Neoadjuvant Immunotherapy for Melanoma

Abstract

Clinical trials have demonstrated the efficacy of immunotherapy, especially checkpoint blockade inhibitors, in the treatment of patients with metastatic melanoma. More recently, improvements in survival have been reported in patients with high risk resectable melanoma when these agents are used in the adjuvant setting. Increasing interest in neoadjuvant immunotherapy for high risk resectable melanoma has been fueled by early reports of significant efficacy. We review the rationale and data behind utilizing neoadjuvant immunotherapy.

Introduction

The advent of effective systemic therapy in the form of checkpoint blockade immunotherapy (CBI) for patients with melanoma has revolutionized modern melanoma care. Traditional therapeutics offered little to patients with recurrent disease after surgical management, with surgical resection and chemotherapy as curative approaches in only a small percentage of patients. The only approved adjuvant therapy, high dose interferon-α, was toxic and of borderline efficacy [1].

The discovery of anti-CTLA4-Ig and the subsequent approval of ipilimumab for systemic therapy in melanoma patients ushered in this new era [2]. Progress has been rapid, with newer agents demonstrating improved efficacy and less toxicity. Anti-PD1 agents have rapidly become standard approaches in the adjuvant setting and in patients with unresectable recurrent disease [3]. Combination approaches using both anti-CTLA-4 and anti-PD1 strategies have demonstrated increased efficacy in refractory patients with stage IV disease, although at a substantial cost of increased immune toxicity [4]. Much progress has been made in understanding some of the immune predictors of response and outcome, and this will continue as more data become available from clinical trials.

This foundation has led to perhaps the most exciting recent trend in melanoma therapeutics - the increasing promise of neoadjuvant therapy using CBI. Our aim in this review is to present the most current information available to surgeons who should carefully consider this approach to patients with high-risk resectable melanoma.

Rationale for neoadjuvant therapy in melanoma

Several immunotherapeutic agents, particularly immune checkpoint inhibitors, have demonstrated efficacy in the setting of unresectable or metastatic melanoma [2, 4, 5], as well as in the adjuvant setting for high-risk stage III patients [6–8]. These agents, used as monotherapy or in combination, have produced durable improvements in survival and even potential cures in patients with advanced stage III and stage IV disease. A logical next step is to consider the use of these agents in the neoadjuvant setting for patients with advanced but resectable disease. There is rapidly accumulating evidence that neoadjuvant therapy can be safely utilized in patients with melanoma and that it may improve outcome when compared to resection alone or resection followed by adjuvant therapy.

There are several important potential advantages to a neoadjuvant approach. These include 1) an opportunity to determine the efficacy of an immune therapy while disease is in situ, 2) the ability to study immune parameters pre- and post-treatment that may predict treatment response, 3) the potential to eradicate clinically occult disease at an earlier time point, and 4) the potential to decrease surgical morbidity by shrinking the tumor prior to resection. The concept of increased efficacy in a neoadjuvant context is based on the theory that having an abundance of tumor antigen present at the time of treatment will facilitate an anti-tumor T-cell response and expansion of effector cells. The precise mechanism for this, and indeed the validity of this, remains unclear and will depend on ongoing and future studies. Nonetheless, there is an increasing body of published experience to support the immune efficacy of a neoadjuvant approach.

There are also some potential disadvantages to a neoadjuvant approach. These include 1) the risk of disease progression despite treatment that may result in increased surgical morbidity or progression to unresectable disease and 2) the risk of immune-related adverse events (irAEs) which can delay surgical resection and are often permanent. In addition, there is still uncertainty about whether neoadjuvant immunotherapy is as efficacious as resection followed by adjuvant treatment.

