Skip to content

Neoadjuvant therapy for resectable melanoma

TL;DR — Two randomised trials have moved neoadjuvant immunotherapy from concept to standard-of-care candidate in resectable macroscopic stage III melanoma. SWOG S1801 (phase 2, 313 patients) found 2-year event-free survival 72% (95% CI 64–80) with neoadjuvant-plus-adjuvant pembrolizumab versus 49% (41–59) with adjuvant-only pembrolizumab (P = .004), with grade ≥3 treatment-related adverse events of 12% versus 14% (Patel 2023, PMID 36856617; NCT03698019). NADINA (phase 3, 423 patients) found 12-month event-free survival 83.7% (99.9% CI 73.8–94.8) with two cycles of neoadjuvant ipilimumab–nivolumab versus 57.2% (45.1–72.7) with surgery then 12 cycles of adjuvant nivolumab, HR 0.32 (99.9% CI 0.15–0.66), P < .001; 59.0% achieved a major pathological response, and among those the estimated 12-month recurrence-free survival was 95.1% (Blank 2024, PMID 38828984; NCT04949113). The advantage is not merely an efficacy gain: pathological response is available within weeks and permits treatment de-escalation, which adjuvant therapy structurally cannot offer. The counter-argument is that both trials used event-free survival, that S1801's analysis has been challenged on time-related biases (Olivier 2024, PMID 38621314), and that no trial has compared modern neoadjuvant against modern adjuvant therapy with overall survival as the primary endpoint.

The trial sequence

Trial Phase n Design Result
OpACIN-neo (Rozeman 2019, PMID 31160251; NCT02977052) 2, randomised 89 enrolled / 86 treated Three neoadjuvant dosing schedules of ipilimumab + nivolumab in resectable stage III nodal melanoma Grade 3–4 immune-related AEs within 12 weeks: 40% (group A, ipi 3 + nivo 1), 20% (group B, ipi 1 + nivo 3), 50% (group C, sequential ipi then nivo; accrual closed early by the DSMB). Difference B vs A −20% (95% CI −46 to 6, P = .158). Radiological objective response 63%, 57% and 35%. One late immune encephalitis death. Group B identified as the tolerable schedule
PRADO (Reijers 2022, PMID 35661157) 2, extension cohort of OpACIN-neo 99 6 weeks of neoadjuvant ipilimumab 1 mg/kg + nivolumab 3 mg/kg in clinical stage IIIb–d nodal melanoma, then response-directed management: major pathological response (≤10% viable tumour) in the index lymph node → omit therapeutic lymph-node dissection and adjuvant therapy; pathological partial response (>10–≤50%) → dissection only; pathological non-response (>50%) → dissection plus adjuvant systemic therapy ± radiotherapy First demonstration that pathological response can be used prospectively to de-escalate surgery and adjuvant therapy
SWOG S1801 (Patel 2023, PMID 36856617) 2, randomised 313 Clinically detectable, measurable, resectable stage IIIB–IVC: three doses neoadjuvant pembrolizumab + surgery + 15 adjuvant doses vs surgery + 18 adjuvant doses Median follow-up 14.7 months. Landmark 2-year EFS 72% (64–80) vs 49% (41–59), P = .004. Grade ≥3 treatment-related AEs 12% vs 14%
NADINA (Blank 2024, PMID 38828984) 3, randomised 423 Resectable macroscopic stage III: two cycles neoadjuvant ipilimumab + nivolumab then surgery (adjuvant only for partial/non-responders) vs surgery then 12 cycles adjuvant nivolumab Median follow-up 9.9 months. 12-month EFS 83.7% (99.9% CI 73.8–94.8) vs 57.2% (45.1–72.7); restricted mean survival time difference 8.00 months (4.94–11.05), P < .001; HR 0.32 (0.15–0.66). Pathological response distribution: 59.0% major, 8.0% partial, 26.4% non-response, 2.4% progression, 4.2% surgery omitted or not yet done. 12-month RFS 95.1% among major responders

NADINA is the only phase 3 trial in this list, and its design encodes the de-escalation logic: major pathological responders received no adjuvant therapy at all, so the comparison is neoadjuvant-with-de-escalation against a full year of adjuvant nivolumab.

