With an ever expanding precision oncology landscape, standard of care treatments often change following the American Society of Clinical Oncology (ASCO) Annual Meeting. The Biofidelity team returned from another transformative ASCO and I wanted to share some of our observations, reflections and learnings.
Despite advances in targeted treatments, the field of genomic testing lingers behind with the heavy burden of interpretation placed squarely on practicing oncologists. Continuing on a theme that has woven through previous ASCOs, this year’s meeting kicked off with a “Genomics 101 for Oncologists” workshop. The objective was to help educate oncologists about the correct selection of broad gene panels and train them on genomic result interpretation and clinical utility.
The workshop highlighted the currently available genomic testing tools and use cases for each depending on the clinical context and what needs to be tested including PCR, Sanger sequencing and next-generation sequencing (NGS).
The comprehensive coverage provided by NGS is accompanied by a high complexity of understanding test limitations and interpretation support. Multidisciplinary teams are needed to effectively order and interpret testing for precision oncology. In resource equipped settings, there is often a molecular tumor board to assist with genomic result interpretation.
In simple words, a key question for oncologists is “How do you decide what to do with a genomic variant that was detected by NGS?”
To help answer that question, some of the key takeaways from the workshop were:
Each NGS test has its own coverage of genes and assay limitations that are important for oncologists to understand. A targeted panel is often desirable compared to comprehensive pan-tumor panels that yield a lot of unactionable information that can be hard to interpret. For RNA based biomarkers, RNA-based NGS is preferable and an orthogonal testing strategy is encouraged. However, the required depth of sequencing and input nucleic acid required for RNA analysis by NGS can be prohibitive. Continued optimization of different testing methodologies is necessary to improve detection rates of both DNA and RNA biomarkers for NSCLC patient treatment.
Majority of current NGS-based tests detect variants in a target gene that is included in the panel rather than only therapeutically actionable variants. This leads to NGS-based testing yielding inconclusive results (in the form of variant(s) that are likely pathogenic, variants of uncertain significance (VUSs), and variants that are likely benign. VUSs can comprise up a high fraction of total variants identified by NGS and may be misinterpreted as causal, particularly by clinicians with less genomics experience.
In the case of NGS-based testing of blood specimens, another byproduct is the detection of Clonal Hematopoiesis of Indeterminate Potential (CHIP) variants that are unrelated to the cancer of interest. CHIP variants are somatic mutations that represent an expansion of hematopoietic cells without an underlying malignancy and are often associated with increasing age, tobacco use, and prior therapies and can increase false-positives when using liquid biopsies. Paired white blood cell sequencing can help subtract CHIP variants but add costs to NGS.
Newer NGS assays also offer paired normal and tumor NGS with suspected/incidental germline findings to offer one test that can do it all. This is often undesirable as the somatic and germline variant calling algorithms are distinct with experts advocating use of distinct somatic tumor testing and germline testing panels. For hereditary cancer testing, there is not yet a good panel for both somatic and germline variant detection. Some of these NGS tests rely on an approximately 50% variant allele frequency as a major determinant for a variant being suspected germline, yet experts caution that the % VAF does not always obey that rule and can land anywhere in the 40-80% range depending on the sample. Most clinicians are concerned about appropriate clinical decision making (including need for referral to medical geneticists) for these suspected germline findings. Also, while there are laws like the Genetic Information Nondiscrimination (GINA) Act to protect patients, having incidental (and potentially) unactionable genetic information as part of patient record was deemed unnecessary.
Due to complexity of variant reporting and interpretation, the experts urge oncologists to verify primary sources of data using different clinical variant interpretation databases such as OncoKb, CIVIC, COSMIC, Clingen, OMIM, GeneReviews, Clin var, Cbio portal, etc. While a useful work-around solution, it is a difficult task given current physician workload, burnout rates and training required to interpret this data. Another consideration with the excessive information that results from NGS testing is the cost of data storage increasing and still being untenable
Newer technologies that can process patient samples with sensitivity and specificity while reducing bioinformatics requirements are needed quickly so treatment decisions for patients can be expedited and real improvements be realized in the delivery of precision oncology to all.
Our technology, Aspyre, is focused on actionable biomarkers and does not detect unactionable biomarkers such as VUSs and/or CHIP variants and may therefore reduce unnecessary follow-up and reduce patient distress. Additionally, due to the targeted testing for pre-determined actionable biomarkers using PCR-based methodologies, there is no need for specialized bioinformatics and genomic interpretation using multiplex PCR-based tests. Most PCR-based genomic analyses involve turnkey data analysis with no specialized bioinformatics required.
Learn more here.
