How to Choose the Right Cancer Knowledge Base for Variant Interpretation
We know we need to overcome the challenges of oncological interpretation in clinical laboratories. But how exactly do we do that?
Sara Patterson, Cara Statz | | 6 min read | Opinion
When clinical laboratory teams first began performing variant analysis for genetic test results, it wasn’t too onerous. Most genetic tests were limited to a single gene, and the amount of relevant information from the scientific and clinical literature that had to be reviewed in the interpretation process was manageable.
Times have certainly changed. For cancer testing alone, clinical lab teams are expected to analyze and interpret variants in dozens of genes for a comprehensive panel – or even manage the output of whole exome or whole genome sequencing workflows, sometimes comparing results from both tumor and normal tissue. This more comprehensive approach has certainly been a tremendous boon to patients because they are more likely to be matched to the most effective treatment or the most promising clinical trial. But the underlying process has become a workflow nightmare for molecular pathologists and variant interpretation specialists. For one thing, the universe of potential variants is enormous. Even if each variant could be reviewed completely in a few minutes, sifting through all of the variants found by one of these tests can take hours or days. That’s hardly scalable for laboratory teams running tests for many cancer patients each day.
One option, of course, is to outsource the entire process. There are several companies that specialize in genomic testing and interpretation in cancer that are eager for new business. But for clinical laboratories, the benefits of outsourcing are outweighed by longer turnaround times and high costs. Many teams would prefer to offer genomic testing capabilities in-house, where they can better manage reporting times and expenses.
To make in-house variant interpretation feasible, though, clinical labs need access to more comprehensive information about cancer variants, targeted therapies, clinical trials, and drug approval status – ideally in a one-stop shop to make the process as fast as possible. Though variant interpretation tools have struggled to meet these requirements in the past, the latest generation of cancer knowledge bases could give clinical laboratory teams what they need to produce the best results for their patients.
Before searching for the ideal solution, it’s important to take a clear-eyed look at the current challenges facing clinical lab teams in the variant interpretation process. Only with that foundation can scientists carefully assess potential approaches to address those issues.
Too many tools
As demand for interpreting cancer variants exploded in the past couple of decades, academic scientists and corporate developers alike jumped in to help. The result of these well-intentioned efforts is a cluttered landscape of tools. Some cancer databases were abandoned after launch; they linger online, with no warning for users that the information they contain is hopelessly obsolete. Certain tools were built for very specific uses, making them valuable, but they force users to jury-rig interpretation workflows and cobble together enough niche tools to produce even the most broadly relevant of results. Well-known and widely used databases contain slivers of information, so each new report of variants must be checked across multiple databases for the most comprehensive results.
As the community’s understanding of cancer-associated variants has improved over the years, the scientific and clinical literature has been filled with case reports, large genome-wide association studies, and everything in-between. The number of papers that must be reviewed to ensure a clear interpretation of any given variant can be overwhelming. In addition, there’s the ongoing challenge of paywalled content. No single clinical laboratory can maintain subscriptions to all possible scientific journals, leaving teams with the high costs of paying for access to individual articles ad hoc or with the fear of possibly missing important but inaccessible content.
Factors to Consider
Next-generation cancer knowledge bases claim to address some or all of these challenges by pulling many discrete sources of information into a single portal. These tools offer a vast improvement to the hodgepodge workflows that have arisen in many clinical laboratories. But not all tools perform equally well, so it is essential to assess certain critical features when choosing the best option.
Plays well with others
Ease of use is one of the most important features of any cancer knowledge base that aims to serve as a single source for variant interpretation. The user interface should be seamless and search capabilities should be intuitive. Any reputable tool should allow prospective users to try it out for free – even on a limited data set – to get a real feel for how it operates.
For the most accurate and actionable patient results, variant interpretation must be performed with the latest data – including drug regulatory status, scientific reports, professional guidelines, and more. Ideally, a cancer knowledge base should be updated on a daily basis to ensure that clinical laboratory scientists are working from the most up-to-date information possible. It’s also important to know who is populating and maintaining the knowledge base. Scientific evidence and other information should be curated by experts, such as PhD- or MD-level scientists and clinicians, who have a clear understanding of the data they’re reviewing and how it will be incorporated into variant interpretation results.
To avoid the disaster of missing critical information that would have changed the interpretation of a variant, a cancer knowledge base must be comprehensive. It should cover the hundreds of genes and thousands of variants that have already been linked to cancer, with thorough molecular content – pulled from the complete scientific and clinical literature – made available at the level of the gene or the variant. Ideally, the cancer knowledge base should also represent complex molecular profiles. In addition, no knowledge base can be considered comprehensive without exhaustive data about targeted therapies and their current regulatory status, clinical trials with clear enrollment information, and evidence of therapeutic efficacy.
A tool that meets all of these requirements can empower clinical laboratory teams to make better recommendations for patient care by linking the molecular profile of variants – including complex variants, such as gene fusions – to their likely effect on a patient’s response to specific therapies. Even closely related variants may have very different functional outcomes – but a sophisticated cancer knowledge base can help molecular pathologists and other clinical scientists tell them apart.
For variant interpretation, clinical lab teams need a resource that can mine the continuous stream of new cancer data. Fortunately, the best cancer knowledge bases have finally reached the point that they can be considered reliable resources for analyzing results of genomic tests for patients with cancer.
Credit: Images sourced from Unsplash.com