Genotype–phenotype correlations and protein domain-level predictors of cerebrovascular malformations in hereditary hemorrhagic telangiectasia

Background: Hereditary Hemorrhagic Telangiectasia (HHT) exhibits marked phenotypic heterogeneity. Although gene–organ associations are well established for visceral involvement, predictors of cerebrovascular malformations (CVMs), particularly brain arteriovenous malformations (bAVMs), remain incompletely defined. This study aimed to investigate genotype–phenotype correlations and identify predictors of bAVMs in a genetically confirmed HHT cohort. Methods: We conducted a retrospective analysis of 142 Caucasian patients with genetically confirmed HHT. Clinical manifestations were systematically assessed and correlated with the mutated gene (ENG, ACVRL1, and SMAD4), variant type (truncating vs. non-truncating), and protein domain location. Multivariable logistic regression was performed to identify independent predictors of bAVMs. Results: The cohort included 83 (58.5%) ACVRL1 and 53 (37.3%) ENG mutation carriers. Bivariate analysis demonstrated distinct phenotypic patterns. ENG mutations were strongly associated with pulmonary AVMs (p < 0.001) and bAVMs (p < 0.001), with bAVMs observed in 35.8% of ENG carriers compared with 3.6% of ACVRL1 carriers. In contrast, hepatic AVMs were more frequent among ACVRL1 carriers (44.6%), although this did not reach statistical significance (p = 0.086). In the multivariable logistic regression model (overall p < 0.001), younger age emerged as the sole independent predictor of bAVMs (OR 0.968, p = 0.040), whereas the mutated gene did not retain independent significance. Conclusion: ENG mutation carriers display a markedly increased cerebrovascular burden, confirming a gene-specific susceptibility to bAVMs. Younger age independently predicts bAVM presence. These findings support age- and genotype-informed risk stratification and may help refine screening strategies in HHT patients.