Pathological response as the operative endpoint

The endpoint that makes the neoadjuvant approach distinctive is not survival but the resection specimen. The International Neoadjuvant Melanoma Consortium convened pathologists in 2016 and 2017 and produced consensus guidelines for handling and reporting resection specimens after neoadjuvant targeted or immune therapy in AJCC 8th-edition stage IIIB/C/D or oligometastatic stage IV melanoma, defining pathological response categories, reporting standards and the spectrum of histopathological response patterns (Tetzlaff 2018, PMID 29945191). Updated pan-tumour scoring guidance was issued jointly by SITC and INMC in 2026 (PMID 41469296).

Category Definition Consequence in PRADO / NADINA
Major pathological response (MPR) ≤10% viable tumour Omit lymph-node dissection (PRADO) and adjuvant therapy (both)
Pathological partial response (pPR) >10% to ≤50% viable tumour Dissection only (PRADO); adjuvant therapy (NADINA)
Pathological non-response (pNR) >50% viable tumour Dissection + adjuvant systemic therapy ± radiotherapy (PRADO)

Sources: PMID 35661157; PMID 38828984; PMID 29945191. In NADINA, 12-month recurrence-free survival among major pathological responders was 95.1% (PMID 38828984) — the number that justifies de-escalation, and simultaneously the number most vulnerable to short follow-up.

Regimen effects on pathological response

Response rate is strongly regimen-dependent, which is why the trials above are not interchangeable.

Regimen Setting Pathological response
Single-agent anti-PD-1 Macroscopic stage III Major pathological response 20–30% (van Akkooi 2023, PMID 36738540)
Ipilimumab + nivolumab Macroscopic stage III MPR 50–60% (PMID 36738540); 59.0% in NADINA (PMID 38828984)
Nivolumab + relatlimab (anti-LAG-3) Resectable clinical stage III or oligometastatic stage IV, 30 patients, two neoadjuvant doses then surgery then 10 adjuvant doses Pathological complete response 57%, overall pathological response 70%, RECIST response 57%; no grade 3–4 immune-related adverse events in the neoadjuvant setting; 1- and 2-year RFS 100% and 92% with any pathological response versus 88% and 55% without (Amaria 2022, PMID 36289334; NCT02519322). Updated at median 47 months: MPR rate 63%, and 80% of patients event-free at 4 years from the start of neoadjuvant therapy (Burton 2025, PMID 40638872)
Dabrafenib + trametinib (BRAF V600-mutant) NeoCombi, single-arm, single-centre, 35 patients: 12 weeks neoadjuvant then resection then 40 weeks adjuvant At median 60 months (95% CI 56–72), 21 of 35 (60%) recurred — 12 (57%) first in locoregional sites with 6 later distant, 9 (43%) first distant including 3 brain (Long 2019, PMID 31171444; Menzies 2024, PMID 38754780; NCT01972347)

The NeoCombi five-year result is the sharpest available caution about pathological response as a surrogate. Targeted therapy produces high pathological complete response rates, yet 60% of patients recurred by five years. A high pathological response rate is regimen-specific evidence of drug activity; it is not automatically evidence of cure, and its prognostic meaning may differ between immune and targeted mechanisms.

A 2026 meta-analysis of neoadjuvant checkpoint cohorts found a qualitative pattern of increasing both pathological efficacy and toxicity when anti-CTLA-4 was added to, or dosed more intensively with, anti-PD-1. Its exploratory baseline IFN-γ subgroup analysis suggested that higher-dose anti-CTLA-4 increased efficacy mainly in IFN-γ-low tumours, while increasing toxicity without efficacy gain in IFN-γ-high tumours; this is hypothesis-generating stratification, not a prospectively validated treatment rule (Lucas 2026, PMID 41478279).

Durability of the pathological response, and biomarkers of achieving it

Longer follow-up of the two Dutch trials supplies the maturity that NADINA and S1801 lack.

Cohort Follow-up Result
OpACIN (n = 20, randomised split-course neoadjuvant vs adjuvant ipilimumab–nivolumab) Median 69 months 1 of 7 patients with a pathological response recurred. Estimated 5-year RFS/OS 70%/90% (neoadjuvant arm) versus 60%/70% (adjuvant arm) (Versluis 2023, PMID 36681299)
OpACIN-neo (n = 86) Median 47 months Estimated 3-year RFS 82%, OS 92%; 3-year RFS 95% in pathological responders (PMID 36681299)
OpACIN + OpACIN-neo pooled biomarker analysis Median 4 years (OpACIN) 2-year RFS 84% overall, 97% in pathological responders versus 36% in non-responders (P < .001). High tumour mutational burden and high IFN-γ-related gene expression signature were each associated with pathological response: pathological response rate 100% with high IFN-γ/high TMB, 91% and 88% with one high marker, and 39% with low IFN-γ/low TMB (Rozeman 2021, PMID 33558721)

The IFN-γ/TMB combination is the closest thing to a predictive biomarker in this setting, and it separates pathological response rates from 39% to 100% — a range wide enough to change whether neoadjuvant therapy is the right strategy for an individual. It has not been prospectively validated as a selection tool.

