative tool in grading and differentiating specific tumor types.

Material and Methods

This was an institutional review board approved retrospective review of brain MRI scans of children aged < 16 years with posterior fossa tumors, over a period of three years. The Departments of Radiology, Neurosurgery, and Pathology in a 2800-bed tertiary care center undertook the study. All children had a preoperative MRI scan and histologically proven diagnosis, by either surgical excision or biopsy. The tumors included were medulloblastoma, pilocytic astrocytoma, ependymoma, ATRT, and anaplastic varieties of the above. Grading was performed according to the World Health Organization (WHO) criteria.

The MRI was performed using a 1.5-T scanner (Siemens, Avanto, Germany). The examination protocol included TSE T2-weighted (T2W) images, fluid-attenuated inversion recovery images (FLAIR), unenhanced and contrast-enhanced T1-weighted (T1W) images carried out in three planes, and diffusion-weighted images (DWI). The following parameters were used: matrix = 256 × 256 and 256 × 192; field of view (FOV) = 220–230 mm; slice thickness = 3–5 mm; and no slice gap. DWI was performed with the following parameters: TR = 2857 ms; TE = 102 ms; matrix = 116 × 85, FOV = 230 mm; slice thickness = 5 mm; b values = f 0 and 1000 mm²/s. ADC map was generated using the b values.

Image analysis

Two radiologists with 6 and 14 years of experience in diagnostic radiology, the latter being a specialist neuroradiologist, reviewed the MRI studies on a PACS workstation (provided by GE Healthcare, Barrington, IL, USA). The reviewers were blinded to the histopathological report and surgical notes. The following aspects of the tumor were recorded in consensus between the radiologists: size; location (cerebellar hemisphere, midline [fourth ventricle, vermis], brain stem or cerebello-pontine angle); qualitative assessment of signal intensity on T1W, T2W images, and DWI (restricted, isointense, or no restriction of diffusion); contrast enhancement (mild, moderate, intense, or no enhancement); presence of cystic components; hemorrhagic components (on susceptibility-weighted imaging [SWI]); and extension through foramina of Luschka and Magendie. The qualitative assessment was done using gray matter as the reference standard and in case of discrepancy a consensus reading was carried out.

ADC analysis using parametric map

For the measurement of ADC values, three elliptical regions of interest (ROIs) were placed in different locations to include most of the solid components of the tumor and their mean value was calculated. In case of enhancing tumors, the ROI was placed in the solid enhancing region. The size of the ROI was uniformly maintained in the range of 75–100 mm². Cystic, hemorrhagic, necrotic areas and tumor-related edema were avoided while placing the ROI. In doubtful cases, the ROI placement was done in consensus between both radiologists. Fig. 1 depicts a sample image of the ROI measurement

Fig. 1.

T1W axial post-contrast (a), DWI (b), and corresponding ADC (c) of a medulloblastoma depicting an example of region of interest measurement on the ADC map. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; T1W, T1-weighted.

Statistical analysis

Data were screened for outliers and extreme values using Box-Cox plot and histogram (for shape of the distribution). Summary statistics were used for reporting demographic and clinical characteristics, t-test was used for analysis of continuous data with normal distribution and Mann–Whitney U test for data with non-normal distribution. Chi-square test was performed for categorical variables and the outcome variables grade. Differences were considered significant at P < 0.05. All the statistical analyses were performed using SPSS 16.0 (SPSS Analytics, Chicago, IL, USA).

Results

A total of 82 children were included (57 boys, 25 girls; mean age = 8.24 years; age range = 1–16 years).

The distribution of the tumors was as follows: medulloblastoma (n = 40, 48.8%); pilocytic astrocytoma (n = 30, 36.6%); ependymoma (n = 10, 12.2%); and ATRT (n = 2, 2.4%). All the medulloblastomas and ATRT were WHO grade IV tumors. Of the pilocytic astrocytomas, 29 (96.6%) were WHO grade I while one had anaplastic features. Eight (80%) ependymomas were of the anaplastic type classified as WHO grade III and 2 (20%) were WHO grade II.

