Is Diagnostics the New Biotech...and Will Pharma Embrace It?

By Mark Ratner / Email the Author / IN VIVO September 2011, Vol. 29, No. 8
Feature Articles / Word Count: 4357 / Article # 2011800145

Executive Summary

The early biotech industry thrived on a combination of hope and hype while diagnostics, which evolved using many of the same tools, was traditionally viewed -- and priced -- almost as a commodity. Now we're seeing the advent of complex, high-value diagnostics. As the techniques underlying tests increase in their biological complexity and the knowledge base of their developers about specific disease areas deepens, is the value proposition becoming more biotech-like?

Pharma has always eyed diagnostics warily, as it did early biotech. As the techniques underlying tests increase in their biological complexity, is the value proposition becoming more biotech-like?

• The early biotech industry thrived on a combination of hope and hype while diagnostics, which evolved using many of the same tools, was traditionally viewed – and priced – almost as a commodity.
• The influence of genomics in the 1990s effectively expanded the definition of a biotech company, driving deal values even as significant product revenues from biologics were first appearing.
• The advent a decade later of complex, genomics-driven diagnostic technologies, often using sophisticated algorithms, has put the spotlight on the role of the CLIA lab as results interpreter as well as the insurer of test accuracy.
• The franchise-oriented, specialist CLIA infrastructure focuses deeply on a narrow therapeutic area, making it one of potentially great interest to some pharmas.

During a meeting at her office last February, the CEO of a next-generation diagnostics start-up suggested that high-value diagnostics companies are "the new biotech." Her argument: the field is facing many of the same adoption challenges as the biotech industry did in its early days. The thought resonated on a common-sense level. The underlying technologies defining biotech medicines and diagnostics are similar; in fact they had common starting points, first with the use of antibodies as targeting agents and probes, and later with genomics. In both cases, the same core assumptions existed around using those tools to unlock the power of the biological perspective. Biotech faced an initial hurdle of developing efficient and reproducible manufacturing processes for production scale-up; diagnostics, similarly, needed to automate. And in each case, there was a need for physicians, potential development partners, and regulators to get their arms around a unique and innovative set of promising technologies.

The conversation with the diagnostics CEO led to another, with Risa Stack, PhD, a partner at Kleiner Perkins Caufield & Byers, a leading venture investor in diagnostics in the last 10 years along with West Coast brethren TPG and Mohr Davidow Ventures. ("CardioDx: Bringing Molecular Diagnostics Into the Cardiovascular Arena" — IN VIVO, March 2010.) Over the summer, in collaboration with IN VIVO, Stack and Jenine Marie Golueke, a recent graduate of the Stanford Graduate School of Business, gathered data around early biotech companies looking for parallels to the emergence of high-value diagnostics. We defined a universe of early biotech companies and their products, looked at financing and alliance trends across a broad swath of the industry, then did the same for a selection of companies that have developed, or are developing or distributing, high-value diagnostics content. Most of these testing companies provide molecular diagnostics and related services through their own specialty laboratories, which are regulated under Clinical Laboratory Improvement Amendments (CLIA), designed to ensure the quality of tests developed by and only to be used in those individual labs. We also included some protein-based tests requiring complex algorithms for interpretation. Excluded, however, were firms that may have or aim to develop such content, but are principally instrument/platform developers (e.g., Qiagen NV , Illumina Inc., and Life Technologies Corp.), as well as large lab operations like Quest Diagnostics Inc. and Laboratory Corp. of America Holdings. However, we did include a handful of platform companies that are primarily associated with high-value diagnostics content of note: Gen-Probe Inc. and Cepheid, for instance.

We quickly discovered both through eyeballing our data and through interviews that direct parallels are weak. But analogies between the emergence of high-value diagnostics and the early days of biotech are nonetheless instructive. Both feature a tentative-at-best embrace by pharma – including its careful support of genomics-driven biotech in the 1990s. Indeed, the drug industry has danced a tango with both potential partners, simultaneously touching and keeping a distance between each of them. Illuminating diagnostics using the light of early biotech also serves as a reminder of the uniqueness of the biotech phenomenon and how expectations around – and demands for – the delivery of health care have changed, potentially inuring to the benefit of the new, complex diagnostics.

