Neoadjuvant, Adjuvant Data Provide Evidence of Curative Potential in Advanced Melanoma

October 1, 2020

Partner | Cancer Centers

Michael B. Atkins, MD, discusses the clinical impact of key adjuvant and neoadjuvant data in the melanoma space, as well as the risks and benefits of both approaches.

Long-term data presented at the 2020 ASCO Virtual Scientific Program demonstrated sustained relapse-free survival (RFS) benefit with adjuvant therapy in patients with advanced melanoma, providing further evidence of the curative potential of postoperative therapy, explained Michael B. Atkins, MD.

For example, findings from a 5-year analysis of the phase 3 COMBI-AD trial demonstrated a RFS rate of 52% (95% CI, 50%-60%) with the combination of dabrafenib (Tafinlar) and trametinib (Mekinist) versus 36% (95% CI, 34%-43%) with placebo in patients with resected, stage III BRAF-mutant melanoma.1

Additional data from the meeting suggest that the role of neoadjuvant therapy is becoming more relevant, explained Atkins, displaying encouraging, long-term pathologic complete response (pCR) rates and high RFS rates.

In this regard, 2-year follow-up data from the phase 2 OpACIN-neo trial demonstrated that 2 cycles of neoadjuvant ipilimumab (Yervoy) and nivolumab (Opdivo) induced an RFS rate of 83.6% in patients with stage III melanoma.2

Moreover, at 2 years, the RFS rate in patients who achieved a pCR remained significantly higher, at 96.9% versus 35.5% in patients without a pCR (P < .001).

“It is important to get an objective measure of whether the patient is responding, which we cannot get in the adjuvant setting,” said Atkins with regard to the utility of neoadjuvant therapy. “[We may] see the tumor shrink or pathologic evidence that the treatment is working [in the neoadjuvant setting]. This could reassure patients and provide them a chance to stop treatment earlier.”

“Also, if the treatment isn’t working, this may give us the opportunity to choose a different therapy earlier,” added Atkins. “We could [add] another drug, switch to another approach, or potentially study the tumor to identify how it is escaping that particular therapy.”

In an interview with OncLive® during the 2020 Institutional Perspectives in Cancer webinar on Melanoma, Atkins, deputy director of the Georgetown-Lombardi Comprehensive Cancer Center, and the Scholl Professor and vice chair of the Department of Medical Oncology at Georgetown University Medical Center, discussed the clinical impact of key adjuvant and neoadjuvant data in the melanoma space, as well as the risks and benefits of both approaches.

OncLive®: Starting with the adjuvant space, could you discuss the data from the COMBI-AD trial that were presented during the 2020 ASCO Virtual Scientific Program?

Atkins: Initial [findings] from the COMBI-AD study showed that patients with stage IIIA to IIIC BRAF-mutant melanoma resected to no evidence of disease (NED) had a significant benefit in RFS with the combination of dabrafenib and trametinib relative to placebo. This has been previously presented and published, but at the [2020 ASCO Virtual Scientific Program], the 5-year update was presented, [showing] RFS curves out to 60 months. There continued to be a 16% difference in RFS [between arms], favoring the dabrafenib and trametinib arm.

In addition, the forest plot showed a significant difference [between arms], favoring the [investigational] treatment [across all patient subgroups]. The difference between the RFS curves were flattening out at the 5-year time point, suggesting that the difference was substantial and likely to be sustained. [Patients] in that difference were likely those who were cured of what otherwise would have been relapsed melanoma.

We’ve also seen RFS data with pembrolizumab (Keytruda) from the KEYNOTE-054 trial. How did this trial compare with COMBI-AD?

The KEYNOTE-054 trial had a similar population in relation to the COMBI-AD trial. If we look at the 2 studies, at the 36-month time point, we see pretty similar results with about a 20% difference in terms of RFS, favoring the treatment arm versus the placebo arm. Similar to the COMBI-AD trial, the KEYNOTE-054 trial [demonstrated] benefit in all the different subsets in the forest plot. So, those studies have shown comparable results to date.

