Investigators Tackle Resistance, Poor Response in Metastatic Melanoma

Mario Sznol, MD

Immune checkpoint inhibitors (ICIs) combined with the anti—CTLA-4 agent ipilimumab (Yervoy) have dramatically improved survival in metastatic melanoma, but resistance and lack of response remain obstacles to wider efficacy, Mario Sznol, MD, said in a presentation during the 4th Annual International Congress on Immunotherapies in Cancer^®. Multiple efforts are under way to understand these issues better and develop improved biomarkers for response, Sznol added.

The ICIs nivolumab (Opdivo) and pembrolizumab (Keytruda) individually are frontline standards in metastatic melanoma and yield long-term survival rates in the range of 40% to 45%, said Sznol, codirector of SPORE in Skin Cancer at Yale Cancer Center in New Haven, Connecticut. Although these immunotherapies are typically administered alone in the first line, they can also be given with ipilimumab, which can extend 5-year survival by an additional 7% to 10%, Sznol said.¹

Data from the phase III CheckMate 067 study (NCT01844505) demonstrated that sustained survival benefit is possible by combining PD-1 inhibition with ipilimumab, Sznol said. The trial enrolled 1296 patients with metastatic treatment-naïve stage III or IV BRAF V600—mutant melanoma who were randomized 1:1:1 to nivolumab monotherapy, single-agent ipilimumab, or combination nivolumab and ipilimumab.²

Findings showed that both the doublet therapy and nivolumab monotherapy resulted in longer progression-free survival (PFS) and overall survival (OS) than single-agent ipilimumab. The median PFS was 11.5 months (95% CI, 8.7-19.3) in the nivolumab/ipilimumab arm and 6.9 months (95% CI, 5.1-10.2) and 2.9 months (95% CI, 2.8-3.2) in the nivolumab and ipilimumab groups, respectively.²

The median OS exceeded 60.0 months (median not reached [NR]; 95% CI, 38.2-NR) among those treated with the combination. In the nivolumab arm, the median OS was 36.9 months (95% CI, 28.2-58.7); in the ipilimumab arm, 19.9 months (95% CI, 16.8- 24.6). The 5-year OS rates were 52%, 44%, and 26% for the doublet therapy, nivolumab monotherapy, and single-agent ipilimumab, respectively.²

For patients with metastatic melanoma who receive nivolumab with ipilimumab or pembrolizumab, the chance of being alive at 5 years is about 50%, Sznol said. “That’s amazing considering that when we started all of this, the chance of being alive at 5 years was 5% with interventions like chemotherapy,” he said.

Pembrolizumab parallels nivolumab’s efficacy in patients with advanced disease, Sznol said. Data from a 5-year follow-up from the phase III KEYNOTE-006 study (NCT01866319) of front-line pembrolizumab versus ipilimumab in 834 patients with advanced melanoma demonstrated that pembrolizumab offered a better survival advantage than the anti—CTLA-4 agent. The 5-year OS was 38.7% with pembrolizumab versus 31.0% with ipilimumab.³

These results show that pembrolizumab elicits an outcome similar to nivolumab, Sznol said. “Really, they’re equivalent drugs in the front-line setting,” he added (Table).^2,3

ICIs can offer promise to patients with advanced disease—if they respond to the anti—PD-1 therapy. There are many reasons a patient might not respond to immunotherapy, Sznol said, such as insufficient T-cell activity at the tumor site. “Why are there no T cells there? It’s probably because they never primed; they never generated an immune response, and the T cells are [consequently] being excluded from the tumor,” he said.

T cells’ failure to prime can be attributed to a low mutation burden that causes no or few antigens to be present. A host factor such as a genetic inability to respond to the antigens can also impair T-cell priming, Sznol said.

Some patients will have T cells present and still not react to ICIs. “There are a lot of things that might keep those T cells, which are the right T cells, from actually working [even if] they’re there,” Sznol said.

Like host factors, tumor factors can obstruct T-cell activity. For example, tumors that express VEGF are known to prevent T cells from entering the tumor microenvironment, Sznol said. Hostile, hypoxic microenvironments can also keep T cells from working, and some tumor cells carry signaling defects that enable them to evade T-cell recognition.

The Search for Biomarkers

Biomarkers predictive of patient response are necessary to select patients for anti— PD-1 therapy and identify alternative approaches for those who do not respond to immune checkpoint blockade, Sznol said. Ongoing efforts to identify genetic indicators of immunotherapeutic efficacy include a mass spectrometry analysis that Sznol coauthored. Seeking to define baseline protein signatures associated with outcome in metastatic melanoma, Sznol and fellow authors evaluated serum from a set of 119 patients with stage IV melanoma prior to the initiation of their ICI-based therapy.