Opposite Ends

Diagnostics such as Genomic Health Inc. 's Oncotype Dx and Myriad Genetics Inc.'s BRACAnalysis breast cancer tests and the services offered by specialty labs such as Clarient Inc., a unit of General Electric Co.'s GE Healthcare , are demonstrating the potential for diagnostics to better manage the administration of expensive drugs. But making the case for high-value diagnostics has been a decade-long struggle. ("How to Earn the Economic Payback Diagnostics Companies Deserve" — IN VIVO, March 2009.) Biotech, on the other hand, started out with the expectation that it would be the be-all and end-all for curing many if not all diseases. And as a result, investment dollars flowed into it.

"The investment community was completely sold on the new science of using recombinant proteins to modulate biological mechanisms," says Kenneth Kaitin, PhD, director of the Tufts Center for Drug Development. The belief was that it would lead to products that could be developed quickly, would have very few side effects, and would be amazingly effective because they were customized to human needs and the disease state.

"When people have an inability to define how big big is, they can let their imagination run wild," says Stelios Papadopoulos, PhD, former vice chairman of Cowen & Co. And they did. One need look no farther than the IPO of Genentech Inc. (now a division of Roche [See Deal]) in 1980: the company filed to raise a million shares at $35, which it did, and the first day the stock traded in the $80s and closed at $77 – long before the dot-com era when huge one-day IPO gains became commonplace.

It was largely retail investors who bought into the IPO, recalls investment banker Fred Frank, vice chairman of Peter J. Solomon, motivated by the belief that the new world of molecular biology would cure everything from cancer to Alzheimer's. Cetus Inc. was second out of the gate, raising $122 million on the thesis that it would be the last financing it would have to do. "The theory of the analyst community at the time," says Frank, "as well as the companies' management teams, was that these are natural molecules and therefore would be safer and get through the FDA faster." Of course, safety issues did crop up, products failed, and by the time FDA brought the biologics review process within CDER in 1997, the bloom had very much fallen off the rose.

It's no surprise that the initial hope – and hype – of biotech was bound to ratchet back. Products developed at biotech firms were reaching the market, mostly replacement therapies like recombinant human growth hormone. ( See Exhibit 1.) But as Kaitin notes, these replacements "were not what the biotech revolution was supposed to bring." The promise of biotech was actually attacking receptors and genetic mechanisms of disease, he says. And before the approvals of the first noteworthy therapeutic monoclonal antibodies – Rituxan (rituximab) in 1997 and Herceptin (trastuzumab) at the end of 1998, both from Genentech – aggregate sales of that early universe of products were steady, but slow growing. ( See Exhibit 2.) Data from IMS Health Inc. (not shown) for all recombinant products, including those developed and sold by Big Pharma (such as recombinant insulin), show a similar pattern, with acceleration only beginning in the 2000s, a trend that is continuing today.

The biotech industry nonetheless found support from investors throughout the 1980s and the first part of the 1990s, mostly from a combination of public and private financing and alliances with Big Pharma. ( See Exhibit 3., showing these data from companies developing large molecules, cell or gene therapies or oligonucleotides.) And importantly, with the increasing focus on biologically driven target identification strategies for drug development, the notion of a biotech company was expanding to include any research-driven drug developers, especially those using genomics, whether the end product was a biologic, oligonucleotide, or small molecule. Indeed, the yearly tallies of biotech drugs in development issued by the trade group BIO and its predecessor organization in the late 1980s and early '90s routinely included a range of small molecules, to the point where BIO discontinued the lists because they were so muddy.