Written by Anjana Bhattacharya, PhD, Vice President of Strategic Marketing
The American Society of Clinical Oncology (ASCO) Annual Meeting 2023 kicked off with informative sessions including a “Genomics 101 for Oncologists” workshop that shed light on the current state of biomarker testing and the gaps therein. The currently available molecular testing tools highlighted were: PCR, Sanger sequencing and next-generation sequencing (NGS) depending on the clinical context and what needs to be tested. Traditional Sanger sequencing is not very sensitive; however, it is still often used for single gene analysis in samples of high tumor purity. Allele specific PCR can improve sensitivity over Sanger sequencing but has limited multiplexing capabilities and thus is quite restricted when multiple genes and biomarkers need to be detected, e.g. in metastatic non-small cell lung cancer (NSCLC) patients.
Notably, a large fraction of NSCLC patients present when the disease has already metastasized making time to treatment extremely critical for overall survival. There are numerous FDA-approved first-line targeted therapies now available for these patients but only if they can get timely biomarker testing. Some clinicians suggested PCR molecular reflex testing at diagnosis for these patients so they can be put on treatment as fast as possible. In an effort to improve the turnaround time of NGS-based biomarker testing, there have been efforts by academic medical centers and larger centers to provide NGS in a patient proximal setting such as local labs. However, so far, bringing testing closer to the patient has been exclusive to resource-equipped settings and is further hampered by the high cost of NGS data storage, bioinformatics and lack of standardization in variant interpretation, leaving many patients out.
NGS is currently the most commonly used method for biomarker testing but it comes with heavy costs to patients who are already dealing with financial toxicity. The cost aspect is further exacerbated by long turnaround times, complicated and lengthy non patient-friendly reports, even with targeted panels. Comprehensive genomic profiling (CGP) assays that are commercially available often yield a lot of unactionable information such as off-label therapies and Variants of Uncertain Significance (VUSs) that are not useful for treatment decision planning. Even if actionable clinical trial results are reported in some CGP reports, clinicians quoted going down a long windy path of figuring out options for their patients in terms of eligibility and geographic location.
Furthermore, as variant interpretation evolves in response to newer clinical study readouts, a lot of unactionable variants detected over time can become actionable, presenting ethical and logistic dilemmas on how to communicate the new clinical significance of pathogenicity to existing patients who were tested previously. 95% of the information presented in lengthy NGS reports was therefore cited as not useful and clinicians are urged to not solely rely on these but also verify primary sources of data using different clinical variant interpretation databases such as OncoKB, CIVIC, COSMIC, Clingen, etc. While a useful work-around solution, it is a tall ask given the current physician workload and burnout rates.
Newer NGS assays offer paired normal and tumor NGS with suspected germline findings in an effort to offer one test that can do it all. Yet, most clinicians voiced concerns about being able to do anything regarding these suspected germline findings. This is often undesirable as the somatic and germline variant calling algorithms are distinct with experts advocating use of distinct somatic tumor testing and germline testing panels. For hereditary cancer testing, there is not yet a good panel for both somatic and germline variant detection. Germline testing and pre-test counseling aspects in cases of suspected germline findings are quite elaborate and separate workflows with implications of cascade testing in the patient’s family.
While resource-equipped settings such as academic medical centers report increased use of NGS for biomarker testing often funded by alternative means, the MYLUNG pragmatic study reported contemporary (Dec’2020- Sep’2022) community-wide biomarker testing rates for both early-stage (ES) and metastatic NSCLC (mNSCLC) patients enrolled from 12 community practices (69 sites) across The US Oncology Network. It is sobering that there are still wide gaps in biomarker testing rates prior to treatment initiation, 64% of ES and 83% of mNSCLC pts had at least one molecular biomarker result; and 37% of mNSCLC pts had results for all nine biomarkers that are recommended for testing in the metastatic setting currently. NGS-based testing was conducted in 37% and 57% of the ES and mNSCLC cohorts respectively. Reasons for not testing included: barriers to ordering (42%, 25%), insufficient tissue (18%, 18%), clinical deterioration/crisis (5%, 12%), and other reasons (37%, 49%), respectively.
Barriers to ordering are commonly both financial and due to limited tissue for testing, which is a known issue in NSCLC patients. Oftentimes, clinical samples from patients can be mucinous, of low sample purity and poor diagnostic yield with current technologies unable to provide satisfactory biomarker testing results. Newer technologies that can process such samples while reducing bioinformatic requirements are needed as upfront as possible so treatment decisions for patients can be expedited and real improvements be realized in the delivery of precision oncology to all.
Written by Anjana Bhattacharya, PhD, Vice President of Strategic Marketing