Alexander Eggermont, MD, PhD, discusses the impact of the 3-year follow-up data from the phase 3 double-blind EORTC 1325/KEYNOTE-054 trial in high-risk stage III melanoma.
Alexander Eggermont, MD, PhD
Adjuvant pembrolizumab demonstrated a sustained, clinically meaningful improvement in recurrence-free survival (RFS) versus placebo in patients with resected high-risk stage III melanoma, according to updated results from the phase 3 EORTC 1325/KEYNOTE-054 trial (NCT02362594) presented during the 2020 ASCO Virtual Scientific Program.
At a median follow-up of 3 years and a total of 473 events (190 in the pembrolizumab arm and 283 in the placebo arm), the RFS rate was 63.7% with pembrolizumab versus 44.1% with placebo (HR, 0.56; 95% CI, 0.47-0.68; stratified log rank P <.001).
“Despite the short follow-up time, we had the number of events needed to conduct an event-driven analysis on RFS,” said Alexander Eggermont, MD, PhD, FASCO, lead author of the study, a professor of oncology at the Paris-Sud Universit, and; professor of oncological surgery and chair in the International Research Network on Cancer, at the Erasmus University of Rotterdam. “Results demonstrated that we had a significant difference [in RFS] between the pembrolizumab and placebo arm.”
In this trial, a total of 1019 patients were randomized to receive pembrolizumab at a flat dose of 200 mg (n = 514) or placebo (n = 505) every 3 weeks for a total of 18 doses or until there was disease recurrence or unacceptable toxicity. The primary end point of the trial was RFS in the overall population and in those with PD-L1—positive tumors.
Notably, the benefit observed with pembrolizumab proved to hold true across all subgroups.
In those with PD-L1—positive disease (n = 853), the RFS rate was 65.3% versus 52.2% with pembrolizumab versus placebo, respectively (HR, 0.57; 95% CI, 0.43-0.74; stratified log rank P <.001). In those with PD-L1—negative disease the RFS was 56.9% with pembrolizumab versus 33.3% with placebo (HR, 0.45; 95% CI, 0.23-0.90; stratified log rank P = .0002).
According to AJCC-7 staging, the RFS was 81.2% versus 66.3% with pembrolizumab versus placebo, respectively in those with stage IIIA (n = 152) disease (HR, 0.50; 95% CI, 0.22-1.16; stratified log rank P = .030); 65.7% versus 47.0%, respectively, in those with stage IIIB disease (n = 472; HR, 0.56; 95% CI, 0.39-0.81; stratified log rank P <.001); and 54.3% versus 32.3%, respectively in those with stage IIIC disease (n = 395; HR, 0.57; 95% CI, 0.40-0.81; stratified log rank P <.001).
Lastly, RFS with pembrolizumab was 62.0% versus 37.1% with placebo in patients with BRAF-mutated disease (n = 440; HR, 0.51; 95% CI, 0.36-0.73; stratified log rank P <.001) and was 61.8% versus 46.5% with pembrolizumab and placebo, respectively, in those with BRAF wild-type disease (n = 448; HR, 0.66; 95% CI, 0.46-0.95; stratified log rank P = .003).
In an exclusive interview with OncLive during the meeting, Eggermont, who is also the Joseph Maisin Honorary Chair of Oncological Surgery at the Louvain Catholic University (Belgium), as well as professor of oncological surgery and prior general director at Gustave Roussy, further discussed the EORTC 1325/KEYNOTE-054 trial and the implications of its findings in the field.
OncLive: Could you provide some background to the phase 3 EORTC 1325/KEYNOTE-054 trial? What did it analyze and what data were reported prior to this year’s virtual meeting?
Eggermont: This trial was conducted in 1019 patients with high-risk stage III melanoma. There was a randomization between a flat dose of pembrolizumab at 200 mg every 3 weeks for 1 year versus a placebo. High-risk stage 3 melanoma was defined in our trial as regional lymph node—positive melanoma. Although, notably, if [a patient] had just microscopic involvement of the regional lymph node, the diameter of the metastasis had to be larger than 1 millimeter; this was to ensure that we had enough events in the stage IIIA portion of patients to get a timely read out of the trial.
Data from this trial were originally published in 2018 in The New England Journal of Medicine and had a median follow-up time of 1.25 years. At that short follow-up time, we already had the number of events needed to conduct an event-driven analysis on RFS. This demonstrated that we had a significant difference between the placebo arm, where there was a RFS of 53.2%, versus 71.4% in the pembrolizumab arm. This difference was highly significant and RFS was the primary end point, which was defined by the number of events.
The co-primary end point was in patients with PD-L1—positive disease; that was the large majority of the overall patient population, namely 953. The HR in that group of patients was basically the same as in the overall group; it was 0.54 versus 0.57, respectively. [Pembrolizumab] was rapidly approved by the European Medicines Agency (EMA) in 2018, followed by the FDA in 2019. Now, we report more mature data on RFS, with a medium follow-up of 3 years. The question now is, “Are the results demonstrating sustainability in terms of impact on FS?” Not surprisingly, they do.
To get the full picture, it’s important to look at the hallmarks of 1-, 2-, and 3-year timepoints in terms of RFS rates. At the 1-year timepoint, the difference is 15%, at 2 years it’s 20%, and at 3 years it’s basically 20%. It’s a completely sustained RFS benefit that we also see at 2 and 3 years at a 20% absolute difference. At 3 years, its 63.7% for the pembrolizumab arm and 44.1% for the placebo arm. This is reassuring but not surprising because it’s what we would expect from checkpoint inhibitors. Moreover, this was already seen in the adjuvant trial with ipilimumab, (Yervoy) which demonstrated a smaller benefit versus the placebo in the same patient group. In this trial, we saw a 10% difference in RFS over the years, translating to the same 10% of distant metastasis-free survival and 10% to 11% overall survival difference over 3, 4, 5 years.
