During the second day of the Medicines for Europe 17th Biosimilar Medicines Conference, held March 28-29 in Amsterdam, the Netherlands, a panel of experts discussed the current state of the regulatory landscape and considerations for tailoring clinical programs in the years ahead.
One of the hottest issues in biosimilar development is the potential to tailor clinical trials for biosimilars as a means to expedite approval and access while reducing the burden of cost and lengthy trials on biosimilar developers. However, stakeholders do not all necessarily agree on the feasibility of streamlining programs or waiving phase 3 clinical trials for biosimilars. During the second day of the Medicines for Europe 17th Biosimilar Medicines Conference, held March 28-29 in Amsterdam, the Netherlands, a panel of experts discussed the current state of the regulatory landscape and considerations for tailoring clinical programs in the years ahead.
The panel’s moderator, Paul Chamberlain, member of the NDA advisory board (who noted that the panelists were expressing personal views), called the discussion a “challenge to the status quo,” and asked the panelists to grapple with the question of how developers can leverage knowledge gained over the past decade, clinical experience, and postauthorization data to provide a more efficient approach to determining the clinical weight of evidence.
First, René Anour, DrMedVet, senior medical assessor in the Austrian Medicines and Medical Devices Agency, opened the panel by describing a notable recent learning from the development process of Samsung Bioepis’ trastuzumab biosimilar, Ontruzant.
During development, “they were nicely similar on a quality level, preclinical level, and on a [pharmacokinetic, PK] level,” he said, “and then something happened that we had not anticipated at all.” In a phase 3 study, there was observed superiority of the biosimilar to the originator in terms of pathologic complete response. “This was a human breast cancer model, and the superiority was significantly shown,” he said. “That struck us in a way, because we thought that, if there was a difference, we would have seen it early on.”
The superiority of the biosimilar was eventually determined to be a result of product shift in the reference trastuzumab, Herceptin, that resulted in changes to its antibody-dependent cell-mediated cytotoxicity (ADCC). At the time, ADCC was not understood to be key to the efficacy of trastuzumab, said Anour; “the learning we can derive from that is, probably, we do not always know what attributes are of interest for the clinical profile of the product. Not yet, at least.”
The case of the change to ADCC in reference trastuzumab, he added, has generated some uncertainty about how far comparability studies can be trusted to make phase 3 trials unnecessary.
Additionally, with pegfilgrastim, Anour said that the factors influencing PK were as not well understood as regulators thought when the first, early biosimilars were being developed. “We found out the PK, [pharmacodynamic, PD] correlation wasn’t very good,” and “we cannot always trust plasma levels of this molecule,” either. In some cases, he emphasized, there remains enough uncertainty that tailored approaches to approval are not warranted. “It might seem like we as the regulators are slowing the process down,” but decisions need to be based on the evidence, not the desire to speed up the process of approval.
Niklas Ekman, PhD, head of the biological section at the Finnish Medicines Agency and vice chair of the European Medicines Agency (EMA) biosimilar medicinal products working party, added that it is difficult to link quality attributes to clinical outcomes and challenging to understand at what level these differences in attributes need to be present in order to generate clinical differences.
In the trastuzumab case, said Ekman, part of the problem was related to assays for ADCC. Some assays are so insensitive that “you can’t see anything, no matter what is happening.” One key challenge is to select the right assay to determine what constitutes a relevant difference between products.
Another case in which some differences have been observed between biosimilars and reference products concerns etanercept; Samsung Bioepis’ biosimilar, Benepali, is associated with fewer injection-site reactions (ISRs) than reference etanercept, and that fact is related to lower levels of high—molecular weight aggregates in the biosimilar than in the reference, Enbrel.
“On the analytical side,” said Ekman, we try to understand as [well] as possible all the differences we see, but obviously, there are situations where we don’t understand the meaning of the difference,” although it is “very seldom” that specific differences in quality attributes can be linked with a clinical observation like reduced incidence of ISRs.
Alex Kudrin, MD, MBA, PhD, an independent biopharmaceutical consultant, added that, for some classes, there exists sufficient experience to potentially waive some data. With products as well understood as growth hormones, for example, it is even questionable whether PK data are necessary. Anour added that insulin could be a good candidate for a tailored approach; if good comparability has been demonstrated, a developer would not necessarily need more data than those derived from a PK/PD study.
However, Kudrin pointed out, some products, like rituximab, are surrounded by more questions. For example, said Kudrin, the FDA has signaled that it is not yet clear what the mechanism of action is for rituximab in the indication of vasculitis. “There are uncertainties around this molecule,” he said, and it is not a product that is suited for testing in healthy subjects.
Additionally, from a clinical perspective, the next wave of biosimilars will be difficult to assess; anti—programmed death-ligand 1 agents will be a particular challenge, because overall survival—the most reliable measure of their efficacy—will not be practical as an end point for phase 3 clinical studies.
Chamberlain raised the question of drugs to treat orphan indications and whether it will be feasible to tailor approaches for clinical development. Anour indicated that tailoring would be necessary for these rare disease drugs, and currently, programs for these products are being addressed on a case-by-case basis as regulators try to find a suitable way forward. Anour encouraged developers to seek early scientific advice about their programs from the EMA.
Finally, Uwe Gudat, MD, head of clinical safety and pharmacovigilance, Fresenius Kabi, called into question whether phase 3 clinical trials are the most fit for understanding differences between biosimilars and their references. “Are the difference we see substantial? Are they robust and reproducible, or is it more or less a chance effect?” he asked. For clinical trials in samples sizes usually used for biosimilars, “it ends up with a maybe.” While these studies are well suited to detect large, dramatic differences, it is unlikely that these would slip through during the preclinical process, and studies can raise as many questions as they answer, he argued.
The human body can be a clinical tool, Gudat said, used alongside the other tools at the developer’s disposal. But “If I have a high degree of confidence, and I have a PK/PD study that adds to that confidence, how much more confidence do I need?”
In response, Anour said, “It has come up a couple of times that clinical trials end up with more questions than we have before. That’s true in some examples, but the answer cannot be ‘let’s not ask any more.’ We need to ask better.”