Contrary to the previous meeting two years ago, we focussed on the importance of the framework around the end users of a biomarker and a drug. So we have this current paradigm where you have to use a biomarker to select a patient to give the patient the correct drug and then all this full concept moves to the daily practice healthcare setting where society needs to decide on the reimbursement, society needs to control that the drugs are available and given to the best patients that are selected using the best assays. We encounter by experience that the framework on how these assays and these biomarkers are selected is maybe reaching its limit.
So this framework was developed around 15-20 years ago while I was a resident in training and at that time it was very valid for the simple reason that there were not that many pathology labs, there were not that many targeted specific drugs and there weren’t that many assays. So at that time it was very clear that you could use the paradigm and that’s where we centred this meeting on: one drug, one assay.
Over the past 20 years the framework has remained the same. Currently we start to encounter more often that this framework is maybe not applicable anymore to the current reality. This poses a lot of issues in terms of is the evidence to reimburse a drug by a government agency or regulatory agency, is that indeed the highest level of evidence that we tended to use previously for approval of drugs? The same occurs for assays – are the assays that have been approved by the regulatory authorities? Is it, indeed, the case that these assays are still the best ones that we should use in our daily practices? The evidence starts to point out that this may not be the case.
So without invalidating or saying that we have to erase, like a tabula rasa, the concept of companion diagnostic development and drug and biomarker development, we need to think of reframing the framework because, in my humble view, it just doesn’t fit reality anymore.
How do these themes contribute to the future of cancer research and translational medicine?
We had this paradigm, we had this paradigm that if you had a positive phase III clinical trial, so a prospectively designed trial where you randomised your patients according to the biomarker, and if that trial becomes a positive trial then we always assumed that the assay that was used in the trial would be the best assay to select the patients because the trial was positive. That’s an important paradigm which we pathologists know that this is not the case but we need to raise the awareness amongst clinicians and patients and regulatory and also industry that this model is, indeed, reaching its limits.
The current paradigm and the new paradigm that has been proposed by, for example John Bartlett at the latest IBCD meeting, was that the clinical trial is not focussed on but does need to be focussed on validating the assay but validating the biomarker. If that trial validates the biomarker, and that’s a new concept, and leaves the validation of the assay to the community, and that community includes regulatory, it includes industry, it includes also patients, it includes the professional academic organisations, then we can more safely evaluate which assay is the best to evaluate that biomarker that has been evaluated in that particular trial. If you do this then we do know that the research stemming out of this positive trial will be more valid. Because what happens now is that you get trials using an assay of which we tend to think, ‘This is not the best assay available,’ and you see a lot of research projects using that same assay. So you start to replicate and induce a framework using probably suboptimal data because the assay that was used in those research projects is probably not the best one to use for that particular biomarker. So in that sense the current paradigm does affect the research.
That’s one item. The second item relates to the current paradigm, again, that if the trial is positive you have a companion diagnostic assay which is approved and what we start to see now more often is that the analytical validity, so all the work done by the academics to evaluate the performance of that assay, is done after the trial, not before. So the paradigm that we propose or we plan to propose or we plan to think about, is that the analytical validity, so meaning how good is the assay and how good compares the assay that is going to be used with the drug to other assays, that all this work should be done before the trial starts so that we can ensure that when the trial starts that we have the best assay available, while this is done now after the trial with all the misery. Why? Because if we do find indeed that the assay is not perfect we cannot go back because the drug is approved and the assay is approved.
That’s one thing and the other thing is about education of the society. I was raised, again 20 years ago, that the assay approved by the regulatory agency is the best assay available. This evidence at the time was endorsed by the fact that in daily practice pathologists were quite, let’s say, creative in terms of using assays. But nowadays we have a tremendous amount of professional accreditation bodies like the CAP accreditation, the clinical pathology accreditation, ISO15189 accreditations, obligatory to use external quality control schemes, most labs do have an obligation to impose quality standards to use particular assays. So nowadays we have much more opportunities to control the quality of the assay used in a particular trial and in a particular healthcare setting.
Using this new framework where we can control the quality of assays using daily practice and based on the evidence of the quality of the performance of the assay before the trial was started, if this framework gets implemented I personally think that the biomarker driven identification of patients for targeted treatments will be much more optimal than it is today and less biased towards an industry setting.