Circulating tumour cells (CTCs) provide a minimally invasive platform for disease monitoring and molecular profiling in colorectal carcinoma (CRC). This study aimed to identify prognostically relevant DNA methylation biomarkers and assess their translational potential in CTC-based liquid biopsy. Candidate CpG sites were first screened in silico using differential methylation analysis, LASSO regression, and Cox proportional hazards modelling based on The Cancer Genome Atlas-Colon Adenocarcinoma (TCGA-COAD) dataset, which included 430 tumour and 39 non-tumour colon mucosa samples. Three representative CpG sites (GPR125, GDNF, and ADCY9) were subsequently validated in an independent cohort of 62 CRC patients using quantitative methylation-specific polymerase chain reaction (PCR), with paired comparisons between CTCs and tumour tissues available in 32 cases. A five-CpG prognostic model derived from TCGA data was significantly associated with overall survival (hazard ratio = 4.23, p < 0.001), with time-dependent area under the curves (AUC)s of 0.718, 0.765, and 0.787 at 1, 3, and 5 years. In tumour tissues, GDNF methylation showed significant associations with age, tumour site, histological grade, T stage, and lymphovascular invasion, while GPR125 methylation was associated with lymphovascular invasion, and ADCY9 showed no significant clinicopathological correlations. Network analysis further identified GDNF as a distinct methylation module linked to multiple adverse pathological features. These findings suggest that DNA methylation biomarkers, particularly GDNF, are closely associated with tumour aggressiveness in CRC and support their potential application in both tissue-based assessment and CTC-driven molecular profiling.
Chen et al. (Wed,) studied this question.