HECTOR: A Web-Based Tool for Automated BRCA1/BRCA2 Variant Classification Under the ClinGen ENIGMA Specifications
Background: The ClinGen ENIGMA BRCA1/BRCA2 Variant Curation Expert Panel (VCEP) has adapted the ACMG/AMP framework into gene-specific specifications. However, applying these specifications manually remains labour-intensive and prone to inconsistency, requiring integration of population, computational, functional, and clinical evidence through gene-specific decision trees and a points-based classification system. Methods: We developed HECTOR, a free web-based tool that implements the complete ENIGMA VCEP v1.2 specifications for BRCA1 and BRCA2. HECTOR automatically populates all evidence codes derivable from public data, routes curator-dependent evidence to a manual input layer and returns a transparent five-tier classification with code-level evidence. We validated HECTOR against two independent reference datasets: the 143-variant ENIGMA Evidence Repository, used as a clinical-grade reference standard, and 134 manually curated in-house variants of uncertain significance. HECTOR was then applied to the complete ClinVar BRCA1/BRCA2 catalogue (n = 34,077). Results: At the criterion level, HECTOR exactly reproduced 326 of 413 VCEP-assigned criteria (78.9%). The discordance arising predominantly from curator-dependent evidence rather than implementation errors whereas computationally accessible criteria showed perfect concordance. Across ClinVar, HECTOR classified 33,913 variants (99.5%). Agreement with definitive ClinVar classifications was 96.7% for pathogenic variants overall. Among variants for which HECTOR generated a definitive classification, directional concordance reached 99.7% for pathogenic and 99.9% for benign variants. HECTOR also resolved a substantial proportion of variants classified as uncertain (67.3%) or conflicting (88.7%), predominantly toward benign classifications. Conclusions: HECTOR provides a faithful, transparent implementation of the ENIGMA VCEP v1.2 specifications for BRCA1 and BRCA2, enabling rapid, standardized, and reproducible application of gene-specific variant classification guidelines while reducing the burden of manual curation.