Abstract P5-03-03: Variant Classification Discordance: A real-world experience of genetic test results in a community-based setting

Abstract

Abstract BACKGROUND Accurate interpretation of hereditary cancer germline genetic variants is critical to ensuring appropriate care. Myriad Genetics has developed tools that are instrumental in the accurate classification of variants, including a previously described history-weighting algorithm (Pheno), mutation co-occurrence statistical analysis (MCO), and in trans haplotype analysis. Differences in classification are known to occur among commercial testing laboratories, however the rate at which a single provider may observe variants with a different classification has not been reported. Here, we compared genetic test results from multiple laboratories ordered by a single surgical community-based practice with the classifications from Myriad’s testing laboratory. METHODS Variants initially reported as a “variant of uncertain significance” (VUS) on hereditary cancer test results from multiple commercial laboratories ordered by a single surgeon at a community-based, comprehensive breast center from June 2013 to May 2021 were evaluated and compared to the classifications from Myriad. In total, 212 variants were submitted for comparison. After review, 42 variants were excluded because they had not been observed previously in Myriad-tested patients. Therefore, 170 variants were eligible for comparison. Variants were classified as pathogenic/likely pathogenic, VUS, and benign/likely benign for comparison. Descriptive statistics were used for analysis. RESULTS Discordant classification was observed between Myriad and other testing laboratories for 28.2% (48/170) of the variants compared. Initially, all 170 variants were classified and reported by other testing laboratories as VUS, however, 15 variants (8.8%) were subsequently reclassified by other testing laboratories (N=13 were reclassified as benign/likely benign; N=2 were reclassified as pathogenic/likely pathogenic). Among all variants compared, 23.5% (40/170) were definitively classified by Myriad as benign/likely benign. For 90.0% (36/40) of these variants, evidence driving the classification relied upon Pheno (32 variants), MCO (4 variants), and in trans haplotype analysis (8 variants), with some variants having multiple lines of evidence. Other in-house specific rules were used for the remaining classifications (4/40;10%). Fifty-five percent (22/40) of the discordant classifications were seen in high-risk genes including APC, BRCA1/2, MLH1, MSH2, MSH6, PMS2, PALB2 and STK11, and 45% (18/40) were in moderate-risk genes including ATM, BARD1, BRIP1, CHEK2, and RAD51C. Of the 13 variants reclassified by other testing laboratories as benign/likely benign, six were classified as VUS by Myriad. Notably, two variants classified by Myriad as VUS were reclassified by other laboratories as pathogenic/likely pathogenic. CONCLUSIONS These data indicate that even in a single practice, significant discordance in variant classification exists based on the chosen laboratory. Myriad definitively classified nearly one-quarter of variants classified as VUS by other laboratories, likely due to the use of Myriad’s laboratory-developed classification tools. The degree of discordance observed here reflects the need for continuous laboratory investment in variant classification tools and evaluation of genetic variants, enabling physicians and patients to receive accurate results to facilitate appropriate medical management decisions. Citation Format: Shelly Cummings, Erin Mundt, Ann Marie Miller, TinaMarie Bauman, Thomas Slavin, Robert Maganini. Variant Classification Discordance: A real-world experience of genetic test results in a community-based setting [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P5-03-03.