"If you asked the question of those who were putting the money to work at these companies," says Papadopoulos, "in the eyes of Wall Street, biotech encompassed aggressive, fast-moving young companies working in the life science space doing cutting-edge science with commercial implications." It included anything you could cast within such a company, including small molecules and even diagnostics. "That's why it was a no-brainer for Wall Street to embrace small molecule companies," he says. "If you were aggressive, losing money, but what you were doing could be a big thing, you were a biotech."

As the biotech industry graduated from its initial focus on the low-hanging fruit of replacement proteins to the use of proteins as biological response modifiers and then a greater awareness of the power of targeting cell surface receptors by whatever means, diagnostic tests were similarly evolving in concept. Both industries depended on many of the same tools: antibodies that could bind to targets either for diagnostic assessment or for drug targeting; the use of gene sequence data and the correlation of SNPs with disease; and more recently, broader-based DNA/RNA sequencing of samples. The biological complexity of diagnostics was increasing from the atomistic measurement of apparent single gene defects and later, single gene anomalies like overexpression of Her2 (the target of Herceptin), to less obvious gene mutations discovered by analyzing SNP patterns (using PCR and arrays), and ultimately to the complex algorithmic assessments of gene expression patterns seen today, made possible by rapid and more efficient gene sequencing. Both drug and diagnostics development also advanced as the explosion in information technology allowed for the capture of a lot more information and therefore the assessment of more complex systems.

That said, significant differences dominated the competitive landscape for drugs and diagnostics, as did assumptions going in about the power of diagnostic tools to revolutionize health care. Most importantly, the business model for diagnostics was horizontal – a manufacturer wanted to sell as many test kits as possible to run on its installed base.

That's changing with the advent of more complex tests. It is a fundamental shift that aligns the diagnostics field more with the management of therapy and the potential for cost savings and efficiencies that entails. In essence, figuring out how to best administer the targeted drugs inspired by the biotech revolution, and to whom, now provides a clinical and economic imperative for the development of complex diagnostics.

A Vertical Model Emerges

"Before 2000, diagnostics was an adventure in seeking accuracy," says Brook Byers, partner at Kleiner Perkins. Tests measured one analyte at a time as a way to detect the presence of disease, in the case of infections, or to quantitate, in the case of blood chemistries or hormone levels. The major breakthroughs were in cytology and in blood chemistries, typified by companies like Technicon (acquired by Bayer Diagnostics, which in turn was acquired by Siemens AG in 2006 [See Deal]), Coulter (which became Beckman Coulter Inc. ), Becton Dickinson & Co. , and Abbott Laboratories Inc. Then, service businesses sprung out of that: LabCorp and Quest and hundreds of other labs, plus all the pathology labs in hospitals and doctors' offices.

Biotech methods very much influenced diagnostics, beginning with Hybritech, which was conceived in the 1980s around the idea of manufacturing pure antibodies using monoclonal antibody technology, for use as probes. Hybritech's tandem assay, which combined two antibodies in one test, increased test accuracy, sensitivity, and specificity, leading to the company's acquisition by Eli Lilly & Co. Gen-Probe was formed around a discovery of a way to detect RNA. And Biosite Inc. (now part of Alere Medical Inc. [See Deal]) formed by a group that left Hybritech after the Lilly acquisition, also appeared on the scene, using antibodies to do point-of-care testing, starting with drugs of abuse and then ER applications.

Gen-Probe and Biosite in the 1990s were foundational and hinted at the vertical, specialist business models for CLIA lab-based complex diagnostics to come. But it was only after 2000 that the notion of diagnostic complexity broadened, owing to genomics and the introduction of TaqMan by Applied Biosystems Inc. (now part of Life Technologies [See Deal]). With the rise of rt-PCR, it was finally possible to automate the interrogation of multiple genes in a reproducible and timely way. "It took PCR from being an interesting research tool to being good enough in terms of quality to be used in a clinical setting," says Byers. Bioinformatics was maturing at the same time, becoming its own discipline in the late 1990s, improving biostatistics.

