New oncogenicity classification guidelines for somatic variants

Franklin by Genoox
6 min readMar 29, 2022

New standards for somatic variant classification have recently been developed by Clinical Genome Resource (ClinGen), Cancer Genomics Consortium (CGC), and the Variant Interpretation for Cancer Consortium (VICC) to help establish the oncogenicity of variants, bringing more clarity to the interpretation of cancer mutations.

Why was there a need for more classification guidelines?

Well-known guidelines already exist for assigning somatic variants based on their clinical interpretation and actionability (diagnostic, prognostic, and therapeutic aspects). However, a critical gap was present regarding standards for oncogenicity classification.

As cancer research evolves and new somatic variants are increasingly being uncovered through tumor sequencing, the lack of consensus in this field may lead to the inconsistent classification of rare somatic variants in cancer, variability in their clinical interpretation, and, consequently, inadequate patient care. In particular, it is fairly common for FDA approvals for drugs and clinical trial inclusion criteria to mention the deleteriousness of a variant. Even when using other guidelines designed to classify variants in cancer, experts are encouraged to take into account its oncogenicity. Nevertheless, up until now, there was no standardized instruction on how to classify a somatic variant as oncogenic.

Therefore, these joint recommendations, published last January in Genetics in Medicine, aim to address this need, by providing explicit and consistent point-based criteria to classify the oncogenic function of variants as a standard operating procedure (SOP).

What’s the difference between the new and previous classification guidelines?

The standards were inspired by the American College of Medical Genetics and Genomics (ACMG) germline pathogenicity guidelines, and adapted to systematically categorize evidence of oncogenicity for somatic single nucleotide variants and small insertions/deletions, resulting in 5 categories: Oncogenic, Likely Oncogenic, Variant of Uncertain Significance (VUS), Likely Benign and Benign.

It is important to emphasize that this SOP does not aim to replace the Association for Molecular Pathology (AMP) somatic guidelines for clinical actionability assessment. On the contrary, as represented in the publication, the guidelines should be used in conjunction with the AMP classification system. Here’s the suggested step-by-step workflow for somatic variant classification and interpretation with this new tool:

Suggested step-by-step workflow for somatic variant classification and interpretation with this new tool

How are the classification standards structured?

This work defines variant oncogenicity as the pathogenicity of the variant in the context of neoplastic disease and relies on different lines of evidence, such as population frequency, functional data, recurrence algorithms, and predictive tools under a quantitative comprehensive framework for somatic variant classification. Similar to what is stated in the ACMG germline guidelines, evidence strength can be rated Very Strong, Strong, Moderate, or Supporting, and in this case, add up to the different amounts of points. The scoring system ranges from less than -7 (benign) to more than 10 (oncogenic), as the image below shows.

Evidence strength can be rated Very Strong, Strong, Moderate, or Supporting, and in this case, add up to the different amounts of points

The article by P. Horak et al, which was made possible through a collaborative community effort across the three notable consortia, establishes 17 criteria for evidence of oncogenicity of somatic variants, that can be grouped into 6 different categories, such as:

  • Effect on Protein (OVS1, OM1, OM2, and SBP2): Rules that depend on the impact the variant has on gene functionality. The major difference with the ACMG guidelines is that, for loss-of-function related evidence, these oncogenicity rules are advised to only be applied in the case of tumor suppressor genes.
  • Cancer hotspots (OS3, OM4, and OP3): Different from the ACMG guidelines, this standard uses the database cancerhotspots.org as the main resource to determine whether or not the variant lies on a genetic hotspot, with very specific details on the number of samples reported with the same variant needed in order to apply the rule.
  • Previously classified variants in the region (OS1, OM3)
  • In silico prediction tools (OP1, SBP1): Evaluation of bioinformatic predictors is fairly similar to what is described in the ACMG guidelines.
  • Frequency data (OP4, SBVS1, SBS1): The treatment of normal frequency databases is comparable to the one advised in the ACMG germline guidelines, with a slightly reduced contribution to overall oncogenicity.
  • Functional studies (OS2 and SBS2): Similar to the ACMG guidelines, in vitro or in vivo studies have a significant impact on the classification of variants. What’s more, since cancer is a highly prolific field for these types of analysis, it is expected for these rules to be more applicable in comparison to the germline workflow.
  • Etiology (OP2): The standard takes into account the genes that are involved in single-cause malignancies, such as bi-allelic RB1 inactivation in retinoblastoma.

How do the new classification guidelines work?

There are several examples provided by the consortia in the publication, especially in the Supplementary Material. These variants, classified by the new SOP, help provide future users with more clarity on how the guidelines should be applied.

For instance, let’s examine the BRAF:c.1780G>A variant. Mutations in the BRAF are the most frequently identified cancer-causing variants in melanoma, and have been identified in various other cancers as well, including non-Hodgkin lymphoma, colorectal cancer, thyroid carcinoma, non-small cell lung carcinoma, hairy cell leukemia, and adenocarcinoma of the lung. In particular, the oncogenic effect of this variant is supported by two strong criteria (4 points each) and two supporting ones (3 points each), leading to an Oncogenic classification. OS2 was applied due to the existence of well-established functional studies that support the oncogenicity of the variant (such as Smalley et al and Wan et al), while the OS3 rule is met according to cancerhotspots.org count, which is 51 samples for that position and 24 for the same amino acid change. As for supporting rules, the variant is predicted to have pathogenic/deleterious effects on FATHMM and CADD (OP1), and it is also absent in the gnomAD normal population database (OP4). This variant is additionally classified as Tier I by the AMP guidelines.

The authors of the paper have stated that the main clinical application of these guidelines would be to classify somatic variants that are not 100% aligned with the previous AMP guidelines’ Tier I classification. This interpretation tool can offer clinicians further assessment of oncogenicity for variants that are not specifically listed in professional guidelines or FDA approvals, impacting their decision-making process. What is more, the new standard might help geneticists gather information regarding the therapeutic, diagnostic, or prognostic significance of somatic variants. A likely oncogenic or oncogenic classification of a variant might even result in a reassignment of an AMP Tier III (variant of unknown significance) to Tier I or II (variants of strong or potential clinical significance).

It is also important to note that for some variants, the ACMG germline classification might offer more insights than this new SOP. An interesting example of this is FLT3:c.2503_2504delGAinsAT, a two-base pair change in the well-known FLT3 gene, which is widely associated with acute myeloid leukemia (AML). In this case, the variant receives at least a Likely Pathogenic classification following the ACMG criteria (PM1, PM2, and PM5), and a Tier I classification by the AMP guidelines (several Level A therapeutic and diagnostic evidences). However, using the new SOP for somatic interpretation, it would only amount to a VUS (meeting OM1, OP4, with a total score of 3 points).

When will the new classifications be available on Franklin?

You can expect to see the new SOP fully integrated with Franklin in the upcoming weeks. In addition, the Genoox team will be conducting training sessions for new and old users to review these new guidelines and how they can be streamlined into the day-to-day variant interpretation workflow. If you would like us to contact you once the new SOP is up and running, please leave your information here in a comment or contact us.

--

--

Franklin by Genoox

The world’s largest, most diverse real-world genetic evidence database. We believe that community data-sharing is the future of genomics