Johnson & Johnson
From exon 20 insertions to subcutaneous delivery and inclusive trial design, Joshua Bauml explains why the biology demanded a different approach, and why first-line tolerability is now an efficacy argument.
The EGFR exon 20 insertion mutation has long been one of oncology’s more frustrating puzzles: a clear driver, no clean oral solution, and patients running out of road before second-line therapy could even enter the picture. Johnson & Johnson’s amivantamab (Rybrevant) was built around that frustration. At the World Conference on Lung Cancer, J&J presented a cluster of datasets spanning overall survival updates from PAPILLON, tolerability findings from COPERNICUS, and subcutaneous delivery data from PALOMA-2, each adding a layer to an increasingly coherent clinical argument. Joshua Bauml, M.D., Vice President and Disease Area Stronghold Leader for Lung and Head and Neck at J&J, spent a decade treating these patients before crossing to industry. His read on the data carries both registers.
Why did the biology of EGFR exon 20 insertions point toward a bispecific approach rather than another TKI?
Joshua Bauml: When you think about why EGFR exon 20 insertion mutations tend to respond poorly to tyrosine kinase inhibitors, you have to understand the mechanism. That mutation creates a pocket on the inside of the cell that is very difficult to design a pill to target. And when you do design a pill that targets it, you end up hitting a lot of wild-type EGFR, which is why historically you see a lot of wild-type toxicity, most notably diarrhea. But it doesn’t mean that EGFR is not a driver. And beyond that, when EGFR is a driver, we know MET is coming right behind.
So we felt that amivantamab was uniquely positioned to address this. And in our phase one study, one of the most dramatic moments in the development of amivantamab was seeing the first response in an exon 20 insertion patient who had literally no options outside of chemotherapy. Seeing a response there, we felt we were in a position to really help patients who need better therapy so desperately. That’s how we got going with this.
Why does the crossover-adjusted hazard ratio from PAPILLON tell a different story than the ITT estimate, and how should clinicians weigh them together?
Joshua Bauml: We know that crossover has a substantial impact on outcomes. But we also know that amivantamab has very meaningful activity in this disease in the second line. J&J is very committed to improving outcomes for patients, not just hitting a p-value. So we constructed the trial to ensure that at the time of progression, patients could get access to amivantamab.
But in the real world, not everyone is able to get a second-line therapy. Barriers exist within markets around the world. We’ve seen consistently across multiple studies that the number of patients who actually receive second-line therapy is rather limited. What the crossover-adjusted analysis helps provide is a perspective on what outcomes might look like for patients in that real-world setting.
“We’re not just going to hit a p-value. We want to make sure we’re improving outcomes for patients worldwide.”
How does COPERNICUS change the way the field should read amivantamab’s tolerability profile compared to what earlier individual trials showed?
Joshua Bauml: I treated lung and head and neck cancers for a decade before I joined J&J. And one of the aspects I’m most proud of in the development of amivantamab is that J&J has been committed to improving outcomes throughout. Through the process of developing this drug, we identified a side effect, figured out how to improve it, then identified another, and improved that too. What we’ve ended up with is quite an armamentarium of ways to improve the treatment experience for patients.
What COPERNICUS allows us to do is take everything we saw in MARIPOSA and PAPILLON, which were fantastic results, and say: we actually know now, because of the iterative learnings from each of those trials, that the experience going forward doesn’t have to look like what was seen in those earlier studies. No single one of those trials allows you to comprehensively characterize that. If you look at the adverse event profile in the longer follow-up COPERNICUS data, I think it’s very clear that amivantamab-based regimens are emerging as a reasonable and feasible option for the front line in EGFR-mutant lung cancer.
Why does the 25 to 30 percent of patients who never reach second-line therapy change how you think about tolerability in the first-line setting?
Joshua Bauml: That number of about 25 to 30 percent is one of the most consistent data points from everywhere you look. But the most interesting thing is, if you ask oncologists how many of their patients don’t get to second line, they say all of them do. The data tells us that can’t be true.
What that means to me is we need to make sure patients get the best treatment in the front line. It does two things. First, it ensures patients get access to the best possible therapies. Second, if you give them the best treatment in the front line, I truly believe you increase the likelihood that they will get a second-line treatment, because you’re not driving aggressive progression from giving an inadequate therapy upfront. We need to make sure we balance quality of life and duration of life. We want patients to live longer and better. Lung cancer is a deadly, horrible disease, and we’re here to stand side by side with patients and physicians as they fight it.
Why is pharmacokinetic noninferiority from PALOMA-3 sufficient to support subcutaneous adoption, without a direct randomized head-to-head comparison?
Joshua Bauml: When we designed PALOMA-3, the goal was clear: you’re going from an infusion that lasts a couple of hours to an injection that lasts about five minutes. That is going to improve quality of life and patient experience objectively. And then we were very pleased to see that in making that shift, you also markedly diminish the rate of administration-related reactions, down to about 15 percent, with reduced severity as well.
So now I have something that improves quality of life in terms of time in the chair and improves quality of life in terms of tolerability. And if you look at the PALOMA-3 data, there’s a hint that it might even be better on efficacy. Obviously, the study isn’t powered for that, but it’s a suggestion. From a pharmacokinetic standpoint, we’ve already demonstrated noninferiority, alongside clear improvements in the treatment experience. What PALOMA-2 does is let us say, okay, the majority of patients are going to be receiving subcutaneous. What does this look like in that setting? It’s similar to COPERNICUS in spirit: what does the picture look like when you apply the best approach? It gives people more data to help guide their discussions.
How did COPERNICUS address the structural barriers that have historically kept underrepresented patients out of clinical trials?
Joshua Bauml: I can speak to this as a physician who treated patients, and now as a leader at J&J. When I was treating patients, you want to get them on trial, get them access to treatment. At the same time, you do not want the trial to be structured in a way that it could lead to misleading conclusions. This led to some overly conservative inclusion/exclusion criteria being used.
One great example is prior malignancy. There was fantastic work done by David Gerber at UT Southwestern looking at this, and his team demonstrated that outcomes are identical regardless of a history of prior malignancy. So why were some trials excluding those patients? Our original trials already incorporated that knowledge. J&J has done a lot of work to improve eligibility criteria throughout our development program.
But we also know unexpected barriers can emerge. Enrollment of minority populations into clinical trials is a real problem. There are trust issues. There are also biologic issues, for instance in the way kidney function is measured, which can vary by group and affect downstream eligibility in ways that are significant. What we did with COPERNICUS was aim for a very pragmatic design. We looked at each eligibility criterion and asked: is this going to directly affect the interpretation of the data? If the answer was no, we removed it. We also reduced the burden of trial participation itself, cutting down the number of visits required for PK and scans, so patients could spend more time with their friends and loved ones than in a clinic chair. That was the whole principle. And if you look at the COPERNICUS data, we enrolled much older patient populations. To demonstrate that degree of tolerability and outcomes in a higher-risk population is something we’re really proud of.
“We identified a side effect, figured out how to improve it, then identified another, and improved that too.”
Joshua Bauml, M.D., is Vice President and Disease Area Stronghold Leader for Lung and Head and Neck, Johnson & Johnson.
Moe Alsumidaie, MBA, MSF, is founder and Chief Editor of Vanguard Publications, which publishes Clinical Trial Vanguard, Pharma Vanguard and BullScope, and Head of Research at CliniBiz. He has two decades in clinical trial operations and data science, with earlier roles at Genentech, Abbott Vascular and Stanford University Medical Center, and is a guest lecturer in clinical trial sciences at Rutgers University.
