Child's Nervous System

, Volume 22, Issue 11, pp 1435–1439 | Cite as

The role of diffusion-weighted magnetic resonance imaging in pediatric brain tumors

  • Peter Kan
  • James K. Liu
  • Gary Hedlund
  • Douglas L. Brockmeyer
  • Marion L. Walker
  • John R. W. Kestle
Original Paper

Abstract

Objectives

Diffusion-weighted imaging (DWI) may enhance the radiographic diagnosis of pediatric brain tumors. This study reviews the DWI properties of pediatric brain tumors at our institution and examines their relationship to tumor grade and type.

Materials and methods

The preoperative DWI and apparent diffusion coefficient (ADC) characteristics of brain tumors in 41 children were compared with histologic diagnosis. Signal characteristics on DWI and ADC maps correlated well with tumor grade. High-grade lesions were hyperintense on DWI and hypointense on ADC maps. Sensitivity, specificity, positive predictive value, and negative predictive value were 70, 100, 100, and 91%, respectively. Signal characteristics did not differ among different tumors of the same grade. All primitive neuroectodermal tumors showed diffusion restriction whereas none of the ependymomas did.

Conclusions

The signal characteristics on DWI and ADC maps appeared to be strongly correlated to grade in pediatric brain tumors and they may assist with preoperative diagnostic predictions.

Keywords

Pediatric brain tumors Magnetic resonance imaging Diffusion-weighted imaging Apparent diffusion coefficient 

Introduction

The optimal treatment and prognosis of pediatric brain tumors depend on the involved cell types and grades. Conventional magnetic resonance (MR) imaging is excellent in characterizing the location and extent of brain tumors but often provides limited preoperative information regarding their grades and types. Recent studies have suggested that diffusion-weighted MR imaging (DWI) may enhance the radiographic diagnosis of brain tumors [1, 4, 7, 8, 9, 15, 16]. The purpose of our study was to review the DWI properties of pediatric brain tumors seen at our institution and examine the relationship between those DWI properties and tumor grade and type.

Materials and methods

Institutional Review Board guidelines were followed. All pediatric tumor operations, excluding reoperations between January 2001 and August 2005, were identified from our operative database. The use of DWI was instituted at Primary Children’s Medical Center in January 2001. Data were extracted from the medical records and the MR imaging characteristics on DWI were reviewed.

Tumors were separated into two groups, “low grade” and “high grade,” based on their histology. High-grade tumors were defined as malignant tumors or tumors with World Health Organization (WHO) grade of three or higher. These consisted of primitive neuroectodermal tumors (PNETs), malignant germ cell tumors, grade III/IV astrocytomas, leukemic/lymphomatous tumefaction, metastatic osteosarcomas, and esthesioneuroblastomas. Low-grade tumors were defined as benign tumors or tumors with WHO grade of two or less. These included juvenile pilocytic astrocytomas (JPAs), grade II astrocytomas, germinomas, craniopharyngiomas, benign meningiomas, nonanaplastic ependymomas, acoustic neuromas, gangliogliomas, low-grade oligodendrogliomas, and dysembryoplastic neuroepithelial tumors. The signal characteristics of each lesion on DWI and apparent diffusion coefficient (ADC) maps were compared with that of the surrounding normal brain and were graded as hyperintense, isointense, or hypointense. The analysis was based on the preoperative interpretation by the neuroradiologists who did not know the final pathology. A positive test (true diffusion restriction) was defined as a hyperintense lesion relative to the surrounding brain on DWI and hypointense on the corresponding ADC map. A negative test was defined as any other combinations of DWI and ADC characteristics. The sensitivity, specificity, positive predictive value, and negative predictive value of DWI as a test to differentiate between high-grade and low-grade tumors were calculated.

DWI technique

DWI was performed using a three-gradient protocol (TR/TE=4,000/110). Heavily diffusion-weighted (b=1,000 s/mm2) images and automatically generated ADC maps were generated and studied.

