From a regulatory standpoint I’m going to bring up two points that I feel are important to consider for anyone who is involved in oncology drug biomarker development. Both points are also important for the future of good drug biomarker development.

First is about the investigational use of the assays and the need for appropriate regulatory requirements. When a therapeutic product sponsor is planning to use the results from any future diagnostic to enrol subjects or to manage subjects in a therapeutic product clinical trial, they should keep in mind that the IVD used in that context would be investigational unless the IVD used for that purpose has already received marketing authorisation for that specific intended use.

So if an investigational IVD is to be used in a therapeutic product clinical trial these clinical investigations are subject to Investigational Device Exemption Regulations 21CFR parts 8.12 and also 50 and 56. These IDE requirements that apply to an investigational IVD, they depend on the risk presented by the device. So the therapeutic product sponsor should assess the risk percentage to study subjects by use of the investigational IDE in the context of the therapeutic product clinical trial. Whether the use of the device in the trial is exempt or is of significant risk or of non-significant risk. Each category has specific requirements under the IDE regulation and these requirements of the IDE regulation would need to be addressed. These regulations also outline good clinical practices for studies conducted in the United States and they include requirements for review of clinical investigation by an IRB and of obtaining informed consent from study participants.

The second topic I will discuss is about the need for proper planning from a therapeutic product sponsor to bring in an in vitro diagnostic test for contemporaneous approval if a companion diagnostic is deemed essential for the safe and effective use of the therapeutic. So when a therapeutic product sponsor is planning a therapeutic product clinical trial, and if the trial is relying on information provided by an IDT, it could be for enrolment, it could be stratification, dose or other uses, the product sponsors should consider whether the IVD companion diagnostic development strategy is aligned with the approved goals for the therapeutic product.

The agency is recently seeing that therapeutic products and IVDs are typically developed on different schedules. We see that more education is needed regarding appropriate pre-planning at the stage of initial enrolment for registration trials in order to develop a high quality CDx. So sponsors should also be aware that when they use exploratory testing that is not sufficiently analytically validated or if it is validated with inappropriate analysis methods that could produce spurious associations.

So in clinical trials the recommended approach is to use a single analytically validated test for the study, either to screen all patients or a single test should be used for central confirmation of patient results prior to trial enrolment. For this, ideally if available, a CLIA or a approved IVD could be used. If the test used for trial enrolment is not a CLIA or an approved IVD then the enrolment test should be analytically validated. There should be a fully specified cut-off established and locked down.

For rare biomarkers the agency understands that sponsors’ preference has been to enrol based on local tests. So what one should remember is that use of a variety of clinical trial assays, what is called a CTA, with difference performance characteristics could result in a unique intent to diagnose population at each test site. So a minimal standard for analytical performance at local sites that will support a successful CDx market submission while safeguarding patient safety.
So a couple of points I would like to emphasise before I conclude this answer that will help for properly pre-planning at this stage of initial enrolment for registration trials in order to develop a high quality CDx. First, the pharma should provide minimum analytical validation requirements as minimal acceptable performance characteristics and acceptance criteria to enrolment sites and should only enrol patients from sites that meet the acceptance criteria.

Pharma should also ensure that all clinical sites should follow the same protocol for specimen collection banking and processing. Information on clinical trial assays includes pre-test methodology, test name, cut-off, specimen type used, a brief analytical validation, those should be available to evaluate if test results are comparable and if they are identifying the same group of patients that are receiving the drug. Varying classification rules if it’s there for the assay, those should be pre-specified, for example, biomarker rules to establish fusion positivity.

Pharma should make every effort, this is very important, pharma should make every effort to bank all marker positive samples and sufficient representative of marker negative patients for retesting with the market-ready CDx to demonstrate that the CDx is safe and effective for use for the corresponding therapeutic product.

All these points will make sure that a contemporaneous marketing authorisation of the therapeutic product and the accompanying IVD companion diagnostic will happen.

What is your opinion about the current issue of non-standardised assay policies leading to imprecision medicine and what has been do so far to standardise assays?

A good question. Standardisation and harmonisation of biomarkers, that’s key to the success of precision medicine. The FDA has been instrumental in several harmonisation standardisation efforts. For example, in the case of PD-L1 the assays for use with pembrolizumab, nivolumab, atezolizumab, durvalumab, they all have been developed independently and they all use different antibody clones, different immunohistochemistry staining protocols, scoring algorithms, cut-offs, for determining PD-L1 status. So there were questions about whether these tests can be used interchangeably and can they inform treatment decisions for the various PD-L1 and PD-1 inhibitors.

Given the variability in PD-L1 testing paradigms and the lack of available resources to perform multiple tests within laboratories, harmonisation methods were crucial. The FDA hosted a workshop along with ASCO and AACR in 2015 and highlighted the FDA/AACR/ASCO complexities in personalised medicine harmonising companion diagnostics across a class of the targeted therapeutics workshop with the unveiling of a blueprint proposal developed by four pharmaceutical companies and two diagnostic companies. Those efforts are still ongoing to harmonise PD-L1… harmonisation of PD-L1 efforts with these pharmaceutical companies and the diagnostic companies.

In the case of complex biomarkers such as TMB, MSI and HRD, NGS-based assays are currently available. They use different approaches to measure these biomarkers. So currently there is no agreement on what parameters contribute to the determination if a sample is positive or negative status for these. The FDA is working with the ? for Cancer Research, in standardising harmonising complex biomarkers such as TMB, tumour mutational burden, and HRD, homologous recombination deficiency. The FDA has been in the forefront of these several harmonisation standardisation efforts of biomarkers which we consider is important for the success of precision medicine.