Towards an External Quality Assessment for Next Generation Sequencing in the Diagnosis of Rare Inherited Anaemias

The diagnosis of patients with inherited anemias is increasingly made in conjunction with high-throughput 'next-generation' sequencing (NGS) analysis, largely using targeted resequencing panels validated for clinical use in diagnostic labs. While there is a joint UK NEQAS and European Molecular Quality Network pilot scheme for Molecular Genetics to assess NGS quality control, this is not disease-specific. Patients with inherited anaemias can have multiple mutations with complex genotype phenotype interactions therefore a scheme assessing interpretation of these results could be of value. Likewise, guidelines for variant reporting (eg ACMG- American College of Medical Genetics and Genomics) provide excellent advice on how to interpret the likely pathogenicity of genetic variants, but no disease-specific guidance exists to assist in the clinical interpretation of NGS findings for individuals with rare inherited anemias. The European Hematology Association Scientific Working Group on Red Cells and Iron carried out a survey among its members to investigate variability in current practice and determine the need for, and feasibility of, a formal external quality assessment (EQA) scheme. Surveys and two clinical vignettes with sample variant call files (VCFs) were distributed among 14 participating labs from 9 countries; 13/14 labs used a targeted panel and 1/14 lab (8%) used a 366-gene virtual panel derived from whole exome sequencing data. Accreditation: 12/14 labs had ISO (International Organisation for Standardisation) accreditation for their NGS; 2/12,used local accreditation schemes. The number of genes per targeted panel ranged from 18 to 215 genes (median 64), covering: congenital dyserythropoietic anemias, Diamond-Blackfan anemia, sideroblastic anemia, red cell membrane and enzyme disorders. Some panels included other related bone marrow failure or iron metabolism disorders. 7/13 labs with targeted panels sequenced only exons, with variable padding into introns, while 6/13 routinely sequenced 3' and 5' untranslated regions. Capture methods were variable between labs and 11/14 labs used Illumina platforms for sequencing and 3/14 Ion Torrent. Sanger sequencing was used for confirmation of NGS variants in 12/14 labs, but used for gap-filling of uncovered regions in 10/14 labs. Reporting of variants followed ACMG guidelines in 10/14 labs, ACGS (Association for Clinical Genomics Science) guidelines in 2/14 and no published guidelines in 2/14. Reporting of Tier 3 (variants of uncertain significance): 8/14 labs reported strict adherence to ACMG guidelines, including only Class 4 and Class 5 variants in clinical reports, while 6/14 labs admitted looser adherence and reporting of Class 3 variants depending on circumstances. The number of samples analysed per year was highly variable between labs (10-600, median 60). Two mini-EQA VCFs were sent with clinical vignettes, for which 100% of labs correctly identified a case of autosomal recessive sideroblastic anaemia due to compound heterozygous mutations in SLC25A38, and 100% correctly identified a case negative by NGS. Class 3 variants were not reported at all in 50% of clinical reports, reported in the main body of the report in 20% and in a separate table in 30% of labs. In conclusion, we have identified common approaches to NGS sequencing in 14 diagnostic laboratories but highlighted variability in accreditation, use of Sanger sequencing and adherence to ACMG guidelines. The feasibility of carrying out an EQA scheme has been established and work will continue with UK NEQAS to formally create such a scheme, with the aim of ensuring improved patient care through the use of objective quality assessments.