Dorothy didn't think much of the scrape in the beginning. Inher 82 years, she'd had plenty of bumps and bruises, so she went about herbusiness, preparing dinner. It wasn't until she turned around, however, and sawthe droplets of blood trailing down her apartment hallway that she realizedsomething was wrong.
Dorothy had a history of atrial fibrillation, so hercardiologist had put her on warfarin to reduce her stroke risk. Unfortunately,the very thing that makes warfarin useful—its ability to prevent clots in amalfunctioning heart—is also the thing that can make it incredibly dangerous tosomeone like Dorothy, whose paper-thin skin breaks so easily. It can lead touncontrolled bleeding.
Dorothy's cardiologist didn't make a mistake. Dorothy'spharmacist didn't make a mistake. When the warfarin was prescribed, Dorothy waslikely given the right dose. But so many factors can complicate the "right"dose of warfarin for any individual that it can sometimes feel like you'replaying Russian roulette.
Same old, same new
Dorothy's case is not unique. Every year, millions ofpatients are given drugs or drug doses that may not suit their condition. In anow decade-old report, it was suggested that 40 to 75 percent of a patientpopulation may not benefit from a particular drug for a given treatment,largely because community medical practice ultimately comes down totrial-and-error based on the best medical knowledge arising from highlycontrolled clinical trials and not highly variable real life.
And despite advances in more targeted drug development anddiagnostic technologies, things haven't necessarily improved over the lastdecade. In a recently published study, Eitan Emir and colleagues at Toronto'sPrincess Margaret Hospital showed that while several new anticancer drugs havedemonstrated significant improvements in efficacy endpoints such as overallsurvival and progression-free survival over standard care, these efficacyimprovements may have come at a significant price for patient safety andtolerance of treatment.
In a meta-analysis of 38 randomized clinical trials of solidtumor agents approved by the U.S. Food and Drug Administration (FDA) between2000 and 2010, the researchers found that when compared to the control agentsin each study, the odds of toxic death were 40 percent higher for patientsreceiving the new drug. Similarly, in patients on the new drug, the odds oftreatment discontinuation increased by a third and the odds of a serious (Grade3 or 4) adverse event increased by more than half.
The researchers were quick to point out that these resultscould have arisen for a variety of factors in the trial design, as well asbecause of the molecular activity of the drugs, but it did reinforce theimportance for a solid understanding of any patient's baseline medical statuswhen selecting a therapy.
The need to more clearly define the connection oftherapeutic decision-making to patient need and performance promptedorganizations like the FDA to issue guidance about the development of companiondiagnostics. In April 2005, the FDA released its draft preliminary discussiondocument, "Drug-Diagnostics Co-Development Concept Paper," to start theconversation.
"Ideally, a new diagnostic intended to inform the use of anew drug will be studied in parallel with early drug development (Phase I or IItrials), and diagnostic development will then have led to prespecification ofall key analytical and clinical validation aspects for subsequent (Phase II andIII) clinical studies," the report read.
The principle was—and is—sound. The execution has been morechallenging.
Changing the approach
According to specialists in the diagnostics industry, thefirst part of the challenge was convincing pharmaceutical companies it was agood idea.
"Companion diagnostics are obviously a great idea," saysStephen Little, vice president of Qiagen's personalized healthcare division in Hilden,Germany. "The physician benefits, the patients benefit and the payors benefit,but to move the field forward, pharma had to understand that it, too, wouldbenefit from diagnostics co-development."
The perfect example of this, Little continues, was in theearly days of Vectibix and Iressa, "two drugs that were not on the market inEurope because they initially couldn't show enough efficacy [in broader patientpopulations]," Little says. "Once their efficacy could be linked to specificbiomarkers and patients could be targeted, however, European regulators acceptedtheir applications—clearly an example of a closed market that was openedthrough the development of diagnostics."
The education was a two-way street, however, and Littleadds, "I came from DxS, a small company that struggled to engender interest inthe pharma industry to combine diagnostic tests with drugs, and we quicklydiscovered that if you don't bring pharma along, you're not going to get veryfar."
By the same token, he continues, pharma initially assumed itknew how the diagnostics industry worked, but it didn't.
"Diagnostics are not drugs. Pharma companies aren't at allequipped to develop diagnostics, get them approved and manufacture them, soonce they realized they could benefit from companion diagnostics, they began toadopt partnership models with diagnostics companies," says Little.
Even for traditional diagnostics companies, there has been alearning curve, according to Henrik Winther, vice president of corporatebusiness development at Dako in Glostrup, Denmark.
"The market-value drivers and the regulatory burdens arevery different between routine diagnostic assays and companion diagnosticassays—prognostic and predictive assays—but at the same time, these assays aretightly interconnected through full diagnostic system solutions—hardware,software, reagents and analyzing tools—adapted to a standardized and optimizedworkflow of the diagnostic laboratory," he explains.