Their age-wise distribution was as follows: medulloblastoma 2–15 years (mean age = 9.05 years); pilocytic astrocytoma 2–16 years (mean age = 7.8 years); ependymoma 1–16 years (mean age = 7.4 years); and ATRT 2–3 years (mean age = 2.5 years).

Of the 40 medulloblastomas, 37 were midline (27 [67.5%] in the fourth ventricle, nine vermian, one brainstem), two at the cerebellopontine (CP) angle, and one in the cerebellar hemisphere. Of the pilocytic astrocytomas, 14 (46.7%) were from the vermis, 8 (26.7%) were in the cerebellar hemisphere, and four each in the brainstem and fourth ventricle. Most of the ependymomas, i.e. 8 (80%) were in the fourth ventricle and 2 (20%) in the cerebellar hemisphere. The ATRT were located in the fourth ventricle and vermis.

MRI features of the pediatric infratentorial tumors

Medulloblastoma: T2W MRI revealed intermediate signal intensity tumors in 19 (47.5%) patients and hyperintense tumors in 21 (52.5%) patients. Enhancement was variable and not a distinguishing feature. Thirty-nine (97.5%) had cystic components. Extension through foramina was present in 21 (58%) patients.

Ependymoma: Eight (80%) were T2W hyperintense while two were isointense. Of the tumors, 6 (60%) showed moderate enhancement and had cystic components. Extension through foramina was present in most tumors (90%).

Pilocytic astrocytoma: All were hyperintense on T2W MRI. Contrast enhancement was present in all cases, with 18 (60%) showing intense enhancement. All tumors had cystic areas; 20 (66.7%) had cystic components within the tumor while 10 (33.3%) were predominantly cystic with a mural nodule.

ATRT: There were two tumors, both of which were T2W hyperintense, with moderate-intense enhancement and cystic components. One showed extension across the foramen of Luschka.

Fig. 2 depicts examples of characteristic appearances of pilocytic astrocytoma, ependymoma, and medulloblastoma.

Fig. 2.

Characteristic appearances of pilocytic astrocytoma, ependymoma and medulloblastoma on conventional MRI. (a–c) Pilocytic astrocytoma - cerebellar/vermian cyst with a mural nodule on T1W and T2W MR showing intense enhancement of the nodule (⋆). (d–f) Ependymoma – T2W hyperintense tumor in the fourth ventricle, extending into the foramen of Luschka and Magendie (arrows). Moderate contrast enhancement and cystic spaces (Δ) are present within. (g–i) Medulloblastoma – intermediate to low T2W signal intensity tumor centered in the roof of the fourth ventricle showing variable contrast enhancement. MRI, magnetic resonance imaging; T1W, T1-weighted; T2W, T2-weighted.

DWI characteristics of the pediatric infratentorial tumors

On visual assessment, most of the medulloblastomas (n = 35, 94.6%) and all ATRTs demonstrated restricted diffusion. None of the pilocytic astrocytomas showed restriction of diffusion; 22 (84.6%) were not restricted while four were isointense to the brain. Among the ependymomas, 3 (30%) showed restricted diffusion and 7 (70%) were isointense.

The mean ADC values of medulloblastoma, high grade/anaplastic ependymoma, WHO grade II ependymoma, pilocytic astrocytoma, and ATRT were 0.592 × 10⁻³mm²/s, 0.916 × 10⁻³mm²/s, 1.127 ×10⁻³mm²/s, 1.590 × 10⁻³mm²/s, and 0.546 × 10⁻³mm²/s, respectively. Table 1 shows the mean ADC values of these tumors with examples of corresponding tumors depicted in Fig. 3.

Table 1.

Histopathological type and grade of the different infratentorial tumors and their mean ADC value of 82 children included in the study.

