Analysis of spatial atlas of ADC targets in immunotherapy-treated NSCLC to link compartment-specific target expression with tumor immune contexture, clinical outcome, and on-treatment remodeling.

J Jean-Philippe Guegan (Explicyte, Bordeaux, France) Z Zachary Cooper (AstraZeneca, North Potomac, MD) K Krista Kinneer (1AstraZeneca, Gaithersburg, United States) J Jixin Wang (State Key Laboratory of Bioinspired Interfacial Materials Science, College of Chemistry, Chemical Engineering and Materials Science) F Fabrice Barlesi D David Planchard B Benjamin Besse J Jordi Remon Masip (Gustave Roussy, Paris, France) A Alban Bessede (Explicyte, Bordeaux, France) L Laura Dillon (1Fralin Biomedical Research Institute, Virginia Tech FBRI Cancer Research Center, Washington, United States) A Antoine Italiano (Gustave Roussy, Villejuif, France)

Abstract

3004 Background: The spatial distribution, co-expression patterns, immune contexture associations, and on-treatment evolution of ADC targets in NSCLC is unknown. Methods: The samples of 100 NSCLC patients treated with PD-1/PD-L1–based immunotherapy (STING NCT04932525) were analyzed using multiplex immunofluorescence with three 7-plex panels assessing ADC targets (Claudin18.2, TROP2, B7-H3, B7-H4), receptor co-expression (HER2, EGFR, MET, STEAP2), and immune contexture (CD8, PD-1, TIM3, LAG3, TIGIT), with PanCK and CD45 for compartment assignment. Digital pathology enabled single-cell segmentation into tumor (PanCK+), stromal, and immune (CD45+) compartments. Immune contexture was defined as Desert, Excluded, or Infiltrated by CD8 distribution. ADC target expression was quantified as the proportion of PanCK+ tumor cells expressing each target or combination. Associations with objective response (ORR), progression-free survival (PFS), and overall survival (OS) were assessed. Paired baseline/on-treatment biopsies (n = 21 pairs) evaluated treatment-induced remodeling. Results: Spatial profiling quantified 30 tumor-cell ADC target phenotypes. Across NSCLC, epithelial target coverage was broad, with > 80% of tumor cells expressing EGFR or MET and > 60% co-expressing both. Co-expression patterns were histology-dependent: EGFR/MET (±HER2) predominated in adenocarcinoma, whereas TROP2/B7-H3 co-expression was enriched in squamous carcinoma. Immune contexture shaped the ADC landscape: Infiltrated tumors showed higher tumor-cell fractions expressing EGFR, MET, TROP2, HER2, STEAP2, and B7-H3 compared with Desert tumors (all p≤0.01), while PanCK+/B7-H4 and PanCK+/CLDN18 were not contexture-associated. Clinically, B7-H4 showed a striking compartment-specific association with outcome. High abundance of B7-H4–expressing immune/stromal cells was the only phenotype replicated across ORR, PFS, and OS (stromal B7H4+: ORR 65.0% vs 38.6%, p = 0.010; PFS 12.0 vs 4.0, p = 0.013; OS 36.4 vs 13.1, p = 0.016). In contrast, tumor-cell B7-H4 predicted resistance and poor survival (tumoral B7H4+: PFS 3.5 vs 10.2, p = 0.007; OS 10.9 vs 35.6, p = 0.047). On treatment, paired biopsies demonstrated target remodeling, with increased B7-H3 expression and a concordant reduction of tumor CLDN18 by composition and membrane intensity. Conclusions: ADC target abundance and co-expression in NSCLC tumor cells are immune-contexture dependent and dynamically remodeled under ICI. Distinct, histology-specific co-expression patterns inform rational mono- and dual-ADC strategies. A novel, clinically actionable finding is the opposite prognostic impact of B7-H4 by spatial compartment, supporting spatially resolved biomarkers and histology- and timepoint-aware ADC±ICI selection for precision treatment in NSCLC.

Article Details

Volume / Issue Vol. 44, Issue 16_suppl
Published June 01, 2026
Pages 3004-3004
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (11)

J

Jean-Philippe Guegan

Explicyte, Bordeaux, France

Z

Zachary Cooper

AstraZeneca, North Potomac, MD

K

Krista Kinneer

1AstraZeneca, Gaithersburg, United States

J

Jixin Wang

State Key Laboratory of Bioinspired Interfacial Materials Science, College of Chemistry, Chemical Engineering and Materials Science

F

Fabrice Barlesi

D

David Planchard

B

Benjamin Besse

J

Jordi Remon Masip

Gustave Roussy, Paris, France

A

Alban Bessede

Explicyte, Bordeaux, France

L

Laura Dillon

1Fralin Biomedical Research Institute, Virginia Tech FBRI Cancer Research Center, Washington, United States

A

Antoine Italiano

Gustave Roussy, Villejuif, France