Tuesday, May 3, 2011

Webinar Report: Impact of Next Generation/Whole-Genome Sequencing on Companion Diagnostics

Biomarkers are increasingly part of pharmaceutical and clinical strategy.  By some estimates, the success rate of FDA approval of new cancer drugs is 75% if mechanism-of-action and predictive or prognostic biomarkers are clearly defined, whereas it is 25% without the biomarker information.  However, identifying new biomarkers for companion diagnosis (CDx) remains a challenge—the identification of KRAS-type biomarkers is rare, there is a double regulatory hurdle and revenue issues hamper pharmaceutical investment in this area.  Whole-genome sequencing is an important tool in the discovery of biomarkers. 

Today, Richard Resnick, CEO of GenomeQuest, moderated a panel discussion on the potential impact of whole-genome sequencing technology on companion diagnosis, and pharmaceutical drug discovery and development [...,...].  The webinar event titled, "The Impact of Next Generation/Whole-Genome Sequencing on Companion Diagnostics" was organized by NGS Leaders (part of Cambridge Healthtech Associates) and held at the Biomarker World Congress in Philadelphia.  Joining Richard on the panel were Glenn Miller, Vice President and Head of Personalized Healthcare & Biomarker Strategy, Portfolio & Alliances, at AstraZeneca,  Dom Spinella, Executive Director of Translational Medicine, at Pfizer, and Iya Khalil, Co-Founder and Executive Vice President, GNS Healthcare.  Eric Glazer, Managing Director, NGS Leaders, introduced the panel members.  Note: Next Generation sequencing (NGS) and whole-genome sequencing (WGS) terms were used apparently interchangeably during the panel discussions. 

The big questions 

Whole-genome sequencing first became available in 2006-2007.  Within a span of three-to-four years, WGS has evolved as an important discovery and potentially a diagnostic tool.  The maturity of this technique is reflected in the list of questions posed for the panel in the agenda: "Will whole-genome sequencing in clinical diagnostics be a practical, high ROI genomics investment for pharma? Can whole-genome sequencing in clinical diagnostics be transformational in drug rescue and how? How will whole-genome sequencing in clinical diagnostics affect the design and operation of clinical trials? What practical steps and investments should pharma be making in this area? What advances in whole-genome sequencing are creating opportunities? How will this impact the strategies of diagnostic companies and research labs? Which therapeutic areas offer the most promise in the near-term?" 

The doors opened by the adoption of whole-genome sequencing 

Richard Resnick said that NGS can provide an unbiased approach for the discovery of cancer-associated mutations and structural information.  He gave the examples of mutations in DNA methyltransferases (DNMT3a, DNMT3b and DNMT1) and NADP+-dependent isocitrate dehydrogenase 1 (IDH1) genes.  IDH1 catalyzes the oxidative decarboxylation of isocitrate yielding α-ketoglutarate (α-KG), and is a recurrent mutation in glioblastoma multiforme (GBM) genomes.  Richard stressed that both biomarkers were discovered because of NGS, and neither was ever the "usual suspect."  IDH mutations have diagnostic and therapeutic implications for the treatment of gliomas and AML (for example, 1, 2, 3 ,4)

Richard also summarized recent work from his group and that by Daniel C. Link and colleagues, which was published in the April 20 issue of Journal of American Medical Association (link to Resnick et al., Link et al., accompanying editorial, and here].  Richard's team was refered an acute promyelocytic leukemia patient with no X-RARA fusion identifiable by routine metaphase cytogenetics or interphase-chromosome fluorescence in-situ hybridization (FISH).  His team performed whole-genome sequencing and identified an insertion of a large chunk of chromosome 15 in the second intron of RARA gene on chromosome 17.  The discovery of resulting bcr3 PML-RARA fusion (atypical AML) allowed attending physician to recommend highly effective ATRA course of treatment, rather than an invasive and an expensive bone marrow transplantation which would have been the choice based on cytogenetics/FISH diagnosis alone.  The second paper by Link et al. provides an example of the discovery of de novo mutation made possible by WGS in a retrospective analysis.  This was a case of a breast cancer patient, who later had an ovarian cancer (a.k.a. second cancer); the patient relapsed with ovarian cancer and acute leukemia, and eventually died.  Whole-genome sequencing, done retrospectively, led to the discovery of p53 gene deletion.  This discovery prompted the healthcare team to return to the patient's family, communicate risk and advise proactive testing and surveillance in genetically related family members.  These two JAMA papers together, exemplify two areas where NGS may make a significant impact, molecular presentation of a disease and discovery of de novo mutations with familial implications.  Richard predicted that one of the future implication of NGS is in diagnosis, for instance, in molecularly defining lung cancer where we do not see EGFR mutations, or breast cancer where we do not see BRCA mutations. 