Histopathology	WHO grade	n	Mean ADC(×10^–3 mm²/s)	Range (×10^–3mm²/s)
Pilocytic astrocytoma	I	30	1.590	0.921–1.881
Ependymoma	II	2	1.127	0.957–1.297
Ependymoma (anaplastic)	III	8	0.916	0.672–1.178
Medulloblastoma	IV	40	0.592	0.392–0.932
ATRT	IV	2	0.546	0.437–0.655

ADC, apparent diffusion coefficient; ATRT, atypical rhabdoid teratoid tumors; WHO, World Health Organization.

Fig. 3.

DWI and corresponding ADC maps of pilocytic astrocytoma (a, b), ependymoma (c, d), medulloblastoma (e, f) and ATRT (g, h). Medulloblastoma (⋆) and ATRT (Δ) demonstrate maximum diffusion restriction and corresponding low ADC values. ADC, apparent diffusion coefficient; ATRT, atypical rhabdoid teratoid tumors; DWI, diffusion-weighted imaging.

Statistically significant differences (P < 0.05) were found between the mean ADC values of the following tumors: pilocytic astrocytoma and medulloblastoma (P = 0.001); pilocytic astrocytoma and anaplastic ependymoma (P < 0.001); medulloblastoma and anaplastic ependymoma. The ADC value of the medulloblastomas and ATRT were not statistically different.

For further analysis, the tumors were sub-classified into low grade (WHO grades I and II) and high grade (WHO grades III and IV) tumors (7). The low-grade tumors comprised 29 pilocytic astrocytomas and two ependymomas. The high-grade tumors included one anaplastic astrocytoma, eight anaplastic ependymomas, all of the 40 medulloblastomas, and two ATRT.

There were significant differences between the two groups when the parameters listed below were assessed.

Diffusion restriction (P < 0.001): only one of the low-grade tumors showed restricted diffusion (grade II ependymoma), 5 (18.5%) were isointense and 21 (77.8%) showed no restriction of diffusion. Among the high-grade tumors, there was only one tumor with no restriction of diffusion (anaplastic pilocytic astrocytoma), while 39 (97.5%) and 8 (61.5%) were restricted on DWI and isointense, respectively. The mean ADC value of the low-grade tumors was 1.567 × 10⁻³mm²/s ± 0.27 and of the high-grade tumors was 0.661 × 10⁻³mm²/s ± 0.2, this difference being statistically significant (P < 0.0001). When 0.9 × 10⁻³mm²/s was used as the cut-off value, the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for differentiating the grades was 87%, 100%, 100%, and 81.8%, respectively.

Signal intensity of the solid component on T2W images (P < 0.001): all of the low-grade tumors were hyperintense (n = 31), while 23 of the high-grade tumors were isointense and 28 were hyperintense.

Contrast enhancement (P = 0.012): all of the low-grade tumors showed enhancement. Among the high-grade tumors, there was variable to no enhancement, the numbers in each group being too small to be clinically significant.

Presence of cystic components (P < 0.001): 19 (61.3%) of the low-grade tumors had cystic areas and 11 (35.5%) tumors were predominantly cystic with mural nodule. None of the high-grade tumors were cysts with mural nodule; three had no cystic components while 48 (71.6%) had cystic areas.

Extension through foramen of Luschka and Magendie (P < 0.001): only 4/31 low-grade tumors extended across the foramina, while 30/47 high-grade tumors showed extension across the foramina.

Hemorrhagic components on SWI (P = 0.021): 6 (26.1%) of the low-grade tumors and 24 (55.8%) of the high-grade tumors showed hemorrhagic components.

Subgroup analysis of the medulloblastomas revealed that the majority (37/40) were midline tumors while three were laterally placed in the cerebellar hemisphere, CP angle, and cerebral peduncle. No difference was found between the mean ADC value between the midline (0.594 × 10⁻³mm²/s ± 0.20) and laterally placed (0.559 × 10⁻³mm²/s ± 0.26) tumors, t(33) = 0.518, P = 0.17. Foraminal involvement was a feature of all the lateral tumors, although not assuming statistical significance (P = 0.18).

