Functional data for LDLR variant classification: comparative insights from high-content microscopy and flow cytometry assays

Background and aims FH VCEP specifications for familial hypercholesterolemia variant interpretation assign a higher evidence weight to functional data from flow cytometry than microscopy assays. This restricts the use of microscopy assays for LDLR variant classification. We aimed to compare functional data obtained by both assays to determine whether microscopy-based assays could support a higher evidence level. Methods Fifty LDLR variants with available flow cytometry and microscopy data were compared for LDL uptake activity, including 21 newly characterized variants by microscopy. Variants were grouped by FH VCEP functional thresholds and results integrated with UK Biobank data to assess associations with lipid traits. Results We validated our microscopy assay with FH VCEP-classified control variants and compared functional activities measured by microscopy and flow cytometry assays, which showed significant correlation (r = 0.66, 95% CI: 0.473–0.795, p=1.45 × 10 -7 ) and a close average agreement (Bland–Altman bias=–0.05, 95% LoA: -0.69 to +0.60). Functional FH VCEP variant groups for both assays resulted in fair categorical agreement (Cohen's κ = 0.26, 95% CI: 0.018–0.479). Integration with UK Biobank data showed that carriers of variants with reduced LDLR activity (<70% and <50%) had higher LDL-C, total cholesterol, and ApoB levels compared to those with normal activity (>90%) for both microscopy and flow cytometry assays, with more pronounced differences observed at the <50% LDLR activity threshold. Conclusion Microscopy assays show concordance with flow cytometry, consistent associations with lipid phenotypes and our OddsPath calculations support re-evaluation of evidence weights for microscopy assays.