As Dako has shifted its focus more and more toward companiondiagnostics development, it has seen its operating procedures and quality assurancesignificantly impacted. In addition, the company's commercial competencies havebeen challenged to expand the customer field from pathologists to also includeoncologists and increase interactions with pharma companies.
The education and early inroads proved vital to getting themarket going. Improvements in biomarker identification and technical leaps inassay development have proven fruitful for the evolution of the companiondiagnostics market, which generated sales of $1.3 billion in 2010 and will growto almost $3.5 billion by 2015, according to an August 2011 report fromvisiongain.
"The market for companion diagnostics has enjoyed stronggrowth in recent years, driven by adoption of marketed products and also due tothe increase in the number of theranostics," says visiongain analyst Rupali Vadhera."Over the course of the next decade, there will be greater demand for companiondiagnostics as they improve the safety-efficacy profile of drugs and helpreduce healthcare costs."
Even the decision of how to develop a companion diagnosticcan take two approaches. The first is a straight co-development project where throughoutthe drug development process, relevant biomarkers are identified, assaysdeveloped and validated and drug and diagnostic submitted for simultaneousregulatory approval.
The other form could be called a "catch-up companion," wherea diagnostic assay already on the market is used to facilitate drugdevelopment. Winther describes this latter form as validation and registrationagreements (VRAs), where commercialized companion diagnostics or regulardiagnostics are used in clinical trials to select patients, and hence thecompany's task, in collaboration with pharma, is to validate and register theassay for the new indication or the new drug.
The biggest differences between the two approaches,according to Winther? VRAs offer shorter timelines, lower risks and lowercosts.
Of course, simply having a test for a given biomarker or setof biomarkers isn't enough. The biomarker and test have to be validated toensure that they actually help identify the right patient for treatment. Andvalidation costs money.
"Validation is one of the big challenges we face, which iswhy we work with pharma companies," says Qiagen's Little. "Let's say you knowyou'll be reimbursed $200 to $300 per test and you have a good sense of themarket size. When you compare that market value to the cost of putting the testthrough clinical trials, you quickly realize that it's not a very compellinginvestment case."
"That's why it's been so important for pharma to becomeinvolved," he adds. "Basically, pharma runs and pays for the clinical trial andthe diagnostics company piggybacks on the trial."
The decision to identify and validate a companion diagnosticadds entirely new layers of complexity, however, to an already complex clinicaltrials process.
When you develop a cancer drug targeted at a mutant form ofa protein (e.g., EGFR, KRAS), it isobvious what you need to sample, says Little. But more often, you've discoveredin your Phase II studies that your drug works on some patients and not onothers, and you need to identify what biomarker is different between those twopopulations.
"Ideally, we would love to develop and validate a testbefore a drug goes into Phase III but that isn't always possible," he adds. Forthis reason, Little feels it is essentially impossible to validate thebiomarker and the drug in the same clinical trial.
In 2011, Cambridge, Mass.-based AVEO Pharmaceuticalsinitiated two such Phase II biomarker studies for its investigational drugtivozanib, a VEGF inhibitor currently in Phase III studies itself.
The first study was to evaluate biomarkers in blood andarchived tissue samples and correlate them with clinical activity and/ordrug-related toxicity from a trial of patients receiving tivozanib for renalcell carcinoma. The more recent randomized study, undertaken with AstellasPharma, will look for biomarkers in patients with advanced metastaticcolorectal cancer receiving mFOLFOX6 and either tivozanib or Roche's Avastin.
"It is critical that we develop strategies that will enhanceour understanding of patient populations that will best respond to specifictherapeutic regimens, and I believe that BATON-CRC is a step forward on thisfront," says the study's lead investigator, Al Benson of Chicago's NorthwesternUniversity Feinberg School of Medicine. "I am excited to be involved in thistrial and to learn more about how this treatment regimen with tivozanib maybenefit patients living with colorectal cancer."
Qiagen's product line also highlights another challenge invalidating diagnostic tests: regulatory variability.
"Unlike drug approvals where most regulatory agencies arelooking for the same information, the regulatory process for diagnosticsapproval is quite variable around the world," Little says. "Things are muchquicker in Europe, which is why we've marketed the KRAS-Erbitux test forseveral years in Europe, but it is new in the U.S."
In July, Qiagen announced it had received FDA approval tomarket its Therascreen KRAS kit to provide guidance to physicians in the use ofErbitux for patients with metastatic colorectal cancer. Although it was Qiagen'sfirst approval in the United States, the company has had this and tests for 30other biomarkers approved throughout Europe, Asia and Japan.
"The bar is set higher at the FDA, and you could have awhole discussion about whether it's too high or too low," Little says.
Dako's Winther highlights the extra challenges affordedcompanion diagnostics, suggesting these tests hold the highest regulatoryclassification level, typically requiring investigational device exemptions(IDEs) and premarket approvals (PMAs). This is in contrast to routinediagnostics that only require "self-declaration" at the in-vitro diagnostic (IVD) Class I level.
