肺癌
计算生物学
小RNA
癌症
肿瘤科
靶向治疗
医学
分子诊断学
临床实习
工作流程
聚合酶链反应
疾病
诊断准确性
实时聚合酶链反应
计算机科学
减法
伴生诊断
比较器
生物信息学
肺病
内科学
癌症研究
诊断试验
临床决策
临床意义
生物
临床诊断
分子成像
作者
Jiajia Song,Ruiting Liu,Linguo Gu,Kemin Wang,Zhenkun Xia,Qiuping Guo,Jin Lin Huang
出处
期刊:Small
[Wiley]
日期:2026-11-17
卷期号:: e76763-
摘要
Early and precise cancer diagnosis is crucial for effective treatment. MicroRNAs (miRNAs) are promising diagnostic targets, but translating their complex, multi-dimensional expression signatures into clinical practice remains challenging. Here, we develop a concise molecular comparator using relative expression levels of specific miRNA pairs. The miRpair-based strategy demonstrates superior diagnostic performance compared to methods relying on the absolute quantification of single miRNAs. For non-small cell lung cancer (NSCLC) diagnosis, we identify hsa-miR-148a-3p/hsa-miR-30d-5p as the optimal diagnostic pair from The Cancer Genome Atlas (TCGA) data, with an area under the curve (AUC) of 0.975. The silicon-based model achieves 90.3% accuracy in 93 clinical serum samples using quantitative analysis. We subsequently translate the comparator into a DNA-based molecular workflow via asymmetric polymerase chain reaction (PCR) and polymerase-assisted hybridization subtraction (PAH-Sub). This process performs threshold-based comparison of dual-miRNA relative abundance directly at the molecular level. Integration with catalytic hairpin assembly (CHA) amplification generates fluorescent diagnostic outputs. Validation with 40 clinical serum samples yielded 87.5% accuracy, demonstrating the feasibility of molecular-level implementation of the miRpair decision rule. This study represents a proof-of-concept evaluation for NSCLC diagnosis. Further validation in independent cohorts incorporating clinically relevant disease controls will be required to establish broader clinical specificity.
科研通智能强力驱动
Strongly Powered by AbleSci AI