January 1, 1970 (Vol. , No. )
Peter Collins, Ph.D. GlaxoSmithKline
This third chapter of a five-part series talks about coordinating diagnostic test kit development between pharmaceutical companies and diagnostic manufacturers.
The previous chapter described how the search for predictive biomarkers is likely to change drug development research. This chapter talks about turning biomarkers into companion diagnostics. Once biomarker tests have been turned into diagnostic test kits, we would be able to routinely classify patients in clinics. But the development of those diagnostic test kits has to be coordinated between two entities—pharmaceutical companies and diagnostic manufacturers—with rather different regulation.
As we have seen in the previous chapters, biomarkers can play a critical role in classifying patients into subpopulations. A biomarker can be used as the basis for creating a routine diagnostic test for clinical use, after it has been approved as an in-vitro diagnostic (IVD) test: i.e. a certified product, reagent, application or other tool that analyses human material and helps to diagnose patients.
There is a tendency for regulators to require that drugs and companion diagnostics are developed in tandem.
A companion diagnostic test is essentially a biomarker test that enables better decision making on the use of a therapy. In other words: it is a diagnostic test that is specifically linked to a therapeutic drug. The goal here is to increase the safety and the efficacy of the drug.
Pharmaceutical and biotechnology companies are trying to change their tack now and are working hard on integrating the co-development concept but true co-development has been a rare phenomenon until now. There are two reasons for this. Firstly, clinically useful biomarkers are usually established late in the drug validation process. Secondly the worlds of drug development and diagnostics, although both part of health care, are parallel universes in many ways.
In the next paragraphs we will elaborate on these issues.
Clinically Useful Biomarkers Found Late in The Drug Validation Process
It is quite difficult to find clinically useful predictive biomarkers early on in a drug development programme, simply because they can only be determined on the basis of the patients’ responses to the drug. A number of biomarkers, such as KRAS and EGFR mutations, could only be established after a sufficient number of patients—well beyond the usual number for a Phase III trial—had elicited better understanding of differential drug response.
Maybe, in the future, if the method of clinical trials changes thanks to our new understanding of diseases and classifying patients (as we described in chapter one and two), we might be able to find these biomarkers at an earlier stage. But as long as still 30% of the drugs fail during Phase III, diagnostic manufacturers will not be able to afford huge investments in the development of companion diagnostics.
The worlds of drugs and diagnostics are parallel universes: in general they have different development timelines, product lifecycles, return on investment, customers, and regulations.
Drugs are valued and reimbursed as products, typically of high value. Diagnostics are valued and paid for as services, typically at a much lower value. Relatively few if any models exist for valuing a drug diagnostic combination. This issue will be discussed in the next chapter.
Drugs are protected by patents, but in the companion diagnostic arena, biomarkers are considered to be within the public domain and there is less emphasis on intellectual property. It is even debated whether biomarkers should be patented at all, because biomarkers are not invented but already exist in cells.
Moreover, drugs and diagnostics are regulated differently. It takes many years and a large amount of data from expensive clinical trials to get a drug approved. Under current rules, however in the current IVD Directive there is little mention of clinical evidence.
However, the EU IVD Directive, which regulates in-vitro diagnostics, is under revision. The current IVD classification does not take scientific and technological evolution into account which is expected to be changed towards a new risk-based approach.
A revised IVD Directive (which may even become a Regulation) will improve safety and efficacy, but will also raise difficulties. It is likely to state that an IVD assay must have the performance characteristics required for fulfilling a clinical purpose. Moreover, requirements may be added on how to demonstrate clinical validity, possibly proportionately with the risk level of the test.
The intention to increase safety will inevitably lead to higher costs and greater efforts for manufacturers. It will take IVD manufacturers years and unusually high investments to take an IVD past the regulators. Undoubtedly, Phase III trials would benefit from well validated biomarker tests. But it is almost impossible to comply with regulations that require a clinically validated diagnostic test, simply because clinical evidence will only be provided by a clinical trial itself, and this is unlikely to be required for many IVDs.
Another point is that more regulation can severely impact innovation. The larger investments that are needed might limit the ability of (especially small) innovative companies to discover and develop biomarkers.
The Quality of Diagnostic Tests
In order to gain more insight into the consequences of stricter regulation, we must dive into the universe of diagnostic practice in the EU a little deeper, because there is another challenging reality.
In the diagnostic arena there are, basically, two types of tests in use: manufactured diagnostic products, which are regulated in the EU, and lab developed tests (LDT’s), which are not regulated. Thus, in the EU, there is a broad range of methodologies, often driven by the individual laboratory’s capabilities, access to diagnostic testing platforms, reimbursement and preferences of the individual team.
How this situation works in practice was pointed out by a multicenter study.1 Various laboratories were asked to select patients with a positive EGFR biomarker or a KRAS biomarker. Some of the tests were in close agreement but, alarmingly, others were not. Due to the poor quality of so-called ‘home brew tests’ or inadequately validated tests in hospital laboratories, patients with a positive EGFR biomarker might be excluded from an effective therapy on false grounds.
The result of this industry dynamic is a lack of standardisation of testing and ultimately, inconsistent patient selection for therapy. Improper patient selections will lead to poor therapy outcomes, or worse, to adverse reactions to drugs.
This situation is not likely to change under a new EU IVD Regulation. IVD manufacturers may have to make greater efforts in the future, including financially, to get an IVD past the regulators. However, there is no legally enforced requirement to use the companion diagnostic test approved in the clinical trial.
So, for as long as regulatory bodies do not sanction laboratories that do not conduct tests in line with pharmaceutical clinical trials, the quality of the testing will not improve. Diagnostic partners cannot justify large investments in bringing products onto a market, in which there are no rules for laboratories to perform substitute testing. Regulation for testing laboratories will be a more effective first step towards increasing quality and safety than sharpening the rules for IVD manufacturers.
The industry needs time to bring diagnostic and therapeutic research together. The regulation should leave room to develop companion diagnostics in various business models, instead of only requiring co-development. Let’s not forget: many drugs have proven to be effective without companion diagnostic tests. We would not want to rule those out.
Regulation to increase safety is in itself a good thing, but it is also important to provide opportunities for small innovative enterprises to enter the market.
Click here for the next chapter in this series, where you’ll learn about the economic aspects of personalized medicine.
Peter Collins, Ph.D., is vp, diagnostics at GlaxoSmithKline.
This article first appeared as chapter 3 of a white paper published by the EuropaBio Personalised Medicine Taskforce entitled “Personalised Medicine: Status Quo and Challenges”.
EuropaBio represents 56 corporate members and 14 associate members and BIO regions, and 19 national biotechnology associations who in turn represent some 1800 small and medium sized biotech companies in Europe. Members of EuropaBio are involved in research, development, testing, manufacturing and commercialisation of biotechnology products and processes. Our corporate members have a wide range of activities: human and animal health care, diagnostics, bio-informatics, chemicals, crop protection, agriculture, food and environmental products and services.
1 Whitehall V, Tran K, Umapathy A, et al. A multicenter blinded study to evaluate KRAS mutation testing methodologies in the clinical setting. J Mol Diagn. 2009;11:543–552.