Perfect companions: using genetic tests to tailor drug treatment to the patient
This article was originally published in SRA
A recent PricewaterhouseCoopers report estimated that the core diagnostic and therapeutic segment of the US personalised medicine market is worth $24bn. This figure is expected to nearly double by 2015, reaching $42bn. Peter Charlish investigates the rapidly emerging field of companion diagnostics.
Ever since the introduction of modern medicines, researchers and clinicians have wondered why the same drug can be effective in one person, but not in another. Or why some patients experience side effects, whereas others do not. But over the past couple of decades, advances in pharmacodynamics, genetics and other disciplines have brought about the realisation that genetic variations between individuals are the probable cause of this phenomenon.
With that realisation came the recognition that, if patients with a particular genotype could be identified, it might be possible to avoid treating them with a medicine that would cause an adverse event, and use an alternative drug instead. Or that patients who are not genetically programmed to respond to a particular medicine could be spared the cost and inconvenience of being treated with an ineffective product, and instead receive a medicine that can confidently be predicted to be effective. Again, this has the potential to improve patient care while at the same time reducing costs.
A further implication of this understanding is that, if patients who are genetically disposed to fail to respond to a particular experimental medicine can be identified, they can be eliminated from clinical trials. Thus the success rate of trials will increase, the reasoning goes, raising the chances of the product being approved, and making clinical trials more cost-efficient.
Thus was born the notion of personalised medicine, which the US National Cancer Institute defines as "a form of medicine that uses information about a person's genes, proteins, and environment to prevent, diagnose, and treat disease". Alberto Gutierrez, director of the US Food and Drug Administration's Office of In Vitro Diagnostic Device Evaluation & Safety (OIVD), has said that new genetic tests together with other diagnostic tools are making it possible to target drugs "so that we'll get the right drug to the right person at the right time".
What precisely is driving the growth of demand for companion diagnostics? There are a number of factors. First, the sector could never have emerged without the advances in genomic and proteomic science already mentioned. For example, many differences in the way individuals respond to drugs are thought to be the result of single nucleotide polymorphisms (SNPs) – variations in DNA sequence that involve a single nucleotide. Until fairly recently, the technology for analysing a patient's DNA was both cumbersome and expensive, which effectively ruled out the routine or bedside screening of patients for particular SNPs. However, the advent of DNA microarrays has made it possible for the presence of specific SNPs to be identified rapidly and cost-effectively. At present, this is still a laboratory-based technology, but eventually it will be simple and cheap enough to use in the doctor's office.
At the same time, the pharma sector has had to rethink how it operates. With the blockbuster model now widely acknowledged as doomed and, in any case, with product pipelines now moribund, pharmaceutical firms have been searching for a new way to maximise returns on their investments. Instead of simply encouraging the use of their products in as many patients as possible, companies are seeking ways of maintaining their profit levels by targeting their products at a smaller, select group of patients to whom they can charge a premium price.
The breakdown of the blockbuster business model is the result of cost pressures as much as a paucity of new products. Ever-increasing regulatory demands have pushed the price of developing a new pharmaceutical product skywards, while at the same time there has been unrelenting pressure from third-party payers to hold down prices. Meeting these additional regulatory requirements has led to longer development times and consequently shorter pay-back periods before patent expiry.
Another positive influence on the growth of companion diagnostics has been the FDA's Critical Path Initiative (CPI), launched in 2004 in an attempt to streamline the process by which a discovery or proof of concept is translated into a drug, biological product or medical device. In 2006, the FDA published a list of 76 critical path opportunities, grouped into six categories, one of which was biomarker development. Biomarkers are broadly defined as molecules that can be measured and used as indicators of a biological process.
A great deal of effort has been expended on understanding the pharmacogenomic variations that partly determine how an individual responds to a particular drug, and in identifying biomarkers that indicate particular pharmacogenomic states. The FDA in particular has encouraged pharmaceutical companies to incorporate pharmacogenomic studies into their drug discovery and development efforts, and a considerable number of biomarkers have been studied with the aim of identifying surrogate endpoints for use in clinical trials.
