Diffuse Intrinsic Pontine Gliomas. PaweÅ Buczkowicz1,3,7, Dong-Anh Khuong-Quang4, Patricia Rakopoulos3,7, Eric Bouffet2, Andrew Morrison3, Ute Bartels2,.
Clinical Relevance of the Histopathological Spectrum of Paediatric Diffuse Intrinsic Pontine Gliomas Paweł Buczkowicz1,3,7, Dong-Anh Khuong-Quang4, Patricia Rakopoulos3,7, Eric Bouffet2, Andrew Morrison3, Ute Bartels2, Stefan M. Pfister6, Nada Jabado4,5 and Cynthia Hawkins1,3,7
Background Tumours of the central nervous system constitute the most common type of solid paediatric tumours and account for the majority of mortality and morbidity in paediatric oncology. Diffuse intrinsic pontine gliomas (DIPG) are among the most devastating paediatric malignancies for which no effective therapy exists and currently DIPGs are the main cause of brain tumour related death in children.1 Virtually all children with this disease die within 1-2 years of diagnosis. The mean age of diagnosis is approximately 7 years. These pontine lesions are the most common type of brainstem tumour, accounting for 58-78% of brainstem glioma. Due to the location of these tumours, diagnosis is mainly based on imaging and surgical intervention is extremely rare. Clinical investigations into the effects of adjuvant chemotherapy on patient survival have not thus far shown any benefit and radiotherapy is palliative, offering only symptom control.2
Materials & Methods Fifty-one infants and children diagnosed with diffuse intrinsic pontine glioma between 1984 and 2011 were included in the study. Thirty-eight were diagnosed and treated at the Hospital for Sick Children and 13 cases were referred from other institutions. Clinical presentation was consistent with DIPG including history of progressive headache, vomiting, diplopia, unsteady gait, slurred speech, papilledema, VI nerve palsy an signs of ataxia. Twenty-six of the patients were female and 25 were male. The mean age of diagnosis for our cohort was 6.64 years and the median was 6.47. The age of patients ranged from 0 days to 15.24 years. More than 80% of patients were under the age of 10 at the time of diagnosis. The median survival was 0.85 years. Any available surgical and autopsy material was examined by routine haematoxylin and eosin (H&E) staining and reviewed according to WHO guidelines, as well as immunohistochemical staining (IHC) for selected antibodies using immunoperoxidase detection. 15/51 cases had available surgical pre-treatment material. Thirtysix patients were studied post-mortem (one pre-treatment autopsy). Coding exons of H3F3A was sequenced using Sanger fluorescent sequencing after amplification by polymerase chain reaction using standard methods. Thirty-seven samples were hybridized to the GenomeWide Human SNP Array 6.0 and three to the Human Mapping 250 SNP Nsp Array from Affymetrix (Santa Clara, CA). Data was analyzed using Partek Genomics Suite v6.4 (Partek Incorporated, St. Louis, MO), Genotyping Console 4.1 (Affymetrix), and GISTIC2.0 (Broad Institute, Cambridge, MA).
(1) Division of Pathology - (2) Division of Haematology and Oncology – (3) The Arthur and Sonia Labatt Brain Tumour Research Centre, Hospital for Sick Children, Toronto, ON, Canada (4) Department of Human Genetics – (5) Department of Paediatrics, McGill University Health Center, Montreal, Canada (6) Division of Pediatric Neuro-oncology, The German Cancer Research Center (DFKZ), Heidelberg, Germany (7) Department of Laboratory Medicine and Pathobiology, Faculty of Medicine, University of Toronto, Toronto, Canada
Results WT-H3.3
K27M-H3.3 WT FatI
-
+
K27M -
ESPNP
+
ASAP2 MYCN
300bp 200bp 75bp
WDR1 PDGFRA
ZNF718
PVT-1/MYC NCRNA00168 OR52N4 PGR
C11orf58
CLK3
LYSMD4
SS18L1
Figure 1. K27M-H3.3 is associated with worse overall survival in DIPG – Kaplan-Meier curve of overall survival for DIPG patients shows significantly worse overall survival for patients carrying the K27M-H3.3 mutation as compared to wild-type (Log-rank, p = 0.0027). Notably, all long term survivors were wild-type for H3F3A.
Figure 2. K27M-H3.3 and WT-H3.3 DIPGs show different focal copy number alterations – Focal recurrent amplifications determined by GISTIC2.0 analysis (q≤0.5) show significant differences in focal gains between sample carrying K27M-H3.3 and samples wild-type for H3.3. This included PDGFRA (4q12), MYC/PVT1 locus (8q24.21) gains and amplifications, which were exclusively found in K27M H3.3 mutants. ASAP2 (2p25.1) and MYCN (2p24.3) gains and amplifications were exclusively identified in wild-type patients.
DIPG03 (MRI Dx – DIPG; autopsy Dx – GBM; WT-H3.3)
DIPG01 (MRI Dx – DIPG; autopsy Dx – PNET; WT-H3.3) Sag T1
Pons – H&E 50X
Pons – H&E 200X
Pons – H&E 600X
Pons – GFAP 200X
5-yo female presenting with focal tumour in pons on MRI. Treated with focal radiation (54Gy over 30 fractions). Overall survival was 5 months. On H&E the tumour is PNET-like with very round uniform nuclei, prominent nucleoli, not directly invasive and sharply demarcated from adjacent brain. There was no vascular endothelial proliferation and no pseudopalisading necrosis typical of high-grade astrocytomas. The tumour appeared more neurocitic. There was some leptomeningeal dissemination present. On immunohistochemistry the tumour was positive for nestin and MAP2, but negative for GFAP.
