A Personalized Reality
Next generation sequencing holds much promise for personalized cancer diagnosis, treatment and management, but how is this being realized and what does the future hold?
Matthew Smith , George Burghel |
At a Glance
- Molecular tumor profiling of clinically actionable mutations using NGS guides the delivery of anti-cancer therapies
- Fast, efficient and cost-effective, targeted NGS is becoming increasingly embedded into the clinical laboratory
- Certain factors are vital for accurate clinical data, such as quality control, FFPE sample compatibility and an optimized target capture assay
- Targeted panels are emerging and evolving in response to the latest genetic discoveries
Clinically actionable mutations lying within certain driver genes are central to tumor development, and hold much utility for cancer medicine. While these mutations carry diagnostic, prognostic or predictive implications, a subset are also deemed ‘druggable’ – able of identifying cancers that can be treated with targeted therapies acting against the subsequent protein product or disturbed pathway.
With more of these mutations coming to light all the time, this exciting field is developing very rapidly. For example, until this year inherited variants within the BRCA1 and BRCA2 genes indicated an increased risk of breast, ovarian and prostate cancer, but were actionable only in the sense that the disease risk could be managed (e.g. through mastectomy and oophorectomy). However, we now regard these mutations as druggable in the sense that ovarian cancers containing the mutations respond to a new class of drugs called PARP inhibitors (1).
The ability to identify such actionable or druggable mutations in tumors holds the key to personalized cancer therapy, informing clinicians and helping to guide treatments. This has many implications; patients will only receive the most appropriate treatment dependent on the underlying molecular profile of the tumor. Personalized therapy for cancer is therefore proving to be safer and more effective than traditional approaches.
A variety of genetic testing technologies are available for profiling these targets, both well-established and emerging. Since the breadth of testing is currently limited to a handful of targets, the majority of routine diagnostics utilize well-established techniques – such as Sanger sequencing, pyrosequencing and qRT-PCR – which enable tests to be turned around in a clinically actionable timeframe, and provide a cost-effective strategy. However, this is only true as long as the number of tests per individual or sample is limited, and with their restricted capacity for multiplexing applications, these techniques are not wholly compatible with ongoing trends.
As the list of actionable genetic markers and targeted therapies expands with the latest research, these types of tests are becoming less feasible. For example, in the last few years a set of actionable mutations in non-small cell lung cancer have been identified, requiring a combination of sequencing, fluorescence in situ hybridization (FISH) and immunohistochemistry for molecular characterization from very limited amounts of tumor material (2,3). With a reduction in cost and improvements in library preparation and sequencing, NGS now has the capability for testing larger, multi-gene panels.
Needle in the haystack
The data load generated by NGS is well-known as a bottleneck, requiring time and expert knowledge to extract meaningful results. This is particularly true within the clinical genetics workflow, where turnaround times are a major priority. As a highly efficient alternative to whole genome sequencing, targeted sequencing is well suited to the clinical laboratory. By capturing specific genomic regions of interest from DNA samples prior to sequencing, only the regions of interest are analyzed. Focusing in on relevant areas of the genome, targeted sequencing panels significantly reduce the sequencing and data load, in turn reducing both time and cost.
Importantly, this approach also enables an increased depth of coverage, providing the sensitivity needed for heterogeneous samples and overcoming many of the challenges typically faced in NGS. Cancer-specific gene panels and enrichment methods are becoming increasingly popular, and a number of laboratories and commercial companies have recently developed and validated these for clinical use (see “Which Capture Method for Targeted NGS?”).
Designing NGS panels
When choosing content for a new panel, the current focus of molecular pathology labs is on delivering results that can be translated into meaningful clinical action. However, this can be complex. The content of any panel is a balancing act between trying to maximize the utility of the panel with expected sample numbers and desired throughput. In general, for a diagnostic panel the focus is often very narrow, maximizing cost-efficiency and sample throughput, while limiting the amount of surplus sequencing data that, as yet, has no recognized clinically actionable relevance. Without any known effect on treatment, variants of unknown significance (VOUS) therefore tend not to be covered. The breadth of content can range from mutational hotspots through to full exons, and for each laboratory this will depend on the target genes and the clinical literature. For example, KRAS and BRAF carry mutational hotspots with well-characterized effects on drug response, and can therefore be specifically targeted. For other mutations, such as those in KIT and PDGFRA genes, which have implications for the etiology of gastrointestinal stromal tumors, diagnostic labs need to look for mutations spread over specific exons. Sometimes known as ‘hot exons’ exhibiting high levels of actionable mutations throughout the entire exon, these can provide a wealth of information.
Another point to consider is that investigations evolve with new discoveries, and when mutations within certain genes, such as the tumor suppressor TP53, become more clinically actionable, it will then become important to look for variants spread over the whole gene. However, there is a lag between discoveries in research and their clinical application. Interpreting novel variants provides a significant challenge, and requires bringing together in silico analysis, literature review, current drug trials and other approaches, and the panel must then be re-evaluated following the addition of any new content. Additional content must therefore be carefully considered, and provide very strong evidence for a tangible difference in patient treatment. Moving forward, one model would be to review the content after set time periods and add additional content, if required, in batches. Moreover, a particularly interesting way that targeted NGS technology has adapted in response to this challenge is with the emergence of custom panels, which enable the user to select a chosen pool of relevant hybridization probes. The flexibility of such systems facilitates researchers in investigating variants relevant to their specific study, increasing the speed at which new content can be validated and decreasing the time lag from the laboratory to the clinic.