The FDA also encourages drug developers to submit the results of pharmacogenomic studies as part of their investigational new drug applications, new drug applications and biologics license applications, although it only obliges them to do so under certain circumstances. For example, pharmacogenomic data must be included in an IND where they are used to support safety in an animal study or for decision-making in a clinical study (to help determine an appropriate dose, for instance). In NDAs and BLAs, pharmacogenomic data are required when they will be used to support the application or incorporated into the drug labelling.
However, just because the conditions are right for the development of the companion diagnostics sector does not mean that its continued growth is assured. There are a number of obstacles to be overcome, the biggest of which are probably the conflicting interests and differing expectations of pharma and diagnostics companies. Put simply, a pharmaceutical company is used to spending many years and many millions of dollars to develop a product that is safe and effective, that meets an unmet medical demand, and that meets all regulatory requirements. In return, it expects a reasonable period of market exclusivity during which it can market the product to doctors and hopefully recoup its investment.
A diagnostics company, in contrast, is usually able to develop a diagnostic product in a matter of months once the utility of a particular biomarker has been confirmed. It is able to do this in a fraction of the time needed to develop a pharmaceutical product, and with commensurately much lower expenditure. The regulatory requirements a company has to meet, if any, before it is able to market its product to laboratories are much lower than for a pharmaceutical product. The downside is that the life of a diagnostic product is comparatively short, and rapid advances in technology may soon render it obsolete.
One particular area where there can be a mismatch is pricing. While patients and healthcare providers realise that a new therapeutic product that claims a high success rate (because the companion diagnostics is able to identify suitable patients) will attract a premium price, there is sometimes also an expectation that the accompanying lab test will be relatively cheap, or even free. This, of course, is not the way that diagnostics companies or, for that matter, testing laboratories will see the situation. As a result, setting a realistic price for the whole package – drug and test – can be a complex business, and it is probably true to say that no standard business model with which all parties feel comfortable has yet emerged.
There are a number of therapeutic areas where companion diagnostics may be of particular use. In this context, the FDA has published a list of approved drugs that contain information relating to pharmacogenomic biomarkers in their labels, and where, therefore, the use of companion diagnostics may be appropriate. Such information may appear in a number of places on the label, including the indications and usage section, the dosage and administration section, in boxed warnings and so on, and may relate to clinical response variability, risk for adverse events and mechanism of drug action, among other things.
The biggest therapeutic area represented in the list, by far, is oncology, which accounts for 30% of the total number of entries (21 out of 70). For example, reference is made to the patient's EGFR (epidermal growth factor receptor) status in the labelling of four products: cetuximab (Lilly's Erbitux, indicated for colorectal and other cancers), erlotinib (Astellas's Tarceva, for a number of cancers including non-small cell lung cancer [NSCLC]), gefitinib (AstraZeneca's Irissa, for NSCLC and other cancers) and panitumumab (Amgen's Vectibix, for colorectal and other cancers). Philadelphia chromosome status is referred to in the labelling for busulfan, dasatinib, imatinib and nilotinib.
A total of 12 psychiatric drugs have labelling that refers to specific biomarkers, mainly the drug metabolising enzyme complex CYP2D6, which may have a bearing on such properties as clinical pharmacology, drug interactions and, in some cases, drug interactions or contraindications. Figure 1 shows the therapeutic groups currently associated with pharmacogenomic data.
Figure 1. Therapeutic categories of drugs with pharmacogenomic data in the label
It is worth noting that the presence of pharmacogenomic information on the label is not restricted to new products. The labelling for warfarin (Bristol-Myers Squibb's Coumadin, which has been marketed since 1954) notes that the drug is mainly metabolised by the polymorphic enzyme, CYP2C9. Studies have indicated an increased bleeding risk for patients carrying either the CYP2C9*2 or CYP2C9*3 alleles: patients carrying at least one copy of the CYP2C9*2 allele required a mean daily warfarin dose 17% less than for patients homozygous for the CYP2C9*1 allele. In addition, certain SNPs in the VKORC1 gene have been associated with lower dose requirements for warfarin. A patient's CYP2C9 and VKORC1 genotype information can assist in selection of an appropriate starting dose of warfarin.
