Abstract
We conducted a randomized phase 3 trial to evaluate whether adjuvant pembrolizumab for one year (648 patients) improved recurrence-free survival (RFS) or overall survival (OS) in comparison to high-dose interferon alfa-2b for one year or ipilimumab for up to three years (655 patients), the approved standard-of-care adjuvant immunotherapies at the time of enrollment for patients with high-risk resected melanoma. At a median follow-up of 47.5 months, pembrolizumab was associated with significantly longer RFS than prior standard-of-care adjuvant immunotherapies (hazard ratio [HR] 0.77; 99.62% confidence interval [CI] 0.59–0.99; P=0.002). There was no statistically significant association with OS among all patients (HR, 0.82; 96.3% CI 0.61–1.09; P=0.15). Proportions of treatment-related adverse events of grades 3 to 5 were 19.5% with pembrolizumab, 71.2% with interferon alfa-2b, and 49.2% with ipilimumab. Therefore, adjuvant pembrolizumab significantly improved RFS but not OS compared to the prior standard-of-care immunotherapies for patients with high-risk resected melanoma.
Introduction
Several therapies have demonstrated an improvement in OS when used as initial treatment for patients with unresectable metastatic melanoma, including ipilimumab (1) and nivolumab alone or in combination (2), pembrolizumab (3), and combinations of BRAF and MEK inhibitors (4–7). In recent years, several of these approaches have been utilized in the adjuvant setting to prevent recurrence of melanoma after surgical excision (8). Adjuvant immunotherapy with high-dose interferon alfa-2b, ipilimumab (cytotoxic T lymphocyte antigen 4 [CTLA-4] blocking antibody, administered at 10 mg per kilogram of body weight every three weeks), or with nivolumab or pembrolizumab (programmed cell death 1 [PD-1] blocking antibodies), improved RFS in randomized clinical trials enrolling patients with completely resected melanoma at high risk of recurrence (9–13); and all are approved for adjuvant use in melanoma by the United States Food and Drug Administration (FDA). High-dose interferon alfa-2b and ipilimumab have each been shown to improve OS compared to an observation or placebo control group in randomized adjuvant therapy trials (9,10). These adjuvant immunotherapy trials enrolled patients with different but partially overlapping stages of melanoma. The European Organization for Research and Treatment of Cancer (EORTC) 1325 (KEYNOTE-054) clinical trial involved a high-risk patient population with stage III melanoma, and demonstrated a 43% lower recurrence or death in patients receiving adjuvant pembrolizumab compared to placebo (12,13). CheckMate 238 enrolled patients with resected higher-risk stage III and stage IV metastatic melanoma; nivolumab showed a 35% improvement in RFS compared to ipilimumab (11). The combination of nivolumab and ipilimumab (the latter given at 1 mg per kilogram every six weeks) was tested in a similar population and did not improve RFS over adjuvant nivolumab (14). It is currently unknown if the use of adjuvant anti-PD-1 therapy improves OS in patients with high-risk resected melanoma. In parallel to the testing of adjuvant immunotherapies for melanoma, adjuvant therapy with the BRAF inhibitor dabrafenib together with the MEK inhibitor trametinib for patients with resected stage III melanoma with a BRAF V600E/K mutation demonstrated an improvement in RFS and distant metastasis-free survival compared to placebo(15). In this trial, OS was reported with the first interim analysis, but the protocol prespecified number of events had not been reached by year 5 (15).
We conducted a randomized phase 3 clinical trial in patients with completely resected melanoma at high risk of recurrence comparing adjuvant pembrolizumab with either of the two standard-of-care adjuvant immunotherapies at the time, which had previously been shown to improve OS over observation or placebo: the high-dose interferon alfa-2b regimen (9) and ipilimumab at 10 mg per kilogram of body weight every three weeks (10). RFS and OS were both primary endpoints in order to determine if adjuvant pembrolizumab therapy would change the course of disease permanently, or if active therapies at the time of recurrence could attenuate differences between patients receiving adjuvant pembrolizumab or the prior standard-of-care adjuvant immunotherapies.
Results
Patients and Trial Regimen
A total of 1,301 eligible patients were randomized from 211 sites in the United States, Canada and Ireland: 647 patients were assigned to the pembrolizumab group, and 654 to the standard-of-care group (Fig. 1). The characteristics of the patients at randomization were similar in the two groups (Table 1). When the trial first opened, standard-of-care adjuvant therapy consisted of high-dose interferon alfa-2b (9). After the FDA approval of adjuvant ipilimumab (10), the protocol was amended to add ipilimumab as a treatment choice. Before the protocol amendment in April 2016, 74 patients had been randomized: 35 to interferon alfa-2b and 40 to pembrolizumab. Among all eligible patients, 8 randomized to pembrolizumab and 89 randomized to standard-of-care did not receive the assigned regimen. Among the patients in the standard-of-care group who received the assigned treatment, 420 received ipilimumab and 145 received interferon alfa-2b (Fig. 1).
Figure 1. Enrollment, randomization, and follow-up.
All eligible patients who were randomized were included in the intention-to-treat population. The safety population included all patients who received at least one dose of treatment that they were randomly assigned. In total, 24 patients in the pembrolizumab group and 20 patients in the standard of care group were found to have major eligibility violations and were excluded from analyses.
Table 1.