If we look further at the hazard ratios for benefit in the subsets of patients with stage IIIA, IIIB, and IIIC melanoma, they appear similar. The KEYNOTE-054 trial has slightly better hazard ratios, but that could potentially be explained by a better outcome on the placebo arm. The placebo arm in the KEYNOTE-054 trial included a large percentage of patients who were BRAF wild-type. Those patients have a better prognosis in terms of RFS than the BRAF-mutant patients on the placebo arm of the COMBI-AD study.

How do the findings from the KEYNOTE-054 trial compare with those of the CheckMate-238 trial?

The CheckMate-238 study had a different design [than the KEYNOTE-054 study]. It included high-risk patients who had stage IIIB, IIIC, or stage IV disease who had been resected to NED. Patients received ipilimumab rather than placebo as the control. Ipilimumab had been previously studied in a trial compared with placebo, and it showed better RFS with a hazard ratio of 0.7.

In order to see how nivolumab would have done relative to placebo, [one should look at] the hazard ratio of the CheckMate-238 study, which was 0.68, and multiply it by the 0.7 [hazard ratio] from the adjuvant ipilimumab study to come up with a hazard ratio of 0.5 versus placebo. This would be similar to what we have seen in the other 2 trials I mentioned.

As with the other studies, nivolumab was essentially better than ipilimumab across all subsets on the forest plot. However, if we look at the RFS curves comparing nivolumab with ipilimumab in the CheckMate-238 and COMBI-AD studies, we see that for the BRAF-mutant population, the COMBI-AD study showed a lower relapse rate in the first 12 months than the immunotherapy studies. However, when we look out 4 or 5 years, there is slightly better RFS on the nivolumab arm than the dabrafenib and trametinib arm. This suggests that there may be a more durable RFS benefit with immunotherapy, particularly with nivolumab than dabrafenib and trametinib; however, the curves run pretty close and it is hazardous to compare [trials] head to head.

How are you applying these data to clinical practice?

For the BRAF wild-type patients, the choice becomes adjuvant nivolumab versus adjuvant pembrolizumab. There isn’t really any difference we can see [with regard to] efficacy or toxicity between those therapies. The decision is then based on other factors such as [physician] comfort in using a particular agent, whether one’s pharmacy has that agent, the price, and maybe marketing.

Increasingly, scheduling is influencing decision making and the 2 regimens are playing leapfrog with each other. It used to be that nivolumab was given every 2 weeks and pembrolizumab was given every 3 weeks. Therefore, people favored pembrolizumab because of the longer treatment interval. However, nivolumab switched to an every 4-week schedule, and it was preferred. Then, recently, pembrolizumab switched to an every 6-week schedule, and it [again] became the preferred approach.

During the webinar, we asked the question, “What would physicians do for a patient with stage III disease, regardless of BRAF status?” Everybody selected pembrolizumab, possibly based on that scheduling difference.

With regard to adjuvant targeted therapy versus adjuvant checkpoint inhibitor therapy, I think it is a discussion that needs to be had with patients. These agents have different toxicity profiles. Targeted therapy has more low-grade toxicities, but they tend to be more persistent during the 1-year course of the therapy. Targeted therapy has a more reliable antitumor effect, but immune therapy would likely produce a more reliable RFS benefit. The toxicities [of checkpoint inhibitors], although intermittent, can be severe. Unlike targeted therapy toxicities, which are always reversible, toxicities [associated with checkpoint inhibitors can be] permanent. Though, they rarely are.

Most of us are more comfortable giving adjuvant immunotherapy because it is easier for the patient. However, there are patients who do not like intravenous drugs, have autoimmune conditions, or live far away [from the infusion center]. In these cases, targeted therapy makes more sense.

Shifting to the neoadjuvant space, what was exciting about the results of the OpACIN-neo trial, which were presented during the 2020 ASCO Virtual Scientific Program?

What has been exciting with the OpACIN-neo trial, where patients got nivolumab and ipilimumab in different schedules prior to surgery at week 6, was that about 60% of patients had pCRs or near pCRs despite the fact that around 40% of patients had radiographic partial response (PR) prior to surgery. Therefore, we learned that the actual antitumor effect is greater than the radiologic images suggest.

Even more important was that in the patients who had a pCR or a near CR, 1 had a subsequent disease recurrence. So, pCR or near pCR may be tantamount to cure.

