Hudson Institute of Medical Research, Clayton, Victoria, Australia; Department of Molecular and Translational Science, Monash University, Clayton, Victoria, Australia; Victorian Clinical Genetics Services, Parkville, Victoria, Australia.
Hudson Institute of Medical Research, Clayton, Victoria, Australia; Department of Molecular and Translational Science, Monash University, Clayton, Victoria, Australia; Australian and New Zealand Children's Haematology/Oncology Group, Clayton, Victoria, Australia.
J Mol Diagn. 2023 Oct;25(10):709-728. doi: 10.1016/j.jmoldx.2023.06.013. Epub 2023 Jul 28.
DNA methylation array profiling for classifying pediatric central nervous system (CNS) tumors is a valuable adjunct to histopathology. However, unbiased prospective and interlaboratory validation studies have been lacking. The AIM BRAIN diagnostic trial involving 11 pediatric cancer centers in Australia and New Zealand was designed to test the feasibility of routine clinical testing and ran in parallel with the Molecular Neuropathology 2.0 (MNP2.0) study at Deutsches Krebsforschungszentrum (German Cancer Research Center). CNS tumors from 269 pediatric patients were prospectively tested on Illumina EPIC arrays, including 104 cases co-enrolled on MNP2.0. Using MNP classifier versions 11b4 and 12.5, we report classifications with a probability score ≥0.90 in 176 of 265 (66.4%) and 213 of 269 (79.2%) cases, respectively. Significant diagnostic information was obtained in 130 of 176 (74%) for 11b4, and 12 of 174 (7%) classifications were discordant with histopathology. Cases prospectively co-enrolled on MNP2.0 gave concordant classifications (99%) and score thresholds (93%), demonstrating excellent test reproducibility and sensitivity. Overall, DNA methylation profiling is a robust single workflow technique with an acceptable diagnostic yield that is considerably enhanced by the extensive subgroup and copy number profile information generated by the platform. The platform has excellent test reproducibility and sensitivity and contributes significantly to CNS tumor diagnosis.
DNA 甲基化阵列分析用于小儿中枢神经系统 (CNS) 肿瘤的分类是组织病理学的重要辅助手段。然而,缺乏无偏倚的前瞻性和实验室间验证研究。涉及澳大利亚和新西兰 11 个儿科癌症中心的 AIM BRAIN 诊断试验旨在测试常规临床检测的可行性,并与德国癌症研究中心的分子神经病理学 2.0 (MNP2.0) 研究同时进行。269 名小儿患者的 CNS 肿瘤前瞻性地在 Illumina EPIC 阵列上进行了测试,其中 104 例同时参加了 MNP2.0。使用 MNP 分类器版本 11b4 和 12.5,我们报告了在 265 例中的 176 例(66.4%)和 269 例中的 213 例(79.2%)的分类,概率评分≥0.90。在 11b4 中,130 例中的 130 例(74%)获得了显著的诊断信息,174 例中的 12 例(7%)的分类与组织病理学不一致。前瞻性地共同参加 MNP2.0 的病例给出了一致的分类(99%)和评分阈值(93%),证明了测试的良好可重复性和敏感性。总的来说,DNA 甲基化分析是一种稳健的单一工作流程技术,具有可接受的诊断产量,该平台生成的广泛亚组和拷贝数谱信息极大地提高了产量。该平台具有出色的测试可重复性和敏感性,对 CNS 肿瘤诊断有重要贡献。