Demographic and Clinical Characteristic of Patients at Randomization.*
| Factor | Standard of care (n=654) |
Pembrolizumab (n=647) |
P-value |
|---|---|---|---|
| Sex | |||
| Female | 259 (40) | 264 (41) | 0.69 |
| Male | 295 (60) | 383 (59) | |
| Age (years) | 54 (18, 86) | 53 (20, 82) | 0.11 |
| Age | |||
| < 50 years | 201 (31) | 234 (36) | 0.057 |
| 50 to < 65 years | 272 (41) | 263 (41) | |
| ≥ 65 years | 181 (28) | 150 (23) | |
| Body-mass index | |||
| < 25 | 151 (23) | 154 (24) | 0.68 |
| 25 to < 30 | 231 (35) | 239 (37) | |
| ≥ 30 | 272 (42) | 254 (39) | |
| Standard of care group choice (pre-randomization) | |||
| High dose interferon | 154 (24) | 153 (24) | 0.81 |
| Ipilimumab | 465 (71) | 454 (70) | |
| High dose interferon pre-choice | 35 (5) | 40 (6) | |
| Disease stage (AJCC v7) | |||
| Stage IIIA (N2a) | 58 (9) | 71 (11) | 0.59 |
| Stage IIIB | 353 (54) | 336 (52) | |
| Stage IIIC | 208 (32) | 209 (32) | |
| Stage IV | 35 (5) | 31 (5) | |
| Type of lymph node involvement | |||
| Macroscopic | 280 (43) | 281 (43) | 0.81 |
| Microscopic | 269 (41) | 271 (42) | |
| Not reported | 105 (16) | 95 (15) | |
| Number of positive lymph nodes | |||
| 0 | 8 (1) | 7 (1) | 0.48 |
| 1 | 249 (38) | 224 (35) | |
| 2 or 3 | 208 (32) | 223 (34) | |
| 4 or more | 85 (13) | 99 (15) | |
| Not reported | 104 (16) | 94 (15) | |
| Ulceration | |||
| Yes | 257 (39) | 245 (38) | 0.83 |
| No | 266 (41) | 274 (42) | |
| Unknown | 131 (20) | 128 (20) | |
| PD-L1 expression status ¶ | |||
| Positive | 536 (82) | 534 (82) | 0.66 |
| Negative | 93 (14) | 94 (15) | |
| Indeterminate | 25 (4) | 19 (3) | |
| BRAF mutation status # | |||
| Wild type | 174 (27) | 171 (26) | 1 |
| Mutated | 138 (21) | 138 (22) | |
| Unknown | 343 (52) | 338 (52) |
No significant difference was detected in comparing the two groups.
PD-L1 staining was performed centrally before randomization on tumor biopsy from the primary or metastatic site
BRAF mutation status was determined by the testing available to the local institution and was not required by the protocol, but was collected where available.
Of the 639 patients who received pembrolizumab, 108 (16.9%) discontinued the regimen due to an adverse event. Among the 565 patients who received standard-of-care therapy, 308 (54.4%) discontinued the regimen due to an adverse event. A total of 134 (21.0%) in the pembrolizumab group discontinued the regimen because of disease recurrence, as compared to 92 (16.3%) in the standard-of-care group. A total of 365 patients (57.1%) in the pembrolizumab group and 50 (8.8%) in the standard-of-care group completed the regimen as stipulated in the protocol (Fig. 1). The median duration of follow-up for the patients alive was 47.0 months among patients randomized to pembrolizumab and 45.7 months randomized to standard of care.
Efficacy
Primary Endpoints
The protocol established that the final analysis for RFS and OS would occur after the planned number of events had happened, or at 3.5 years after the last patient was randomized if the anticipated number of events had not yet been reached. As 3.5 years had passed without reaching the planned number of events, the clinical trial was analyzed for the primary endpoints, at which point 98% of the anticipated RFS events and 57% of the anticipated OS events had occurred. In the intent-to-treat population, RFS was significantly longer in the pembrolizumab group than in the standard-of-care group; HR for recurrence or death 0.77 (99.62% CI, 0.59 to 0.99; log-rank P=0.002) (Fig. 2A). The RFS was also longer favoring pembrolizumab in the subgroup of patients with PD-L1 positive melanoma, with a HR of 0.69 (95% CI, 0.58 to 0.84) (Supplementary Fig. 1A). In the intent-to-treat population, the HR for OS was 0.82 (pembrolizumab:standard-of-care) with 96.3% CI, 0.61 to 1.09; log-rank P=0.15 (Fig. 2B). In the subgroup of patients with PD-L1 positive melanoma, the HR for OS was 0.84 with 97.8% CI, 0.59 to 1.22; log-rank P=0.29 (Supplementary Fig. 1B).
Figure 2. Kaplan-Meier estimates of main time-to event endpoints.
A) recurrence-free survival (as assessed by local investigators), B) OS in the intention-to-treat population, and C) exploratory analysis of post-recurrence OS. Cox regression models were stratified by randomization stratification factors: PD-L1 status, intended standard of care regimen choice, and stage of disease. Hazard ratios report pembrolizumab versus standard of care (reference). In the intention-to-treat analysis for RFS there were 524 events (252 in the pembrolizumab group, and 272 in the control group). The HR for recurrence was 0.77 (99.62% confidence interval [CI], 0.59 to 0.99). OS analysis was completed at the protocol specified time of 3.5 years with 57% of events. There was not a statistically significant difference between the pembrolizumab group and the control group (0.82, 96.3% CI 0.61, 1.09). The post-recurrence analysis of OS was not a protocol-specified endpoint; it is included to provide information to evaluate post-recurrence outcomes in both study groups.
