Scholarlink

<p>Immunohistochemical (IHC) profiling is essential for ovarian tumor subtyping, but comprehensive data on antigen and protein expression frequencies is fragmented across the literature. We developed a structured, harmonized database of IHC expression profiles in ovarian tumors extracted from published studies. The dataset was constructed using large language models to identify and extract relevant data from abstracts and tables from systematically retrieved publications. Screening of 5,961 studies identified 100,149 raw tumor-stain pairs, which were processed through a multi-step curation pipeline to filter, standardize, and harmonize the records. The final dataset comprises 12,212 tumor–stain frequency records derived from 1,450 studies, with all tumor nomenclature standardized to the 2020 WHO Classification of tumors. Records include details such as expression type, total sample counts, expression percentages, subcellular localization and expression distribution. The resource provides a consolidated, evidence-based reference to support research, analyses, and the development of computational models in diagnostic pathology.</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

Ovarian Stainology: Database of evidence-based immunohistochemical antigen expression in ovarian tumors

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月30日

查看全文
  • Jan Gruszczyński,
  • Kacper Trębacz,
  • Emilian Małek,
  • Michał J. Dopierała,
  • Aleksander Gostyński-Glina,
  • Patryk Kraiński,
  • Hanna Pelant,
  • Miłosz Kadziński

摘要

Immunohistochemical (IHC) profiling is essential for ovarian tumor subtyping, but comprehensive data on antigen and protein expression frequencies is fragmented across the literature. We developed a structured, harmonized database of IHC expression profiles in ovarian tumors extracted from published studies. The dataset was constructed using large language models to identify and extract relevant data from abstracts and tables from systematically retrieved publications. Screening of 5,961 studies identified 100,149 raw tumor-stain pairs, which were processed through a multi-step curation pipeline to filter, standardize, and harmonize the records. The final dataset comprises 12,212 tumor–stain frequency records derived from 1,450 studies, with all tumor nomenclature standardized to the 2020 WHO Classification of tumors. Records include details such as expression type, total sample counts, expression percentages, subcellular localization and expression distribution. The resource provides a consolidated, evidence-based reference to support research, analyses, and the development of computational models in diagnostic pathology.