Investigators developed a pretreatment serum test that used mass spectrometry to identify a group of serum proteins that distinguish the patients with metastatic disease who experienced prolonged survival following treatment with an anti— PD-1 regimen. The test, which measures the ubiquity of 209 circulating proteins or protein fragments, classifies samples as either “sensitive” or “resistant” to immune checkpoint blockade.⁴

Authors said the test has the potential to identify patients with “sensitive” versus “resistant” serum designations who would have better OS with PD-1 blockade alone or with a doublet therapy consisting of ipilimumab and nivolumab, possibly streamlining treatment for these patients.⁴

Moreover, in their analysis, Weber et al also found that the serum of patients who are resistant to ICIs carries acute phase, complement, and wound-healing molecules. The serum test might aid development of treatments that overcome primary anti— PD-1 resistance, authors said. Such therapies might involve inhibition of complement activation and the suppression of components of wound healing, they added.⁴

Biomarkers could guide the development of therapies that would enable a patient to overcome primary resistance to immune checkpoint blockade, according to Sznol. “We’re trying to improve patient outcomes [and] responses, but we don’t actually know which of these patients need the T-VEC [talimogene laherparepvec; Imlygic], the tumor-infiltrating lymphocytes, the anti-CD40, and all of these other agents that are being developed,” he said. “Without biomarkers, it’s hard to rationally develop therapies.”

Driving Future Development

“There is a huge number of clinical trials that are currently exploring [just about] anything that you can think of, from agents that create T cells in the microenvironment to agents that increase their function,” Sznol said. “Many agents are being combined with anti—PD-1 drugs in phase II trials, and the danger is [that] despite all of this work, we don’t have good biomarkers to select the patients who would respond to these individual agents.”

Anti-CD40, FLT3, TLR agonists, costimulatory agonists, cytokines, modified cytokines, and stimulator of interferon genes (or STING) agonists are among the modalities being explored in combination immunotherapy trials, Sznol said.

Table. Nivolumab and Pembrolizumab: Efficacy of First-line Options^2,3 (Click to Enlarge)

Sznol cited the negative phase II/ III ECHO-301/KEYNOTE-252 trial (NCT02752074) as an example. The study was initiated after preliminary data from the open-label phase I/II ECHO-202/ KEYNOTE-037 trial (NCT02178722) of combination pembrolizumab and epacadostat showed that the doublet therapy induced objective responses in 55% (30 of 54) of evaluable patients with advanced melanoma. This included 8 complete responses (CRs) and 22 partial responses (PRs).⁵

Early-stage data presented at the 2017 European Society for Medical Oncology Congress suggested that the combination could be an efficacious newcomer to the melanoma treatment armamentarium: Findings indicated that 25 of 45 (56%) of treatment-naïve patients responded to pembrolizumab plus epacadostat. There were 6 CRs and 19 PRs in this untreated subset, investigators said.⁵

“Response rates were close to 60% in the front-line setting, which is higher than you would expect with PD-1 alone,” Sznol said.

The doublet therapy also appeared to confer a valuable survival benefit. The median PFS was 12.4 months (90% CI, 6.2-23.8) for all per-protocol patients. At 6, 12, and 18 months, the PFS was 70%, 54%, and 50%, respectively. In the treatment- naïve group, the median PFS was 68% at 6 months and 52% at both 12 and 18 months.⁵

These encouraging results prompted further study of the combination in KEYNOTE-252/ECHO-301, which enrolled 706 patients with histologically confirmed unresectable stage III or IV melanoma who had not received prior treatment for advanced or metastatic disease. Patients with BRAF V600—mutant melanoma were also permitted to participate and allowed to have prior BRAF- or MEK-targeted therapy. The trial population was randomized to receive epacadostat at 100 mg twice daily combined with pembrolizumab at 200 mg every 3 weeks (n = 354) or the same dose of single-agent pembrolizumab (n = 352).⁶

After a median follow-up of 12.4 months, the median PFS was 4.7 months with epacadostat and pembrolizumab versus 4.9 months with pembrolizumab monotherapy (P = .517). At 12 months, the PFS rate was 36.9% in the combination arm and 36.6% with single-agent pembrolizumab. The objective response rate with the combination was 34.0% versus 32.0% with pembrolizumab.⁶

Without a biomarker, investigators cannot select patients who will respond to the experimental therapy. This is one of “the dangers of development,” Sznol said. “If the therapy only affects a small population of patients, when you go to phase III trials, you are going to see results like these with some of the combinations that are moving forward.”

However, this complication is not unique to melanoma. “The challenge in the future is to identify which patients are going to respond to which [intervention], and I think that’s going to be true not only in melanoma but in almost every other disease [for which] we study these immunotherapy combinations,” Sznol said.

Biomarkers, he added, could aid selection for clinical trials and help investigators ensure that they are testing the experimental intervention in the appropriate patient population. “There are 50 different mechanisms of response and many different reasons why patients do not respond,” Sznol said. “If you do a clinical trial without selection and you see only a 5% to 10% response rate, maybe only a few of those patients had the right mechanism of response for the combination [being tested].”