A Look at the KEYNOTE-054 and COMBI-AD Data

Axel Hauschild, MD, PhD: In the following, we particularly want to discuss the KEYNOTE-054 and the COMBI-AD studies because they have quite comparable eligibility criteria, and therefore they are not equally mature, but they might have the chance to compare, indirectly at least, the data. Would you like to give us a brief update on KEYNOTE-054 and COMBI-AD?

Dirk Schadendorf, MD: Yes. That’s quite important, Axel. The trial design for both studies was comparing pembrolizumab and dabrafenib/trametinib. Both used a placebo arm as a control, and both used a 1:1 randomization of patients who got included. These were patients truly in stage III, so their AJCC [American Joint Committee on Cancer staging system,] classification, 7th edition, was used, which included patients in stages IIIA, IIIB, and IIIC.

Patients in stage IIIA are the sentinel node positive patients. In the stage IIIA category, both studies decided not to include all IIIA patients but only a small subset of patients who had the sentinel nodes. The tumor deposit in the node needed to be more than 1 mm in order to be included.

It was dependent upon the different countries. For example, in Germany, this is less than 20% of the patients in the IIIA category, according to the 7th edition of the AJCC staging system. In other countries where the primary detection of the primary tumor is not so frequent, parts of Eastern Europe for example, the frequency of patients who are in that category of more than 1 mm of sentinel node burden is much higher.

This has consequences for the patient characteristics’ inclusion in that clinical trial. The interesting part is that the majority of the patients were in stages IIIB and IIIC. Looking at the relapse-free survival, it was quite interesting to see that both treatments were given for 1 year without any rationale, just 1 year.

If you look at the relapse kinetics, you see that, of the patients who got treated with pembrolizumab, which is also true also for nivolumab at the end, roughly 25% of the patients who are relapsing are relapsing while on treatment within this 1 year. That’s completely different to targeted therapy in the dabrafenib/trametinib arm, and in the COMBI-AD study, only a small fraction, 5% to 7%, of the patients progressed while on treatment. 

One has to say that, for patients who got controlled nicely for this 1 year of treatment with dabrafenib/trametinib, if they relapsed, they relapsed after stopping the targeted therapy at month 12. After roughly 30 months, the relapse rate between targeted therapy and checkpoint blockade was very comparable and very stable at that point. It stabilized, and it’s stabilizing over time now. 

Axel Hauschild, MD, PhD: Dirk, we didn’t discuss the magnitude of the benefit or that the hazard ratio was 0.50, 0.47 in all of this. It’s around a 50% improvement of relapse-free survival, and it has the same benefit for distant metastasis-free survival. For COMBI-AD, we already have the first interim analysis available. The final will be done with a final evaluation of the entire study. All of them are showing a 0.50 hazard ratio for a benefit. Don’t you think this is a breakthrough?

Dirk Schadendorf, MD: That’s definitely a major step forward. If you think about 20 years of discussion regarding interferon, the hazard ratios were in the range of 0.8, a little bit more or less. It’s a completely different ball game here. When we talk about the effects on distant-metastasis-free survival as well, the hazard ratios were shrinking from relapse-free survival to distant-metastasis-free survival. 

We see a constant effect in both clinical trials. The hazard ratio is also protecting against the occurrence of distant metastasis in the same magnitude, cut roughly by half. That’s an important message not just for the physicians; it’s also good news for the patients and families.

Transcript Edited for Clarity