评估循环游离 DNA (cfDNA) 中 5-羟甲基胞嘧啶 (5hmC) 对鼻咽癌 (NPC) 的诊断价值并开发诊断模型。来自 174 名鼻咽癌患者和 146 名非癌症患者的 cfDNA 中的全基因组 5hmC 谱使用 5hmC-Seal 技术对个体进行分析。使用最小绝对收缩和选择算子 (LASSO) 逻辑回归开发了基于 cfDNA 5hmC 的诊断模型来识别 NPC 患者，并使用受试者工作特征 (ROC) 曲线和混淆矩阵评估性能。来自 NPC 患者的 5hmC-Seal 数据NPC 显示出与非肿瘤样本不同的全基因组分布。我们的初步分析表明，基于 12 个基因的 5hmC 标记物组合是一种准确的诊断模型，可有效区分 NPC 样本和非癌性样本（训练集：曲线下面积 (AUC)= 0.97 [95% CI: 0.94-0.99] ；测试集：AUC= 0.93 [95% CI：0.88-0.98]）优于 EBV DNA 测试。诊断评分在区分非癌症受试者和早期 NPC 方面表现良好（训练集：AUC=0.99 [95% CI：0.98-1]；测试集：AUC=0.98 [95% CI：0.95-1]）和晚期 NPC（训练集：AUC=0.96 [95% CI：0.93-0.99]；测试集：AUC=0.93 [95% CI：0.88-0.98]）。值得注意的是，在 EBV 阴性患者中，在训练集 (AUC= 0.94 [95% CI: 0.88-1]) 和测试集 ( AUC=0.91 [95% CI: 0.81-1]).5hmC cfDNA 修饰是有前景的 NPC 非侵入性生物标志物，具有高灵敏度和特异性，特别是对于早期和 EBV 阴性 NPC。版权所有 © 2024。由 Elsevier Ltd 出版。

To evaluate the diagnostic value of 5-hydroxymethylcytosines (5hmC) in circulating cell-free DNA (cfDNA) for nasopharyngeal carcinoma (NPC) and to develop a diagnostic model.Genome-wide 5hmC profiles in cfDNA from 174 NPC patients and 146 non-cancer individuals were analyzed using the 5hmC-Seal technique. A cfDNA 5hmC-based diagnostic model to identify NPC patients was developed using least absolute shrinkage and selection operator (LASSO) logistic regression, and performance was evaluated with receiver operating characteristic (ROC) curves and confusion matrices.The 5hmC-Seal data from patients with NPC showed a different genome-wide distribution than non-tumor samples. Our initial analysis revealed a 12-gene-based 5hmC marker panel to be an accurate diagnostic model effectively distinguishing between NPC samples and non-cancerous samples (training set: area under curve (AUC)= 0.97 [95 % CI: 0.94-0.99]; and test set: AUC= 0.93 [95 % CI: 0.88-0.98]) superior to EBV DNA testing. The diagnostic score performed well in differentiating the non-cancer subjects from early-stage NPC (training set: AUC=0.99 [95 % CI: 0.98-1]; test set: AUC=0.98 [95 % CI: 0.95-1]), and advanced-stage NPC (training set: AUC=0.96 [95 % CI: 0.93-0.99]; test set: AUC=0.93 [95 % CI: 0.88-0.98]). Notably, in EBV-negative patients, the diagnostic scores showed excellent capacity for distinguishing EBV-negative patients with NPC from non-cancer subjects in both the training set (AUC= 0.94 [95 % CI: 0.88-1]) and test set (AUC=0.91 [95 % CI: 0.81-1]).5hmC modifications in cfDNA are promising noninvasive biomarkers for NPC, offering high sensitivity and specificity, particularly for early-stage and EBV-negative NPC.Copyright © 2024. Published by Elsevier Ltd.