Guideline position

ASCO's systemic-therapy guideline update for melanoma identified 21 additional randomised trials in its updated systematic review (Seth 2023, PMID 37579248), and an influential European commentary argued in 2023 that neoadjuvant immunotherapy had already become best medical practice and would become the standard of care for macroscopic stage III melanoma (van Akkooi 2023, PMID 36738540). The European guideline treatment update covers the setting (Garbe 2025, PMID 39709737); synthesis is in guidelines.

The critique, presented as a live disagreement

The pre-S1801/NADINA evidence base was genuinely uncertain. A Cochrane review searching through 10 August 2021 included eight randomised trials and 402 participants; the pooled overall-survival estimate was HR 0.43 (95% CI 0.15–1.21; 2 trials, 171 participants; very-low-certainty evidence), and adverse events were 26% versus 16% (RR 1.58, 95% CI 0.97–2.55; 2 trials, 162 participants; very-low-certainty evidence). The review found no comparative quality-of-life data and explicitly called for adequately powered trials and validation of pathological response (Gorry 2023, PMID 36648215). That dated synthesis is partly superseded by the later phase 2 and phase 3 efficacy trials, but its endpoint and certainty limitations remain relevant.

Three specific objections have been raised to the SWOG S1801 analysis, and they are content rather than noise (Olivier 2024, PMID 38621314):

  1. An arbitrary rule in event assignment that may have affected the distribution of events over time.
  2. Different rates of early censoring between groups, raising the possibility of informative censoring producing an artefactual event-free-survival advantage.
  3. Phase 2 results carry a risk of chance findings and should be confirmed.

The first two objections are specific to S1801's EFS definition, which counted as events: disease progression or toxicity precluding surgery; inability to resect all gross disease; progression, surgical complications or toxicity preventing adjuvant therapy within 84 days of surgery; post-surgical recurrence; or death (PMID 36856617). Several of those events can only occur in the neoadjuvant arm's timeline or the adjuvant arm's timeline but not symmetrically, which is precisely the structure that generates time-related bias. NADINA's phase 3 design and its independent confirmation of direction and magnitude answer objection 3 but not objections 1 and 2, since NADINA's primary endpoint is also event-free survival (PMID 38828984).

Neither trial reports overall survival as a primary endpoint, and follow-up in NADINA at publication was 9.9 months.

Real-world implementation

Single-institution prospective-database analysis of 159 patients with clinically detectable stage IIIB–IIID or oligometastatic stage IVA–IVC melanoma treated 2020–2024 (102 neoadjuvant, 57 adjuvant-only, median follow-up 28.7 months in the neoadjuvant group) found landmark 1-year event-free survival 79.2% (95% CI 71.6–87.5) versus 54.7% (43.2–69.3), and 2-year EFS 68.9% (59.9–79.2) versus 46.3% (34.7–61.8), P = .05 (Dougherty 2026, PMID 42360644). The effect size reproduces outside the trials, in a non-randomised comparison with the confounding that implies. Implementation research has also documented how centres operationalise the shift: a modified Delphi consensus in a European-accredited cancer centre in Ireland addressed practical implementation of neoadjuvant immunotherapy in stage III melanoma (Cronin 2025, PMID 40857557).

What neoadjuvant therapy changes beyond efficacy

  • It supplies an early, individual readout of drug activity. Pathological response at 6 weeks stratifies subsequent risk more sharply than any pre-treatment variable (PMID 35661157; PMID 38828984).
  • It permits surgical de-escalation. PRADO omitted therapeutic lymph-node dissection in major responders (PMID 35661157); a dedicated trial of omitting therapeutic lymph-node dissection in patients with a major pathological response in the index node is now recruiting (NCT06754904, 213 planned, primary completion 2027).
  • It permits systemic de-escalation. NADINA gave no adjuvant therapy to major responders (PMID 38828984).
  • It shortens total treatment. Two neoadjuvant cycles versus twelve adjuvant cycles in NADINA (PMID 38828984).
  • It creates new failure modes: toxicity or progression precluding surgery, and delay to an operation that would otherwise have happened immediately (PMID 38621314).

