Scholarlink

<p>Prostate-specific antigen (PSA) testing lacks specificity due to benign conditions. This study assessed whether integrating DNA methylation signatures with PSA metrics improves discrimination of prostate cancer (PCa) from non-PCa cases.&#xa0;Targeted bisulfite sequencing of 74 CpG sites in 9 genes was performed on urine sediment from 194 patients (89 PCa, 105 non-PCa). Patients were stratified by total PSA (tPSA) and free-to-total PSA ratio (%fPSA) into risk groups. Random forest identified discriminatory methylation markers, and support vector machine (SVM), k-nearest neighbor (KNN), and Bayesian models were constructed and validated.&#xa0;PSA-based stratification alone could not reliably distinguish PCa from non-PCa, with 26.80% of patients falling into a diagnostic “gray zone”. Twenty CpG sites showed significant differential methylation (<i>p</i> &lt; 0.01). In addition, several CpG loci exhibited methylation differences across PSA risk groups. Models based on random forest-selected markers achieved strong diagnostic performance, with KNN yielding the highest accuracy (AUC &gt; 0.95, 100% specificity), and consistently high sensitivity (&gt; 80%), maintaining strong discriminative power even in PSA “gray-zone” cases.&#xa0;Urine-based DNA methylation profiling enhances the diagnostic accuracy of PCa beyond PSA testing, particularly in cases within the PSA gray zone. Validation in larger independent cohorts is warranted to establish its clinical utility.</p> Graphical Abstract <p></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

Urinary DNA Methylation Profiling Improves Discrimination of Prostate Cancer Across PSA-Defined Risk Strata

  • 期刊论文

  • 发表日期：

    2025年11月17日

查看全文
  • Wei Zhu,
  • Yi Qian,
  • Xiaokai Zhao,
  • Zhenxuan Fang,
  • Zeyu Luo,
  • Wenhua Xie,
  • Yifang Cao,
  • Wei Chen,
  • Huiyu Fu,
  • Jiayu Peng,
  • Lijun Zhang,
  • Jieyi Li,
  • Siyu Lei,
  • Jing Jin,
  • Ziying Gong,
  • Daoyun Zhang,
  • Yi He

摘要

Prostate-specific antigen (PSA) testing lacks specificity due to benign conditions. This study assessed whether integrating DNA methylation signatures with PSA metrics improves discrimination of prostate cancer (PCa) from non-PCa cases. Targeted bisulfite sequencing of 74 CpG sites in 9 genes was performed on urine sediment from 194 patients (89 PCa, 105 non-PCa). Patients were stratified by total PSA (tPSA) and free-to-total PSA ratio (%fPSA) into risk groups. Random forest identified discriminatory methylation markers, and support vector machine (SVM), k-nearest neighbor (KNN), and Bayesian models were constructed and validated. PSA-based stratification alone could not reliably distinguish PCa from non-PCa, with 26.80% of patients falling into a diagnostic “gray zone”. Twenty CpG sites showed significant differential methylation (p < 0.01). In addition, several CpG loci exhibited methylation differences across PSA risk groups. Models based on random forest-selected markers achieved strong diagnostic performance, with KNN yielding the highest accuracy (AUC > 0.95, 100% specificity), and consistently high sensitivity (> 80%), maintaining strong discriminative power even in PSA “gray-zone” cases. Urine-based DNA methylation profiling enhances the diagnostic accuracy of PCa beyond PSA testing, particularly in cases within the PSA gray zone. Validation in larger independent cohorts is warranted to establish its clinical utility.

Graphical Abstract