Variant curation and interpretation in hereditary cancer genes: An institutional experience in Latin America

Background: Variant curation refers to the application of evidence-based methods for the interpretation of genetic variants. Significant variability in this process among laboratories affects clinical practice. For admixed Hispanic/Latino populations, underrepresented in genomic databases, the interpretation of genetic variants for cancer risk is challenging.

Methods: We retrospectively evaluated 601 sequence variants detected in patients participating in the largest Institutional Hereditary Cancer Program in Colombia. VarSome and PathoMAN were used for automated curation, and ACMG/AMP and Sherloc criteria were applied for manual curation.

Results: Regarding the automated curation, 11% of the variants (64/601) were reclassified, 59% (354/601) had no changes in its interpretation, and the other 30% (183/601) presented conflicting interpretations. With respect to manual curation, of the 183 variants with conflicting interpretations, 17% (N = 31) were reclassified, 66% (N = 120) had no changes in their initial interpretation, and 17% (N = 32) remained with conflicting interpretation status. Overall, 91% of the VUS were downgraded and 9% were upgraded.

Conclusions: Most VUS were reclassified as benign/likely benign. Since false-positive and -negative results can be obtained with automated tools, manual curation should also be used as a complement. Our results contribute to improving cancer risk assessment and management for a broad range of hereditary cancer syndromes in Hispanic/Latino populations.