Open questions in dosing remain: SWE-NEO is randomising monotherapy against combination immunotherapy in resectable stage III melanoma (NCT06794775, 128 planned, primary completion 2032). The competing standard is described in adjuvant therapy; reviews of the evolving management have been published (Papadopoulos 2026, PMID 42352510; Reddish 2026, PMID 41990381; Kuijpers 2024, PMID 38610925; Long 2023, PMID 37104746).

What is being tested next

Trial Question Status (verified 2026-09-01)
NCT06754904 Omission of therapeutic lymph-node dissection in patients with a major pathological response in the index node — the PRADO strategy tested prospectively Recruiting, n = 213, primary completion 2027-04-01
NCT06794775 (SWE-NEO) Neoadjuvant monotherapy versus combination immunotherapy in resectable stage III melanoma — the OpACIN-neo dosing question at phase 3 Recruiting, n = 128, primary completion 2032-04-15
NCT03567889 Neoadjuvant intratumoral Daromun in clinical stage IIIB/C/D melanoma — a non-checkpoint neoadjuvant approach Recruiting, n = 186, primary completion 2027-12
NCT03698019 (SWOG S1801) The trial whose analysis is contested Active, not recruiting, n = 313
NCT04949113 (NADINA) The phase 3 result Active, not recruiting, n = 423

The full landscape is in clinical trials landscape. The trial that is not on this list is the one comparing neoadjuvant against adjuvant therapy with overall survival as the primary endpoint — a gap re-checked against the active phase 3 registry on 2026-09-01.

Interpretation rules for this page

  • Event-free survival is not overall survival, and its definition differs between S1801 and NADINA (PMID 36856617; PMID 38828984).
  • State the follow-up. NADINA's headline result comes from a median 9.9 months (PMID 38828984).
  • NADINA compares strategies, not drugs. Neoadjuvant ipilimumab–nivolumab with de-escalation versus adjuvant nivolumab is a comparison of two whole pathways (PMID 38828984).
  • Toxicity depends on schedule, not just on the combination. OpACIN-neo grade 3–4 immune toxicity ranged 20% to 50% across three schedules of the same two drugs (PMID 31160251).
  • Pathological response requires standardised specimen handling to be comparable across centres (PMID 29945191; PMID 41469296).
  • The S1801 bias critique is unresolved, and NADINA does not address its first two objections (PMID 38621314).

Open questions

  • Does neoadjuvant sequencing improve overall survival, or only the endpoints trials have chosen to measure (PMID 36856617; PMID 38828984; PMID 38621314)?
  • Is de-escalation safe in the long term for major pathological responders, whose 95.1% RFS figure rests on 12 months of follow-up (PMID 38828984)?
  • Can therapeutic lymph-node dissection be omitted after major pathological response? NCT06754904 is testing exactly this.
  • Is combination ipilimumab–nivolumab necessary, or does anti-PD-1 monotherapy suffice? NCT06794775 is randomising it.
  • Should neoadjuvant therapy extend to stage IIB/IIC, where adjuvant therapy has already moved (PMID 40198940)?
  • What is the right event-free-survival definition for a neoadjuvant trial, given that several component events are asymmetric between arms (PMID 38621314)?