Discussion

Infratentorial tumors comprise 45–60% of brain tumors in children (13). JPA, medulloblastoma, and ependymoma are the commonest pediatric posterior fossa tumors that need to be distinguished preoperatively (1). Among these, JPA is usually seen in older children aged > 5 years, while medulloblastoma, ependymoma, and ATRT are seen in children aged < 5 years (2). Differentiation from other tumors such as brainstem glioma, cerebellar gangliocytoma, and other extra-axial tumors like epidermoid, schwannoma, and meningioma is usually possible by their location and characteristic features.

Conventional neuroimaging is accurate in localizing and determining the extent of tumor. Although several differences exist in their imaging characteristics such as signal intensity, homogeneity, cystic nature, and extension through foramina, it is not always possible to predict the tumor type or grade reliably (14). Moreover, many of these tumors demonstrate anaplastic dedifferentiation, which adds to the complexity of imaging appearance.

DWI has shown to provide additional information in grading these tumors. The principle of DWI is based on the differences in the diffusion characteristics of water molecules within various tissues. Diffusion is restricted in densely packed tissues with high cellularity. Higher-grade tumors have a greater cellularity and hence are likely to demonstrate diffusion restriction (15,16). A few studies have shown DWI to be useful in characterizing these tumors (1,5,7 –11,17). With this background, we aimed to assess the role of DWI in differentiating and identifying the grade of pediatric infratentorial tumors.

Some of the earlier studies have evaluated both visual assessment of diffusion restriction as well as quantitative assessment using the ADC parametric map and have shown varying results. Erdem et al. (17) found that diffusion was abnormally restricted in all the PNET examined (n = 7) but was restricted in non-PNET in only 6% of the cases. According to Rumboldt et al. (8), ADC values were significantly higher in pilocytic astrocytomas (1.65 × 10³ mm²/s) when compared to ependymomas (1.10 × 10³ mm²/s) and medulloblastomas (0.66 × 10³ mm²/s). In another retrospective study of 50 patients by Jurkiewicz et al. (1), ADC values of pilocytic astrocytomas (1.54 × 10³ mm²/s) differed significantly from medulloblastomas (0.75 × 10³ mm²/s) and anaplastic ependymomas (0.99 × 10³ mm²/s).

Jaremko et al. (11), in a retrospective study of 40 children, described how a threshold ADC value of 0.8 × 10³ mm²/s differentiated JPA and medulloblastoma. Pierce et al. (9) studied DWI characteristics in 103 children and found similar results with an optimal threshold of 0.66 × 10³ mm²/s for identifying medulloblastomas with 86% PPV, 97% NPV, and 93% accuracy. Poretti et al. (14) found similar results in a study comprising 24 children with ADC values being higher in low-grade than higher-grade tumors, however with overlap of values between WHO grade II and grade III tumors. These findings were also confirmed by Tuntiyatorn et al. (18) in another study comprising 15 patients.

On the other hand, conflicting results have also been noted. A combination of ADC values and MRS was found to be useful when used in conjunction, rather than by using each parameter alone by Schneider et al. (19). Jaremko et al. (11) observed that while JPA and medulloblastoma could be differentiated by DWI alone in 88% of cases, ependymoma could not be reliably differentiated from either of these and the overlap of ADC values did not allow accurate prediction of tumor grade. Tzika et al. (20) observed that relative cerebral blood volumes (rCBVs) and ADC mapping complemented MRS, rather than when used alone in predicting tumor grade.

Our results are in concordance with some of the previous studies and confirms that the ADC values of higher-grade tumors is significantly lower than the low-grade tumors (7,18). A cut-off value 0.9 ×10⁻³mm²/s aids in identifying the higher-grade tumors (WHO grades III and IV) with a sensitivity, specificity, PPV, and NPV of 87%, 100%, 100%, and 81.8%, respectively.