DIPG19 (MRI Dx – DIPG; autopsy Dx – GBM; K27M-H3.3)
Sag T2
Pons – H&E 50X
Pons – H&E 200X
Pons – H&E 600X
Midbrain – H&E 200X
5-yo boy presenting with chocking episodes, head-tilt and ataxia. A focal lesion was present in the pons on MRI. Treated with focal radiation (56Gy over 30 fractions) and Nimotuzumab. The patient succumbed to the disease after 18 months. On H&E the tumour histology was consistent with that of GBM; including pseudopalisading necrosis, vascular endothelial proliferation. There was diffuse intraparenchymal infiltration of the tumour. On immunohistochemistry the tumour was negative for H2AX, MGMT, VEGF and P53 with weak positivity for EGFR.
DIPG26 (MRI Dx – DIPG; autopsy Dx – LGA; K27M-H3.3)
Pons – H&E 20X
Pons – H&E 100X
Pons – H&E 600X
Pons – H&E 20X
Pons – H&E 200X
Pons – MIB1 100X
Frontal Lobe
Frontal Lobe – H&E 40X
Frontal Lobe – H&E 200X
Leptomeningeal Spread – H&E
Leptomeningeal Spread – H&E 4X
Leptomeningeal Spread – H&E 10X
Conclusions 1. K27-H3.3 mutation confers worse overall survival in DIPG patients. All long term survivors are wild-type for H3F3A. 2. K27M-H3.3 and WT-H3.3 DIPGs have different focal copy number alterations. PDGFRA and PVT-1/MYC gains/amplifications are exclusively found in K27MH3.3 DIPGs with frequencies of 48% and 30%, respectively. Gains/amplifications exclusively found in WT-H3.3 DIPG include MYCN. 3. Pontine brainstem gliomas diagnosed as DIPG on MRI can have histologic features of astrocytomas and PNET. Histological examination of tissue from 51 DIPGs (21 pre-treatment samples, 37 autopsy samples) found 49 with astrocytomas (45 high-grade, 4 grade-II) and two PNET. 4. Histology alone is not a predictor of survival. Two patients with high-grade pathology were long term survivors, while two patients had low-grade astrocytomas on autopsy yet had clinical course considered “classic” for DIPG. The two high-grade long term survivors were wild-type for H3.3, while the two low-grade patients with aggressive clinical course and short term survival were positive for K27M-H3.3. 5. 1/3 of DIPG patients exhibit leptomeningeal spread of their tumour. Diffuse tumour invasion of the brainstem was common among all of the DIPG, with the majority diffusely involving the spinal cord and thalamus. Some patients exhibited distal spread as far as the frontal lobes. These patients may benefit from new treatment strategies targeted towards inhibiting tumor infiltration rather than proliferation, especially since some of these very infiltrative tumors exhibit few mitoses and scarce MIB1 positivity. Conversely, on autopsy two patients had severe radiation necrosis of the brainstem with several other patients showing mild to moderate radiation necrosis. Based on prognostic and therapeutic implications of K27MH3.3 and the histologic spectrum of DIPG, our findings argue for biopsy and H3.3-mutation testing at diagnosis, which could be rapidly integrated into the clinical decision-making algorithm, particularly in atypical DIPG. Further, the frequency of leptomeningeal disease in this patient group suggests that focal radiation may be inadequate for some of these patients. New therapeutic approaches need to incorporate new molecular and histological data in order to achieve maximum benefit for DIPG patients.
References 6.5-yo female with typical MRI presentation. Treated with focal radiation (54Gy over 30 fractions) and chemotherapy (Nimotuzumab). Patient succumbed to the disease after 5.5 months. On H&E this tumour appears very diffuse with leptomeningeal spread and tumour with GBM features present in the subarachnoid space invading into the brain. There is presence of Wallerian degeneration in the cervico chord. There are tumour cells present as far as the frontal lobe.
4-yo female with typical presentation diagnosed with DIPG. Treated with focal radiation (56Gy over 31 fractions). Overall survival was 7 months. On H&E this is an infiltrative brain tumour with no particular highgrade features. There is no vascular endothelial proliferation, necrosis and very few mitoses. Less than 1% of tumour cells are MIB1 positive. There is leptomeningeal spread and exophytic component involving both the pons and the medulla.
1. Hargrave D, Bartels U, Bouffet E. (2006) “Diffuse brainstem glioma in children: critical review of clinical trials” The Lancet Oncology 7: 241-248. 2. Donaldson SS, Laningham F, Fisher PG (2006) “Advances toward an understanding of brainstem gliomas” Journal of Clinical Oncology 24: 12661272. 3. Khuong-Quang DA, Buczkowicz P, Rakopoulos P et al. (2012) “K27M mutation in histone H3.3 defines clinically and biologically distinct subgroups of pediatric diffuse intrinsic pontine gliomas” Acta Neuropathologica 123(3): 439-447.
