Diagnostics and their role in personalised medicine

Collaborations between diagnostic and pharmaceutical industries are expected to grow driven by the development of personalised medicine

This trend is evidenced by Pfizer’s August 2009 collaboration with Abbott’s diagnostics division to develop a test to screen non-small cell lung cancer tumours for gene rearrangements. Pfizer has developed a novel investigational agent, PF-02341066, that selectively targets cancer-causing genes implicated in the progress of many cancers. The test will be used for patient selection for future clinical trials of the agent.

1. The future of pharma is inextricably linked with smart diagnostics

The molecular level of personalised medicine

Personalised medicine – or the use of information about a person’s genes, proteins, and environment to prevent, diagnose, and treat disease – has been much talked about in recent years. And some observers are wondering what the excitement is all about. In the words of Roche’s CEO, Severin Schwann: “Personalised healthcare is nothing new. Doctors have always tried to fit the therapy to the patient if possible. But what’s happened more recently is that we’ve begun to go a level deeper (...). We’re now exploring the biology of disease and treatment at the molecular level.”

Molecular medicine does not per se define personalised medicine but molecular tools are important as they should enable greater relevance in the information provided by diagnostic tests.

Personalised medicine as a spectrum

As personalised medicine means different things to different people, additional complementary ways of characterising diagnostics may further help distinguish different shades of grey in the personalised medicine spectrum.

In the strictest sense, personalised medicine diagnostics may consist exclusively of companion diagnostics, which are by definition geared towards supporting a therapy decision for a particular drug, patient by patient. At the more permissive end of the spectrum, personalised medicine tests may include also early diagnostics, prognostics and possibly all other types of diagnostics. You may indeed argue that if a diagnostic were not designed to inform treatment decisions for individual patients – one way of defining a personalised medicine diagnostic – it would not have much sense.

Companion diagnostics and the future of pharma

The concept of companion diagnostic (CD), and its role in the future of healthcare was also mentioned in Pharma 2020: Marketing the future, a report published by PricewaterhouseCoopers (PwC) in February 2009. This report discusses the key forces reshaping the pharmaceutical marketplace and the changes required to create a sales and marketing model that is better adapted to the stakeholder priorities expected for 2020. Of particular relevance to the diagnostics industry is the move from mass market therapies to specialist therapies, which is highlighted in the report. Many specialist therapies are very costly (e.g. almost $300,000 annually for Fabry’s or Gaucher’s disease) and are used to treat smaller target patient populations with specific disease subtypes. In this context, there is thus a growing imperative, both clinical and budgetary, to accompany therapies with diagnostic tools of increasing sensitivity and specificity to better enable the identification of those patients in the relevant disease subtype and most likely to benefit from the therapy. The marketing model foreseen for most specialist therapies in 2020 will include a companion diagnostic as a key component.

Highlighting a movement towards personalised medicine is not straightforward as a concept as it is so widely open to interpretation. It is however important that we try as the response rates on drugs are still unsatisfactory, varying widely from “20% to 75%.

2. Companion diagnostics partnerships with the pharma industry

If diagnostics are so important to improving the value of therapeutics to patients, we should expect the pharmaceutical industry to be entering significant CD partnerships with the IVD industry. However, this is not the case and diagnostics collaborations with pharma have yet to become an established practice.

Diagnostics collaborations with pharma have yet to become an established practice

Only, seven partnerships were announced in 2008 between pharmaceutical and diagnostic companies to develop a companion diagnostic. This represents a significant drop from the 14 collaborations announced in 2007 but there is no clear up or downward trend over the period 2004-2008, with annual deal numbers varying between 6 and 14 throughout. Companion diagnostics partnerships with pharma have yet to become an established industry practice.

Most 2008 deals focused on cancer and involved a big pharma partner

Three key themes were reflected in the partnerships announced in 2008:

• Cancer attracted strong interest as the disease area of choice for the development of companion diagnostics. In 2008, all deals focused on diagnostics for cancer.
• Big pharma was dominant as pharmaceutical partner for collaborations with the diagnostics industry. In 2008, all pharmaceutical partners were top 20 companies by sales of prescription pharmaceuticals (2008 ranking from IMS health). OSI Pharmaceuticals was the only exception but big pharma company Roche was a co-partner in OSI’s collaboration with Abbott’s diagnostics division to develop a pharmacogenomic test to identify patients most likely to respond to cancer drug Tarceva (erlotinib) for non-small cell lung cancer. In this study, we counted any deals involving Genentech as a big pharma deal due to Roche’s majority ownership in Genentech at the time of the deal and which became a full ownership during 2009.
• Niche specialists dominated as in vitro diagnostics partner for collaborations with the pharmaceutical industry. In 2008, Abbott was the only IVD major involved in collaborations with third-party pharmaceutical companies for companion diagnostics. None of the other diagnostics partners announcing deals in 2008 – Aureon, Celera, Dako and DxS – were ranked among the ten largest in vitro diagnostics companies.