Post-recurrence Outcomes
We performed an exploratory analysis of post-recurrence OS, which was not significantly different between the two arms; HR for death 1.20, (95% CI, 0.90, 1.61). PD-1 blocking antibodies as a single drug were the most commonly used regimen after recurrence in the standard-of-care group (39% of patients); 14% of patients in the pembrolizumab group received single-agent PD-1 antibody therapy (Supplementary Table 1). Other post-recurrence therapies included anti-PD-1 in combination with anti-CTLA-4 or other immune-modulating agents, anti-CTLA-4 alone, intralesional oncolytic therapy, interleukin-2, interferon alfa-2b, BRAF and MEK inhibitor targeted therapies, radiation and surgery. Thirty percent of the patients in both groups did not have data available regarding post-recurrence treatment. The predominant site of first recurrence was distant (60% with pembrolizumab, 71% with standard-of-care). Proportions of local, in-transit, and regional recurrences in the pembrolizumab group were 16%, 12%, and 11%, respectively; proportions in the standard-of-care group were 9%, 8%, and 13%, respectively (Supplementary Table 1).
RFS and OS According to Other Variables
The between-group associations in RFS observed in the overall population were consistently observed across subgroups based on baseline characteristics, including according to sex, age, standard-of-care therapy received, and stage of disease, with the possible exception that the group with macroscopic lymph node involvement may derive more improvement from pembrolizumab compared to microscopic lymph node involvement (Fig. 3; RFS macroscopic/microscopic interaction P = 0.08, all other interaction P > 0.22). Patients with PD-L1 positive tumors in the standard-of-care group may derive improvement in post-progression OS compared to the pembrolizumab group (Supplementary Fig. 2; interaction P = 0.09). Otherwise, between-group associations for overall population OS and post-recurrence OS were similar across subgroups (overall population OS: Supplementary Fig. 3, all interaction P > 0.28; post-recurrence OS: Supplementary Fig. 2, all other interaction P > 0.54).
Figure 3. Forest plot for recurrence-free survival according to subgroup in the overall population.
An unstratified Cox regression model was used to estimate the hazard ratios of recurrence or death in the pembrolizumab group as compared to the standard of care group among all the patients. A stratified Cox regression model stratified by the randomization stratification factors (PD-L1 status, control arm choice, and stage of disease) is reported for the overall population. 95% confidence intervals are presented.
Among the 97 patients who did not receive therapy as assigned by the study randomization, 52 (54%) withdrew consent for further follow-up or were lost to follow-up within 30 days of randomization, and RFS and OS were censored at this time. There was no evidence of a difference in RFS or OS comparing patients who received therapy per protocol (n=1204) versus the patients who did not (n=97); HR for recurrence or death (not treated:treated) 1.44 with 95% CI, 0.89 to 2.13 and HR for OS 1.41 with 96.3% CI, 0.66 to 3.02. RFS and OS results in the subset of patients who received therapy per protocol (n=1204, safety population) were similar to results in the intent-to-treat population (n=1301); HR for recurrence or death 0.75 with 99.62% CI, 0.58 to 0.97 and HR for OS 0.78 with 96.3% CI, 0.58 to 1.04. In a final exploratory analysis, we analyzed the RFS data according to the two treatment options in the standard-of-care group, high dose interferon or ipilimumab (Supplementary Fig. 4). In a Cox regression model for RFS comparing the interaction between randomized group with pre-randomization standard-of-care choice is P = 0.49, indicating no strong evidence of heterogeneity in treatment arm association with RFS by pre-randomization standard-of-care treatment choice.
Safety
Adverse events of any grade related to the trial regimen occurred in over 90% of patients: in 583 patients (91%) treated with pembrolizumab, 401 (95%) treated with ipilimumab, and 143 (99%) treated with interferon alfa-2b (Table 2). Proportions of grade 3 and higher adverse events related to treatment were 17%, 43%, and 66% among patients treated with pembrolizumab, ipilimumab, and interferon alfa-2b respectively. Fatigue (59%) and maculo-papular rash (29%) were the most common events among patients treated with pembrolizumab, while fatigue (51%) and diarrhea (48%) were most common among patients treated with ipilimumab; and fatigue (88%) and aspartate aminotransferase increase (70%) were most common among patients treated with interferon alfa-2b (Table 2). Pneumonitis was an uncommon event in all of the treatment arms, and since its frequency was less than 10%, it is not included in Table 2. Among patients treated pembrolizumab pneumonitis of any grade was observed in 4.1 % of patients, and grade 3 pneumonitis in 0.6%. Among patients treated with ipilimumab, the proportions of pneumonitis were 5.0% (all grades) and 1.4% (grade 3), and among patients treated with interferon alfa-2b, they were 2.0% (all grades) and 0% (grade 3). Supplementary Table 2 provides information on any-causality adverse events in 10% or more of patients with any of the three treatment regimens. There was one death due to myocarditis in the pembrolizumab group attributed by the investigators to the treatment, and two among patients receiving ipilimumab, one due to colitis and one due to pneumonitis.
Table 2. Adverse events with a frequency of 10% or greater in any category, determined to be related to the treatment by the treating physician; reported among patients who received protocol therapy.
To be included in the safety analysis, patients must have received at least one dose of protocol therapy (n=1204). Adverse event severity was scored using National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0.
