Scholarlink

<p>Immune checkpoint inhibitors (ICI) have improved clinical outcomes for some patients with advanced NSCLC, however a substantial proportion of patients remain treatment resistant. Here we analyze the NSCLC tumor microenvironment (TME) using multiplexed immunofluorescence (mIF) of biopsies taken from patients prior to ICI treatment. We apply a deep-learning model to classify the cellular phenotypes and probe functional and metabolic states of both tumor and immune cells, aiming to reveal predictive features of response to ICI. Tissue neighborhoods are generated to allow geometric profiling of spatial densities and interactions at a range of scales. Multivariate modelling of ICI response yields a model that predicts progression-free survival (PFS) over 24 months (AUC = 0.8). The selected features in the model imply a role for cell-cell proximities within discrete metabolic contexts. These tissue insights may supplement our understanding of the current paradigms around classical immunology in the NSCLC TME and its influence on immunotherapy outcomes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Metabolic characterization of tumor-immune interactions by multiplexed immunofluorescence reveals spatial mechanisms of immunotherapy response in non-small cell lung carcinoma (NSCLC)

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月3日

查看全文
  • James Monkman,
  • Aaron Kilgallon,
  • Clara Lawler,
  • Rafael Tubelleza,
  • Thazin Nwe Aung,
  • Jonathan H. Warrell,
  • Ioannis Vathiotis,
  • Ioannis P. Trontzas,
  • Niki Gavrielatou,
  • Nay Nwe Nyein Chan,
  • Rotem Czertok,
  • Shai Bookstein,
  • Ken O’Byrne,
  • Ettai Markovits,
  • David L. Rimm,
  • Arutha Kulasinghe

摘要

Immune checkpoint inhibitors (ICI) have improved clinical outcomes for some patients with advanced NSCLC, however a substantial proportion of patients remain treatment resistant. Here we analyze the NSCLC tumor microenvironment (TME) using multiplexed immunofluorescence (mIF) of biopsies taken from patients prior to ICI treatment. We apply a deep-learning model to classify the cellular phenotypes and probe functional and metabolic states of both tumor and immune cells, aiming to reveal predictive features of response to ICI. Tissue neighborhoods are generated to allow geometric profiling of spatial densities and interactions at a range of scales. Multivariate modelling of ICI response yields a model that predicts progression-free survival (PFS) over 24 months (AUC = 0.8). The selected features in the model imply a role for cell-cell proximities within discrete metabolic contexts. These tissue insights may supplement our understanding of the current paradigms around classical immunology in the NSCLC TME and its influence on immunotherapy outcomes.