An Inside Look at Diagnostics
This case study illustrates the evolution of the diagnostic industry across different continents, diagnostic companies' evolving business models, stak...
The FDA’s decision to approve what it terms a complementary diagnostic may not have been a groundbreaking event after all. In fact, it might make the testing landscape even more complex to navigate. Ewelina Golebiewska, Associate Director, Diaceutics asks who, if anyone, will be motivated to drive us towards ‘complementary Dx’ adoption. More importantly, is this ‘new’ concept just an unnecessary complication?
In October 2015, the FDA approved the first drug with what it termed a complementary diagnostic in the label. It was Dako’s PD-L1 IHC 28-8 IVD test kit that underwent the same rigorous regulatory and review process as companion diagnostic, the difference being that it is not essential for determining which patients should receive the drug. Rather, it can help predict the response to a therapy, in this case Opdivo. In May 2016, another such pairing was approved – Genentech’s Tecentriq with Ventana’s PD-L1 SP142 Assay. These approvals, together with the announced plans to bring that ‘new’ category of tests under FDA regulation, are seen as part of the agency’s increasing efforts to keep pace with the scientific advancements in an area that is already developing beyond the existing guidelines. But was it really such a ground-breaking event? And isn’t the FDA a bit late to the party anyway?
Been there before
In 2006, the FDA produced the first list of pharmacogenomics biomarkers in drug labelling, including around 60 drug-biomarker combinations that can affect clinical decision-making. In 2016, more than 130 drugs have actionable biomarkers in their labels1, adding up to almost 170 unique drug-biomarker combinations, yet only a handful of those are catered for by FDA-approved companion diagnostic devices2 (Figure 1). In addition, a further 25 labels ‘require’ genetic testing but the companion tests used are generic LDTs rather than devices3. That leaves us with a further 85 drugs that recommend testing with actionable or informative biomarkers that can still guide therapy, but ‘are not essential to the safe and effective use of a therapeutic product’ (FDA definition of a CDx4).
Figure 1. Based on FDA Pharmacogenomics list and Diaceutics Research we identified 169 unique drug-actionable biomarker pairs. Out of these only 20, all of which are oncology therapies, are linked to FDA-approved companion diagnostics (CDx).
And this is where the concept of ‘complementary’ diagnostics comes in. It is testing that can help to guide therapy but is not essential to prescribe the drug, a type of testing that has long been part of therapy labelling (at least implicitly) (Figure 2), and in fact internalised by pharma for at least half a decade. So what is new? There is one key difference, namely the insistence that such testing is now to be regulated alongside, and linked to, the drug.
Figure 2. Examples of labels that, implicitly, require testing of patients to determine appropriate dosing or patient population that may benefit from the therapy. None of these therapies is associated with a ‘companion diagnostic’.
These labelling examples illustrate how inconsistent FDA regulation can be. Take fluorouracil, for instance: three to six per cent of people have dihydropyrimidine dehydrogenase (DPD) deficiency which correlates with very severe, sometimes lethal side effects from fluorouracil and related chemotherapy. The label indicates the drug should not be used in patients with DPD deficiency, but to date the FDA has not actively enforced this. One could argue that for these patients the DPD status is essential information to guide treatment, so should the therapy come with a companion diagnostic? Or would that classify as complementary? Or is it up to someone else to decide?
In these situations to date, the FDA has left the decision to test with physicians, and it is unlikely that the new regulation would deal retrospectively with over a hundred therapies that use genetic and other tests to guide therapy decisions.
For the sake of argument let’s forget about those and look to the future. The obvious question is what value does the new category of regulation hold? More importantly, which stakeholders will drive adoption of any newly-regulated complementary testing?
Will pharma drive complementary testing?
There have been instances when pharma has sought to highlight the role of a complementary test, as was the case with Novartis and Gleevec. Novartis launched a campaign in 2011 to standardize monitoring and dosing in AML patients using a BCR-ABL test in collaboration with Asuragen and Cepheid. Better monitoring in that case resulted in a more effective therapy, improved dosing and greater trust in the brand, while at the same time ensuring a growing market for Gleevec’s follow-up drug, Tasigna.
Despite the fact that Opdivo was approved with the PD-L1 complementary testing which triggered this dialogue, BMS has sought to emphasize the fact that, as opposed to Keytruda, their therapy did not require testing (see the recent Opdivo TV ad here: https://www.ispot.tv/ad/AL_Z/opdivo-longer-life. That message resonated with clinicians, as a recent Diaceutics survey revealed, with a significant percentage indicating they did not see the need to order the test before prescribing Opdivo for NSCLC. Put simply, that means no test sales for Dako.