Results

Between January 2001 and August 2005, a total of 133 children underwent a craniotomy for newly diagnosed brain tumors at Primary Children’s Medical Center. Of these, 41 children (22 boys and 19 girls) had preoperative DWI. They ranged in age from 2 months to 18 years (mean 7.7 years). Ten patients had high-grade tumors [four PNETs, one malignant germ cell tumor, two high-grade (III/IV) astrocytomas, one leukemic/lymphomatous tumefaction, one metastatic osteosarcoma, and one esthesioneuroblastoma] and 31 patients had low-grade tumors (ten JPAs, four grade II astrocytomas, one germinoma, three craniopharyngiomas, two benign meningiomas, six nonanaplastic ependymomas, one acoustic neuroma, one ganglioglioma, two low-grade oligodendrogliomas, and one dysembryoplastic neuroepithelial tumor).

Using our three-tiered grading system (hyperintense, isointense, or hypointense), signal characteristics on DWI and ADC maps correlated well with tumor grades. Of the ten high-grade tumors, seven were hyperintense to brain on DWI (four PNETs, one malignant germ cell tumor, one grade IV astrocytoma, and one leukemic tumefaction) and three were isointense (metastatic osteosarcoma, anaplastic astrocytoma, and esthesioneuroblastoma). High-grade lesions generally appeared to be hyperintense to brain with diffusion restriction on DWI and hypointense to brain on ADC maps (Fig. 1). Among the remaining 31 low-grade tumors, all were isointense to brain on DWI. Low-grade lesions appeared to be isointense to brain on both DWI and ADC maps.
Fig. 1

PNET arising from the left cerebellopontine angle with extensive supratentorial involvement. Axial T1-weighted imaging (a), T1-weighted imaging with contrast (b), T2-weighted imaging (c), and DWI (d) revealed a lesion that is hyperintense to brain

Sensitivity, specificity, positive predictive value, and negative predictive value of DWI as a diagnostic test to differentiate between high-grade and low-grade pediatric brain tumors were 70, 100, 100, and 91%, respectively (Table 1). The signal characteristics on DWI and ADC maps did not appear to be different among different tumors of the same grade. It is interesting to note that all four PNETs (3 fourth ventricular and 1 cerebellopontine angle/subtemporal) in our series showed evidence of diffusion restriction whereas none of the 6 fourth ventricular ependymomas were diffusion-positive (Figs. 2 and 3).
Table 1

Evaluation of DWI as a diagnostic test to differentiate between high-grade and low-grade pediatric brain tumors

 

High-grade tumors

Low-grade tumors

Positive

9 (TP)

0 (FP)

Negative

3 (FN)

31 (TN)

Sensitivity

Sn=TP/(TP+FN)

75%

Specificity

Sp=TN/(TN+FP)

100%

Positive predictive value

PPV=TP/all positives

100%

Negative predictive value

NPV=TN/all negatives

91%

Sn Sensitivity, Sp specificity, PPV positive predictive value, NPV negative predictive value, TP true positive, FN false negative, FP false positive, and TN true negative

Fig. 2

Fourth ventricular medulloblastoma. Axial T1-weighted imaging with contrast (a), sagittal T1-weighted imaging (b), and DWI (c) demonstrated a lesion with true diffusion restriction that is hyperintense on DWI and hypointense on ADC maps

Fig. 3

Fourth ventricular nonanaplastic ependymoma. Axial T1-weighted imaging with contrast (a), sagittal T1-weighted imaging (b), and DWI (c) revealed a lesion that is isointense to brain without evidence of diffusion restriction

Discussion

In our study, simple signal characteristics on DWI and ADC maps correlated well with tumor grades. High-grade lesions tended to be hyperintense on DWI with true diffusion restriction. The high specificity and positive predictive value suggest that DWI is a good test in identifying high-grade brain tumors. In other words, a positive finding on DWI (hyperintense on DWI and hypointense on ADC) is highly predictive of a high-grade lesion. Our results agree with previous reports. Klisch et al. [7] reported three cases of supratentorial PNETs with diffusion restriction. Kono et al. [8], Bulakbasi et al. [1], and Yamasaki et al. [16] found that high-grade tumors tended to exhibit diffusion restriction with low ADC values.