| Pembrolizumab (n=639) | Ipilimumab (n=420) | High-dose Interferon (n=145) | ||||
|---|---|---|---|---|---|---|
| Any Grade | Grade 3–5 | Any Grade | Grade 3–5 | Any Grade | Grade 3–5 | |
| Any (regardless of treatment attribution) | 628 (98%) | 204 (32%) | 415 (99%) | 240 (57%) | 145 (100%) | 105 (72%) |
| Any related to treatment | 583 (91%) | 111 (17) | 401 (95%) | 179 (43%) | 143 (99%) | 96 (66%) |
| Led to Discontinuation | 108 (17%) | 108 (17%) | 272 (65%) | 272 (65%) | 36 (25%) | 36 (25%) |
| Lead to Death | 2 (<1%) | 2 (<1%) | 2 (<1%) | 2 (<1%) | 0 | 0 |
| Toxicities observed in > 10% of patients any grade | ||||||
| Fatigue | 374 (59%) | 3 (<1%) | 214 (51%) | 13 (3%) | 127 (88%) | 17 (12%) |
| Rash - maculo-papular | 187 (29%) | 9 (1%) | 183 (44%) | 31 (7%) | 27 (19%) | 0 |
| Diarrhea | 144 (23%) | 18 (3%) | 200 (48%) | 51 (12%) | 63 (43%) | 3 (2%) |
| Hypothyroidism | 135 (21%) | 0 | 56 (13%) | 2 (<1%) | 10 (7%) | 0 |
| AST increased | 110 (17%) | 14 (2%) | 92 (22%) | 18 (4%) | 101 (70%) | 15 (10%) |
| Arthralgia | 109 (17%) | 3 (<1%) | 31 (7%) | 3 (1%) | 35 (24%) | 1 (1%) |
| ALT increased | 95 (15%) | 19 (3%) | 108 (26%) | 24 (6%) | 97 (67%) | 20 (14%) |
| Myalgia | 72 (11%) | 1 (<1%) | 29 (7%) | 3 (1%) | 53 (37%) | 2 (1%) |
| Anemia | 55 (9%) | 1 (<1%) | 36 (9%) | 1 (<1%) | 40 (28%) | 0 |
| Anorexia | 50 8%) | 1 (<1%) | 74 (18%) | 2 (<1%) | 78 (54%) | 0 |
| Lymphocyte count decreased | 46 (7%) | 3 (<1%) | 8 (2%) | 0 | 35 (24%) | 14 (10%) |
| Abdominal pain | 34 (5%) | 2 (<1%) | 60 (14%) | 5 (1%) | 20 (14%) | 1 (1%) |
| Dizziness | 27 (4%) | 0 | 27 (6%) | 0 | 45 (31%) | 1 (1%) |
| White blood cell count decreased | 27 (4%) | 0 | 6 (1%) | 0 | 83 (57%) | 20 (14%) |
| Hypertriglyceridemia | 25 (3.9%) | 1 (0.2%) | 8 (1.9%) | 0 | 18 (12.3%) | 8 (5.5%) |
| Colitis | 18 (3%) | 13 (2%) | 52 (12%) | 35 (8%) | 0 | 0 |
| Neutrophil count decreased | 14 (2%) | 2 (<1%) | 14 (3%) | 0 | 91 (63%) | 51 (35%) |
| Anxiety | 9 (1%) | 0 | 8 (2%) | 0 | 16 (11%) | 2 (1%) |
| Depression | 8 (1%) | 0 | 5 (1%) | 0 | 26 (18%) | 4 (3%) |
Discussion
In this randomized phase 3 trial comparing pembrolizumab to either ipilimumab or interferon alfa-2b for the treatment of patients with resected, high-risk stage III and stage IV melanoma, pembrolizumab was associated with a statistically significant 23% improvement in RFS. For OS, the HR of 0.82 was not statistically significant. A HR of 0.84 without statistical significance was also seen in the subgroup of patients with tumor PD-L1 expression. The RFS benefit observed in this trial is in the range of what has been shown with nivolumab compared to ipilimumab in a similar population of patients with resected stage III or IV melanoma in the Checkmate-238 trial (11). The KEYNOTE-054 study compared pembrolizumab to placebo in patients with resected stage III cutaneous melanoma and also demonstrated an improvement in RFS (12,13). Therefore, three large adjuvant clinical trials have shown consistent improvement in RFS with the adjuvant administration of an anti-PD-1 therapy. The consequent question of whether adjuvant anti-PD-1 therapy improves OS, has not been reported as final results in the KEYNOTE-054 trial (12,13) and showed a non-significant 2% difference at 4-years in the Checkmate-238 trial, analyzed at 73% of the survival events needed for significance (16). In the current trial, pembrolizumab improved OS by 18%, which also did not meet the pre-specified level of significance when analyzed at 57% of the originally planned survival events. The difference in RFS and OS benefit in our trial may reflect the use of other widely available effective therapies for patients with melanoma recurrence (2–7), or may in part reflect the assessment of OS at a fixed time per protocol, leading to a smaller number of events than would have been necessary for a fully powered event-driven analysis. However, since OS is highly dependent on the availability of post-recurrence therapies, which tend to improve over time, it is unlikely that this result would change with later evaluations.
One of the study findings was that the first recurrence at distant sites was 11% higher in the standard-of-care treatment group. This was due to an apparent increase in local and in-transit recurrences in the pembrolizumab arm (12% higher), while both groups had a similar proportion of recurrence events in regional lymph nodes. These are exploratory findings and their significance is hard to determine with this dataset. We also note here that the early relapse rate at the first scan visit was very high in our study as has been observed in previous adjuvant therapy melanoma trials (11,13), and was consistent between both arms. This finding suggests that the ability to detect distant disease at baseline is limited using current imaging scans and physical exam alone. We are hopeful highly sensitive blood-based markers for minimal residual disease will improve our confirmation of baseline disease free status.
Efforts testing neoadjuvant therapy with immune checkpoint inhibitors in melanoma are currently underway. Multiple regimens have been tested in small feasibility studies including a single dose of pembrolizumab, varying strengths and doses of nivolumab plus ipilimumab, and monotherapy versus combination of immune checkpoint blockade (17–20). However, it is currently not known whether it is superior to give systemic therapy before or after surgery of resectable melanoma. A large, randomized trial powered to detect differences in event-free survival known as S1801 is currently enrolling in the US cooperative groups, and compares adjuvant pembrolizumab given as described in the current study versus 3 doses of neoadjuvant pembrolizumab followed by 15 doses of adjuvant pembrolizumab given every 3 weeks.