There are a number of ways in which diagnostics companies approach the companion diagnostics sector. Probably the classic approach is to develop a test that can be used to identify candidates for a particular therapy. The best known example of this is Dako's HercepTest, which is indicated as an aid in the assessment of breast and stomach cancer patients for whom treatment with trastuzumab (Genentech's Herceptin) is being considered. HercepTest is an immunohistochemical assay for HER2 protein (c-erbB-2 protein), which is overexpressed in some of these patients (and who are likely to respond to trastuzumab treatment).
Dako now markets a number of tests for use in assessing cancer patients. The EGRF pharmDx kit is used to identify colorectal cancer patients eligible for treatment with cetuximab or panitumumab; the ER/PR pharmDx kit is indicated to help identify patients eligible for treatment with antihormonal or aromatase inhibitor therapies; and the TOP2A FISH pkarmDx kit is used to identify type II topoisomersases, which are the target of anthracycline drugs such as doxorubicin or epirubicin.
Another company that focuses on oncology tests to support targeted disease management is bioMérieux, whose bioTheranostics subsidiary offers real-time, RT-PCR-based assays that can be applied to formalin-fixed, paraffin embedded needle core biopsies, excisional biopsies or surgical samples. Among the kits available from bioTheranostics are tests for the KRAS and BRAF oncogenes, which can help determine anti-EGFR responsiveness, as well as a test for mutations in the tyrosine kinase domain of the EGFR gene.
Alternatively, some diagnostics companies take the route of establishing partnerships with pharma companies relatively early in the R&D process. This has the benefit that the pharma partner can take advantage of pharmacogenomic or other biomarker information to select patients for clinical trials, which should accelerate the development and approval processes, as well as assisting the diagnostics partner to tailor the companion diagnostic product exactly to the characteristics of the drug.
Celera Diagnostics is a good example of this type of company. It says its partnerships take advantage of its ability to validate biomarkers developed either in-house or by its partners, which may then be incorporated into clinical trial programmes carried out at its CLIA-certified testing laboratory, Berkeley HeartLab. Celera has particular expertise in the areas of gene expression, SNP identification and proteomics, and these may be used to distinguish responders from non-responders, resulting in shorter clinical trials and improved time to market.
Alternatively, Celera can support internal research efforts and clinical studies conducted through another clinical partner. Furthermore, it can apply its expertise to the identification of disease subtypes or related diseases that may be suitable for a particular drug, thereby leading to market expansion.
An example of a company whose focus is mainly on developing tests for use with drugs still in development is the Almac Group, which is based in Northern Ireland. One of the ways it works with pharma companies is to develop predictive tests through the identification of gene signatures, patterns of gene expression that can predict the response to a particular therapeutic agent, or indicate the likely disease outcome. Gene signatures are investigated by studying the expression of various combinations of gene transcripts. Not only can such information contribute to the design of clinical trials, but it can also lead to the development of a test that can be presented to regulatory agencies for approval and marketed as a test or companion diagnostic for a therapeutic agent.
Table 1, which lists significant partnerships between diagnostics and pharma companies in the area of companion diagnostics over the past few years, gives an impression of the pace of deal-making in this area.