Compatibility with FFPE tumor samples
Solid tumor samples present two primary challenges. Firstly, because of tumor heterogeneity and the presence of DNA derived from non-tumor cells, a variant of interest may only occur at a relatively low allele frequency in the sample. Since detecting these is facilitated through deep sequencing, researchers are particularly interested in NGS platforms that allow considerable depth (i.e. targeted panels). In addition, the use of DNA extracted from formalin-fixed, paraffin-embedded (FFPE) tissue can present many technical challenges, impacting on both DNA integrity and yield. Due to the fixative process, the DNA can be degraded and clinical scientists often have to work with very small amounts of DNA. Because of this, a number of quality control metrics are analyzed, including the accurate measurement of low level DNA. Targeted capture methods are also carefully considered to ensure the uniform representation of all regions of interest.
Quality control for clinical application
Quality control procedures are vital to ensure accurate NGS data. All new tests undergo extensive validation initially with well characterized samples, including a variety of positive, no-mutation and no-template controls, while the laboratory based work and the bioinformatics pipelines themselves are also validated in order to estimate the sensitivity, specificity, reproducibility and repeatability of the tests. Following the initial validation, positive and negative controls are included in each assay, with the sequencing quality, coverage, depth and mutant allele frequency all determined and data analyzed and validated by two scientists.
The test report summarizes the interpreted results clearly, and is once again checked and authorized by a second experienced scientist. Not only do the discovered variants need to be reported, but the report must also be able to verify that the reason a variant was not detected in a region of interest was because it was generally not there or below the reported sensitivity of the test, and not because of lack of sequencing depth. In addition, the challenge of tumor heterogeneity is also considered. If the test’s detection sensitivity has not reached the level required to detect low allele frequencies, then this needs to be fed back to the clinician so additional testing can be performed if desired. Consideration of the latest research discoveries is also important, and published literature and known databases (such as COSMIC) are frequently used in interpretation and reporting.
The future of NGS in molecular pathology
The fundamental premise of personalized cancer therapy is to ensure the right treatment for the right person at the right time, and with the area of genomic medicine growing at an unprecedented rate, it is becoming clear that targeted NGS is playing a vital role in this. Indeed, this technology is becoming increasingly embedded within the clinical laboratory, with new panels emerging and evolving in response to the latest genetic discoveries. These panels provide the capability to detect low-level mutations from the ever increasing catalog of clinically actionable aberrations and markers for directing cancer therapy, and in fact, many of these genetic markers are already in use today. However, it is also clear that a certain level of consideration is necessary in order to accommodate the particular needs of the clinical laboratory, including the requirement for accuracy and sensitivity. Along with existing and emerging testing strategies, NGS has an extremely important role to play in future cancer characterization and treatment.
George Burghel is HCPC registered clinical scientist, Genomic Diagnostics Laboratory at The Manchester Centre for Genomic Medicine, Central Manchester University Hospitals NHS Foundation Trust, UK.
Matthew Smith is principal clinical scientist, Molecular Pathology Diagnostic Service, Cellular Pathology, University Hospitals Birmingham NHS Foundation Trust, Queen Elizabeth Hospital Birmingham, UK.
Which Capture Method for Targeted NGS?
The type of capture method is of utmost importance for targeted NGS, and the two main approaches fall into either the hybridization or amplicon-based categories — each with its own set of advantages and drawbacks.
Utilizing PCR, amplicon strategies tend to be quick and easily integrated into existing laboratory workflows.
Data quality tends to be less robust when compared with a hybridization based approach, as it is very hard to determine and remove bias introduced by PCR (e.g. polymerase errors, formation of secondary structures and preferential amplification of some fragments due to differences in GC content).
With the ability to easily capture larger target regions, this is the method of choice for larger panels.
Traditionally, more DNA input was required, and the library preparation tends to be longer when compared with PCR-based methods. The technology has been improving, however, and advanced hybridization-based technologies, such as the SureSeq Solid Tumor Panel (Oxford Gene Technology), use extensive research validation of lower input DNA, and focus on making the whole process more streamlined.
- CC Gunderson, KN Moore, “Olaparib: an oral PARP-1 and PARP-2 inhibitor with promising activity in ovarian cancer”, Future Oncology, 11, 747–757 (2015). PMID: 25757679.
- NI Lindeman, et al., “Molecular testing guideline for selection of lung cancer patients for EGFR and ALK tyrosine kinase inhibitors. Guideline from the College of American Pathologists, International Association for the Study of Lung Cancer, and Association for Molecular Pathology”, J Thorac Oncol, 8, 823–859 (2013). PMID: 23552377.
- PJ Roberts, TE Stinchcombe, “KRAS mutation: should we test for it, and does it matter?”, J Clin Oncol, 10, 1112–1121 (2013). PMID: 23401440.