On occasion, a therapeutic development makes a meaningful difference in patient outcomes without involvement of a new, expensive drug. An outstanding example of this in melanoma is the adoption of neoadjuvant checkpoint blockade for patients with clinically evident stage III disease. The SWOG S1801 and NADINA randomized clinical trials both demonstrated marked improvement in event-free survival (EFS), with hazard ratios of 0.58 and 0.32, respectively.^1,2 The simplicity of the elegant design of the S1801 trial is perhaps the purest proof of the neoadjuvant concept in that the patients in both arms of the study received the same surgical and systemic treatments, only in a different sequence, a simple trick indeed.

This remarkable benefit appears to derive from simply initiating immune checkpoint blockade with the target (the tumor) in situ. Although the Food and Drug Administration has not formally approved a neoadjuvant regimen for clinical stage III melanoma patients to date, undisputedly, a new standard has been established by these two trials, which confirmed the promising findings of predating trials.^3,4

An unanswered question, however, is whether similar benefits can be obtained for patients with clinical stage II melanoma. This is an important group because despite the absence of clinically evident disease in their regional lymph nodes, stage II patients have a high risk for recurrence and death from melanoma, worse, in fact, than many stage III patients. It also should be noted that about one third of clinical stage IIb/c patients are found to have pathologic stage IIIc/d disease at operation because they harbor clinically undetectable sentinel lymph node (SLN) metastases. It is challenging, however, to evaluate the value of neoadjuvant treatment in this population for several reasons. First, we generally have used pathologic response as an early readout of efficacy in neoadjuvant studies, which is not straightforward when there is no disease to measure. Similarly, the dysfunctional tumor microenvironment that is the theoretical scientific basis for improved efficacy with neoadjuvant sequencing is harder to visualize with no clinically evident tumor. And finally, what benchmarks can be used to compare outcomes for neoadjuvantly treated clinical stage II patients?

These questions are posed as we eagerly await the results of a recently completed single-arm study using a single preoperative dose of pembrolizumab followed by operation and 1 year of adjuvant pembrolizumab for patients with clinical stage IIB/C cutaneous melanoma.⁵ The primary endpoint of that trial is the SLN metastasis rate benchmarked against historical controls, which was selected as a surrogate for pathologic response rate. That study has been presented in abstract form, but full results are still pending, which will help to determine whether SLN metastasis rate is a robust endpoint.

The second issue of the immunologic advantage for neoadjuvant treatment does have theoretical support. First described by Alistair Cochran et al.,⁶ It also is the most likely reservoir of tumor-specific T-cells which might be activated by preoperative immunotherapy. It is certainly a scientifically sound hypothesis that checkpoint inhibition with this the SLN appears to be selectively immune-suppressed secondary to factors released from the primary tumor site delivered as effluent into draining regional nodes.⁷ It also is the most likely reservoir of tumor-specific T cells, which might be activated by preoperative immunotherapy. It certainly is a scientifically sound hypothesis that checkpoint inhibition with this node in place will be more effective than adjuvant therapy after it has been removed.

The final issue is the topic of the article from Farooq et al.⁸ that accompanies this editorial, which reports an analysis with the express purpose of benchmarking outcomes for the population included in the prospective study mentioned earlier. These authors specifically excluded SLN status from their multivariable analysis because that information was not available at the time of neoadjuvant treatment and potentially became confounded by the treatment itself. Possibly due to limited statistical power, they found only one factor, lymphovascular invasion, that was independently associated with recurrence (it is important to note that SLN status was the most significant prognostic variable in the study group, but was omitted from the multivariable analysis). This lack of statistical significance for any other clinicopathologic variable known at the time of diagnosis is not unexpected, however, because analytics from even our largest and best databases struggle to consistently characterize prognosis for patients with stage II disease. For example, the International Melanoma Database and Discovery Platform (the basis of the most recent American Joint Committee on Cancer staging system) and the German Malignant Melanoma Registry, two quite robust platforms, produce survival estimates for stage IIB/C melanoma that differ by as much as 10–15%.^9,10

We will gain insights from further analysis of the recently completed single-arm trial analyzing both clinical outcomes and immune correlates in this important population. Additional studies, such as the Neo ReNi II trial at the Melanoma Institute Australia, will probe combination immunotherapy with PD-1 and LAG-3 inhibition in a similar stage II population.¹¹ In the end, to change clinical practice, the question of efficacy of neoadjuvant immunotherapy in clinical stage II melanoma will need to be answered in the same way we have tackled similar questions for melanoma in the past, with randomized, prospective clinical trials.

Source:

Faries, M.B., Hieken, T.J. Clinical Stage II Melanoma: Is There a “Simple Trick” to Significantly Improve Event-Free Survival?. Ann Surg Oncol 33, 8–9 (2026). https://doi.org/10.1245/s10434-025-18616-8

https://link.springer.com/article/10.1245/s10434-025-18616-8