Early Neoadjuvant Immunotherapy with Interferon alfa-2b

The first modern neoadjuvant immunotherapeutic approach was a trial of high dose interferon α−2b (HDIFN) in patients with clinical stage III melanoma (Table 1) [9]. This phase II trial enrolled 20 immunotherapy-naïve patients from 2001–2005. After a surgical lymph node biopsy confirmed nodal metastases, patients were given four weeks of neoadjuvant HDIFN (20 MU/m2/d IV 5 days/week × 4 weeks), a regimen identical to that used in the adjuvant therapy trials of HDIFN [1]. Patients were then assessed for clinical response and underwent therapeutic lymphadenectomy followed by subcutaneous maintenance HDIFN (10 MU/m2/d three times/week) for the remaining year. An objective clinical response rate (cRR) of 55% was observed and a pathologic complete response (pCR) was seen in 15% of patients. Grade 3/4 irAEs were noted in 25% of patients. Responders were noted to have increased tumor T-cell infiltrates and a distinct endotumoral T cell repertoire, characterized by the presence/absence of distinct T cell subsets [9]. This was the first trial to support a potential benefit to neoadjuvant therapy using an immune activating agent and to identify evidence of immune modulation on a cellular level.

Table 1.

Clinical Trials of Single Agent Neoadjuvant Immune Therapy

Therapeutic Agent	Patients	Regime	N	Outcomes	Grade ¾ irAEs
IFNα-2b Moschos et al., 2018	stage IIIB/C	IV HDIFN 20 MU/m2 5x/week × 4 weeks → surgery → subcutaneous HDI 10MU/m2 3x/week × 48 weeks	20	cRR – 55% pCR – 15%	25%
Pembrolizumab Huang et al. 2019	Resectable stage IIIB/C/or IV	Pembro 200mg × 1 dose → surgery → pembro × 1 year	29	pCR/near pCR: 30%	7 events

cRR – clinical response rate, pCR – pathologic complete response, HDIFN – high-dose interferon alfa-2b, irAE – immune-related adverse events, pembro - pembrolizumab

Neoadjuvant Immune Checkpoint Blockade: Early Reports

Song and colleagues at the University of Pennsylvania reported their experience of clinical stage III patients undergoing adjuvant CBI with or without neoadjuvant CBI. They reported an experience with 59 patients, of whom 18 received adjuvant therapy only and 41 received neoadjuvant and adjuvant therapy. They reported improved distant disease-free survival (DDFS) in the neoadjuvant followed by adjuvant CBI group (HR 0.38, p=0.028) but no statistically significant difference in disease-free survival (DFS) (HR 0.56, p=0.17), although the trend favored the neoadjuvant group. In the 39 patients who underwent neoadjuvant therapy and were evaluable for pathologic response, 13% had a pCR and 59% had a partial pathologic response [10]. Notably, patients who had either partial or complete pathologic response experienced improved 3-year disease-free survival, loco-regional recurrence-free survival, and distant recurrence-free survival compared to those with no response. This study, albeit retrospective, provided some evidence of an improvement in outcome in patients treated in the neoadjuvant setting compared to standard surgical resection followed by adjuvant therapy.

The first CBI to receive FDA approval for metastatic or locally advanced/unresectable melanoma was ipilimumab, approved in 2011. Ipilimumab is a recombinant human monoclonal antibody that binds to cytotoxic T-lymphocyte-associated antigen 4 (CTLA4) and blocks the interaction of CTLA-4 with its ligands, CD80 and CD86. Although overall response rates in patients with stage IV melanoma were only 11.1% [11], overall survival was improved compared to patients randomized to GP100 vaccine alone (10 months vs. 6.4 months, p<0.001) [2]. Toxicity was common and sometimes irreversible, with autoimmune colitis the most common immune related adverse event (irAE). The real promise of this agent was the observation that responses were durable, and complete responders were potentially cured [12].

Building on the studies of ipilimumab in the unresectable/metastatic setting as well as HDIFN in the neoadjuvant setting, Tarhini and colleagues evaluated ipilimumab in combination with HDIFN for patients with locally or regionally advanced melanoma [13]. Two cycles of Ipilimumab were given prior to resection at 3 or 10 mg/kg, 3 weeks apart. After resection, patients were given an additional two cycles of Ipilimumab 3 weeks apart, then every 12 weeks for four cycles. HDI was given concurrently with pre and post-op Ipilimumab. This was a small trial (n= 30, 28 evaluable), but they reported a pre-operative radiologic response rate of 36% and pCR of 32% [13]. They observed that higher T-cell clonality in the primary tumor and/or loco-regional disease post-treatment was associated with an increase in recurrence-free survival (RFS). Combination therapy was well tolerated with no delays in planned surgery, but the 10mg/kg dose of ipilimumab was associated with more irAEs.

