The accurate detection of colorectal cancer (CRC) at its initial stage can reduce mortality. However, the broad application of endoscopy has been limited due to the invasive procedure and patient noncompliance. Liquid biopsy with subsequent mapping of methylation in specific cell-free DNA (cfDNA) may represent an alternative approach for early diagnosis. In this study, we have developed a minimal-invasive blood-based test for detection of precancerous lesions and early-stage CRC. Using TCGA M450K methylation data, we identified candidate methylation sites with the highest Fold Change (FC) for three genes (SEPTIN9, SDC2 and VIM), which were selected from previous studies. Based on logistic regression models, we developed a 3-gene methylation signature for CRC diagnosis with high accuracy (Sensitivity =0.959, Specificity =1, AUC =0.997). Using independent public databases and data from blood samples, this model has demonstrated superior performance. The AUC was 0.919-1 and 0.905-0.916 in public tissue database for CRC and blood sample data, respectively. Thus, our proposed 3-gene methylation signature has a more reliable performance than other methods. Furthermore, signal enhancement effect of 3-gene methylation signature can improve the accuracy of early diagnosis for CRC, which demonstrates the potential for clinical application.

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC9360219PMC

Publication Analysis

Top Keywords

methylation signature
16
early diagnosis
12
3-gene methylation
12
colorectal cancer
8
methylation
6
crc
5
septin9-sdc2-vim methylation
4
signature
4
signature biomarker
4
biomarker early
4

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!