Thoreson, Emily K

Abstract Background DPYD testing can identify patients at risk for severe toxicity when a standard fluoropyrimidine antineoplastic dose is prescribed. Sequencing-based methods detect rare variants, which require classification. Methods DPYD sequencing results over a 9-year period through to 2024 were analyzed. The dataset was filtered to calculate the positive test rate for different genotyping strategies, including targeted analysis of the Association for Molecular Pathology (AMP) Tier 1 and 2 recommended alleles, and as compared to sequencing. Variant classifications based on the laboratory’s clinical protocol were compared to ClinVar classifications and designations by the Clinical Pharmacogenetics Implementation Consortium (CPIC) expert panel. Results A DPYD variant was reported in 12.79% (896/7003) of those tested, with 2.53% (170/7003) of all reports including a variant of uncertain significance. Had only AMP Tier 1 or both Tier 1 and 2 been tested, 9.30% and 9.44% of patients would have received a positive report, respectively. Among the 281 unique, non-Tier 1/Tier 2 variants encountered during sequencing, >80% were not classified by the CPIC expert panel, while >48% were not reported in ClinVar. Variant classifications by our laboratory and those in ClinVar were generally concordant with the CPIC expert panel, despite the lack of a standardized classification system. Conclusions While AMP Tier 1 genotyping would have identified most patients at risk for fluoropyrimidine toxicity, sequencing is more comprehensive, identifying 26% more reportable variants compared to targeted genotyping. A standard framework for pharmacogenomic variant classification could facilitate consistent variant classification and a transition from targeted genotyping to sequencing.

AbstractPathogenic variants in COL1A1 and COL1A2 are involved in osteogenesis imperfecta (OI) and, rarely, Ehlers‐Danlos syndrome (EDS) subtypes and OI‐EDS overlap syndromes (OIEDS1 and OIEDS2, respectively). Here we describe a cohort of 34 individuals with likely pathogenic and pathogenic variants in COL1A1 and COL1A2, 15 of whom have potential OIEDS1 (n = 5) or OIEDS2 (n = 10). A predominant OI phenotype and COL1A1 frameshift variants are present in 4/5 cases with potential OIEDS1. On the other hand, 9/10 potential OIEDS2 cases have a predominant EDS phenotype, including four with an initial diagnosis of hypermobile EDS (hEDS). An additional case with a predominant EDS phenotype had a COL1A1 arginine‐to‐cysteine variant that was originally misclassified as a variant of uncertain significance despite this type of variant being associated with classical EDS with vascular fragility. Vascular/arterial fragility was observed in 4/15 individuals (including one individual with an original diagnosis of hEDS), which underscores the unique clinical surveillance and management needs in these patients. In comparison to previously described OIEDS1/2, we observed differentiating features that should be considered to refine currently proposed criteria for genetic testing in OIEDS, which will be beneficial for diagnosis and management. Additionally, these results highlight the importance of gene‐specific knowledge for informed variant classification and point to a potential genetic resolution (COL1A2) for some cases of clinically diagnosed hEDS.