The results of these initial studies of neoadjuvant immune therapy were compelling. Subsequent trials were pursued in order to identify more efficacious and better-tolerated CBIs and dosing schedules.

Immunotherapy using anti PD-1 strategies

A second class of CBI was identified that was more efficacious and less toxic than ipilimumab. These target the programmed cell death 1 (PD-1) receptor or programmed cell death ligand 1 (PD-L1). PD-1 is present on T cells and PD-L1 is found on tumor cells and responding immune cells. Binding of PD-1 to its ligand causes down-regulation of the T cell response. Anti-PD-1 and anti-PD-L1 antibodies work by blocking this interaction, effectively boosting immune response to tumor cells. These CBI agents include nivolumab (anti-PD-1), pembrolizumab (anti PD-1), and atezolizumab (anti-PD1-L). Anti PD-1 and anti PD-L1 antibodies have shown improved efficacy in stage IV disease compared to ipilimumab and have largely replaced ipilimumab as initial monotherapy in patients with recurrent melanoma [5, 14]. These agents are associated with fewer irAEs and are more efficacious than monotherapy with ipilimumab or dacarbazine[3].

Efficacy of neoadjuvant therapy with single agent checkpoint blockade

Interferon-a2b and pembrolizumab have been used in the neoadjuvant setting as single agents (Table 1). Pathologic CRs are 15–30%. In addition to the data on single agent HDIFN from Moschos and colleagues presented above, Huang and colleagues used Pembrolizumab at 200 mg as a single pre-operative dose in 27 patients with resectable stage III and IV melanoma. They observed a “major” or complete pathologic response in 8/27 patients (30%) [15]. These data are based on very small numbers of patients, but suggest that the benefit of single agent neoadjuvant therapy with CBI may be limited to less than half of patients treated.

Efficacy of Neoadjuvant Trials using combination checkpoint blockade

Combination CBI using an anti-PD1 and anti-CTLA4 strategy improves outcome with increased toxicity in stage IV disease[16] [4]. Interestingly, there is recent data suggesting that upregulation of CTLA-4-mediated suppression after neoadjuvant pembrolizumab may identify patients in whom use of both anti-PD1 and ipilimumab may be more effective [15]. Including the neoadjuvant trial by Tarhini and colleagues using combination HDIFN and ipilimumab, there have been four trials using combination CBI as a neoadjuvant approach (Table 2).

Table 2.

Clinical Trials of Neoadjuvant combination Therapy

Therapeutic agent/authors	Included Patients	Regime	N	Outcomes	Grade ¾ irAEs
Ipilimumab, interferon alfa-2b Tarhini et al. 2018	Locally or regionally advanced	Arm A: ipi 3mg/kg q3 weeks × 2 doses + HDIFN (20 MU/m2/day) → surgery → ipi 3mg/kg q3 weeks × 2 doses then q12 weeks × 4 doses + HDIFN	15	cRR – 36% pCR – 32%	8 events
		Arm B: ipi 10mg/kg q3 weeks × 2 doses + HDIFN (20 MU/m2/day) → surgery → ipi 10mg/kg q3 weeks × 2 doses then q12 weeks × 4 doses + HDIFN	15		17 events
Ipilimumab, nivolumab Amaria et al. 2018	Resectable clinical stage III or oligometastatic stage IV	Arm A: nivo 3mg/kg q2 weeks × 4 doses → surgery → nivo 3mg/kg q2 weeks × 13 doses	12	Arm A:  ORR 25%  pCR 25%	8%
		Arm B: ipi 3mg/kg + nivo 1 mg/kg q3 weeks × 3 doses → surgery → nivo 3mg/kg up to 13 doses	11	Arm B:  ORR 73%  pCR 45%	73%
Ipilimumab, nivolumab OpACIN - Blank et al. 2018	Resectable clinical stage III	Arm A: surgery → ipi 3mg/kg + nivo 1mg/kg q3 weeks × 4 doses	10	NR	90%
		Arm B: ipi 3 mg/kg + nivo 1mg/kg q3 weeks × 2 doses → surgery → ipi 3 mg/kg + nivo 1mg/kg q3 weeks × 2 doses	10	cRR 78%  pCR 30%	90%
Ipilimumab, nivolumab OpACIN-neo - Rozeman et al. 2019	Resectable clinical stage III	Arm A: ipi 3mg/kg + concurrent nivo 1 mg/kg q3 weeks × 2 doses → surgery	30	cRR 63%  pCR 80%	40%
		Arm B: ipi 1mg/kg + concurrent nivo 3mg/kg q3 weeks × 2 doses → surgery	30	cRR 57%  pCR 77%	20%
		Arm C: ipi 3mg/kg × 2 doses → nivo 3mg/kg q3 weeks × 2 doses → surgery	26	cRR 35%  pCR 65%	50%