In conclusion, pembrolizumab as adjuvant therapy for patients with high-risk resected stage III and stage IV melanoma improves RFS in comparison to the prior standard-of-care adjuvant immunotherapies, but a similar association with OS was not observed among all patients or in the subset of patients with PD-L1 positive tumors. The safety profile of pembrolizumab is consistent with the toxicity spectrum that has been already defined for this agent. Single-agent anti-PD1 antibody treatment should be considered the standard-of-care for adjuvant immunotherapy of high-risk melanoma.
Methods
Patients
The SWOG clinical trial S1404 enrolled patients from December 2015 to October 2017 who were 18 years of age or older with histologically confirmed cutaneous, acral or mucosal melanoma, and were within 98 days from definitive surgery. Patients had resectable stage IIIA (N2a), IIIB, IIIC, or IV disease defined by the American Joint Committee on Cancer 2009 classification, 7th edition (21). A complete regional lymphadenectomy was required for all patients with stage III disease. Imaging studies to document melanoma-free status at enrollment included either a total body positron emission tomography combined with computed tomography (CT) of diagnostic quality or a CT of the chest, abdomen and pelvis with contrast, and a magnetic resonance imaging or CT scan of the brain with intravenous contrast. Main exclusion criteria included prior immunotherapy in any setting for melanoma, active autoimmune disease that had required systemic treatment within two years of study entry, uveal melanoma, and a history of brain metastasis. For full information on eligibility criteria see the protocol in the Appendix. Tumor tissue was required for central pathological evaluation for PD-L1 expression by immunohistochemistry using the 22C3 antibody as previously described (22), and was scored on a scale of 0 to 5 (with higher numbers reflecting a higher level of PD-L1 expression); a score of 2 or higher (staining on greater than 1% of cells) was considered to indicate PD-L1 positivity. The clinical trial was registered as ClinicalTrials.gov number NCT02506153.
Trial Design and Regimen
This was an open-label randomized phase 3 study. Patients were randomized to receive either intravenous infusion of 200 mg of pembrolizumab every three weeks for a total of 18 doses, or an approved standard-of-care adjuvant immunotherapy. When the trial was first opened, the standard-of-care consisted of high-dose interferon alfa-2b, given intravenously at 20 million units per meter squared of body surface area per day, 5 days per week for four weeks followed by 10 million units per meter squared of body surface given subcutaneously three times per week for 11 months (9). After the FDA approval of adjuvant ipilimumab in October of 2015, the protocol was amended in April 2016 to add the choice of ipilimumab dosed at 10 mg per kilogram of body weight intravenously every three weeks for four doses followed by the same dose as maintenance every 12 weeks for up to 11 doses (10). Block randomization (1:1 ratio; block sizes varied: 6, 8, 10; sequence pre-specified by SWOG) was performed centrally using the National Cancer Institute (NCI) web-based OPEN platform and stratified according to stage, PD-L1 melanoma staining, and the planned standard-of-care regimen.
Assessments
Investigator-assessed recurrence was based on imaging or physical exam, with biopsy confirmation whenever possible. Clinical assessment and whole-body imaging occurred every 3 months for the first 2 years beyond randomization, and then every 6 months. Brain imaging was performed annually. Beyond year 5, no study-specific imaging was required, but RFS and OS status were to be monitored up to 10 years. Adverse events were scored using NCI Common Terminology Criteria for Adverse Events, version 5.0.
Trial Oversight
The trial was sponsored by SWOG, the NCI, and Merck Sharp & Dohme Corp. The initial protocol and all amendments were reviewed and approved by SWOG, the NCI, the NCI Central Institutional Review Board, and at each institution. Each study subject provided voluntary, written, informed consent as approved by the human subject protection committee of each institution. The work was conducted in compliance with all ethical guidelines including good clinical practice standards and the Declaration of Helsinki. The data were collected by staff at each site and monitored by SWOG. The data were analyzed and interpreted by the authors. All authors had access to the full data used in the manuscript, and attest that the study as reported here follows the protocol.
Statistical Analysis
The three primary endpoints were: RFS, OS and OS in the subgroup with PD-L1-positive tumors. RFS was measured from date of randomization to date of first of recurrence or death from any cause; patients last known to be alive without recurrence were censored at date of last contact. OS was measured from date of randomization to date of death from any cause; patients last known to be alive were censored at the date of last contact. Models for power and sample size calculations are detailed in the protocol. Full information was 536 RFS and 374 OS events in the overall population. Per protocol, the final analysis would occur at 3.5 years after the last patient was randomized if the anticipated number of events had not yet been reached. Alpha allocation to control the study-wise error to under 0.05 is detailed in the Supplemental Materials. The alpha levels for these analyses are 0.0038 for RFS, 0.037 for OS in the overall population, and 0.023 for OS in the PD-L1 positive subgroup (the alpha in the PD-L1 positive subgroup accounts for the correlation between the subgroup and overall population, and maintains the study-wise alpha below 0.05 with calculations per Speissens and DuBois) (23). Baseline characteristics were compared between groups with Wilcoxon rank sum tests and Fisher’s exact tests. All three primary analyses were based on stratified log-rank tests in the intent-to-treat population. RFS and OS were estimated using the Kaplan-Meier method and Cox regression models were fit. A non-randomized comparison of survival from recurrence by randomized treatment was also performed. For primary endpoint analyses, confidence intervals are reported to match the alpha level of the test. For other analyses, 95% confidence intervals are reported. Two-sided p-values are reported. Toxicity was summarized among patients who received at least one dose of protocol therapy. All analyses were performed in R (version 4.0.2) and SAS (version 9.4).
