부산대학교 도서관

Abstract Background Colorectal cancer (CRC) is one of the most common malignancies, and early detection plays a crucial role in enhancing curative outcomes. While colonoscopy is considered the gold standard for CRC diagnosis, noninvasive screening methods of DNA methylation biomarkers can improve the early detection of CRC and precancerous lesions. Methods Bioinformatics and machine learning methods were used to evaluate CRC‐related genes within the TCGA database. By identifying the overlapped genes, potential biomarkers were selected for further validation. Methylation‐specific PCR (MSP) was utilized to identify the associated genes as biomarkers. Subsequently, a real‐time PCR assay for detecting the presence of neoplasia or cancer of the colon or rectum was established. This screening approach involved the recruitment of 978 participants from five cohorts. Results The genes with the highest specificity and sensitivity were Septin9, AXL4, and SDC2. A total of 940 participants were involved in the establishment of the final PCR system and the subsequent performance evaluation test. A multiplex TaqMan real‐time PCR system has been illustrated to greatly enhance the ability to detect precancerous lesions and achieved an accuracy of 87.8% (95% CI 82.9–91.5), a sensitivity of 82.7% (95% CI 71.8–90.1), and a specificity of 90.1% (95% CI 84.3–93.9). Moreover, the detection rate of precancerous lesions of this assay reached 55.0% (95% CI 38.7–70.4). Conclusion The combined detection of the methylation status of SEPT9, SDC2, and ALX4 in plasma holds the potential to further enhance the sensitivity of CRC detection.