How roughly three hundred driver genes converge on a few signalling and checkpoint circuits. One composite cell map, seventeen individual pathways, and how often each circuit is broken.
Faded dashed nodes belong to other pathways and show where this circuit plugs into the rest of the network. Click any gene for its entry on the composite map.
Solid bars are figures stated in the text of Sanchez-Vega et al. 2018 [1]: RTK–RAS is the median across cancer types, NRF2 is the overall rate. Range bars are estimates read from that paper's figures and from gene-level rates. They are not tabulated values. Use cBioPortal for exact counts.
Approximate pan-cancer share of tumours with a likely driver alteration (mutation, focal amplification, deep deletion, fusion or promoter lesion), rounded from TCGA PanCancer studies [1, 2, 9–11] and PCAWG for TERT [7]. Colour gives the pathway.
Qualitative tiers synthesised from the TCGA pathway paper [1] and the tumour-type marker papers. Stated anchors: melanoma 95%, thyroid 84%, pancreas 78%, lung adenocarcinoma 74% RTK–RAS; lung squamous 68% and non-hypermutated endometrial up to 95% PI3K; NRF2 25% in lung squamous [1]. The same pathway is often hit through different genes in different tissues (for example KRAS in pancreas, BRAF in thyroid, EGFR in glioblastoma), which reflects cell of origin [15].
| Pathway | Agent | Target | Biomarker / setting | Status | Evidence |
|---|
FDA year is the first US approval of that agent (or the first of the group) in this pathway context. Indications widen and narrow over time, and approvals elsewhere differ. This is a research summary, not treatment guidance.
Filled dot: a primary, well-established consequence of the pathway's lesions. Ring: a secondary or context-dependent contribution. The last four columns are the 2022 additions (one hallmark and three enabling characteristics or cell types).
Scope and limits. Frequencies come from primary-tumour cohorts (mostly TCGA), which under-represent metastatic, paediatric, haematological and non-European-ancestry cancers. Purity limits detection of deletions, so suppressor loss (for example CDKN2A in pancreas) is undercounted. Gene-level percentages in this atlas are rounded guides for comparison and should not be cited as point estimates. Drug names reflect approvals and trials up to 2026 and are not treatment advice.