On visual inspection, diffusion restriction was a significant feature of the high-grade tumors. There was only one low-grade tumor with restricted diffusion, this being a WHO grade II ependymoma, and one high-grade tumor with no diffusion restriction, which was an anaplastic pilocytic astrocytoma.

The mean ADC values of the individual tumors were comparable with some of the above studies (1,8) and are as follows: pilocytic astrocytoma (1.590 ×10⁻³mm²/s); WHO grade II ependymomas (1.127 × 10⁻³mm²/s); anaplastic ependymomas (0.916 × 10⁻³mm²/s); medulloblastoma (0.592 ×10⁻³mm²/s); and ATRT (0.546 × 10⁻³mm²/s).

The qualitative features that were also different between these groups, although even though the numbers in each category being small to reliably distinguish when used alone were – signal intensity on T2W imaging, contrast enhancement, presence of cystic areas/cyst with mural nodule, extension through foramina, and hemorrhagic components on SWI. These features, when used in conjunction with ADC values, improve the accuracy of diagnosis.

We propose an algorithmic approach to posterior fossa tumors in children that can be used more practically on a day-to-day basis. The age of the child will aid in initial screening. In a child aged > 5 years, the commonly encountered tumors are pilocytic astrocytoma and ependymoma. The two can be differentiated using morphological characteristics that are quite typical and well described for each lesion. In a child aged < 5 years, medulloblastoma, ependymoma, and ATRT are the usual possibilities. ATRT, being an off-midline tumor, can be diagnosed in most cases, while the real diagnostic dilemma lies in differentiating the medulloblastomas from the ependymomas. Morphological characteristics like T2 signal intensity and extension through the foramina frequently overlap in these tumors with added inter-observer variation due to the subjective nature of T2 signal and enhancement assessment. Hence, estimation of ADC values will benefit in differentiating the two. The mean ADC value of medulloblastomas, WHO grade II ependymomas, and anaplastic ependymomas were 0.592 × 10⁻³mm²/s, 1.127 × 10⁻³mm²/s, and 0.916 × 10⁻³mm²/s, with statistically significant differences between these groups. A cut-off value of 0.9 × 10⁻³mm²/s aids in identifying the higher-grade tumors.

The present study has some limitations. Although the total number is a strength of the study, absolute numbers in some of the tumor types (ATRT and grade II ependymomas) are rather small to assume statistical significance. We have assessed the mean ADC value by placing three ROIs in the tumor; however, this does not account for the entire heterogeneity in large tumors. Due to small numbers in the various subgroups of the variables assessed, a multivariate regression analysis could not be performed and the differences based on location of tumor types could not be adequately evaluated. Recent genomic studies have shown that medulloblastoma comprises four different subgroups at the molecular level (WNT, SHH, group 3, and group 4) that are distinctly unique in terms of their survival, demographics, and at the genetic level (21). We have, however, not distinguished these subgroups histologically and hence there is scope for further research in this area.

In conclusion, assessment of diffusion restriction and ADC values allows reliable differentiation between high-grade and low-grade tumors. A reasonably accurate diagnosis of the tumor type and grade can be arrived at using an algorithmic approach that uses the commonly described MRI features and DWI in conjunction; the tumor ADC being useful in differentiating medulloblastomas from ependymomas in difficult cases. Quantitative measurements can be easily performed and we propose a reference ADC value of 0.9 × 10⁻³mm²/s to differentiate between the high-grade and low-grade tumors.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Kirthi Sathyakumar https://orcid.org/0000-0002-7742-5328 Krishna Prabhu

References

Jurkiewicz

Paku\la-Kościesza

Che\lstowska

, et al. Infratentorial tumors in children–value of ADC in prediction of grade of neoplasms. Pol J Radiol 2010; 75:18–23.

Poretti

Meoded

Huisman

TAGM.