cRR – clinical response rate, pCR – pathology complete response, ORR – overall response rate, HDIFN – high-dose interferon alfa-2b, irAE – immune-related adverse events, NR – not reported, ipi – ipilimumab, nivo – nivolumab, pembro – pembrolizumab

Amaria and colleagues conducted a randomized phase II trial evaluating neoadjuvant ipilimumab/nivolumab vs. nivolumab monotherapy in 23 patients with resectable clinical stage III or oligometastatic stage IV melanoma [17]. Endpoints included pCR, overall response rate (ORR), toxicity, and immune correlates. They reported higher response rates to combination therapy compared to single-agent nivolumab (ORR 73% vs. 25%, pCR 45% vs. 25%), but higher toxicity (73% vs. 8% grade 3 irAEs). The planned accrual for this trial was 40 patients, but due to progression of disease in two patients in the monotherapy arm that precluded surgical resection, the trial was stopped early.

In the OpACIN trial, Blank and colleagues evaluated ipilimumab and nivolumab in patients with stage IIIB melanoma in an attempt to determine the immune efficacy, safety, and feasibility of giving combination CBI as a neoadjuvant plus adjuvant treatment compared to adjuvant only. CBI naïve patients with palpable stage III disease were eligible. Ipilimumab 3 mg/kg and nivolumab 1 mg/kg were given in a split schedule (6 weeks pre-op and 6 weeks post-op) and were compared to the same regime given post-op only for 12 weeks. The primary outcome of the trial was determination of neo-antigen T cell response, safety and feasibility. There were 10 patients in each arm, and the pathologic response rate observed in the neoadjuvant arm was high (78%) with 30% of patients experiencing a pCR in the neoadjuvant group. Unfortunately, the toxicity of the regimen was also high with 90% of patients experiencing irAEs ≥ 3 [18].

A follow up trial (OpACIN-neo) was conducted with the goal of exploring different dose combinations of neoadjuvant CBI to decrease toxicity and optimize efficacy. Ipilimumab in combination with nivolumab were given as neoadjuvant therapy in three different dosage/timing combinations (Table 2). A recent update of this trial was presented at ASCO, 2020 [19]. The investigators reported on 86 patients at a median follow up of 24.6 months. Again, a high response rate to neoadjuvant combination CBI was reported (77%) with the observation that while most patients in the trial had ongoing irAEs (68%), only 20% of patients had irAEs ≥ 3 [19]. Importantly, at a median follow up of 24.6 months, median relapse-free and event-free survival (RFS/EFS) were not reached, suggesting that the response to neoadjuvant combination CBI was durable. In this trial, which evaluated 3 different dose combinations, ipilimumab 1 mg/kg and nivolumab 3 mg/kg appeared to be the optimal dose strategy [19]. The estimated RFS at 2 years was 84%. The investigators did report that patients who had a pathologic response (as determined by surgical resection of the nodal basin) experienced a favorable RFS of 97% at a median of 2 years, compared to only 36% in patients without a significant pathologic response. They observed that median survival was not reached for any of the three dose arms. There was no statistically significant difference in response rates between the three dosage arms or incidence of toxicity. These data support a rationale for two cycles of ipilimumab 1mg/kg and nivolumab 3 mg/kg prior to resection of metastatic melanoma to clinically enlarged lymph nodes. Pathologic response rates are high and durable, with reasonable toxicity.

These early reports provide compelling evidence that combination CBI as a neoadjuvant approach in high-risk patients should be strongly considered.