We also learned that patients may not need subsequent therapy after surgical resection if they have had a pCR or near pCR.

Finally, we saw that there was a lot of toxicity in the neoadjuvant setting, but the nivolumab-3 (3 mg/kg) and ipilimumab-1 (1 mg/kg) regimen had the best toxicity profile without costing any efficacy. So, that became the preferred regimen.

What are some of the key takeaways from the PRADO trial?

[The combination of ipilimumab and nivolumab] was tested in the PRADO trial, which evaluated whether we could use a marker node to assess what was going on in the lymph nodes in the whole basin. Potentially, this could not only avoid subsequent adjuvant treatment if patients had a good response in the marker node but avoid completion lymph node dissection or therapeutic lymph node dissection in those patients.

[Investigators] used the nivolumab-3, ipilimumab-1 regimen. Before treatment began, [patients underwent] a biopsy of a palpable lymph node. A marker was inserted into that lymph node so that it could be identified radiographically and at the time of surgery. Then, they gave 2 doses of nivolumab and ipilimumab, and patients underwent surgery at 6 weeks.

A similar pathologic response rate of 71%, as well as a pCR or near pCR rate of about 60%, was observed. The patients with a pCR or near pCR only had the marker node resected, whereas the other nodes were left in place. Patients with a pathologic partial response had a full node dissection. In the other patients who didn’t have a pathologic response, [investigators] did a full node dissection and gave additional adjuvant therapy with either more nivolumab or BRAF and MEK inhibitors if the patient had a BRAF mutation.

At the 2020 ASCO Virtual Scientific Program, investigators reported that [patients’] quality of life was better if the marker lymph node had been removed [versus] if they had undergone completion lymph node resection, which was not surprising. We have yet to see the efficacy data [to determine] how long the benefits will last and whether the patients who had 1 lymph node removed will have the same outcome as the patients in the OpACIN-neo trial who had a therapeutic lymph node dissection despite achieving a pCR. We anticipate that those results will be reported at a meeting in the near future.

What insight can you provide regarding the risks and benefits of neoadjuvant versus adjuvant treatment?

For neoadjuvant treatment, patients may get tumor shrinkage, which might lead to decreased surgical morbidity. As was suggested in the PRADO study, could also decrease the need for surgery. It is conceivable, with a neoadjuvant approach, that if we see a radiologic response that is equivalent to what may be a pCR or near pCR, we may be able to avoid removing the marker node in the future. Patients may also get potentially better destruction of micrometastases [in the neoadjuvant setting] than in the adjuvant setting because we are activating more immune cells in the bulky lymph nodes. [Those cells] have a chance to travel and seek out the micrometastases. Therefore, patients may have superior prevention of distant relapse.

That is what is being tested in the SWOG-S1801 trial (NCT03698019), which is examining neoadjuvant treatment with pembrolizumab versus adjuvant treatment with pembrolizumab and evaluating RFS and overall survival.

From a translational research standpoint, [neoadjuvant therapy] also gives us the opportunity to collect specimens so we can understand mechanisms of response and resistance to various therapies. Then, we might understand how to best treat patients in the stage IV setting. Finally, it is a great tool for potentially examining the effectiveness of various combination treatments where we can get information using pCR or near pCR as a surrogate marker of efficacy. That might tell us whether a particular combination offers advantages over single-agent anti–PD-1 inhibitors, for example, without having to go through a large phase 3 trial in the metastatic setting.


  1. Hauschild A, Dummer R, Santinami M, et al. Long-term benefit of adjuvant dabrafenib + trametinib (D+T) in patients (pts) with resected stage III BRAF V600-mutant melanoma: five-year analysis of COMBI-AD. J Clin Oncol. 2020;38(suppl 15):10001. doi:10.1200/JCO.2020.38.15_suppl.10001
  2. Rozeman EA, Reijers I, Hoefsmit EP, et al. Twenty-four months RFS and updated toxicity data from OpACIN-neo: a study to identify the optimal dosing schedule of neoadjuvant ipilimumab (IPI) and nivolumab (NIVO) in stage III melanoma. J Clin Oncol. 2020;38(suppl 15):10015. doi:10.1200/JCO.2020.38.15_suppl.10015