"The regulatory and clinical affair competences within Dakohave therefore expanded dramatically within both expertise and capacity duringthe last five years," he says. "From a market-value driver point of view, thereis no doubt that companion diagnostics require technologies that can providequantitative results as well as innovative assay controls securing high-qualitystandardized and robust results, whereas the qualitative diagnostics aretypically driven by parameters like turnaround time and workflow. Thesedifferences in value drivers are of course reflected in different demandstowards our R&D functional area and those tools developed by the technologydepartments."
Cancer is key
The one therapeutic area that has likely seen the mostactivity with companion diagnostics is oncology, where the ever-expandingrepertoire of targeted therapeutics and efforts spent in preclinical and earlyclinical discovery and development have already pointed to the key biomarkers.
Little suggests the focus on oncology may revolve aroundsmaller patient populations that one need study because of the nature of thatlink.
"It will be interesting to see what happens as we move intoother therapeutic areas," he suggests.
When you move away from genetic biomarkers into areas suchas metabolomics and proteomics, however, the water gets muddier and therequired patient populations will likely rise significantly. The trick, hebelieves, will be in finding the most important biomarkers, not necessarily themost biomarkers.
"We try to get a test we can use and that we know works, notnecessarily the one with the most data. Just because you can measure 10,000points, you don't have to. Ten points may be sufficient," Little says.
The growing activity in the oncology market is drawing theattention of not just diagnostics shops, but also of more traditionallysupply-side companies looking to expand their markets and leverage theirinternal technology expertise. One such company is Life Technologies inCarlsbad, Calif., which in September announced it was partnering withBristol-Myers Squibb Co. (BMS) to collaborate on companion diagnosticsprojects.
"The pharmaceutical industry is increasingly turning itsfocus to discovering and delivering targeted personalized medications," saidLife Technology's president of medical sciences, Ronnie Andrews. "For oncologyalone, hundreds of agents are currently in clinical trials, and we see strongmarket opportunities in the robust expansion this will mean for the companiondiagnostics space."
For Life Technologies, it's about translating the company'ssuccess in technology platform development into the clinic and joining itsknowledge of genomics and proteomics with BMS' expertise in clinicaldevelopment to generate flexible, cost-effective diagnostic solutions.
Similarly, Qiagen is looking to leverage its long-standingexpertise in laboratory and research diagnostics into the companion diagnosticsmarket, relying on technologies introduced from such acquisitions as DxS andmost recently, Ipsogen. As Little suggests, these acquisitions give the companya solid presence in both solid tumors and blood cancers.
The KRAS approval by the FDA is just the first step in theUnited States. Upwards of 40 percent of patients with metastatic colorectalcancer are unlikely to benefit from drugs like Erbitux and Vectibix becausethey carry a mutant form of the KRAS gene. For these populations, the testshould allow physicians to put patients on other drugs that might have a betteropportunity to provide benefit, thus reducing the loss of valuable treatmenttime.
Given that theirs is the only FDA-approved KRAS test on theU.S. market and that an estimated 110,000 colorectal cancer patients could betested annually, Qiagen sees this test as a $20 million market. The company has15 other such projects underway to co-develop and market companion diagnosticswith major pharma and biotech companies.
A quick survey of abstracts being presented at the EuropeanSociety for Medical Oncology's (ESMO) 2012 meeting in Vienna would suggest thatthe pursuit of companion diagnostics in oncology is not showing any signs ofslowing.
Ingrid CsToth and colleagues in the European Lung CancerWorking Party presented their findings of two microRNA biomarkers that offersignificant predictive power in identifying which non-small cell lung cancer(NSCLC) patients will benefit most from treatment with a cisplatin-vinorelbineregimen.
Similarly, Evelyn McKeegan and colleagues at Abbott presenta retrospective analysis of four clinical trials of patients receiving theinvestigational drug linifanib for relapsed NSCLC. They found a signature oftwo tumor biomarkers was significantly associated with improved survival onlinifanib but not with the other tested drugs, whether the drug was usedfirst-line or later.
Pablo Maroto of Barcelona's Hospital de Sant Pau similarlynoted links between c-Myc expression and renal cell carcinoma in patientsreceiving Sutent. Patients with c-Myc negative tumors experiencedprogression-free survival more than double that of their c-Myc positivecounterparts.
As with Qiagen, Dako's expertise has largely been focused ononcology. In June, Dako announced FDA approval of two HER2 companiondiagnostics for Genentech's new cancer drug Perjeta (pertuzumab) targetingHER2-positive breast cancer. And in February, the company signed acollaboration agreement with Amgen to develop a companion diagnostic for one ofits investigational cancer drug candidates.
"Dako's focus has been on oncology, primarily based on ourhistory within cancer diagnostics and our technology offerings suited for theroutine pathology lab," says Winther. "However, with Agilent's acquisition ofDako, the technology offerings have significantly expanded, and hence we nowhave the potential to both lead the innovations within oncology companiondiagnostics, but also to plan for expanding into other therapeutic areas."
STORY CONTINUESClick here for Part 2
STORY CONTINUESClick here for Part 2