Scholarlink

<p>Recent advances in spatial biology can reveal how tissue organization changes in disease; however, interpreting these datasets in a generalized, scalable way remains challenging. Existing computational approaches rely on pairwise comparisons or unsupervised clustering, which can lack statistical rigor and miss rare, clinically relevant cellular niches. Here we present QUICHE—an automated and scalable statistical framework designed to discover cellular niches differentially enriched in populations, histological structures or acellular regions. Using in silico models and spatial proteomic imaging of human tissues, we show that QUICHE can accurately detect low-prevalence, condition-specific niches, outperforming the next best algorithm threefold. To investigate how tumor structure influences recurrence risk in triple-negative breast cancer, we applied QUICHE to a multicenter spatial proteomics cohort of 314 primary tumor resections. We discovered niches consistently enriched in tumor border and extracellular-matrix-remodeling regions, including those associated with recurrence-free survival. These findings were validated in two independent cohorts, suggesting that antitumor responses are driven by coordinated engagement between innate and adaptive immune cells, rather than any single population. QUICHE is provided as an open-source Python package (<a href="https://github.com/jranek/quiche">https://github.com/jranek/quiche</a>).</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

The automated computational workflow QUICHE reveals structural definitions of antitumor responses in triple-negative breast cancer

  • 期刊论文

  • 发表日期：

    2026年2月18日

查看全文
  • Jolene S. Ranek,
  • Noah F. Greenwald,
  • Mako Goldston,
  • Christine Camacho Fullaway,
  • Cameron Sowers,
  • Alex Kong,
  • Silvana Mouron,
  • Miguel Quintela-Fandino,
  • Robert B. West,
  • Sean C. Bendall,
  • Michael Angelo

摘要

Recent advances in spatial biology can reveal how tissue organization changes in disease; however, interpreting these datasets in a generalized, scalable way remains challenging. Existing computational approaches rely on pairwise comparisons or unsupervised clustering, which can lack statistical rigor and miss rare, clinically relevant cellular niches. Here we present QUICHE—an automated and scalable statistical framework designed to discover cellular niches differentially enriched in populations, histological structures or acellular regions. Using in silico models and spatial proteomic imaging of human tissues, we show that QUICHE can accurately detect low-prevalence, condition-specific niches, outperforming the next best algorithm threefold. To investigate how tumor structure influences recurrence risk in triple-negative breast cancer, we applied QUICHE to a multicenter spatial proteomics cohort of 314 primary tumor resections. We discovered niches consistently enriched in tumor border and extracellular-matrix-remodeling regions, including those associated with recurrence-free survival. These findings were validated in two independent cohorts, suggesting that antitumor responses are driven by coordinated engagement between innate and adaptive immune cells, rather than any single population. QUICHE is provided as an open-source Python package (https://github.com/jranek/quiche).