Abstract
Genome-wide DNA methylation profiling has emerged as an important diagnostic tool that complements histopathology for CNS tumors in children and adults. Literature describing its application in Asian countries is nonetheless limited. Herein, we report the feasibility and utility of adopting such platform for children diagnosed with CNS tumors in Hong Kong. A multi-institutional cohort (n=94, 97% of Chinese ethnicity) with CNS embryonal or high grade neuroepithelial tumors (HGNET) diagnosed in Hong Kong from 1996–2020 was assembled based on tissue availability. DNA was extracted from FFPE tumor material (median 301ng, range 13-1000ng), bisulfite converted and profiled with the Infinium Methylation EPIC BeadChip kit. Raw data were analyzed on the German Cancer Research Center MNP 2.0 classifier and through unsupervised dimensionality-reduction analysis (t-SNE) referencing a published CNS tumor reference dataset (GSE90496). The radio-histologic diagnosis included medulloblastoma (n=65), ATRT (n=9), pineal parenchymal tumors (n=7), ETMR (n=5), CNS-PNET (n=4) and other embryonal tumors/HGNETs (n=4). Methylation class could be assigned based on results from MNP 2.0 (calibrated score ≥ 0.9) in 62 patients (66%, including 2 clustering with control) and t-SNE in 22 (23%), while no-match was encountered in 10 (11%). Methylation-based analysis allowed confirmation of diagnosis and assignment of molecular subgroup in 64 patients (68%), confirmation of histologic diagnosis alone in 5 (5%) and resulted in revision/reassignment of diagnosis in 13 (14%). Among medulloblastoma samples that were assigned molecular tumor classes (n=57), 8 clustered with WNT-activated medulloblastoma, 13 with SHH-activated medulloblastoma, 10 with Group 3 medulloblastoma, 21 with Group 4 medulloblastoma, and 5 with non-medulloblastoma entities (high-grade gliomas=3, ETMR=1, ATRT=1). In conclusion, epigenomic profiling allowed refinement of disease classification for pediatric CNS tumors. Availability of such methodology in Asia sets the stage for international collaborations in molecularly-driven trials.
Published Version (Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have