Table 1. Recent deals in the area of companion diagnostics
MDx to help Pfizer identify and develop biomarkers to predict response to its investigational PARP inhibitor PF-01367338
Development of a PCR−based companion diagnostic test for identifying people with NSCLC that harbours EGFR activating mutations
Development of a novel theranostic test to aid oncologists in choosing the appropriate treatment for metastatic melanoma
Development of a PCR-based companion diagnostic that will be used alongside Pfizer's rindopepimut immunotherapy vaccine for newly diagnosed glioblastoma multiforme
AstraZeneca to use its experience in the development and commercialisation of oncology products to help Dako produce companion diagnostic tests in the cancer area
Abbott's molecular business is to provide Pfizer with a companion diagnostic test for screening NSCLC patients for eligibility to receive Pfizer's PF-02341066 cancer drug
Bristol-Myers Squibb/ImClone Systems (Lilly)
Development of a K-RAS companion diagnostic for use of cancer drug Erbitux (cetuximab) in metastatic colorectal cancer patients
Development and commercialisation of point-of-care influenza test for specific virus strains, using polymerase chain reaction technology
Development of a companion diagnostic test kit for the Boehringer's Tovok (afatinib) lung cancer drug
DxS to provide its TheraScreen EGFR29 mutation kit to AstraZeneca as a companion diagnostic for the latter's Iressa (gefitinib) cancer drug
DxS to provide Amgen with its K-RAS test as a companion diagnostic for Vectibix (panitumumab) in metastatic colorectal cancer patients in the US
The Wellcome Trust
Development of improved test to detect the V600E B−RAF mutation for use in clinical trials
Dako to develop diagnostic tests that will be used to identify patients more likely to benefit from cancer treatment drugs being developed by BMS
Laboratory Corporation of America
Co-development of new diagnostic tests in the areas of companion diagnostics, metabolic syndrome, oncology and diabetes
Abbott to develop a test to indicate which NSCLC patients are most likely to benefit from Tarceva (erlotinib)
Merck & Co
Merck to develop RNAi-based therapeutics using cancer targets identified by Celera, which will be able to develop and commercialise related companion diagnostics specific to therapeutic candidates arising from Merck's programme
Ventana Medical Systems
The Critical Path Institute
Development of standardised evaluation procedures for companion diagnostics and associated targeted cancer therapies
Development of biomarker and pharmacogenomic tests for growth failure patients that could become commercial companion diagnostic tests for Ipsen's short stature therapies
bioMérieux to develop a test to determine which patients are best suited to benefit from Ipsen's breast cancer drug, BN83495
Oxford Genome Sciences
Biosite to establish a companion diagnostic assay to enable OGeS to build a portfolio of personalised cancer therapeutics
Source: Clinica Medtech Intelligence, Scrip Intelligence and company websites
The emergence of companion diagnostics is forcing companies to think about their business in new ways, and about how the pharmaceutical and diagnostics sectors can best collaborate with each other. As the concept of personalised medicine gains wider acceptance, and regulatory agencies demand more and more data about new pharmaceutical products, neither pharma nor diagnostics companies can afford to ignore the companion diagnostics market.
Regulation of companion diagnostics – the US position
The FDA's Office of In Vitro Diagnostic Device Evaluation & Safety (OIVD) regards companion diagnostics as being at the heart of personalised medicine and, therefore, maintains that they carry the same risk profile as the drug itself. This means that proper use of the diagnostic is critical to the proper use of the drug. Failure to use the drug properly based on test results is regarded as unsafe.
OIVD's definition of companion diagnostics includes tests that are developed for drugs that are ready for commercialisation or already in the market, as well as tests that are co-developed with the pharmaceutical. While such "after market" tests could in theory lead to changes in the drug label, there is at present no regulatory mechanism for changing a label unless safety issues arise. Hence, from a regulatory point of view, drug/diagnostic co-development is preferable.
In the case of co-developed combinations, diagnostic data are most frequently submitted as part of the drug's investigational new drug, new drug application or biologics license application. However, in some cases, such as when the diagnostic has more than one intended use, submission as a pre-market authorisation is more appropriate. For "after market" diagnostics the normal route is via a PMA, which would be assessed with reference to the Center for Drug Evaluation and Research or Center for Biologics Evaluation and Research. Tests used to make decisions relating to the management of clinical trials generally require IND/IDE approval.
OIVD encourages companies to consult with it as early as possible in the development of a companion diagnostic, as the best way of ensuring that all critical issues are addressed in a timely manner.
As far as guidance is concerned, the FDA issued a Drug-Diagnostic Co-Development Concept Paper in 2005. The document addressed issues related to the development of a single in vitro diagnostic test for mandatory use in making decisions about a single drug selection for patients in clinical practice. It did not specifically address issues related to pharmacogenomic testing for the purposes of drug dosing determinations or monitoring of drugs, although it did contain principles that the FDA considered may be relevant to the development of these types of tests.
For example, it discussed the processes and procedures for submitting and reviewing a co-developed drug-test product; the in vitro ability to accurately and reliably measure the analyte of interest, including analytical sensitivity and specificity; the ability of a test to detect or predict the associated disorder in patients, including clinical sensitivity and specificity, and/or other performance attributes of testing biological samples; and the factors to consider when evaluating the patient risks and benefits in diagnosing or predicting efficacy or risk for an event such as the response to a drug.