Data availability
All data (to replicate every analysis in the manuscript and any Supplementary Information) will be posted to the NCI NCTN Data Archive per NCTN policy (https://nctn-data-archive.nci.nih.gov). Patient-level data, including a data dictionary, will be available within 6 months of publication through the United States NCTN/NCORP Data Sharing Archive (https://nctn-data-archive.nci.nih.gov) following the Data Sharing Archive policies. De-identified patient-level data, including the numbers, tables and figures in the paper, will be made available. The protocol (including the statistical analysis plan in section 11 of the protocol) and the informed consent form are available in the Supplementary Information.
Supplementary Material
Statement of significance.
Adjuvant PD-1 blockade therapy decreases the rates of recurrence, but not survival, in patients with surgically resectable melanoma, substituting the prior standard-of-care immunotherapies for this cancer.
Acknowledgments
The authors extend their deepest gratitude to the patients, their families and supportive friends/caregivers for participating in the study. We thank Dr. Lawrence Flaherty (Barbara Ann Karmanos Cancer Institute, Detroit, MI) for serving as chair of the SWOG data monitoring committee for oversight of this trial. We thank Danae Campos and Catrina Mireles for logistical and administrative support. The authors wish to thank SWOG Melanoma Committee Patient Advocates, Valerie Guild (deceased) and Samantha Guild, for their invaluable contributions in support of this study.
Support: Support: NIH/NCI NCTN grants CA180888, CA180819, CA180820, CA180863, UG1CA233178, UG1CA233329, UG1CA189821, UG1CA233331, UG1CA233324, and in part by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA. A.R. is supported by NIH/NCI grants R35 CA197633 and P01 CA244118.
Conflicts of Interest
No potential conflicts of interest reported for R. Conry, P. Funchain, K. Gunturu, D. Johnson, K. Kendra, M. Kovacsovics-Bankowski, H. Li, J. Moon, E. Sharon, T. Truong.
K. Grossmann reports planned future employment with Merck, and receiving consulting fees from and serving as a consultant or advisor for Bristol-Myers Squibb, Novartis, Iovance and Array, and receiving grant support from NIH/NCI (U10CA180888, U10CA180819, U10CA180820, U10CA180863, UG1CA233178, UG1CA233329, UG1CA189821, UG1CA233331, and UG1CA233324).
M. Othus reports serving as a consultant for Daiichi Sankyo, Biosight, and Merck, serving on independent data safety monitoring boards for Celgene and Glycomimetics, and reports grant support from NIH/NCI (grant award U10CA180819).
S. Patel reports receiving grants or contracts from Provectus, Ideaya, Bristol Myers Squibb, InxMed, Foghorn Therapeutics – clinical trial contract with institution, honoraria from Merck & Co – (personal) peer discussion group leader (non-promotional speakers bureau), support for attending meetings and/or travel from Merck & Co; Cardinal Health (personal), participation on a data safety monitoring board or advisory board for Immunocore, Reata – IDMC Chair (personal), Castle Biosciences; TriSalus; Cardinal Health – Advisory Board (personal); receipt of drugs for clinical trial from Provectus, Ideaya, Bristol Myers Squibb, InxMed, Foghorn Therapeutics; receipt of catheter device from TriSalus.
A. Tarhini reports grants or contracts (contracted research with institution) from Merck, OncoSec, Genentech-Roche, Bristol Myers Squibb, Clinigen, consulting fees (consulting/advisory board participation) from Merck, Bristol Myers Squibb, Novartis, Genentech-Roche, Array Biopharma, Partner Therapeutics, Sanofi-Genzyme/Regeneron, Pfizer, EMD Serono, NewLink Genetics, BioNTech, Immunocore, and Eisai.
V. Sondak reports serving as on advisory boards for Bristol-Myers Squibb, Eisai, Merck, Novartis, Regeneron, and Replimune, and on independent data safety monitoring boards for Bristol-Myers Squibb, and receiving research funding from Neogene Therapeutics.
M. Knopp reports leadership role on Board of Directors of American College of Nuclear Medicine.
N. Khushalani reports contracted research funding to Institute from Bristol-Myers Squibb, Merck, HUYA, Celgene, Regeneron, Replimmune, Novartis, GSK and Amgen, consulting fees from Advisory Board or Steering Committee Member for Bristol-Myers Squibb, Merck, Regeneron, HUYA, Iovance, Jounce, Array, Immunocore, Sanofi, and NCCN (from Pfizer), support for attending meetings and/or travel from Sanofi; Participation on a Data Safety Monitoring Board for AstraZeneca and Incyte, common stock in Bellicum, Asensus Surgical (formerly. TransEnterix), Amarin, and Mazor Robotics.
J. Cohen reports honoraria from Bristol-Myers Squibb, and participation on a data safety monitoring board or advisory board for Sanofi-Genzyme.
E. Buchbinder serves on advisory board for Bristol Myers Squibb.
K. Lewis reports “all support for the present manuscript” from Merck, grants or contracts from Bristol-Myers Squibb, Roche/Genentech, GSK, consulting fees from Roche and Merck
B. Chmielowski reports that study S1404 was supported by SWOG NCTN, with payments made to his institution, grants or contract (clinical trial support) from Bristol-Myers Squibb, Macrogenics, Array BioPharma, EMD Serono, Daiichi Sankyo, Merck, Karyopharm Therapeutics, Infinity Pharmaceuticals, Rgenix, Biothera, Aeglea Biotherapeutics, Advenchen Laboratories, Idera, Neon Therapeutics, Xencor, Compugen, Iovance Biotherapeutics, PACT Pharma, RAPT Therapeutics, Immunocore, Lilly, Incyte, Ideaya, Tolero Pharmaceuticals, Ascentage Pharma, Novartis, Atreca, Replimune, with payments made to his institution, consulting fees for service on advisory boards for Iovance Biotherapeutics, IDEAY Biosciences, Epizyme, Deciphera, Sanofi Genzyme, OncoSec, Genentech, Nektar, honoraria from Speakers Bureau: Regeneron – Sanofi Genzyme, support for travel to meetings and accommodations from Regeneron, Epizyme, Deciphera, participation on Nektar data safety monitoring board.