The situation may be different with the Genentech/Ventana combo as both companies are owned by Roche. Time will tell whether Tecentriq marketing will leverage the clearly informative role of PD-L1 in patient selection to support sales of Ventana’s test.
Clearly pharma could drive complementary testing and it will when there is a clear ROI, but Opdivo shows that the complementary label creates plenty of ‘wriggle room’ to promote or ignore the test .
Will laboratories drive complementary testing?
Perhaps one reason for the FDA’s insistence that pharma submits complementary testing for approval is the hope that it will support the shift to regulated diagnostic kits in US Labs. The majority of ‘informative’ and ‘actionable’ biomarkers are currently being tested using laboratory developed tests (LDTs), a segment of the IVD market (as of Q2 2016) regulated under CLIA rather than the FDA. In light of this and our observation that already over 70 per cent of commercial labs offering the PD-L1 test have developed an LDT regardless of companion or complementary test availability, suggests that financial considerations are of equal importance to labs in adopting new tests.
As testing evolves to integrate biomarker panels it is not clear that top-down regulation of complementary testing is the way to go. Commercial labs like Caris Molecular Intelligence and OncoPlexDx are developing biomarker panels driven by bottom-up need, providing physicians with comprehensive information to guide their therapy choice. As NGS technology gets ever cheaper, and the market becomes crowded with panel tests or even DTC pharmacogenomic kits, physicians may opt for a disease-specific or even multi-disease ‘complementary’ diagnostic that helps them choose between multiple therapies rather than providing a simple ‘yes-no’ answer for one branded drug before moving on to the next CDx-Rx combo. Would a non-essential, yet expensive (regulated, tied to a therapy) test appeal to the lab manager?
Will physicians drive regulated complementary testing?
In the Diaceutics survey of 30 anti-PD-1 prescribing oncologists, 40 per cent admitted to not understanding the nuanced differences between the two approved PD-L1 tests, and even some who claimed to be knowledgeable went on to provide erroneous definitions of ‘complementary’ and ‘companion’ diagnostics. With adoption of companion diagnostics still remaining suboptimal, the introduction of even more nuanced labelling for complementary tests is not the way to earn doctors’ trust.
Doctors clearly want better testing to aid treatment and patient management. For example, Diaceutics recently asked 20 MS specialists what diagnostic tests would make their life easier. The recurring answers were a call for a test ‘that would predict response to all the MS disease-modifying treatment’ and ‘that would help predict disease progression’. What this says to us is the definition of ‘precision’ or ‘personalized medicine’ is migrating from its original context in oncology with the intimate link between disease, driver mutation and therapy, towards more multifaceted diseases and multiple treatment modalities. And in that new domain the ‘one drug – one biomarker’ model will not fit physician’s needs.
Will diagnostic companies drive complementary testing?
The current predominant business model for the diagnostic industry guides it away from making deep financial investment in driving test adoption. Whilst there are exceptions (Genomic Health with CYTYC), it is perceived that the diagnostic company’s challenge is to make the test available in a disease area where clinical demand or pharma or payers may drive the test uptake. Diagnostic companies like Dako, Ventana and Abbott have actively participated with pharma in a ‘fee for service’ model. Here they are funded to develop a regulated test alongside the therapy in the optimum case scenario where therapy and test can then get approved and launched together. Sometimes the diagnostic company also goes on to work with pharma (where funded) to support companion test commercialization since companion status in a label will drive demand. However, without the guarantee that pharma will encourage testing (witnessed already with BMS and Opdivo), the regulated complementary concept is no moneymaker for the diagnostic industry.
We understand why the FDA wants to leverage pharma to support the shift away from LDTs with the regulated complementary labelling. Occasionally in diagnostics the stakeholders can align around new tests and drive rapid change, as seen with BCR-ABL or HLA-B testing in HIV. But the arrival of a regulated complementary label does not appear to have any real stakeholder alignment and we at Diaceutics fear it may complicate rather than simplify clinical decision-making. For this reason we think it does not really further the cause of personalized medicine.
 http://www.fda.gov/Drugs/ScienceResearch/ResearchAreas/Pharmacogenetics/ucm083378.htm, accessed 28/04/2016
 PharmKGB accessed 28/04/2016
In Vitro Companion Diagnostic Devices Guidance for Industry and Food and Drug Administration Staff, 2014 http://www.fda.gov/downloads/medicaldevices/deviceregulationandguidance/guidancedocuments/ucm262327.pdf