The key to growth in the personalised medicines sector lies in companion diagnostics. Austin Tanney, Scientific Liaison Manager at Almac, explains why a new generation of biomarkers is needed to capture the complex biology involved
More complex assays are being used to stratify patients for enrolment in early phase clinical trials
Personalised medicine has become a mainstay of the pharmaceutical industry. Companion diagnostics, where a specific diagnostic test is performed prior to administration of a drug, is becoming increasingly common. The first companion diagnostic to receive FDA approval was in 1998 and was an immunohistochemistry (IHC) test for the HER2 biomarker linked to the drug Herceptin. HER2 is the receptor that is targeted by Herceptin and, thus, the biological association of the test and drug is clear and unambiguous.
There has been a dramatic increase in the number of drug/diagnostic pairings that have obtained FDA approval. In the last three years alone, a total of nine companion diagnostic tests have been approved, bringing the total to 18 overall. This apparent progress is, however, somewhat misleading, since of the 18 approved diagnostics currently listed on the FDA website, 10 are directed to HER2, either as IHC or in-situ hybridisation (FISH) assays. The remainder are also fairly low complexity tests looking at a single mutation, such as KRAS, and like the HER2 test for Herceptin the biomarker is often the target of the drug.
In recent years it has become clear that the paradigm of the drug target doubling as the biomarker is flawed in many cases. A notable example of this is the case of EGFR inhibitors, such as cetuximab and panitumumab. Here we have seen that the presence of the EGFR receptor is not a strong biomarker and that the mutational status of a different gene in the same pathway, KRAS, is a more important predictor of response. KRAS mutational status is now commonly used as a companion diagnostic for prescribing these drugs, although subsequent studies have shown that KRAS mutational status is not the only determinant of response.
It has become clear that the underlying biology of a disease and the response and resistance mechanisms to chemotherapeutic agents are complex, even when targeted strategies are being employed. A good example of this is the issue facing the continued use of the first generation of angiogenesis inhibitors that were designed to target the VEGF pathway. While a subpopulation of cancer patients clearly benefit from these agents, neither the presence of the VEGF receptor nor activation of the downstream VEGFR pathway predicts response. This has created a dilemma as to how the existing agents – and indeed second-generation angiogenesis inhibitors – can be targeted to improve response rates.
What is now evident is that a new generation of biomarkers is required that can accurately capture the complex biology that underpins sensitivity to these agents. This in turn has forced the companion diagnostics industry to embrace more complex assay designs and technologies, such as multiplex PCR, microarrays and Next Gen strategies, to develop and validate appropriate biomarkers capable of accurately capturing this more complex biology.
To date there are no complex gene signatures that are approved as companion diagnostics. However, this is something that is increasingly being seen in the early development pipeline, with many more complex assays being used to stratify patients for enrolment in early phase clinical trials.
The pharmaceutical industry seems now to be much more open to the use of more complex biomarkers as stratification tools for early phase clinical trial enrichment studies
The pharmaceutical industry seems now to be much more open to the use of more complex biomarkers as stratification tools for early phase clinical trial enrichment studies. Indeed, many pharma companies have built experienced teams internally that are responsible for delivering appropriate biomarkers for new agents in their pipeline. The strategy employed by many of these companies is that the biomarker validation and early clinical trials of the drug/biomarker combination is outsourced to companies such as Almac Diagnostics. The advantage of this approach is that the drug can be killed quickly if efficacy is not seen in the biomarker-selected population.
When a biomarker is discovered in a pre-clinical setting, there are a number of important considerations that must be accommodated prior to its implementation clinically. Most commonly, the first consideration is the migration of the assay to an appropriate tissue and/or platform. Preclinical biomarker studies are frequently carried out on cell line or animal models using fresh tissue, so a migration study is usually required as the tissue to be tested will often be formalin fixed and paraffin embedded (FFPE).
Another form of migration that often has to be considered is a platform migration. When discovery has been carried out on a platform such as microarray or next generation sequencing, the delivery of a clinical diagnostic may be better suited to a qPCR-based platform, depending on the number of genes in the signature. Typically, Almac carries out both a platform and tissue migration, often moving from a signature discovered from fresh tissue using a microarray to a qPCR panel optimised for performance in FFPE.
When applying a biomarker in an early stage clinical trial it is essential that the biomarker be validated to a sufficient standard
When applying a biomarker in an early stage clinical trial, particularly if the marker is being prospectively applied and patient treatment decisions are being made as a result of the test, it is essential that the biomarker be validated to a sufficient standard. For a clinical test to be released as a full companion diagnostic, FDA approval is obviously required. However, for early phase clinical trials where the biomarker hypothesis is still clinically untested, there needs to be an alternative route.
For early phase trial enrichment studies, where patient treatment decisions are being made based on the result of the test, Almac develops all its biomarker assays as ‘laboratory developed tests’ (LDTs), which are delivered from its CLIA and CAP accredited laboratory in the UK. The company finds its geographical location is ideal to support these early phase studies, which are typically run across multiple sites in North America, Europe and Asia.
The upfront development costs for a CLIA validated assay is significantly lower than that of an FDA-compliant assay. There are, however, challenges associated with delivery of these tests.
The assays themselves are unique to the client and are run exclusively for that client. Furthermore, when these assays are used for early phase trial enrichment, a very rapid turnaround time (TAT) is required, with results usually needed in less than one week. Many of these assays take four to five days to run, and to meet the TAT must be processed as soon as they are received, typically as batches of one. This is extremely labour-intensive and, as such, the cost of running such assays is very high compared with simple ‘off the shelf’ high throughput tests that most clients are used to.
Once the early phase trials are complete, if the biomarker hypothesis has held up and the client wishes to develop the test as a full companion diagnostic, there are a number of possible routes. The most common one, and the one that has been taken by most companies to date, is to develop the assay into a kit and submit the kit to the FDA for a full pre-market approval (PMA). However, with the development of more complex assays that are technically challenging the standard kit-based approach may not be appropriate. The platforms used for multiplex gene signatures are not currently as widely available and there is often a significant degree of expertise required to run such assays.
For a kit-based approach the analytical validation must be performed across three independent sites to ensure inter-lab reproducibility
Importantly, the degree of biomarker validation required is similar to the more traditional kit-based approach in terms of both the analytical and clinical performance characteristics. The major difference relates to the number of sites required for the validation. For a kit-based approach the analytical validation must be performed across three independent sites to ensure inter-lab reproducibility, while for a LDT-PMA the validation can be carried out at the single site that will deliver that assay. Furthermore, the LDT-PMA route does not require the same investment by the diagnostic partner in terms of kit manufacture and distribution.
A major risk relates to the timely shipment of samples to the central laboratory and the potential impact on TATs. However, the emergence of niche courier companies that specialise in the global shipment of biological samples mitigates this risk to a large extent.
Another obvious risk of the LDT-PMA route is the reliance on a single site for assay delivery. Almac can, if required, run assay validation across two of its sites to account for this risk, but it does increase the overall validation costs. Key decisions of this type are typically addressed during the risk management phase of the assay design control process.
To conclude, drug/diagnostic pairings are clearly on the increase and it is likely that in the next five to 10 years, most new drugs approved (particularly in the oncology space) will have an associated companion diagnostic. As the complexity of the diagnostics being investigated grows, we are beginning to see a change in the way that these diagnostics are developed and will also expect to see changes to the model by which these tests are delivered commercially.