Now, we are seeing the early signs that anti—PD-1 will provide the same type of consistency, but at a larger, 20% difference with regard to RFS at years 2 and 3. We expect to be able to report on distant metastasis-free survival at the very end of this year. Everything seems to point in a similar direction. In the pembrolizumab trial, we already reported on distant metastasis-free recurrences as first events. Interestingly, for the distant metastasis-free recurrence events, we see it has a hazard ratio of 0.55, [which is] almost exactly the same to the RFS curves. We should be able to report officially on this secondary end point, metastasis-free survival, by the end of this year; events are coming in slowly because of the efficacy of the treatment.
Another important point of fundamental importance in this trial is that, at the time of recurrence, patients were unblinded. If patients were the placebo arm, they had access, as trial medication, to pembrolizumab in an advanced setting for up to 3 years or upon progression. There are 2 phases in this trial; so we will be able, due to cross over phenomenon, to report in the end OS in terms of whether there is a benefit to give it to all patients or whether providing at the time of relapse would salvage the same patient group. In the end, it is the only randomized adjuvant trial with a crossover setup that will be able to answer that second question.
The other primary end point was the 953 patients with PD-L1—positive disease. The 3-year median follow-up data showed a hazard ratio of 0.57. The PD-L1-negative patients do just as well, if not better, because they have a hazard ratio of 0.45. Whether you are PDL-1 positive or negative, you derive the same benefits of adjuvant therapy with pembrolizumab.
Were any differences between the substages of stage III disease observed?
The substages are stage IIIA, which is just microscopic lymph node involvement, IIIB, and IIIC. Again, we noticed a lot of consistency [with pembrolizumab]. The HR for stage IIIA was 0.50 and a difference of 15% in relapse rates at the 3-year mark. For stages IIIB and IIIC, we saw HR of 0.56, 0.57 and we see differences of 18% and 22% in absolute relapse rates, respectively. All subgroups derived similar benefits. This information is important because there are discussions for the stage IIIA group whether this should be provided as an option in the context of the immune-related adverse events (AEs) that one might observe in a patient group that has a risk of recurrence than those with stage IIIC disease.
Interestingly, the HR was 0.50 for the AJCC-7 classification, which was followed for this trial, but we also did a reclassification by AJCC-8. There, in the AJCC-8 classification, we found that the stage IIIA group is an even less risky patient population. The HR in that group was 0.43, so the benefit, if anything was even a little bit better. This is helpful for discussions determining whether this can be recommended to patients in the adjuvant setting. Of course, there is a difference between older and younger patients. For example, 80-year-old patients have a lesser chance of relapse than 22-year-old patients. Younger patients might be more likely to want to receive treatment that will improve prognosis or outcome. Ultimately, these updated results provided additional, very solid information on a very large group of patients.
What does the safety profile look like with this approach?
One element, of course, is what are the AEs of anti—PD-1 adjuvant therapy? The immune-related AEs are mild in terms of grade 3/4 events; these are very rare and you only see them in approximately 7% of patients. Per category, you only see them in 1% of patients. With the exception of endocrine disorder, its mostly hyperthyroidism followed by hypothyroidism, which you see in approximately 23% of patients; that is mostly grade 1/2 in severity and that is mostly only laboratory type of events. However, if it shows up as a clinical problem, then it’s easily counteracted by a thyroid hormone or replacement therapy. The other immune-related AEs are quite rare, so that [toxicity] does not seem to be a major obstacle.
One aspect of the immune-related AEs is that they have a tendency to be chronic. As such, when proposing this to patients with a very low risk of relapse, the therapeutic index (TI) comes into play. For the complete group of patients with stage III disease, it seems as though the TI for adjuvant therapy with pembrolizumab is clearly in favor of the TI that it represents and that the benefits with this treatment outweigh the associated immune-related AEs.
Earlier this year, an observation within this trial was published in JAMA Oncology stating that patients who did experience immune-related AEs, and again more than 90% of those events are grade 1/2 events in severity, have a better outcome than those who do not have those events. That appears to be a reflection of the immune tonus, the tonus of the immune system. With this information, patients can be reminded that although their immune-related AEs are uncomfortable, their presence could actually be a good sign.
What should your colleagues take away from this research?
What we can say with this research is that we expect for this to be confirmed with very similar data with regard to distant metastasis-free survival and that it will also translate to OS benefit in the end; however, it could take years for events to come in. Until then, it’s pretty clear that this is one of the standards of care. Importantly for patients with BRAF-mutant melanoma, it’s important to remember that anti—PD-L1 works equally well, regardless of BRAF mutation status. In our trial, the relapse rate at 3 years is even more favorably impacted in the BRAF-mutant patient population with a difference of 25% compared with the BRAF wild-type population where the difference was 16%; this is at least equal to the BRAF/MEK combination treatment in the BRAF-mutant population.
In general, in patients with BRAF-mutant disease the adjuvant setting, immunotherapies are a bit slower to kick than the BRAF/MEK combinations, but they eventually catch up, the curves cross over, and in the end, immunotherapy seems to be as good, if not better, than the BRAF/MEK combinations. This is true for advanced disease as well as for adjuvant treatments. This is the perfect consistency that we have from advanced disease into the adjuvant setting. These results are very reassuring for both patients and the medical community.
Eggermont AM, Blank CU, Mario M, et al. Pembrolizumab versus placebo after complete resection of high-risk stage III melanoma: New recurrence-free survival results from the EORTC 1325-MG/Keynote 054 double-blinded phase III trial at three-year median follow-up. J Clin Oncol. 2020;38(suppl 15):10000. doi:10.1200/JCO.2020.38.15_suppl.10000