Comparison and combination of mutation and methylation-based urine tests for bladder cancer detection

Abstract

Background and aims

Several non-invasive tests for detecting bladder cancer (BC) are commercially available and are based on detecting small panels of BC-associated mutations and/or methylation changes in urine DNA. However, it is not clear which type of biomarker is best, or if a combination of the two is needed. In this study we address this question by taking a 23-gene mutation panel (GALEAS™ Bladder, GB) and testing if adding a panel of methylation markers improves the sensitivity of BC detection.

Methods

Twenty-three methylation markers were assessed in urine DNA by bisulphite conversion, multiplex PCR, and next generation sequencing in 118 randomly selected haematuria patients with pre-existing GB data (56 BCs and 62 non-BCs), split into training and test sets. We also analysed an additional 16 GB false-negative urine DNAs.

Results

The methylation panel detected bladder cancer in haematuria patients with 69% sensitivity at 96% specificity (test set results, 95% CIs 52-87% and 80-99%, respectively). Corresponding sensitivity and specificity for GB were 92% and 89%. Methylation and mutation markers were highly concordant in urine, with all GB false-negative samples also negative for methylation markers.

Conclusions and limitations

Our data show that, with a comprehensive mutation panel, any gains from adding methylation markers are, at best, marginal. It is likely that low tumour content is the commonest cause of false-negative urine test results. Our study does have a limited sample size and other methylation markers might behave differently to the those studied here.