"I think applications in the therapeutic settings were more straightforward in the 1990s than they were for diagnostics," says Kathy Behrens Wilsey, PhD, former partner at Robertson Stephens Investments and co-founder of Kew Group LLC, which is trying to integrate the use of known cancer biomarkers into the community oncology treatment setting. "The high-value information in diagnostics wasn't as clear." Advancements in information technology were key drivers of diagnostics innovation. "Some of the basic utility features of IT – processing speeds, memory – have given us the ability to look at DNA information or other types of protein information in more complex systems," she says. "It's really taken the Human Genome Project for the research in a broad research enterprise to appreciate the value in all that information."

"This all needed to happen to allow for complex, advanced molecular diagnostics to be enabled," says Byers. Genomic Health, a Kleiner Perkins investment, was the first to integrate those components, and used the know-how to validate its breast cancer diagnostic test with the rigor of a therapeutics clinical trial.

Unlike Beckman Coulter or Becton Dickinson or Abbott, Genomic Health and its brethren were wholly vertical – they featured a narrow and deep focus on a specific disease state, which facilitates adoption and payment of their high-value offerings. There were already hints of this approach in the industry: Biosite had followed that model in cardiovascular, and Cytyc Corp. (now part of Hologic Inc. [See Deal]) made its mark by working closely with Ob-Gyns to drive adoption of its automated pap smear system and, importantly, with payors, to make sure the docs were paid for running the tests. "If you're going to touch the end customers – the doctors or payors – you've got to go vertical and deep," says Byers. It's also wise to remain platform agnostic – a concept anathema to the big box companies – to be in position to choose the best technology for a particular purpose.

Recognition

Biosite and Cytyc were acquired during a period of M&A validation of diagnostics in general, and such deals were a potential harbinger of a reversal in attitude by pharma, which at one point had more dedicated players in diagnostics: Gen-Probe was owned by Chugai Pharmaceutical Co. Ltd., Quest was part of GlaxoSmithKline PLC, and Dade Behring was a cobbling of several pieces over time, including the spin-out of Behring Diagnostics from Hoechst AG in 1995.

William Quirk, analyst at Piper Jaffray, sees a hangover in pharma from those days. "I think part of the historical hesitation [to embrace diagnostics] was that many of these central lab companies were actually spun out of pharma. Right there, you have a financial partner that's eliminating not increasing its exposure."

That began to change when Roche bought Chugai and spun out Gen-Probe for antitrust reasons. [See Deal] Then came Dade Behring's reorganization and IPO in 2003. "You had a very early leader in molecular diagnostics and an early leader in central lab automation that all of a sudden became investable assets for the first time in quite a while," says Quirk. At the same time, Genomic Health was starting to demonstrate that by asking the right questions, start-ups could, indeed, answer high-value questions and hence garner high-value price points. "The three big events helped the unfolding in the early 2000s," says Quirk. "This raised the profile of the industry, which in turn helped generate more interest from the venture standpoint." ( See Exhibit 4.)

Then, in 2006, multinationals with broad interests in diagnostics began to buy. "They were making investments in health care from a spending standpoint," says Quirk: viewing diagnostics as a way of reducing the cost of health care to end users. Siemens rolled up Bayer AG's diagnostics business and also Diagnostic Products and Dade Behring. [See Deal] GE Healthcare made a bid for Abbott's diagnostics business. [See Deal] Qiagen bought Digene Corp. [See Deal] ("Siemens--Stealing GE's Thunder" — IN VIVO, September 2007; "Wake-up Calls in IVD" — IN VIVO, April 2007 and "Siemens acquires Bayer Diagnostics, Creates Integrated Diagnostics Powerhouse " — IN VIVO, July 2006.)

