Scholarlink

Background <p>Ovarian cancer is a highly heterogeneous disease, with substantial variation in patients’ responses to neoadjuvant chemotherapy (NACT) and prognosis. Currently, specific stratification biomarkers for patient selection are lacking. Stratifying patients based on noninvasive imaging data prior to treatment could significantly improve treatment decisions.</p> Methods <p>Consensus clustering of pretreatment radiomic features from computed tomography images identified imaging subtypes in NACT patients. Five classifiers were developed for patient stratification in a subsequent cohort, and the biological behavior behind the radiomics-based clusters was explored.</p> Results <p>Consensus clustering identified four patient clusters, with consistent survival trends across both clustering and stable classifiers. Clusters 3 and 4 showed higher complete gross resection rates, longer overall survival, and more BRCA1 mutations, while clusters 1 and 2 had poorer prognoses and lower complete gross resection rates. Despite a higher tumor stage, cluster 4 was more sensitive to chemotherapies and targeted therapies. Cluster 2 had the lowest proportion of well-differentiated tumors, and upregulated genes were enriched in pathways related to epithelial mesenchymal transition and other classic cancer pathways. Weighted gene co-expression network analysis identified key modules and hub genes in clusters 1 and 2, with RNA signature scoring used for validation.</p> Conclusions <p>Stratification based on noninvasive imaging revealed distinct biological patterns among ovarian cancer patients, helping to identify subgroups with favorable responses to neoadjuvant chemotherapy and better survival outcomes, thereby facilitating more personalized treatment decisions.</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

CT radiomic stratification signature to optimize clinical decisions for ovarian cancer patients receiving neoadjuvant chemotherapy and the underlying biological basis: a multicenter retrospective study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月28日

查看全文
  • Shimin Zhang,
  • Xinyue Li,
  • Siyang Zhang,
  • Meiyao Li,
  • Jiarui Wang,
  • Jiage Zhang,
  • Yutong Liu,
  • Zhuonan Liu,
  • Guan Wang,
  • Yi Zhang,
  • Zhuo Yang,
  • Jiangdian Song,
  • Xin Zhou

摘要

Background

Ovarian cancer is a highly heterogeneous disease, with substantial variation in patients’ responses to neoadjuvant chemotherapy (NACT) and prognosis. Currently, specific stratification biomarkers for patient selection are lacking. Stratifying patients based on noninvasive imaging data prior to treatment could significantly improve treatment decisions.

Methods

Consensus clustering of pretreatment radiomic features from computed tomography images identified imaging subtypes in NACT patients. Five classifiers were developed for patient stratification in a subsequent cohort, and the biological behavior behind the radiomics-based clusters was explored.

Results

Consensus clustering identified four patient clusters, with consistent survival trends across both clustering and stable classifiers. Clusters 3 and 4 showed higher complete gross resection rates, longer overall survival, and more BRCA1 mutations, while clusters 1 and 2 had poorer prognoses and lower complete gross resection rates. Despite a higher tumor stage, cluster 4 was more sensitive to chemotherapies and targeted therapies. Cluster 2 had the lowest proportion of well-differentiated tumors, and upregulated genes were enriched in pathways related to epithelial mesenchymal transition and other classic cancer pathways. Weighted gene co-expression network analysis identified key modules and hub genes in clusters 1 and 2, with RNA signature scoring used for validation.

Conclusions

Stratification based on noninvasive imaging revealed distinct biological patterns among ovarian cancer patients, helping to identify subgroups with favorable responses to neoadjuvant chemotherapy and better survival outcomes, thereby facilitating more personalized treatment decisions.