Neoadjuvant versus adjuvant combination immunotherapy: immune efficacy

The final concern about the neoadjuvant approach is the lack of data available to support its use in lieu of surgical resection followed by adjuvant single agent or combination CBI. The initial study that addressed this question was the original OpACIN trial, reported by Blank and colleagues. As previously discussed, this was a phase 1b randomized trial of ipilimumab (3 mg/kg) and nivolumab (1 mg/kg) as neoadjuvant/adjuvant therapy (two courses before lymphadenectomy and two courses after) vs adjuvant only (four courses after lymphadenectomy)[18]. The trial accrued 20 CBI-naïve patients, with 10 patients in each arm. The primary endpoint of the trial was the ability of the treatment arm (neoadjuvant vs. adjuvant) to generate specific anti-tumor T cell responses. The investigators determined that in this small set of patients, neoadjuvant therapy was superior to adjuvant therapy in generating tumor-specific T cell responses. Although this was a small study, the results are compelling and suggest that neoadjuvant therapy using combination CBI may be more efficacious in terms of clinical outcome than surgical resection and adjuvant therapy.

Immunologic Observations and Predictors of Response in Patients Undergoing Neoadjuvant Immunotherapy

One of the potential advantages of neoadjuvant immunotherapy is the ability to assess immune parameters before and after surgical resection. The hope is that identification of pre-operative predictors of response to neoadjuvant CBI will allow for a more tailored treatment strategy. These observations, which vary significantly among the reported experience, demonstrate trends that are not surprising. Responders to neoadjuvant immune therapy have been reported to have higher baseline CD8+ tumor infiltrates, higher tumor cell PD-L1 expression, higher tumor mutational burden, more cell proliferation of CD45+ infiltrating lymphocytes [17], and more T cell clonality in the tumor infiltrates [18] compared to non-responders.

PD-L1 expression in the melanoma is a logical predictive biomarker in patients undergoing neoadjuvant anti-PD1 therapy. In a prospective randomized trial of chemotherapy vs. nivolumab monotherapy, Weber and colleagues reported a much higher response to therapy in patients with PD-L1 expression (n=77, ORR 34.5% (12.2–49.2)) compared to those who were PD-L1 negative (n=87, ORR 7.3% (−13.4–21.5))[3]. Despite this, research has demonstrated that while there is an association of higher response rates with higher PD1-L expression, patients without expression can also achieve dramatic responses to CBI [20]. Others have observed that patients with PD-L1 negative tumors are more likely to respond to combination CBI than monotherapy [21]. Another potential predictive biomarker of response to neoadjuvant therapy is tumor T-cell infiltration. In the neoadjuvant interferon alfa-2b trial reported by Moschos and colleagues, patients who responded to neoadjuvant therapy were more likely to have increased tumor cell infiltrates in pretreatment biopsies [9].

Melanoma has a high tumor mutational burden, which is generally associated with the production of a higher level of neoantigens available for an immune response than tumors with a lower mutational burden. This can be quantitated for individual tumors and may provide an additional marker of likelihood of response to neoadjuvant therapy. Indeed, Amaria and colleagues reported that in a randomized trial of CBI (nivolumab vs. ipilimumab/nivolumab) in high-risk resectable melanoma, there was a trend toward higher response in tumors with a higher mutational burden [17]. They also reported that higher CD8(+) T cell infiltrates and higher expression of a panel of lymphoid markers were associated with response to neoadjuvant CBI.

Huang and colleagues recently reported a “neoadjuvant response signature” identified in 8/27 patients treated with a single dose of anti-PD-1 therapy [15]. In this trial, patients with resectable stage III or IV melanoma received a single dose of pembrolizumab (200 mg) after a pre-treatment biopsy. Three weeks later they underwent resection, followed by adjuvant therapy. They found that patients whose tumors exhibited this “signature” had an excellent pathologic response and survival outcome. The investigators characterized this signature as evidence of “exhausted” CD8 T cells in the tumor pre-treatment. Patients who responded to neoadjuvant monotherapy with pembrolizumab developed evidence of “immune reinvigoration” within three weeks. The investigators suggest that evidence of this “exhausted” T cell phenotype prior to one dose of CBI may help identify patients most likely to benefit from therapy. In addition, they observed that proliferation of regulatory T cells in response to CBI was associated with recurrence, suggesting that when observed, this may prompt treatment with combination CBI with anti-CTLA4-Ig in addition to an anti-PD-1/PD-1L agent [15].