Similar themes emerge when we analyse the pharmaceutical and diagnostics partners involved in companion diagnostics licensing deals over 2004-2008 in more detail.

2.1 The pharma partners

Roche, Pfizer and Merck were the most active pharma partners over 2004-2008. Roche’s pharmaceutical division, including Genentech, stands out as the most active third-party pharmaceutical licensing partner for the diagnostics sector over 2004-2008 with ten announced deals – an average of two deals per annum. Pfizer, Merck and AstraZeneca followed Roche with 6, 5 and 3 diagnostics partnerships respectively over the same period.

The position of Roche as the leading pharmaceutical partner for collaborations with diagnostics companies is remarkable if we consider that none of the other pharmaceutical companies with a major IVD affiliate – Abbott, Bayer and J&J – announced more than one partnership with a third party diagnostics company over 2004-2008. However, the full picture about these diagnostics majors is not available as there is little visibility about any intra-group diagnostics collaborations with their respective pharmaceutical divisions.

 2.2 The diagnostics partners

The most active diagnostics partners for deals with pharma were all niche players. Serial deal making for companion diagnostics was limited amongst diagnostics companies. Only four diagnostics companies announced at least two partnerships with pharmaceutical companies over 2004-2008 and they were all niche diagnostic specialists: Celera, Dako, Epigenomics and Perlegen. Dako is the largest among these diagnostic companies with $322 million of net sales reported in 2008.

Three observations follow from this analysis and our discussions with selected industry players:

1. The pharmaceutical industry is not currently a priority market for large diagnostics companies. The development risk and time to market associated with drug candidates make the development of a companion diagnostic significantly less attractive to major diagnostics manufacturers than the revenues currently available from its more traditional target market of clinical laboratories.
2. The limited deal flow by company suggests that even for niche diagnostic companies, the prospective economics of developing a companion diagnostic may not always be attractive. Key factors that impact the net present value expected from companion diagnostics projects include the strength of the intellectual property, the pricing and reimbursement coverage and the extent of testing required by regulators to obtain key marketing authorisations. Achieving a positive net present value from a companion diagnostics development project will be a challenge unless some of these factors become more favourable.
3. For those diagnostic companies that do target the pharmaceutical industry, some are developing companion diagnostics without entering a partnership with a pharmaceutical company. When diagnostics companies have the required funding and access to sufficient, high-quality biological samples to conduct such development work without partnering with pharma, this approach can help keep more of the value in-house. In due course, however, it can be helpful to have some form of public support from the targeted drug’s marketer to underline the validity of the test as a companion diagnostic for the drug.

3. Biomarker testing requirements

Drug approval agencies, including the FDA and EMEA, are encouraging greater use of biomarkers and diagnostics in drug development and prescribing decisions, thus promoting the concept of companion diagnostics for drugs. 

The FDA recently started reporting a list of genomic biomarkers that it considers valid to guide the appropriate clinical use of approved drugs. The list is being updated on a quarterly basis and counted 32 valid genomic biomarkers in mid September 2009.

Most drug labels in the list provide pharmacogenomic information with no immediate recommendation for genetic testing. However, testing is “recommended” or “required” in a few cases.  At 20 March 2009, only 4 biomarkers were “required” to be tested for – three for cancer.

We are still at the start of the process if we consider that only 4 biomarkers are “required” to be tested for. However, the FDA was prompted to publish its list following a marked increase over the last decade of approved drugs labels containing pharmacogenomic information. The FDA estimates that 10% of approved drug labels now contain pharmacogenomic information and this is expected to continue increasing.

The EMEA’s communication on the requirement for biomarker testing is less transparent than the FDA’s but its initiatives should not be overlooked. For example, the European agency played a key role in requiring biomarker testing for Amgen’s Vectibix, following the FDA’s accelerated approval without specific testing requirements. The EMEA also has a larger number of drugs for which biomarker testing is required – in mid 2009, we counted at least 11 drugs with the requirement.

We expect greater harmonisation between different regulatory agencies to develop over time through greater consultation but also following pressure from clinician communities as stakeholders in one country push to implement practices already included in drug labels in other countries.

4. Moving towards personalised medicine

Increasingly, pharmaceutical companies will not move a drug candidate to the clinical development stage without a clear biomarker development program. These companies understand the contribution of biomarkers and diagnostics in improving the design and probability of success of clinical trials. In addition, pressure from healthcare payers is putting more emphasis on the availability of a companion biomarker test when deciding on a drug’s reimbursement. These factors will combine to accelerate the development of new diagnostics for personalised medicine. We anticipate that alliances and collaboration will be inevitable as the market need expands.

Notes to the editor

• Early diagnostics: diagnostic products permitting the detection of a disease at very early stages of its development thus giving more treatment options (e.g. early lung cancer detection allowing surgery).
• Prognostics: diagnostics that provide a prediction or estimate the risk of developing a particular condition based on
  • phenotypic (e.g. transcriptomic, proteomic or metabolomic) parameters; or
  • genomic (e.g. hereditary or gene based) characteristics.