Diffuse intrinsic pontine tumors: a study of primitive neuroectodermal tumors versus the more common diffuse intrinsic pontine gliomas

Abstract

Object: The diagnosis of diffuse pontine tumors has largely been made on the basis of MRI since the early 1990 s. In cases of tumors considered "typical," as a rule, no biopsy specimen has been obtained, and the tumors have been considered diffuse intrinsic pontine gliomas (DIPGs). There have been sporadic reports that primitive neuroectodermal tumors (PNETs) of the pons may not be distinguishable from the DIPGs by radiological imaging. This study presents 2 cases of diffuse pontine PNETs with molecular evidence that these are indeed PNETs, distinct from DIPGs, thus supporting biopsy of diffuse pontine tumors as a standard of care.

Methods: Biopsy specimens were obtained from 7 diffuse pontine tumors and snap frozen. Two of these 7 tumors were identified on the basis of pathological examination as PNETs. All 7 of the diffuse pontine tumors were analyzed for gene expression using the Affymetrix HG-U133 Plus 2.0 GeneChip microarray. Gene expression was compared with that of supratentorial PNETs, medulloblastomas, and low- and high-grade gliomas outside the brainstem.

Results: Unsupervised hierarchical clustering analysis of gene expression demonstrated that pontine PNETs are most closely related to PNETs of the supratentorial region and not with gliomas. They do not cluster with the 5 DIPGs in the study. Thirty-eight genes, including GATA3, are uniquely differentially expressed in pontine PNETs compared with other types of pediatric brain tumors, including DIPGs and other PNETs at a false discovery rate statistical significance of less than 0.05.

Conclusions: The cluster and individual gene expression analyses indicate that pontine PNETs are intrinsically different from DIPGs. The 2 pontine PNET cases cluster with supratentorial PNETs, rather than with DIPGs, suggesting that these tumors should be treated with a PNET regimen, not with DIPG therapy. Since diagnosis by imaging is not reliable and the biology of the tumors is disparate, a biopsy should be performed to enable accurate diagnosis and direct potentially more effective treatments.

Figure 1

Radiographic and histopathologic characteristics of two cases of pontine PNET. A. MRI of 16 month old female. Sagittal post-gadolinium T1 midline image shows a non-enhancing expansile pontine mass with anterior extension into the suprasellar cistern, encasing the pituitary stalk (arrow). B. Same female. Axial diffusion (B1000) weighted image show restricted diffusion in the pontine mass (arrowhead) and an additional discontinuous mass in the inferior right sylvian fissure (arrow). C. MRI of 28 month old male. Axial T2 image shows a hyperintense heterogeneous expansile pontine mass distorting the fourth ventricle (arrow). D. Same male. Coronal FLAIR image show superior extension of the heterogeneous pontine mass into the right cerebral peduncle (arrow). E and F. Representative hematoxlyin/eosin stained sections that are diagnostic of PNET (magnification 600x).

Figure 2

Pediatric brain tumor cluster analysis. Colors indicate tumors of interest: Pontine PNETs (PPNET, red), supratentorial PNETs (STPNET, green), and DIPGs (blue).

Figure 3

GATA3 immunostaining in pediatric brainstem tumors. A. Pontine primitive neuroectodermal tumor (PNET) immunostaining shows nuclear GATA3 protein throughout the tumor, original magnification 200x, inset original magnification 400x. B. Diffuse intrinsic pontine glioma (DIPG) immunostaining shows no GATA3 protein, original magnification 200x, inset original magnification 400x. C. Normal pons immunostaing also stains negative for GATA3 protein, original magnification 200x, inset original magnification 400x.