Abstract
Ridofranz / iStock / Getty Images Plus
When targeted therapy first emerged as an option for cancer patients in the late 1990s, next-generation sequencing was not available as it is today. Instead, single gene assays were used to detect somatic molecular alterations, and the only widely available methods included cytogenetics, immunohistochemistry (IHC), fluorescence in situ hybridization (FISH), polymerase chain reaction (PCR), and “Sanger” sequencing of specific regions of genes. Furthermore, it was an era when only a few somatic genes were known to have predictive value and even then, those genes weren't mentioned on drug labels and no companion diagnostics had yet been approved.
At the turn of the 21st Century, diagnostic technology began to catch up and the cost of diagnostic DNA sequencing declined significantly, opening the door to opportunities for more targeted cancer care. For example, the first line treatment for advanced non-small cell lung cancer (NSCLC) patients was chemotherapy until the early 2000s. In 2003 and 2004, the first targeted therapies for NSCLC were approved, targeting EGFR, but the approval did not require detection of a mutation in EGFR, and only a small percentage of the unselected patients treated with the drugs had responses. It was not until 2013 that an EGFR targeted therapy limited usage to patients with EGFR mutations, the population of NSCLC patients that have been shown to benefit preferentially from this therapy.
The identification of new drug targets drove a need for more comprehensive diagnostic testing that would maximize the information gained from a limited amount of sample tissue. The advent of next generation sequencing (NGS) allowed physicians to assay for mutations in many genes in a single test, using a multi-gene “panel” of cancer genes. “Although these NGS panel tests weren't technically companion diagnostics by today's definition, you had a test that could identify variants across many different genes, that could point to targeted therapies that were either FDA-approved or in clinical trials,” says Sheryl Krevsky Elkin, Ph.D., Chief Scientific Officer at N-of-One, a QIAGEN company focusing on molecular decision support. “NGS and testing panels became a game changer for diagnosis and targeted treatment of cancer.”
At this time, additional, potentially targetable genes began to emerge, and sequencing technology continued to move forward. In 2012, Foundation Medicine launched its first commercially available comprehensive genomic profiling (CGP) test, FoundationOne®, which allowed for sequencing of the full coding region of 180 genes. “It was an enormous technological leap,” says Dr. Elkin. “However, now we had this really big test that assayed a lot of genes that were not actually clinically actionable.”
Over the next few years, many participants in the industry scaled back, favoring a 50-gene pan-cancer panel that could be used for any tumor type and that covered the major genes that could be matched to approved drugs as well as the major pathways that were thought to be predictive of response to targeted therapies.
Large Gene Panels Return
Today, sequencing platforms are more affordable and accessible than ever before. In fact, in 2018, the FoundationOne®CDx, now covering 324 genes, received reimbursement coverage by the Centers for Medicare and Medicaid Services and received designation as a companion diagnostic for multiple targeted therapies. In addition, many more targeted therapies have been approved for various cancer indications, and they have become pervasive in clinical trials. The knowledge and literature surrounding cancer biology has grown exponentially, expanding the landscape of potentially targetable molecules. As a result, the allure of larger gene panels for somatic mutations has returned, and the value of this information is now significantly increased, when the panels are used appropriately.
“The advantage of a large panel is you really cover your bases,” says Dr. Elkin. For example, large gene panels, which sequence hundreds of genes, can help identify which clinical trial a patient should join or provide data on genes that may become clinically actionable in the future. In addition, large gene panels can be used to calculate tumor mutational burden (TMB), a statistic that has generated significant interest in oncology as a potential predictor of response to immunotherapy in some types of cancer.
“The disadvantage of large panels is that they are more expensive to run and sometimes they're more expensive to interpret because you get so much information,” says Dr. Elkin. Without good bioinformatics and interpretation, she explains, less common but relevant mutations may simply get missed in the potentially overwhelming amount of information returned. Furthermore, many large panels designed to assess TMB may contain a significant number of genes that are sequenced in the assay in order to have adequate coverage to calculate TMB but for which clinical interpretation may not be necessary and could be confusing, because many of the genes are not themselves yet clinically actionable. The laboratories performing the testing, as well as the oncologists themselves, face a significant challenge staying abreast of all the information. The interpretation of the test itself can add to the turnaround time on a test, delaying potentially important information.