DWI is based on the Brownian movement of water in tissues. In tissues with unrestricted water diffusion, the MR signal is dispersed and attenuated [i.e., low intensity (dark) on DWI]. In contrast, restriction of water diffusion increases the signal intensity on DWI (i.e., bright). Increased cellularity, decreased extracellular space, and high nuclear-to-cytoplasmic ratio are currently believed to be the factors responsible for water diffusion restriction in high-grade tumors [1, 8, 9, 15, 16]. In support of the theory, Chenevert et al. [2], showed that in animal models, ADC increases during chemotherapy when there is a reduction in cellularity and more extracellular space and decreases back to baseline with tumor regrowth when the reverse takes place. Thus, in addition to its diagnostic role, DWI could offer a novel way to monitor tumor response to therapy.

Other investigators found ADCs to be somewhat useful in differentiating cell types of different tumors that share a similar pathological grade [1, 4, 16]. However, in our study, the signal characteristics on DWI and ADC maps did not appear to be different among tumors of the same grade. This may be explained by the fact that unlike ADC, which is a precisely calculated continuous variable, our three-tiered classification system for the signal characteristics on DWI and ADC maps is probably less sensitive. Nevertheless, we believe our simple classification system will be useful at the bedside for neurosurgeons assessing newly diagnosed brain tumors.

It is interesting to note that all of the PNETs/medulloblastomas in our study demonstrated properties of diffusion restriction whereas none of the posterior fossa ependymomas did. This phenomenon was also observed by Yamasaki et al. [16] and is likely related to the high-grade features of PNETs compared with nonanaplastic ependymomas. In our experience, obtaining a DWI in addition to the preoperative MRI was a quick and useful modality to preoperatively differentiate PNETs from other posterior fossa brain tumors. Given the minimal expenditure of time, we suggest obtaining preoperative DWI on all newly diagnosed fourth ventricular tumors in children as an aid to differentiate between ependymomas and medulloblastomas.

DWI is widely used in many clinical conditions, such as the evaluation of acute cerebral ischemia in which cytotoxic edema limits water diffusion [5, 12]. Its advantage in distinguishing epidermoid tumors from arachnoid cysts is also well known [10, 14]. DWI is also useful in differentiating cerebral abscesses from tumors because the high cellularity and viscosity in the purulent abscess fluid restricts water diffusion [3, 6]. Less common uses of DWI in the assessment of traumatic brain injury [13] and demyelination [11] have also been reported.

Our qualitative study is limited by the relatively small number of patients and by the heterogeneous group of tumors. Nevertheless, it appears that simple signal characteristics on DWI correlated well with tumor grades in the pediatric population. The routine use of DWI in all newly diagnosed pediatric brain tumors could give insight into the aggressiveness of the lesion and be helpful to the pediatric neurosurgeon preoperatively in differentiating fourth ventricular tumors and predicting prognosis.

Notes

Acknowledgment

The authors thank Kristin Kraus for her editorial assistance in preparing this paper.

