Non-invasive diagnosis for urothelial carcinoma using a dual-target DNA methylation biomarker panel

Abstract

Background

Urothelial carcinoma (UC) is a common malignancy worldwide. Aberrant DNA methylation is implicated in UC carcinogenesis. This study sought to delineate the DNA methylation landscape in UC and identify DNA methylation-based biomarkers for early detection of UC.

Methods

Whole genome bisulfite sequencing (WGBS) was conducted on bladder cancer tissues and paired normal tissues. By integrating WGBS data with The Cancer Genome Atlas (TCGA) UBC data, a DNA methylation-based biomarker was identified. When combined with a known UC biomarker AL021918.2, the performance of the dual-target test was evaluated in voided urine samples from 224 UC patients and 419 controls.

Results

Notable hypomethylation was observed in UC samples compared to normal samples. Through differential methylation analysis, differential methylation CpG sites, regions, and genes were identified. Of these, Transmembrane protein 106A gene (TMEM106A) was screened as a new UC biomarker. In a dual-target test, using triplex quantitative methylation-specific PCR (qMSP) to examine TMEM106A and AL021918.2 methylation levels, the training set showed a sensitivity of 89.0 %, a specificity of 92.9 %, and an area under the curve (AUC) value of 0.941 (95 % confidence interval [CI]: 0.913–0.969). Similarly, the validation set showed a sensitivity of 90.0 %, a specificity of 91.1 %, and an AUC value of 0.922 (95 % CI: 0.881–0.962). In addition, our dual-target test demonstrated outstanding detection rates for low-grade or early-stage tumors.

Conclusions

We provide a comprehensive analysis of DNA methylation profiles in UC, and highlight the promising clinical potential of dual-target urine tests for UC detection.