Scholarlink

Purpose <p>Malignancy evaluation of thyroid nodules is an important and challenging issue in clinical practice. In this study, we intended to explore novel biomarkers for the differential diagnosis of malignant and benign thyroid nodules.</p> Methods <p>We performed whole-genome DNA methylation profiling by MethylationEPIC BeadChip and expression profiling by RNA-Sequencing, to identify differentially methylated and expressed genes in fresh-frozen tissues from early-stage papillary thyroid cancer (PTC) and benign thyroid nodule (BTN) patients. The results were validated first in two GEO datasets, and then in two new independent studies (Validation Ⅰ and II) with a total of 757 formalin-fixed paraffin-embedded samples using mass spectrometry. BRAFV600E mutation was examined by Sanger Sequencing. Binary logistic regression analysis was performed to calculate odds ratios (ORs) for per 10% decrease in methylation level.</p> Results <p>The <i>PARP4</i> gene was identified in the discovery round and verified in the GEO datasets. Compared to BTNs, PTC cases presented a significant hypomethylation of <i>PARP4</i> in both Validation Ⅰ (odds ratios ORs &gt; 1.72 per 10% decrease in methylation level, all <i>p</i>≤1.733 × 10<sup>− 13</sup>) and Validation Ⅱ (ORs &gt; 1.60 per 10% decrease in methylation level, all <i>p</i>≤1.488 × 10<sup>− 13</sup>). The predictive model was constructed with great performance in Validation Ⅰ (AUC: 0.825) and Validation Ⅱ (AUC: 0.815). Besides, we observed significant correlations between <i>PARP4</i> methylation and clinical characteristics of early-stage PTC. Totally, the AUCs for <i>PARP4</i> methylation alone and in combination with BRAFV600E mutation were 0.819 and 0.873, respectively.</p> Conclusions <p>We disclosed significant <i>PARP4</i> hypomethylation in early-stage PTCs compared to BTNs, and constructed a predictive model as a promising tool for differentiating early-stage PTCs from BTNs.</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

Discovery and validation of PARP4 methylation in distinguishing thyroid cancer from benign nodule

  • 期刊论文

  • 发表日期：

    2025年12月22日

查看全文
  • Mengxia Li,
  • Yifei Yin,
  • Yizhu Mao,
  • Junjie Li,
  • Haixia Huang,
  • Minmin Zhang,
  • Hong Li,
  • Chenxia Jiang,
  • Rongxi Yang

摘要

Purpose

Malignancy evaluation of thyroid nodules is an important and challenging issue in clinical practice. In this study, we intended to explore novel biomarkers for the differential diagnosis of malignant and benign thyroid nodules.

Methods

We performed whole-genome DNA methylation profiling by MethylationEPIC BeadChip and expression profiling by RNA-Sequencing, to identify differentially methylated and expressed genes in fresh-frozen tissues from early-stage papillary thyroid cancer (PTC) and benign thyroid nodule (BTN) patients. The results were validated first in two GEO datasets, and then in two new independent studies (Validation Ⅰ and II) with a total of 757 formalin-fixed paraffin-embedded samples using mass spectrometry. BRAFV600E mutation was examined by Sanger Sequencing. Binary logistic regression analysis was performed to calculate odds ratios (ORs) for per 10% decrease in methylation level.

Results

The PARP4 gene was identified in the discovery round and verified in the GEO datasets. Compared to BTNs, PTC cases presented a significant hypomethylation of PARP4 in both Validation Ⅰ (odds ratios ORs > 1.72 per 10% decrease in methylation level, all p≤1.733 × 10− 13) and Validation Ⅱ (ORs > 1.60 per 10% decrease in methylation level, all p≤1.488 × 10− 13). The predictive model was constructed with great performance in Validation Ⅰ (AUC: 0.825) and Validation Ⅱ (AUC: 0.815). Besides, we observed significant correlations between PARP4 methylation and clinical characteristics of early-stage PTC. Totally, the AUCs for PARP4 methylation alone and in combination with BRAFV600E mutation were 0.819 and 0.873, respectively.

Conclusions

We disclosed significant PARP4 hypomethylation in early-stage PTCs compared to BTNs, and constructed a predictive model as a promising tool for differentiating early-stage PTCs from BTNs.