NEW YORK (GenomeWeb) – A team lead by the Association for Molecular Pathology has developed a set of new standards and guidelines for interpreting and reporting sequence variants in cancer that it hopes will improve genomic testing and precision care for cancer patients.
The guidelines, published in the Journal of Molecular Diagnostics today, were devised by a working group of the AMP Clinical Practice Committee that included representatives from the American College of Medical Genetics and Genomics, the American Society of Clinical Oncology, and the College of American Pathologists. According to the publication, they emulate "the expert consensus opinion of the working group members with input from the stakeholders they represent."
In a nutshell, the guidelines propose a four-tiered system for classifying somatic sequence variants, based on their clinical significance for diagnosis, prognosis, and therapy. Tier I includes variants with the highest level of evidence, such as an associated US Food and Drug Administration-approved drug, inclusion in professional guidelines, or results from well-powered studies.
Tier II contains variants with potential clinical significance, for example those associated with drugs that are FDA-approved for a different tumor type, variants associated with investigational therapies, or results from small studies or preclinical trials. Tier III variants are currently of unknown clinical significance and tier IV variants are considered benign or likely benign. The guidelines also describe a process for variant annotation, classification, and reporting and include a list of commonly used bioinformatics tools and databases.
"It is our hope that we will soon see the widespread adoption of these guidelines to improve communication between molecular pathologists, oncologists, and geneticists" and ultimately, to improve patient care, said Marilyn Li, vice chief of the division of genomic diagnostics and director of cancer genomic diagnostics at Children's Hospital of Philadelphia, who chaired the working group.
Li told GenomeWeb that as more and more clinical laboratories adopted next-generation sequencing for genomic tumor testing over the years, questions arose over how best to interpret and report the results.
AMP's Interpretation of Sequence Variants in Somatic Conditions working group came into being after a 2014 roundtable meeting of AMP, ASCO, and CAP, she said, and was formally established in early 2015. It includes experts from a number of US academic medical centers who also represent the various professional societies. Commercial players were not part of the group but industry lab professionals, through their AMP membership, were able to provide suggestions and feedback, Li said.
The group deliberately confined its work to the interpretation and reporting of cancer variants. Another AMP working group, led by Lawrence Jennings, a pathologist at Lurie Children's Hospital of Chicago, is tasked with developing standards and guidelines for the development, optimization, and validation of NGS-based oncology panels that are still pending. Both groups presented their work at AMP's annual meeting in Charlotte last month.
Li explained that her group started by sending two surveys to AMP members last year, one asking them about the technology they use for cancer NGS testing, the other how they report the results. On the technical side, they found that there was no consensus on how to validate tests, which led to the creation of the second working group. The other major difference was how labs classified and reported variants. In the absence of standards, many labs took a page from the ACMG/AMP guidelines for germline variant interpretation, which were designed for inherited disease genetic testing and focus on the pathogenicity of variants.
However, the interpretation of somatic variants should consider not just their disease-causing potential but also their impact on clinical care, Li explained. The same variant, for example, may have different effects in different types of cancer. Also, cancer genomes often harbor multiple variants, with different allele frequencies because of the heterogeneity of tumors, which affects their clinical significance. Overall, "the existing guidelines on germline variant interpretation do not meet the much more complicated needs of cancer genomics," she said.
To come up with the new guidelines, the working group, whose members have different clinical backgrounds, went on to conduct extensive literature searches and searches of numerous germline, somatic variant, and disease-specific databases. It also consulted professional guidelines for cancer, compiled feedback from several public meetings, and incorporated the professional expertise and experience of its members. The tiered classification system was ultimately chosen, Li said, "to address the complexity of somatic variant interpretation and reporting, with the goal to provide oncologists and patients this complex information in an easily digestible format."
It took a while to get there, though. "There were times when we had major differences," Li recalled, for example on the criteria used for each variant tier, or whether labs should be encouraged to report both somatic and germline variants, but they were able to reach agreement in the end, which sometimes required them to re-evaluate available evidence. "It took us quite some time to make sure that these guidelines are both easily understood and acceptable to all of the medical professions involved," she said, so they can be adopted not only by academic cancer centers but also by community group oncology practices.
Li said the guidelines are mainly directed at two major groups: clinical laboratory professionals, who run genomic cancer tests and interpret the results, and oncologists who read the reports and use them to make informed decisions on patient care.
The next step will be to foster adoption of the recommendations. The working group, together with other AMP members, is already developing webinars, variant assessment challenge material, and other forms of educational sessions to assist with that, Li said.
Interest in the guidelines appears to be strong, she said, judging by the large number of attendees of the early Saturday morning session at AMP's annual meeting last month where the working group presented the new standards. "I think this is very much anticipated and I think a lot of laboratories will adopt these," she said.
Olivier Elemento, head of the laboratory of cancer systems biology and an associate professor at Weill Cornell Medicine, said that his laboratory had already adopted many aspects of the proposed guidelines. "It is likely that we will review our reporting strategy as a result of the publication of the guidelines but I don't anticipate any major changes," he told GenomeWeb. The new standards are similar to existing best practices, he said, and he expects that most labs will review their analytical and reporting strategies "to converge towards the proposed recommendations."
"The oncology genomic testing field has been in dire need of standardization and guidelines that focus on clinical actionability of genomic testing results," he said, and the fact that the guidelines were developed by "highly respected organizations" means they will likely have a lasting impact.
However, he said he was surprised by a lack of emphasis on open, community-driven efforts to capture and catalog context and tumor-type dependent clinical interpretations of somatic alterations, such as CIViC, Weill Cornell's Precision Medicine Knowledgebase, or the Cancer Genome Interpreter. "Categorizing variants into tiers of clinical actionability needs extensive and continuous literature curation — it is critical that such efforts be shared and ideally coordinated across institutions," he said. "Otherwise we will continue to generate broadly discordant clinical reports."
Also, some recent developments are missing from the guidelines, he said, for example how to report mutational burden and neoepitope burden, which are both highly relevant for identifying patients likely to respond to immunotherapy.
Li expects the guidelines will evolve over time, and the working group plans to monitor their adoption and the need for updates. "We understand that cancer genomics is a rapidly evolving field," she said. "Therefore, the clinical significance of any variant in the therapy, diagnosis, or prognosis should be re-evaluated on an ongoing basis."
The guidelines will also go hand in hand with other standardization efforts, for example the minimum variant level data (MVLD) standards for the curation of somatic variants that were recently published by the Clinical Genome Resource (ClinGen) Cancer Somatic Working Group, of which Li is a member.