Fig. 2

Determining rational thresholds for establishing 1p/19q-codeletion and + 7/ − 10 status. A A scatter plot showing the percentage of genes lost on chromosome arm 1p against the percentage of genes lost on 19q for all tumors with all of three TCGA SCNA pipelines showing at least 50% gene loss on 1p and 19q and at most 85% genes loss on 1q and 19p. All tumors that satisfy an 85% threshold on all four chromosome arms are predicted to be 1p/19q-codeleted oligodendrogliomas; all those that do not are predicted to be IDH-mutant astrocytomas. Lines connect points that represent the same patient. B An 85% threshold for 1p/19q-codeletions is optimal for the TCGA training set and an independent validation dataset published by Capper et al. as measured by MCC. It is nearly optimal on an independent validation dataset published by Jonsson et al. C A scatter plot of percent gain/loss of genes on chromosomes 7 and 10, respectively, for 91 histological lower-grade TCGA IDH-wildtype gliomas. All those without EGFR amplification and at least 10% gene gain/loss on chromosomes 7 and 10 also satisfy a 50% threshold. D Histological lower grade TCGA IDH-wildtype gliomas without EGFR amplification who do not meet a 50% threshold for + 7/ − 10 are much younger than those who do. This age difference increases when TERTp mutation status is accounted for. E A 50% threshold for + 7/ − 10 separates + 7 and − 10 from intact 7 and intact 10 across three independent validation sets and the TCGA holdout validation set. F Of all TCGA histological lower-grade IDH-wildtype diffuse gliomas, 21 do not or cannot be confirmed to meet the WHO 2021 criteria for adult diffuse glioma