Reassessing the Need for Radical Lymphadenectomy after Neoadjuvant CBI: the MeMaLoc and PRADO Studies

The high response rates, as well as the durability of major pathologic responses, suggest that modification of surgical approaches to clinically positive nodal disease based on response to neoadjuvant therapy should be further studied. This strategy would have the potential to spare patients the surgical morbidity of radical lymphadenectomy when they have a pathologic response in the clinically apparent disease. In the MeMaLoc study, a small trial conducted by investigators in the Netherlands, patients with regional nodal metastases underwent placement of a magnetic seed in the clinically apparent node(s) (“index disease”). The investigators asked whether the pathologic response in the “index” node was representative of the regional nodal basin at resection. Twelve patients from the OpACIN-neo trial were enrolled in this additional study. The magnetic beads were placed using ultrasound guidance into the largest involved node. The patients were treated with neoadjuvant combination CBI, and at 6 weeks underwent complete regional lymphadenectomy. The index node was removed and analyzed separately. Surgeons found that the technology was intuitive, and all index nodes were successfully retrieved. They also found that the pathologic response in the index node was 100% congruent with the response in the remaining nodal basin[22]. This study suggests that patient selection for nodal basin observation following a significant pathologic response to neoadjuvant immunotherapy may be a safe and potentially preferable management approach following confirmation of a pathologic response in the clinically apparent disease.

The PRADO study [23] was also conducted with a subset of patients from the OpACIN-neo trial with findings reported at ASCO 2020[24]. This study asked whether surgical approaches to the regional lymphadenectomy could be safely modified based on a patient’s response to neoadjuvant immune therapy. In this trial, investigators accrued patients who underwent placement of a magnetic marker at the site of the clinical nodal disease prior to neoadjuvant therapy. They asked whether a complete resection of the regional lymph nodes could be safely omitted when the patient experienced a major pathologic response to neoadjuvant therapy, as determined by surgical resection of the index disease. Not surprisingly, the investigators found that if the patient had a significant pathologic response to neoadjuvant therapy and avoided a radical lymphadenectomy, surgical morbidity was lower when compared to patients who did not experience a major pathologic response and went on to regional node dissection. This suggests that removal of the index node/nodes instead of a complete dissection may be an attractive option in patients who have a major pathologic response at the index site of disease, although further follow up of the patients in this study is required to assess the safety of this approach and the long-term recurrence rates.

Conclusions

Initial trials of neoadjuvant CBI have demonstrated promising results in terms of pathologic response rates, which have in turn correlated with improved recurrence-free survival. The overwhelming majority of patients in these studies were able to complete surgical resection as planned. Additional trials will need to be done to further define optimal dosing regimens and schedules to decrease irAEs. With recent approval for triple therapy (combined BRAF/MEK inhibition with PD-1 blockade) in the metastatic setting[25, 26], we expect that we may also see future trials using triple therapy in the neoadjuvant setting for BRAF-mutant melanoma.

Overall, compared to the adjuvant approach, neoadjuvant treatment strategies have the potential to provide an early assessment of response. A consistent observation reported in most of the neoadjuvant immune therapy trials is the association of pathologic response to improved outcome. The data are so consistent that it may be reasonable to use pCR as a surrogate marker for RFS. This facilitates an assessment of efficacy early on and may prompt a change in strategy for non-responders. In addition, it calls into question the need for additional (adjuvant) immune therapy, although this question should and will be answered in a prospective trial. As we gain more information about potential predictors of response, we can further modify dosing, timing and combinations of agents to optimize outcomes. While there is compelling evidence to move to a neoadjuvant approach for high-risk clinical stage III/IV disease, further studies will determine whether a neoadjuvant vs. adjuvant vs. perioperative approach is the preferred strategy, or at least elucidate which clinical scenarios are best served by each strategy.

Synopsis:

Neoadjuvant immunotherapy for patients with high risk resectable melanoma produces excellent response rates, which are associated with improved recurrence-free survival. Ongoing and future studies will further define optimal dosing and combinations, and more precisely identify patients likely to benefit from a neoadjuvant strategy.

Data availability Statement:

This is a review article. There are no shared data applicable to this submission.