44. Reevaluation of copy number variant (CNV) classifications in the clinical laboratory setting: challenges, insights, and experiences with a laboratory-initiated process

Reevaluation and reanalysis of previously performed genomic tests is becoming more commonplace, through laboratory-initiated processes, clinician and even patient requests. According to recent ACMG policy statements, laboratories share the duty to recontact when variants or gene-disease relationships are reclassified. Establishing processes for reanalysis and recontact represents a significant challenge for clinical laboratories, particularly cytogenetics laboratories with microarray and interpretive data which often predate implementation of more sophisticated databases. We share our experiences from a first attempt at establishing a process for reevaluation of copy number variant (CNV) classifications in an academic reference laboratory, where access to clinical information and clinician-patient recontact represent notable obstacles. Approximately 4000 CNVs identified through microarray testing were reviewed, including CNVs classified as variants of uncertain significance (VUS) or likely benign (LB) and overlapping genes and regions with established haploinsufficiency/triplosensitivity by ClinGen Dosage Sensitivity Map curations (DS Score=3, DS3). Of 45 VUS/LB CNVs overlapping DS3 genes, 10 were reclassified to pathogenic (P) or likely pathogenic (LP) based on new evidence in the literature. Amended reports and verbal communication of the updated result was provided in all cases. In some, updated clinical history obtained during reevaluation aided in re-interpretation of the variant. Additionally, 12/45 CNVs involving DS3 genes remained classified as VUS, 9 of which only partially overlap the gene of interest, the functional effect of which is difficult to predict by microarray. We also identified 44 multiply-encountered CNVs (close-match: ≥99% overlap/size similarity) involving 244 cases with discordant classifications, many represented recurrent CNVs with reduced penetrance or deletions overlapping recessive-disease genes. Such discordances illustrate the need for better standards of classification of CNVs, which likely require refinement beyond the current 5-tier classification system. Although challenges remain, this CNV reevaluation process has informed our evolving curation and data sharing processes, with the overarching goal of improved patient care.