R. Kudchadkar reports research support from Merck, Bristol Myers Squibb, and Regeneron, serves on advisory boards and receives consulting fees Bristol-Myers Squibb, Merck, Array, Regeneron, and Genentech.
Z. Eroglu reports research support from Pfizer and Novartis, and served on advisory boards for Array Biopharma, OncoSec, Regeneron, Genentech, Novartis, SunPharma, and Natera
B. Gastman reports receiving grant from Alkermes to lab 8/24/2018–8/23/2021, grant from Nonimmune Tech to lab 5/4/2018–4/30/2021, grant from InstilBio to lab 9/10/2020–9/10/2025 grant from Quest Imaging to lab 1/12/2020 −11/11/2021, consulting fees from Quest Imaging, honoraria from Castle Biosciences and Merck, support from Quest Imaging for attending meeting, participation on advisory boards for Merck, Alkermes, Castle Biosciences, Questing Imaging, Cardinal Health, Bristol Myers Squibb, leadership roles as Chair of Plastic Surgery Research Council, Surgical Committee Chair of SITC, and Multiple chair/vice-chair of committees at the American Society of Plastic Surgery, stock for stock options from Castle Biosciences.
S. Ebbinghaus reports employment at Merck, and owns stock in Merck.
S. Ahsan reports employment at Merck, and owns stock in Merck.
N. Ibrahim reports employment at Merck, and owns stock in Merck.
T. Petrella reports honoraria for serving on Merck Advisory Board.
L. Korde reports participation on Data Safety Monitoring Board, ECOG-ACRIN DSMC, non-voting member (NCI representative).
J. Kirkwood reports receiving consulting fees from Amgen, Bristol-Myers Squibb, Checkmate Pharm, Harbour BioMed, Iovance Biotherapeutics, Istari Oncology, OncoSec Medical, Novartis, Scopus BioPharma, Pfizer, Elsevier, OncoCyte, Takeda Pharm, Immunocore, Natera, DermTech, honorarium from Bristol-Myers Squibb for unbranded presentation, support for travel from Axio Research/Cytel.
A. Ribas reports receiving honoraria from consulting with Amgen, Bristol-Myers Squibb, Chugai, Genentech, Merck, Novartis, Roche, Sanofi and Vedanta, is or has been a member of the scientific advisory board and holds stock in Advaxis, Apricity, Arcus, Compugen, CytomX, Five Prime, Highlight, ImaginAb, Isoplexis, Kalthera, Kite-Gilead, Merus, PACT Pharma, RAPT, Rgenix and Tango, has received research funding from Agilent and from Bristol-Myers Squibb through Stand Up to Cancer (SU2C).
References
- 1.Hodi FS, O’Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, et al. Improved survival with ipilimumab in patients with metastatic melanoma. The New England journal of medicine 2010;363(8):711–23 doi 10.1056/NEJMoa1003466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Larkin J, Chiarion-Sileni V, Gonzalez R, Grob JJ, Rutkowski P, Lao CD, et al. Five-Year Survival with Combined Nivolumab and Ipilimumab in Advanced Melanoma. N Engl J Med 2019;381(16):1535–46 doi 10.1056/NEJMoa1910836. [DOI] [PubMed] [Google Scholar]
- 3.Robert C, Schachter J, Long GV, Arance A, Grob JJ, Mortier L, et al. Pembrolizumab versus Ipilimumab in Advanced Melanoma. N Engl J Med 2015;372(26):2521–32 doi 10.1056/NEJMoa1503093. [DOI] [PubMed] [Google Scholar]
- 4.Long GV, Stroyakovskiy D, Gogas H, Levchenko E, de Braud F, Larkin J, et al. Combined BRAF and MEK inhibition versus BRAF inhibition alone in melanoma. N Engl J Med 2014;371(20):1877–88 doi 10.1056/NEJMoa1406037. [DOI] [PubMed] [Google Scholar]
- 5.Larkin J, Ascierto PA, Dreno B, Atkinson V, Liszkay G, Maio M, et al. Combined vemurafenib and cobimetinib in BRAF-mutated melanoma. N Engl J Med 2014;371(20):1867–76 doi 10.1056/NEJMoa1408868. [DOI] [PubMed] [Google Scholar]
- 6.Robert C, Karaszewska B, Schachter J, Rutkowski P, Mackiewicz A, Stroiakovski D, et al. Improved overall survival in melanoma with combined dabrafenib and trametinib. N Engl J Med 2015;372(1):30–9 doi 10.1056/NEJMoa1412690. [DOI] [PubMed] [Google Scholar]
- 7.Dummer R, Ascierto PA, Gogas HJ, Arance A, Mandala M, Liszkay G, et al. Encorafenib plus binimetinib versus vemurafenib or encorafenib in patients with BRAF-mutant melanoma (COLUMBUS): a multicentre, open-label, randomised phase 3 trial. Lancet Oncol 2018;19(5):603–15 doi 10.1016/S1470-2045(18)30142-6. [DOI] [PubMed] [Google Scholar]
- 8.Eggermont AMM, Robert C, Ribas A. The new era of adjuvant therapies for melanoma. Nat Rev Clin Oncol 2018;15(9):535–6 doi 10.1038/s41571-018-0048-5. [DOI] [PubMed] [Google Scholar]
- 9.Kirkwood JM, Strawderman MH, Ernstoff MS, Smith TJ, Borden EC, Blum RH. Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: the Eastern Cooperative Oncology Group Trial EST 1684. J Clin Oncol 1996;14(1):7–17. [DOI] [PubMed] [Google Scholar]
