Photo by National Cancer Institute on Unsplash

At the 2023 American Association of Cancer Research (AACR) Annual Meeting going on now, Moderna and Merck released the detailed results from the Phase 2b KEYNOTE-942/mRNA-4157-P201 trial evaluating mRNA-4157 (V940), an investigational individualized neoantigen therapy (INT), in combination with Keytruda in patients with resected high-risk melanoma (stage III/IV). The topline was released early in the year and despite it being viewed as positive, questions remained as to the complete results. Let us go through the major questions.

What is mRNA-4157-P201?

It is a personalized mRNA cancer vaccine that codes for up to 34 neoantigens. The goal of the treatment is to prime the immune system against a patient’s specific tumor and generate a T-cell response that is unique for each patient’s tumor. The combination with Keytruda is to ensure that immune response is not short circuited by inhibitory checkpoints. The history of cancer vaccines is long and filled with failure. The logic of priming the immune system to target a patient’s tumor is compelling but it has proven exceptionally difficult. The results toplined earlier in the year (and the Merck decision to exercise its option for the program in 2022 for $250M) has generated some excitement that the hurdles of cancer vaccine science are being overcome.

Were there any questions about the earlier data?

I think the most significant question revolved around statistical significance. The Hazard Ration (HR) for recurrence free survival (RFS) was an impressive 0.56 indicating a 44% improvement for the combination versus Keytruda alone. That being said the company used a single-side p-value of 0.0266. I do not want to swamp you with statistics, but the more common (and robust) robust method is to use two-side p-values and if they used that industry standard the effect would not have been statistically significant.

 Can the use of a singled-sided p-value be justified?

Perhaps but generally not in this case. The shorthanded way to think about the difference between the two standards is that the two-sided presupposes that the treatment can be beneficial, detrimental, or no effect. In contrast, the one-sided in this scenario assumes that treatment could be beneficial or no effect and excludes the possibility of a detrimental effect. It seems incorrect to assume away one possibility, which is why the two-sided is standard. All of that being said, this does not mean the data are bad, but the results are less robust than they first appear, which adds risk to larger studies that will try to replicate these results. In other words, one should not assume there is no effect but simply increase the risk and uncertainty around larger trials.

Are there any other concerns with the data?

There are some questions as to whether Keytruda underperformed in this trial with an 18-month RFS 62% in the new trial. The closest comparison might be Keynote-054 which showed an 18-month RFS of 71%. The combination inhe new trial had an 18-month RFS of 79% and when compared to the 62% of the Keytruda monotherapy, the benefit seems solid, but the question is whether that monotherapy effect was randomly low. If that is true, then the monotherapy effect in future trials might be closer to 70% in which case the benefit of the cancer vaccine looks less robust. Of course, cross trial comparisons are always difficult but given the questions about the robustness of the statistical significance it would not take a huge increase in the performance of the control arm to make a real difference in the statistical significance.

Is there anything in the data that increases confidence in the treatment arm?

The companies reported the results of a sub-group analysis comparing the effect in tumor mutation burden (TMB) high versus non-high. By way of background TMB is generally thought to be correlated with immune recognition and increased effectiveness of IO treatments (see article). One can think of high TMB tumors as priming the immune system, which should make treatment with Keytruda more effective and that is generally the case.

If this sounds similar to the proposed mechanism of action of the cancer vaccine, then you are paying attention. The idea is the use the neoantigen vaccine to prime the immune system against tumors that are likely hidden from the immune system. If that is the case, then one would expect to see the effect of TMB to no longer be relevant. In other words, treatment with the cancer vaccine would prime the immune system in the same way as a high TMB primes the immune system.

This is sort of what we saw in the subset analysis. The HR in the high TMB was 0.65 and in the non-high subset it was 0.59. Those are pretty similar results and hint that the cancer vaccine is priming the immune system. This is certainly encouraging and from my perspective is the most positive aspect of the data.

Will the cancer vaccine work in larger phase III trials?

That is the ultimate question and there are a couple of moving parts. The statistical significance questions increase risk and uncertainty with the larger trials. The potential underperformance of the control arm also increased risk as that may revert to the mean, which would lower the perceived benefit of the treatment (keeping in mind the risks of cross trial comparisons). The data in the TMB subsets indicate the treatment is priming the immune system, which actually decreases the risk of larger trials.

Ultimately, I suspect that if we look at a series of larger phase III trials in the future, there is going to be a mix of success and failures. It seems that the cancer vaccine does prime the immune system and provides an added benefit. The degree of that added benefit is probably not enough to generate statistically significant results in all future trials but is also not small enough to fail in them all.

As such, I think we eventually will see some successes and failures. Ideally, for the companies the successes will be in the larger indications, which would maximize the commercial potential. I do not think we can handicap whether the first phase III trial will be a success or not, but it seems that the cancer vaccine has enough of an efficacy signal to succeed in some indications.

More By This Author:

Biogen Moves Deeper In Alzheimer's With Denali
Sarepta: More Drama Or Just More Noise?
Amylyx Pharmaceuticals - An Emerging Leader In ALS