Development of a clinically applicable assay for medulloblastoma molecular subgroup classification. (25th February 2016)

• Record Type: Journal Article
• Authors: Mather, Michael; Hicks, Debbie; Clifford, Steve
• Is Part Of: Lancet. Volume 387(2016)Supplement 1
• Page Start: S70
• Publication Date: 2016-02-25
• Subjects: Medicine -- Periodicals; Medicine -- Periodicals; Medicine; Medicine; Electronic journals; Periodicals; 610.5
• DOI: 10.1016/S0140-6736(16)00457-8
• Languages: English
• ISSNs: 0140-6736

Abstract

Abstract: Background: Medulloblastoma is the most common malignant tumour of childhood. Recent work shows that medulloblastoma comprises four molecular subgroups, which can be classified by a 450k DNA-methylation array. However, because this method is not practicable for routine clinical practice, there is a major unmet need for patient care and clinical trials. Work within our group suggests that a clinically applicable DNA-methylation-based, minimal-signature assay can rapidly identify these subgroups. We aimed to validate this new assay against the gold standard subgrouping by methylation array and apply it to tumour samples for which subgroup is unknown. Methods: 100 ng of DNA from both fresh frozen and formalin-fixed paraffin-embedded (FFPE) biopsy samples were chemically treated to induce methylation-dependent single nucleotide polymorphisms. Subsequent amplification at specific loci via multiplex PCR permitted methylation status characterisation via mass spectrometry which enabled subgroup classification using in-house developed algorithms. We then challenged the assay by using scant and degraded DNA extracted from nuclear preparations originally intended for fluorescence in-situ hybridisation. Subgroup classification was correlated with clinical data. Findings: 82·7% of fresh frozen, 62·9% of FFPE, and 73·3% of nuclear preparation samples could be amplified; 94 (91%) of 103 samples that passed quality control had a confidently assigned subgroup which, in every case, matched with the gold standard methylation array subgrouping call. Samples with as little as 50·4 ng of DNA were successfully subgrouped and some samples had been resected over 46 years ago (mean 12·7 years) with DNA quality (260 nm to 280 nm ratio) as low as 1·22 (mean 1·72). Interpretation: We have shown that a low-cost minimal-methylation signature assay can reliably subgroup medulloblastoma, even with archival or degraded DNA, and could be used clinically for routine patient stratification. This assay uses only a tenth of the DNA amount that is typically used in methylation array subgrouping. Further work should investigate use of minimal-methylation signatures to identify methylation subgroups in other cancers. Funding: North of England Children's Cancer Research Fund, Northern Institute for Cancer Research.