Creating a Variant Database for the American Society of Hematalogy By Consensus Variant Classification of Common Genes Associated with Hematologic Malignancies

Background: With the rapid decline of sequencing costs and the introduction of next-generation sequencing (NGS) instruments with higher processing capacity, the rate-limiting step in reporting genetic test results is moving away from sequencing production towards data interpretation. Powerful bioinformatics tools have been created to address this bottleneck. Though substantial advances have been made in clinical variant classification, much work remains. This is particularly true in the evaluation of somatic variants, where there is currently a high degree of variability in how members of the global molecular genetics and pathology community establish and validate bioinformatics pipelines and apply classification guidelines, and in how results are reported to clinicians.

Aim: Create a reliable and widely available resource for variants relevant for hematologic malignancies and make their classification readily accessible to the wider hematology community, facilitating NGS diagnostics and improve the consistency and robustness of testing results to improve patient care .

Methods: The ASH Somatic Working Group (SWG) is comprised of hematologists, hematopathologists, molecular biologists and bioinformaticians. The group was formed under the auspices of ASH two years ago to improve clarity around the various gene panels being used in our community and in the reporting of genetic variants associated with hematologic malignancies. The group collected 1) the list of genes (i.e. entity-specific panels) that are typically used in research or clinical laboratory applications at 8 different and specialized laboratories; 2) the variants within those genes and their respective internally developed tiering and interpretations among 6 laboratories; and 3) the institutional bioinformatics pipelines used for those interpretations. The working group identified 70 genes most commonly used in assays for myeloid and lymphoid diseases.

A three-class system for defining somatic pathogenicity was used, to unify the submissions from 6 institutions. Pathogenic variants are those that have been clearly determined to be associated with tumorigenesis . Variants of unknown significance (VUS) may have evidence supporting or refuting their effect on somatic pathogenicity, but this evidence is not yet strong enough to classify the variants as pathogenic or benign. Benign variants have been determined not to drive tumorigenesis and are typically polymorphisms that are present throughout the (healthy) population at variable frequency.

Individual variant assessment was performed by the contributing institutions using a variety of resources, including ClinVar, COSMIC, cBioPortal/OnkoKB, St. Jude PCGP, ARUP, LOVD, dbSNP and IARC databases, as well as population frequency information from gnomAD/ExAC and in-silico mutation impact prediction tools included SIFT, PolyPhen2, Mutation assessor, LRT, FATHMM, CADD, REVEL and dbNSFP. Additionally, evidence was uncovered through manual literature review.

Results: Through an iterative process, the SWG has initially identified 202 variants of high confidence (with highly concordant interpretation amongst contributing laboratories) in 42 of 70 genes. These results will be publicly available as both as a multi-institutional collaborative manuscript on best practices in this field as well as a searchable web application on the ASH web site (www.hematology.org/research) to aid other clinicians and investigators. On this website variants can be queried using genomic coordinates based on hg19, GRCh38 or Gene and HGVSc nomenclature. With the ongoing process of resolving remaining discrepancies, the high-confidence data will be updated regularly, with evidence provided to support classification.

Conclusions: ASH has created an application to serve as a resource for NGS diagnostics based on a systematic review process, which we anticipate will lead to more consistency in molecular testing reports for patients undergoing evaluation of myeloid and lymphoid malignancies. Regular updates will help hematology professionals to use NGS in routine patient care on an ongoing basis.