- 10.Eggermont AM, Chiarion-Sileni V, Grob JJ, Dummer R, Wolchok JD, Schmidt H, et al. Adjuvant ipilimumab versus placebo after complete resection of high-risk stage III melanoma (EORTC 18071): a randomised, double-blind, phase 3 trial. Lancet Oncol 2015;16(5):522–30 doi 10.1016/S1470-2045(15)70122-1. [DOI] [PubMed] [Google Scholar]
- 11.Weber J, Mandala M, Del Vecchio M, Gogas HJ, Arance AM, Cowey CL, et al. Adjuvant Nivolumab versus Ipilimumab in Resected Stage III or IV Melanoma. N Engl J Med 2017;377(19):1824–35 doi 10.1056/NEJMoa1709030. [DOI] [PubMed] [Google Scholar]
- 12.Eggermont AMM, Blank CU, Mandala M, Long GV, Atkinson V, Dalle S, et al. Adjuvant Pembrolizumab versus Placebo in Resected Stage III Melanoma. N Engl J Med 2018;378(19):1789–801 doi 10.1056/NEJMoa1802357. [DOI] [PubMed] [Google Scholar]
- 13.Eggermont AMM, Blank CU, Mandala M, Long GV, Atkinson VG, Dalle S, et al. Adjuvant pembrolizumab versus placebo in resected stage III melanoma (EORTC 1325-MG/KEYNOTE-054): distant metastasis-free survival results from a double-blind, randomised, controlled, phase 3 trial. Lancet Oncol 2021;22(5):643–54 doi 10.1016/S1470-2045(21)00065-6. [DOI] [PubMed] [Google Scholar]
- 14.Long GV, Schadendorf D, Del Vecchio M, Larkin J, Atkinson V, Schenker M, et al. Adjuvant therapy with nivolumab combined with ipilimumab vs nivolumab alone in patients with resected stage IIIB–D/IV melanoma (CheckMate 915). Annual Meeting of the American Association for Cancer Research (AACR) 2021:Abstract 5091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dummer R, Hauschild A, Santinami M, Atkinson V, Mandala M, Kirkwood JM, et al. Five-Year Analysis of Adjuvant Dabrafenib plus Trametinib in Stage III Melanoma. N Engl J Med 2020;383(12):1139–48 doi 10.1056/NEJMoa2005493. [DOI] [PubMed] [Google Scholar]
- 16.Ascierto PA, Del Vecchio M, Mandala M, Gogas H, Arance AM, Dalle S, et al. Adjuvant nivolumab versus ipilimumab in resected stage IIIB-C and stage IV melanoma (CheckMate 238): 4-year results from a multicentre, double-blind, randomised, controlled, phase 3 trial. Lancet Oncol 2020;21(11):1465–77 doi 10.1016/S1470-2045(20)30494-0. [DOI] [PubMed] [Google Scholar]
- 17.Blank CU, Rozeman EA, Fanchi LF, Sikorska K, van de Wiel B, Kvistborg P, et al. Neoadjuvant versus adjuvant ipilimumab plus nivolumab in macroscopic stage III melanoma. Nat Med 2018;24(11):1655–61 doi 10.1038/s41591-018-0198-0. [DOI] [PubMed] [Google Scholar]
- 18.Huang AC, Orlowski RJ, Xu X, Mick R, George SM, Yan PK, et al. A single dose of neoadjuvant PD-1 blockade predicts clinical outcomes in resectable melanoma. Nat Med 2019;25(3):454–61 doi 10.1038/s41591-019-0357-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rozeman EA, Menzies AM, van Akkooi ACJ, Adhikari C, Bierman C, van de Wiel BA, et al. Identification of the optimal combination dosing schedule of neoadjuvant ipilimumab plus nivolumab in macroscopic stage III melanoma (OpACIN-neo): a multicentre, phase 2, randomised, controlled trial. Lancet Oncol 2019;20(7):948–60 doi 10.1016/S1470-2045(19)30151-2. [DOI] [PubMed] [Google Scholar]
- 20.Amaria RN, Reddy SM, Tawbi HA, Davies MA, Ross MI, Glitza IC, et al. Neoadjuvant immune checkpoint blockade in high-risk resectable melanoma. Nat Med 2018;24(11):1649–54 doi 10.1038/s41591-018-0197-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Balch CM, Gershenwald JE, Soong SJ, Thompson JF, Atkins MB, Byrd DR, et al. Final version of 2009 AJCC melanoma staging and classification. J Clin Oncol 2009;27(36):6199–206 doi 10.1200/JCO.2009.23.4799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Daud AI, Wolchok JD, Robert C, Hwu WJ, Weber JS, Ribas A, et al. Programmed Death-Ligand 1 Expression and Response to the Anti-Programmed Death 1 Antibody Pembrolizumab in Melanoma. J Clin Oncol 2016;34(34):4102–9 doi 10.1200/JCO.2016.67.2477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Spiessens B, Debois M. Adjusted significance levels for subgroup analyses in clinical trials. Contemp Clin Trials 2010;31(6):647–56 doi 10.1016/j.cct.2010.08.011. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data (to replicate every analysis in the manuscript and any Supplementary Information) will be posted to the NCI NCTN Data Archive per NCTN policy (https://nctn-data-archive.nci.nih.gov). Patient-level data, including a data dictionary, will be available within 6 months of publication through the United States NCTN/NCORP Data Sharing Archive (https://nctn-data-archive.nci.nih.gov) following the Data Sharing Archive policies. De-identified patient-level data, including the numbers, tables and figures in the paper, will be made available. The protocol (including the statistical analysis plan in section 11 of the protocol) and the informed consent form are available in the Supplementary Information.