Another challenge with large gene panels is the presentation of the information in the report. For example, a report of 600 genes can be a source of confusion rather than enlightenment. That's why, Dr. Elkin explains, “it's critical that the report be constructed in such a way that the mutations that have been shown to predict response to therapy, or to significantly influence prognosis or diagnosis, are listed up front in a prominent manner. These clinically actionable mutations should be clearly delineated from variants of unknown significance.”
In 2017, the Association of Molecular Pathology (AMP) described a set of guidelines for the reporting of somatic variants in cancer, offering a structure by which laboratories and clinical decision support companies could clearly distinguish the levels of evidence for each of the variants detected in a diagnostic sequencing test (Li et al., 2017). The guidelines recommend four tiers, listed below; these tiers are used in the reporting schema utilized by QIAGEN Clinical Insights.
“Having a report that differentiates between the variants with high levels of evidence and variants with low levels of evidence is extremely important,” says Dr. Elkin. “Physicians need to be able to read the report quickly and understand what it means in order for it to help them to determine treatment.”
Reports can also be made more useful by limiting in-depth clinical interpretation to the subset of genes deemed most relevant. For example, if a panel sequences 600 genes, the top 150 genes that can be potentially matched to an approved therapy or clinical trial can be interpreted, rather than the entire panel. “The interpretation then becomes easier, faster, less expensive, and more useful, so that physicians are not receiving a 700-page report for each test,” says Dr. Elkin.
Accelerating TaT Using Somatic Interpretation Services
To address the challenge of large complex panels and the need to stay informed of the latest treatments, drugs, and trials, QIAGEN has developed and offers QIAGEN Clinical Insights (QCI). Available as a quick turnaround interpretation service with pre-configured reports summarizing variant interpretation, or an up-to-date clinical knowledge base with the latest molecular interpretation information, QCI takes the time out of searching, investigating, and consolidating information, helping to support physicians' decision-making decisions for target therapies based on the patients' unique genetic variants.
“For a large panel where you end up with hundreds of variant calls, you can load your results into the QCI Interpret interface,” says Dr. Elkin. QCI Interpret is a clinical research interpretation and reporting software tool that can be used to assess next-generation sequencing data. It is linked to a curated database, QIAGEN Knowledge Base. “QCI Interpret performs automated classification of those variants and can compute a pathogenicity score (using the guidelines set out by the American College of Medical Genetics, or ACMG) and an actionability score in terms of the four tiers and levels of evidence listed in the AMP guidelines.”
As an additional or alternative service, QCI Precision Insights, powered by N-of-One, can take the filtered and scored variants and perform manual analysis to verify or refine the classification. “QCI Precision Insights also provides detailed information to the physician about which drugs could target the identified variants, the level of evidence for using a drug to target each variant, evidence that a variant predicts resistance to a drug, clinical trials that are available for any of the identified variants, and a summary of the clinical evidence. Evidence from completed clinical trials, published in the literature, can be critical to supporting a physician's decision in cases where the drug is not recommended by the National Comprehensive Cancer Network® guidelines,” says Dr. Elkin.
“In other words, QCI Interpret and Precision Insights can help take all of that genetic information, and all those hundreds of variant calls, and narrow it down to identify the variants a physician should be focused on and the reasons and evidence behind this information. What used to take many hours can now be achieved in minutes to hours,” says Dr. Elkin.
In addition, QIAGEN is actively developing approaches that allow for more customized subset analyses of genes within a large gene panel. For instance, the entire gene panel can be run through QCI Interpret, and the scores generated can identify the top variants for clinical interpretation. The top variants can be submitted to QCI Precision Insights to produce a full, detailed report. If a physician is only interested in variants with Tier I (variants of strong clinical significance) or Tier II evidence (variants of potential clinical significance), QIAGEN could identify and provide a full report on only those variants.
“Focusing on a subset of relevant genes can help clinicians cut out a lot of potential noise in a report and see only the results that are clinically meaningful,” she says.
In summary, while somatic cancer screening panels have become more comprehensive and complex, making it hard to interpret and stay informed, services, technology, and software tools have stayed ahead of this evolution bringing near real-time access to reports that accelerate TaT and help inform targeted therapies.