Neuroimaging of pediatric posterior fossa tumors including review of the literature. J Magn Reson Imaging 2012; 35:32–47.

Becker

LE.

Pathology of pediatric brain tumors. Neuroimaging Clin N Am 1999; 9:671–690.

Okamoto

Ito

Takahashi

, et al. MRI of high-grade astrocytic tumors: early appearance and evolution. Neuroradiology 2002; 44:395–402.

Bulakbasi

Guvenc

Onguru

, et al. The added value of the apparent diffusion coefficient calculation to magnetic resonance imaging in the differentiation and grading of malignant brain tumors. J Comput Assist Tomogr 2004; 28:735–746.

Al-Okaili

Krejza

Wang

, et al. Advanced MR imaging techniques in the diagnosis of intraaxial brain tumors in adults. Radiographics 2006; 26 Suppl 1:S173–189.

Poretti

Meoded

Cohen

, et al. Apparent diffusion coefficient of pediatric cerebellar tumors: a biomarker of tumor grade? Pediatr Blood Cancer 2013; 60:2036–2041.

Rumboldt

Camacho

DLA

Lake

, et al. Apparent diffusion coefficients for differentiation of cerebellar tumors in children. AJNR Am J Neuroradiol 2006; 27:1362–1369.

Pierce

Provenzale

JM.

Evaluation of apparent diffusion coefficient thresholds for diagnosis of medulloblastoma using diffusion-weighted imaging. Neuroradiol J 2014; 27:63–74.

10.

Schneider

Confort-Gouny

Viola

, et al. Multiparametric differentiation of posterior fossa tumors in children using diffusion-weighted imaging and short echo-time 1H-MR spectroscopy. J Magn Reson Imaging 2007; 26:1390–1398.

11.

Jaremko

Jans

LBO

Coleman

, et al. Value and limitations of diffusion-weighted imaging in grading and diagnosis of pediatric posterior fossa tumors. AJNR Am J Neuroradiol 2010; 31:1613–1616.

12.

Kotsenas

Roth

Manness

, et al. Abnormal diffusion-weighted MRI in medulloblastoma: does it reflect small cell histology? Pediatr Radiol 1999; 29:524–526.

13.

Baldwin

Preston-Martin

Epidemiology of brain tumors in childhood–a review.

Toxicol Appl Pharmacol 2004; 199:118–131.

14.

Poretti

Meoded

Huisman

TA.

Neuroimaging of pediatric posterior fossa tumors including review of the literature. J Magn Reson Imaging 2012; 35:32–47.

15.

Thurnher

MM.

2007 World Health Organization classification of tumors of the central nervous system. Cancer Imaging 2009; 9 Spec No A:S1–3.

16.

Kan

Liu

Hedlund

, et al. The role of diffusion-weighted magnetic resonance imaging in pediatric brain tumors. Childs Nerv Syst 2006; 22:1435–1439.

17.

Erdem

Zimmerman

Haselgrove

, et al. Diffusion-weighted imaging and fluid attenuated inversion recovery imaging in the evaluation of primitive neuroectodermal tumors. Neuroradiology 2001; 43:927–933.

18.

Tuntiyatorn

Nantawas

Sirachainan

, et al. Apparent diffusion coefficients in evaluation of pediatric brain tumors. J Med Assoc Thai 2013; 96:178–184.

19.

Schneider

Viola

Confort-Gouny

, et al. Infratentorial pediatric brain tumors: the value of new imaging modalities. J Neuroradiol 2007; 34:49–58.

20.

Tzika

Zarifi

Goumnerova

, et al. Neuroimaging in pediatric brain tumors: Gd-DTPA-enhanced, hemodynamic, and diffusion MR imaging compared with MR spectroscopic imaging. AJNR Am J Neuroradiol 2002; 23:322–333.

21.

Skowron

Ramaswamy

Taylor

MD.

Genetic and molecular alterations across medulloblastoma subgroups. J Mol Med 2015; 93:1075–1084.