Molecular-genetic mechanisms of cerebral arteriovenous malformations: A systematic review of germline, somatic and experimental evidence.

Cerebral arteriovenous malformations (AVMs) are high-flow vascular anomalies that carry a considerable burden of haemorrhagic morbidity. Work over the past decade has shown that both inherited syndromes and apparently sporadic lesions can be traced to alterations in a handful of signalling pathways, but the relevant evidence sits awkwardly between human genetic studies, animal models, and molecular profiling experiments. To bring together the current evidence on the molecular and genetic underpinnings of cerebral AVM formation, clinical course, and candidate therapeutic targets. We searched MEDLINE (PubMed), Embase, Web of Science Core Collection, and Scopus from inception to 1 January 2026. Records were screened, and full texts assessed, by two reviewers working independently. Studies were eligible if they reported genetic, molecular, or epigenetic data on cerebral AVMs. Because of substantial heterogeneity in designs and outcomes, evidence was synthesised qualitatively. Of 343 records, 44 primary studies (28 clinico-genetic and 16 experimental) together with 20 reviews and meta-analyses met inclusion. Inherited AVM syndromes largely reflected germline variants in the TGF-β/BMP network (ENG, ACVRL1, SMAD4, GDF2/BMP9) or in arteriovenous-specification regulators (RASA1, EPHB4); penetrance was incomplete and there was support for local modifying "second-hit" events. Resected sporadic cerebral AVMs often carried somatic activating mutations in the RAS/MAPK cascade-mostly KRAS, less often BRAF-at low variant allele fractions that demanded high-sensitivity assays. Endothelial-specific activation of KRAS or BRAF reproduced AVM-like phenotypes in vivo and showed MEK-ERK dependence. Susceptibility variants (notably at 9p21/ANRIL), inflammatory mediators, extracellular-matrix remodelling, and epigenetic changes appear to fine-tune lesion behaviour and haemorrhagic risk. Syndromic predisposition and clonal somatic events in endothelium converge on a small set of developmental signalling axes that govern angiogenesis and vessel-wall maturation. Diagnostics anchored in this biology, together with pathway-targeted therapy, now need prospective evaluation, paired with standardised approaches for detecting low-level mosaicism.