Spatial characterization and predictive implications of the tumor immune microenvironment in stage III melanoma under neoadjuvant immunotherapy.

Y Yichen Zhang (Institute of Carbon Neutrality, Sino-French Institute for Earth System Science, College of Urban and Environmental Sciences, Peking University) E Emma Wagner (Stanford University) A Alexxandra Hurtado (Charles R. Drew University of Medicine and Science, Los Angeles, CA) J Jay Rajesh Chadokiya (Stanford University School of Medicine, Stanford, CA) P Pauline Funchain A Amanda R. Kirane (Stanford University School of Medicine, Stanford, CA) A Allison Betof (Stanford University School of Medicine, Stanford, CA)

Abstract

e21534 Background: Despite recent trials establishing neoadjuvant immune checkpoint blockade (ICB) as the standard of care for clinical stage III melanoma, nearly half of patients fail to achieve adequate pathologic response and clinical trajectory remains highly variable. These limitations emphasize the need for robust biomarkers to guide patient selection. Here, we used spatial proteomics to characterize the tumor immune microenvironment before and after neoadjuvant ICB, and to identify features of favorable response. Methods: Samples were collected from 10 stage III melanoma patients who underwent neoadjuvant nivolumab/relatlimab (n = 7) or ipilimumab/nivolumab (n = 3). Pre-ICB samples were obtained via core biopsy or fine-needle aspiration at diagnosis, while post-ICB samples were from lymph node dissection. Clinical outcomes were documented by pathologist assessment and systematic chart review. Multiplex immunofluorescence was performed using the Lunaphore COMET. A custom Python-based pipeline was created for data analysis after image acquisition. Results: Spatial mapping of annotated cell types revealed marked differences in immune cell distribution and composition correlating with treatment response. Pre-ICB samples from patients with major pathologic response (MPR, ≤10% residual viable tumor) showed decreased segregation between immune-rich and tumor regions, whereas post-ICB samples revealed response-dependent patterns of immune infiltration: non-MPR patients with rapid clinical progression (n = 3) exhibited sparse immune cells within the tumor bed, non-MPR patients with clinically stable disease (n = 3) showed immune aggregates, and MPR patients (n = 4) demonstrated dense immune infiltration. From a compositional standpoint, MPR patients exhibited a higher relative abundance of immune cells with lymphoid predominance, including increased CD8 + :CD4 + T cell and CD8 + T cell:M2 macrophage (CD68 + CD163 + ) ratios in both pre- and post-ICB samples. Within this group, CD8 + T cells displayed lower post-ICB expression of PD-1 and LAG-3. Notably, CD8 + HLA-DR + NK cells (CD3 - CD56 + ) were also uniquely identified in B cell clusters post-ICB in MPR patients, suggesting an immunoregulatory phenotype with cytotoxic potential. In contrast, non-MPR patients were characterized by a CD4 + T cell predominant lymphoid population and a higher relative abundance of myeloid cells after treatment. Conclusions: Collectively, these findings highlight CD8 + T cell-centric spatial immune features as potential predictive biomarkers of response to neoadjuvant ICB in stage III melanoma. The emergence of CD8 + HLA-DR + NK cells further suggests coordinated innate-adaptive immune crosstalk beyond adaptive enhancement alone. Moreover, in non-MPR patients, persistence of immune aggregates post-ICB may help distinguish a favorable clinical trajectory.

Article Details

Volume / Issue Vol. 44, Issue 16_suppl
Published June 01, 2026
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (7)

Y

Yichen Zhang

Institute of Carbon Neutrality, Sino-French Institute for Earth System Science, College of Urban and Environmental Sciences, Peking University

E

Emma Wagner

Stanford University

A

Alexxandra Hurtado

Charles R. Drew University of Medicine and Science, Los Angeles, CA

J

Jay Rajesh Chadokiya

Stanford University School of Medicine, Stanford, CA

P

Pauline Funchain

A

Amanda R. Kirane

Stanford University School of Medicine, Stanford, CA

A

Allison Betof

Stanford University School of Medicine, Stanford, CA