Although this set of transactions centered on diagnostics sold as kits, start-ups were being funded to develop complex tests following the Genomic Health model, and with a nod to the success of other vertically focused companies such as Cytyc, Biosite, and even Digene in molecular-based HPV testing. It was only a matter of time before CLIA labs, as specialists in both complex test development and delivery of those tests and related services to providers, became acquisition targets by a variety of players, starting in 2009 with LabCorp buying Monogram Biosciences Inc. and Qiagen adding DxS Ltd. (now Qiagen Manchester Ltd. ). [See Deal] [See Deal]

Just as when pharmas acquired biotechs for their already-in-place development and process scale-up capabilities, it was a buy versus build calculation. And the trend has accelerated recently with Quest, LabCorp, GE, and Big Pharma Novartis AG all making notable pick-ups. ( See Exhibit 5.) Novartis' acquisition of Genoptix Inc. , side by side with GE buying Clarient last year, is the most intriguing because it opens up the notion of pharma using CLIA labs as an extension of their internal drug discovery and development capabilities. [See Deal] [See Deal] ("With Genoptix, Novartis Continues Its Diagnostics Build-up" — IN VIVO, February 2011 and "GE Acquires Clarient To Anchor Its Molecular IVD Business" — IN VIVO, December 2010.)

Novartis emphasizes that it sees Genoptix as a strong operating business as well as an extension of its scientific capabilities. A CLIA lab has requirements that ensure the accuracy of most molecular tests. Running rt-PCR on an automated system in a clinical setting using off-the-shelf reagents requires careful quality control. "You can't just take it out of the box and the reagents out of the bottles," says Brook Byers. So a good CLIA lab will have better quality than a lab just running a kit, because with the latter there's no QC on the front end. "It's assumed the kit is accurate," he says, "which is not always the case." Also, a company can continue to improve its tests in a CLIA lab format, as opposed to tests that have secured FDA approval via the 510(k) regulatory framework, for which you have to freeze your design. "It's why the Mayo Clinic is a Mecca for people to go to for esoteric testing – they run a massive system and are continually innovating," Byers says, especially for rare or unusual diseases, which are increasingly important franchises for Big Pharma.

Should Pharma Buy In?

There are some signs of broad pharma interest in diagnostics. Roche and Abbott have long-standing businesses, but these are for the most part walled off from their pharma units – unlike Novartis' Novartis Molecular Diagnostics unit, which reports directly to the head of pharma. ("Novartis Follows Its Own Business Development Model into Molecular Diagnostics " — IN VIVO, May 2010.) Most diagnostics acquisitions by pharmas will be opportunistic, such as Eli Lilly & Co. buying Alzheimer's disease diagnostics developer Avid Radiopharmaceuticals Inc. last year. [See Deal] Avid synergizes nicely with Lilly's Alzheimer's disease drug development, but Lilly says it will primarily use partnerships to develop diagnostic tests rather than build that capability internally as Novartis is doing. And service providers like Roche and Abbott exist for the one-off development of companion diagnostics like Pfizer Inc. 's just-approved Xalkori (crizotinib) for non-small-cell lung cancer or GlaxoSmithKline PLC's Mage-3 cancer immunotherapies in development, for which it has twice turned to Abbott. [See Deal] [See Deal]

Pharma's reticence to support the development of high-value diagnostics thus far, except on a test-by-test basis for its clinical-stage targeted drugs, is understandable.

"An industry sector needs time to incubate and curate before the large companies – the strategics – come in," Byers points out. "While they have a lot of resources, they are really risk averse. You first need champions within large companies, and they have a tendency to want all the answers before they move." Pharma took its time buying into biotech, waiting until there was tangible evidence of products and franchises. ( See "Big Pharma's Leap into Biologics: Bridging both Scientific and Cultural Gaps" — IN VIVO, October 2007.) "The same thing is beginning in diagnostics, led by Novartis, with its acquisitions and stated goals, and the increasing movement of executives into diagnostics within the company," Byers says.