References

  1. 1.
    Bulakbasi N, Guvenc I, Onguru O, Erdogan E, Tayfun C, Ucoz T (2004) 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 28:735–746PubMedCrossRefGoogle Scholar
  2. 2.
    Chenevert TL, McKeever PE, Ross BD (1997) Monitoring early response of experimental brain tumors to therapy using diffusion magnetic resonance imaging. Clin Cancer Res 3:1457–1466PubMedGoogle Scholar
  3. 3.
    Desprechins B, Stadnik T, Koerts G, Shabana W, Breucq C, Osteaux M (1999) Use of diffusion-weighted MR imaging in differential diagnosis between intracerebral necrotic tumors and cerebral abscesses. AJNR Am J Neuroradiol 20:1252–1257PubMedGoogle Scholar
  4. 4.
    Gauvain KM, McKinstry RC, Mukherjee P, Perry A, Neil JJ, Kaufman BA, Hayashi RJ (2001) Evaluating pediatric brain tumor cellularity with diffusion-tensor imaging. AJR Am J Roentgenol 177:449–454PubMedGoogle Scholar
  5. 5.
    Gray L, MacFall J (1998) Overview of diffusion imaging. Magn Reson Imaging Clin N Am 6:125–138PubMedGoogle Scholar
  6. 6.
    Kim YJ, Chang KH, Song IC, Kim HD, Seong SO, Kim YH, Han MH (1998) Brain abscess and necrotic or cystic brain tumor: discrimination with signal intensity on diffusion-weighted MR imaging. AJR Am J Roentgenol 171:1487–1490PubMedGoogle Scholar
  7. 7.
    Klisch J, Husstedt H, Hennings S, von Velthoven V, Pagenstecher A, Schumacher M (2000) Supratentorial primitive neuroectodermal tumours: diffusion-weighted MRI. Neuroradiology 42:393–398PubMedCrossRefGoogle Scholar
  8. 8.
    Kono K, Inoue Y, Nakayama K, Shakudo M, Morino M, Ohata K, Wakasa K, Yamada R (2001) The role of diffusion-weighted imaging in patients with brain tumors. AJNR Am J Neuroradiol 22:1081–1088PubMedGoogle Scholar
  9. 9.
    Krabbe K, Gideon P, Wagn P, Hansen U, Thomsen C, Madsen F (1997) MR diffusion imaging of human intracranial tumours. Neuroradiology 39:483–489PubMedCrossRefGoogle Scholar
  10. 10.
    Laing AD, Mitchell PJ, Wallace D (1999) Diffusion-weighted magnetic resonance imaging of intracranial epidermoid tumours. Australas Radiol 43:16–19PubMedCrossRefGoogle Scholar
  11. 11.
    Ono J, Harada K, Mano T, Sakurai K, Okada S (1997) Differentiation of dys- and demyelination using diffusional anisotropy. Pediatr Neurol 16:63–66PubMedCrossRefGoogle Scholar
  12. 12.
    Rowley HA, Grant PE, Roberts TP (1999) Diffusion MR imaging. Theory and applications. Neuroimaging Clin N Am 9:343–361PubMedGoogle Scholar
  13. 13.
    Smith DH, Meaney DF, Lenkinski RE, Alsop DC, Grossman R, Kimura H, McIntosh TK, Gennarelli TA (1995) New magnetic resonance imaging techniques for the evaluation of traumatic brain injury. J Neurotrauma 12:573–577PubMedCrossRefGoogle Scholar
  14. 14.
    Tsuruda JS, Chew WM, Moseley ME, Norman D (1990) Diffusion-weighted MR imaging of the brain: value of differentiating between extraaxial cysts and epidermoid tumors. AJR Am J Roentgenol 155:1059–1065, discussion 1066–1068PubMedGoogle Scholar
  15. 15.
    Wilke M, Eidenschink A, Muller-Weihrich S, Auer DP (2001) MR diffusion imaging and 1H spectroscopy in a child with medulloblastoma. A case report. Acta Radiol 42:39–42PubMedGoogle Scholar
  16. 16.
    Yamasaki F, Kurisu K, Satoh K, Arita K, Sugiyama K, Ohtaki M, Takaba J, Tominaga A, Hanaya R, Yoshioka H, Hama S, Ito Y, Kajiwara Y, Yahara K, Saito T, Thohar MA (2005) Apparent diffusion coefficient of human brain tumors at MR imaging. Radiology 2(35):985–991Google Scholar

Copyright information

© Springer-Verlag 2006

Authors and Affiliations

  • Peter Kan
    • 1
  • James K. Liu
    • 1
  • Gary Hedlund
    • 2
  • Douglas L. Brockmeyer
    • 1
  • Marion L. Walker
    • 1
  • John R. W. Kestle
    • 1
  1. 1.Department of Neurosurgery, Primary Children’s Medical CenterUniversity of Utah School of MedicineSalt Lake CityUSA
  2. 2.Department of Radiology, Primary Children’s Medical CenterUniversity of Utah School of MedicineSalt LakeUSA

Personalised recommendations