Stratifying cancer using whole-genome sequencing 

Dom Spinella of Pfizer said that the real implication of NGS is in stratifying disease.  Five hundred years ago, everything was labeled as miasma.  Last two centuries, saw histology and cytology defining and stratifying cancer types and stages.  "To me, it’s all about stratifying disease.  It's not lung cancer any more, it's what flavor of lung cancer," he said.  "Now, we are targeting molecular attributes."  NGS is ushering an era of molecular etiology, as opposed to the current cytological or histological defined cancer. 

What NGS can and cannot do

 Since its introduction in 2006-2007, NGS has gone mainstream.  Richard quoted a statistics that Beijing Genomics Institute has sequenced over 50 000 genomes so far.  Today NGS is widely available as a relatively affordable service.  Many recent articles and posts are suggesting the potential of $1000 genome sequencing services within a few years [read here].  Many sequencing providers today have a turnaround time of one week or so.  Richard was optimist about the technology.  He said that WGS will lead to an era where one can identify responders and non-responders, predict drug response, and lead to the development of lab panel tests.  Glenn Miller of AstraZeneca, however, offered a skeptical view.  He said that whole-genomic sequencing is a great discovery tool for patients whose doctors persist in asking the question, what's really wrong.  "It is not a savior, just a tool," he said.  Whole-genomic sequencing does allow generation of new drug compounds to be developed.  According to Glenn, while NGS has tremendous value as a discovery tool, bringing it to a clinical laboratory as a diagnostic tool faces the same difficulties as does the microarray-based methods: NGS-based method is similarly difficult to validateshould the markers or the process be validated, or both; how to convince a Medical Director of a clinical laboratory to sign-off on routine adoption of an NGS-based method; is it cost-effective from the clinical laboratory's perspectiveadding NGS instrumentation and replacing low-cost employees in favor of bioinformatics-driven service will tremendously increase the cost.  In Glenn's view, NGS will remain a discovery tool, particularly a retrospective-analysis tool.  Glenn said that in the end, physicians do not care first-generation, second-generation or third-generation sequencing.  All they want is an answer.  At the end of the day, it economics.

Dom echoed Glenn's sentiments.  Dom said that there are tens-of-thousands of analytics associated with SNPsone has to understand the biology and determine the causative and passenger (reactive) biomarkers.  "NGS only gives a hypothesis," he said, "and by the time we validate [the biomarker for CDx], we lag drug development process; the drug has already moved to third phase."  It's an issue of timeshould we go back [in trials] with new data just now coming from NGS.  It's also the issue of economics; Dom asked, how could one justify additional resources if only two years are left on a patent, and additional resources does not guarantee a change in the reimbursement rate of a drug.  "I'm lagging behind the drug development and I'm not changing [the nature of the] game," he sums, and adds that the only place where NGS will have impact is retrospective analyses.  He further commented that the current approach is empirically driven; it is not clear what patients' cohorts are needed to identify such biomarkers, and this is one of the current hurdle in biomarker development.  Iya Khalil, disagreeing with Glenn and Dom, said that retrospective and prospective analyses will help, and it is increasingly possible to do this during phase 1 and 2 trials due to machine-learning algorithms and super-computing power being applied to NGS.  The demand coming from payers and patients will help bring the methodology mainstream.  As a discovery tool, ability to work with a bank of patient e-records along with biological samples and using the data-crunching power of supercomputers will help identify predictive biomarkers in a clinically-relevant time-frame. 


NGS is a wonderful discovery tool to generate hypothesis which will need to be tested. It will help generate hypothesis early on, e.g., which population is likely to respond. It is also a disease stratification tool.  In this aspect it will become commonplace. There are several groups attempting to bring together individual biomarker tests into a single or fewer tests.  NGS will be an option in the development of such diagnostic tests. 
Twitter Feeds

No comments:

Post a Comment