Indeed, one can argue that on one level, pharma is a logical intersection at which diagnostics and biotech can meet, at a mutual level of complexity that is appropriate for both new biotech (targeted) drug discovery/development and especially to figure out how to use existing drugs. "It feels like 2011–2012 will be tipping-point years when large incumbent diagnostics and pharma companies move aggressively into advanced molecular diagnostics," says Byers. There's also been recognition by new players, including instrument providers Illumina and Life Technologies, which are expressing an interest in obtaining diagnostics content, and even food products giant Nestle SA, which acquired Prometheus Laboratories Inc. to expand its health care-related presence and focus both on treatment and wellness. [See Deal]

But to some, complex diagnostics as currently constructed don't yet mirror what's going on in a disease process sufficiently to be of use to drug developers, and will only have limited impact on patients.

The reason goes back to first principles. Both biotech and diagnostics started out without very much deep thought at all about how you choose targets, says Leroy Hood, MD, PhD, president of the Institute for Systems Biology. The focus was on skillfully using the new tools and techniques in hand: "It was more, can you get a GPCR receptor expressed in a certain tissue and if you block it, does it do something for a disease," he says. But just as it became clear that replacing or manipulating single genes (or the proteins they encode) was too simplistic an approach for the treatment of many complex diseases, it's becoming equally clear that only rarely will manipulation of a single target be effective in treating disease. So in effect, this was not a particularly rational approach, even though there was a logic to it, as it presumed a target was disease causative because modulating it had an effect.

Similarly, in diagnostics, "for very clear conceptual reasons, the idea that a single marker could be great in detecting a disease is utter nonsense," says Hood. "Any single marker has really marginal value except, occasionally, to follow response to therapy or reoccurrence six months out. What we've come to appreciate is that 99% of those potential biomarkers are noise, and you have to have good strategies for sorting out signal from noise."

Operational Value

Hood's comments may be straight-on and a foreshadowing of an integration of such diagnostic methods for disease assessment into drug discovery – an approach Merck & Co. Inc. tried a decade ago through its acquisition of Rosetta Inpharmatics LLC, but found too expensive to maintain. [See Deal] ("A Friend-ly New Resource for Biomarker Validation?" — START-UP, March 2009.) Pathway/systems approaches could become necessary for drug development, superseding single-target-oriented biomarker approaches. But that's down the road.

That's not to say that pharma won't buy into and support complex diagnostics – through M&A or otherwise through strategic relationships – in the near term. But it will do so only to the extent it also sees operational value. Just as with biotech, where it took first seeing that second-generation biotech products like Rituxan, Herceptin, Betaseron (interferon beta, now sold by Bayer AG's Bayer Healthcare Pharmaceuticals AG), and Abbott's Humira (adalimumab) were becoming established, rapidly growing franchises: by many accounts Humira, now nine years post-approval, maintains a 19% growth rate according to IMS Health and may become the number-one selling drug by 2015–2016.

Following the biotech time line, and assuming, as Byers notes, that pharma currently responds mainly to product-focused opportunities, high-value diagnostics today may be where biotech was 10 years ago, with initial products emerging and franchises being built around them. Companion diagnostics are also making their way to approval: Xalkori and Roche/Genentech's Zelboraf (vemurafenib) for melanoma were each approved by FDA in August with simultaneous reviews of the drug and the diagnostic. But the larger opportunities for developers of complex tests aimed at guiding and gauging response to therapies lie with products and services that look at classes of drugs and not individual agents, and so are not tied to a single drug developer's interests. Assets around which to build franchises, such as those of Genomic Health or Agendia BV in breast cancer, both through internal development and potentially in-licensing of tests and targets.

For pharma as well as other players, it boils down to a question of investable assets. As evidenced by the trend in M&A, some believe that the CLIA lab model opens up the opportunity for the development of high-impact, value-priced complex diagnostics, as well as concurrent leverage for related research activities. Indeed, Hood's vision could be implemented more quickly because of the CLIA lab model and the recognition by pharma and others of its importance as an infrastructure and as a sustainable business, as a service provider directly to physicians.

Click image to enlarge