However, the 2005 document was a preliminary concept paper and not intended for implementation. The FDA is expected to issue formal guidance on the co-development of drugs and diagnostic personalised medicine products this year, though its timetable is not yet clear.
Regulation of companion diagnostics – the EU perspective
In the currently applicable EU IVD Directive 98/79/EC, there is no direct mention of companion diagnostics, nor is there any formal guidance available from the European Commission.
But the EU is preparing for regulatory change with the anticipated entry into force of the forthcoming revised IVD Directive (IVDD) – though this will not happen until 2015 at the earliest.
A draft proposal is expected in the first half of this year, following the commission's three-month public consultation on the IVDD that ended in September 2010.
There is still much speculation about what the new requirements will be and how the pharma industry might be involved, as well as whether the pharmaceutical Directive 2001/83/EC might need revising to encompass the changes being made in the IVDD and for cross-referencing purposes.
The scope of the IVDD arguably encompasses companion diagnostics in a generic way. Article 1, Scope, definitions, b) defines an in vitro diagnostic medical device as: "Any medical device which is a reagent, reagent product, calibrator, control material, kit, instrument, apparatus, equipment, or system, whether used alone or in combination, intended by the manufacturer to be used in vitro for the examination of specimens, including blood and tissue donations, derived from the human body, solely or principally for the purpose of providing information:
concerning a physiological or pathological state; or
concerning a congenital abnormality; or
to determine the safety and compatibility with potential recipients; or
to monitor therapeutic measures."
It is not clear whether the phrase "used alone or in combination" should be taken to refer to use in combination with pharmaceuticals, but there is no doubt that the IVDD applies to all medical diagnostics.
What the IVDD fails to do, however, is to address the potential additional issues that arise when the results of a diagnosis are used to influence pharma development or choices, and so the potential higher risks associated with use with drugs are not addressed by the current IVD Directive.
Indeed, at present, the vast majority of IVDs, and that includes companion diagnostics, fall into the general class of IVDs that only require self-certification by the manufacturer without the involvement of notified bodies (third-party certification bodies), who are accredited to test IVDs. This means that, at present in the EU, tests for cancer markers, for example, can be designed and manufactured without the company having to seek any third-party certification.
Only self-tests and high-risk tests in one of two lists (A/B) in the EU IVDD's Annex II – such as tests for Human T-lymphotropic virus, HIV, rubella, toxoplasmosis etc – require notified body involvement.
This is a fundamental concern for the drugs sector, which fears that the credibility of its products could be damaged because of what may be perceived as an overly lax approach.
The revision of the IVDD offers an opportunity to change this. The commission has been proposing the adoption of a risk-based classification scheme similar to that devised by the Global Harmonisation Task Force, an option that has been favourably received by the European Diagnostic Manufacturers Association.
The GHTF risk-based approach features four levels of risk and would result in the majority of companion diagnostics – including cancer markers and genetic testing, which are currently under the EU "general" class with no notified body involvement – being upgraded into Class C (the third most strictly regulated of the four GHTF classes).
Under Class C, if the GHTF model is followed, a notified body would ensure that the quality management system was appropriate, check the adverse event reporting procedure was in place, and conduct a pre-market review of technical documentation to ensure conformity with the essential requirements.
This is all very well, but it will mean considerable upheaval at manufacturers as they seek to meet the new conformity assessment requirements of the new classification, and EDMA has asked for a five-year transition period for products currently on the market.
EU regulators are heavily involved in ongoing GHTF IVD work (such as clinical evidence), which they will aim to influence, and then adopt at EU level within the IVD Directive revision.
In addition, to sell products in Europe, companies must also apply for listing on the individual reimbursement schemes in each of the countries in which they wish to sell their products.
This is already a lengthy and costly process, often involving health technology assessment appraisal too and/or some kind of proof of the medical benefit or cost-benefit balance.
Pricing and reimbursements systems, however, can operate separately in each member state, and also in silos when it comes to pharmaceuticals and IVDs – another major hurdle for industry.
Peter Charlish is a principal analyst for Informa Business Information, the publishers of Regulatory Affairs Pharma. Extra reporting on the EU regulation of companion diagnostics by Amanda Maxwell, EU regulatory affairs editor of Clinica Medtech intelligence, a sister publication of Regulatory Affairs Pharma.