Assessment of Key Indicators of Pathogenicity in PKD1/2 Variants

Background: Autosomal dominant polycystic kidney disease (ADPKD) is primarily caused by pathogenic (P) and likely pathogenic (LP) variants in PKD1 and PKD2, with ~5% of affected individuals remaining genetically unresolved. Due to the high rate of novel variants seen in PKD1/2, many will be variants of uncertain significance (VUS) with insufficient evidence to be classified as either LP or likely benign (LB), and may comprise a significant proportion of the unresolved cases. According to ACMG, evidence including phenotypic and prevalence data in affected and/or unaffected populations can be used to support VUS reclassification. Here we utilized unaffected population prevalence and renal phenotypes to assess VUS in PKD1/2. Methods: A retrospective analysis of PKD1/2 variants identified via a 385 renal gene panel (the RenasightTM test) was performed. The gnomAD database (v2.1.1) was queried for variant population prevalence. Claims information were abstracted from the Komodo Health database to assess phenotypes based on the absence or presence of kidney cystic ICD codes (Q61.0-.9 and/or N28.1). Results: Across 61057 cases with test results, 4002 cases had a P/LP variant (1565 unique variants) and 5409 had a VUS in PKD1/2 (2900 unique variants) with no P/LP in a cystic gene. 97.3% of P/LP PKD1/2 variants were absent from gnomAD and none were present in >3 alleles (excluding 2 hypomorphic variants). Likewise, 70% (n=2025) of VUS were present in ≤ 3 alleles in gnomAD, of which 1310 (64.7%) were absent. 22.8% (n=661) of all VUS were both absent and had ≥1 patient with a cystic ICD code. Patients with the remaining absent VUS (n=649) had incomplete clinical profiles, preventing further evaluation by the testing lab. Conclusion: Absence of a variant from gnomAD and cystic disease ICD-10 codes are examples of evidence supporting pathogenicity, yet they are not sufficient to enable reclassification without additional evidence. Here we demonstrate that less than a quarter of VUS identified in PKD1/2 satisfy these pieces of evidence, edging these variants closer to LP. For an additional 22%, phenotypic information was lacking but could support reclassification. These findings highlight the important role physicians have in providing phenotypic information to the testing lab for patients with rare VUS.