References

  1. Patel SP, et al. Neoadjuvant-Adjuvant or Adjuvant-Only Pembrolizumab in Advanced Melanoma. The New England journal of medicine. 2023;388:813-823. PMID 36856617
  2. Blank CU, et al. Neoadjuvant Nivolumab and Ipilimumab in Resectable Stage III Melanoma. The New England journal of medicine. 2024;391:1696-1708. PMID 38828984
  3. Olivier T, et al. Neoadjuvant followed by adjuvant pembrolizumab in melanoma: time biases in the data analysis of the SWOG S1801 trial. Translational oncology. 2024;45:101959. PMID 38621314
  4. Rozeman EA, 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. The Lancet. Oncology. 2019;20:948-960. PMID 31160251
  5. Reijers ILM, et al. Personalized response-directed surgery and adjuvant therapy after neoadjuvant ipilimumab and nivolumab in high-risk stage III melanoma: the PRADO trial. Nature medicine. 2022;28:1178-1188. PMID 35661157
  6. Tetzlaff MT, et al. Pathological assessment of resection specimens after neoadjuvant therapy for metastatic melanoma. Annals of oncology : official journal of the European Society for Medical Oncology. 2018;29:1861-1868. PMID 29945191
  7. Deutsch JS, et al. Updated pan-tumor guidelines for neoadjuvant scoring of pathologic response: a joint SITC and INMC effort. Annals of oncology : official journal of the European Society for Medical Oncology. 2026;37:141-154. PMID 41469296
  8. van Akkooi ACJ, et al. Neo-adjuvant immunotherapy emerges as best medical practice, and will be the new standard of care for macroscopic stage III melanoma. European journal of cancer (Oxford, England : 1990). 2023;182:38-42. PMID 36738540
  9. Amaria RN, et al. Neoadjuvant relatlimab and nivolumab in resectable melanoma. Nature. 2022;611:155-160. PMID 36289334
  10. Burton EM, et al. Long-Term Survival and Biomarker Analysis Evaluating Neoadjuvant Plus Adjuvant Relatlimab (anti-LAG3) and Nivolumab (anti-PD1) in Patients With Resectable Melanoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2025;43:2856-2862. PMID 40638872
  11. Long GV, et al. Neoadjuvant dabrafenib combined with trametinib for resectable, stage IIIB-C, BRAF(V600) mutation-positive melanoma (NeoCombi): a single-arm, open-label, single-centre, phase 2 trial. The Lancet. Oncology. 2019;20:961-971. PMID 31171444
  12. Menzies AM, et al. Five-year analysis of neoadjuvant dabrafenib and trametinib for stage III melanoma. Annals of oncology : official journal of the European Society for Medical Oncology. 2024;35:739-746. PMID 38754780
  13. Versluis JM, et al. Survival update of neoadjuvant ipilimumab plus nivolumab in macroscopic stage III melanoma in the OpACIN and OpACIN-neo trials. Annals of oncology : official journal of the European Society for Medical Oncology. 2023;34:420-430. PMID 36681299
  14. Rozeman EA, et al. Survival and biomarker analyses from the OpACIN-neo and OpACIN neoadjuvant immunotherapy trials in stage III melanoma. Nature medicine. 2021;27:256-263. PMID 33558721
  15. Seth R, et al. Systemic Therapy for Melanoma: ASCO Guideline Update. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2023;41:4794-4820. PMID 37579248
  16. Garbe C, et al. European consensus-based interdisciplinary guideline for melanoma. Part 2: Treatment - Update 2024. European journal of cancer (Oxford, England : 1990). 2025;215:115153. PMID 39709737
  17. Dougherty KE, et al. Implementing Paradigm Shifting Clinical Trials: Real-World Outcomes of Neoadjuvant Versus Adjuvant Immune Checkpoint Inhibitors in Advanced Melanoma Patients. Annals of surgical oncology. 2026;33:7952-7959. PMID 42360644
  18. Cronin C, et al. Implementation of neoadjuvant immunotherapy in stage III melanoma: a modified Delphi consensus study in a European-accredited cancer center in Ireland. The oncologist. 2025;30. PMID 40857557
  19. Papadopoulos N, et al. Neoadjuvant Therapy and the Evolving Management of Resectable Advanced Melanoma. Cancers. 2026;18. PMID 42352510
  20. Reddish S, et al. Neoadjuvant systemic therapy in stage III and IV resectable melanoma: an update to management and future directions. The New Zealand medical journal. 2026;139:93-99. PMID 41990381
  21. Kuijpers AMJ, et al. Neo-Adjuvant Therapy for Metastatic Melanoma. Cancers. 2024;16. PMID 38610925
  22. Long GV, et al. Neoadjuvant Checkpoint Immunotherapy and Melanoma: The Time Is Now. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2023;41:3236-3248. PMID 37104746
  23. Luke JJ, et al. Pembrolizumab versus placebo as adjuvant therapy in resected stage IIB or IIC melanoma: Long-term follow-up, crossover, and rechallenge with pembrolizumab in the phase III KEYNOTE-716 study. European journal of cancer (Oxford, England : 1990). 2025;220:115381. PMID 40198940
  24. Lucas MW, et al. Immune signature-based uncoupling of checkpoint inhibitor efficacy and toxicity. Immunity. 2026;59:29-33.e2. PMID 41478279
  25. Gorry C, et al. Neoadjuvant treatment for stage III and IV cutaneous melanoma. The Cochrane database of systematic reviews. 2023;1:CD012974. PMID 36648215