Exploiting Urine Derived DNA for the Assessment of Bladder Cancer using High Accuracy Sequencing

PROJECT SUMMARY Many attempts have been made to develop urine-based biomarkers for bladder cancer surveillance and monitoring. However, the most widely used approach, urine cytology, is most effective for high-grade lesions.

To date, none of the current FDA-approved tests have been widely adopted due to low sensitivities (55%?70%) and specificities (71%?83%).

Performance is poor for low-grade tumors due to the low abundance of aneuploid cells and high levels of interobserver variability.

Previous studies identified several recurrently mutated genes occurring in 70-80% of both muscle invasive and non-muscle invasive bladder cancers (MIBC and NMIBC, respectively).

Detection of these mutations could help in early cancer detection, initial stratification for treatment options, detection of minimal residual disease, or identification of emerging chemotherapy resistance.

As with most other solid cancers, accessing tumor tissue either by biopsy or surgical resection is often limited or unobtainable. Furthermore, these characteristically small samples are not necessarily representative of the entire tumor.

For this reason, tumor cells and/or DNA shed into the urine holds the promise of yielding detailed genetic information about a tumor using a simple, non-invasive, urine test.

The advent of next-generation sequencing (NGS) has opened up the possibility of clinically exploiting DNA as a cancer monitoring analyte.

However, high error rates of NGS has proven to be a major impediment for using this technology for low frequency variant detection. To overcome this limitation, we have previously developed Duplex Sequencing (Duplex-Seq). Using this technique we can detect a single variants present in ~5x107 wild-type bases.

We hypothesize that assessment of urine derived DNA (uDNA) by Duplex-Seq will perform better than urine cytology or conventional NGS-based approaches for detecting post-treatment residual cancer. Such information would eventually be used in determining treatment response and clinical decision making.

We propose to develop and validate the use of Duplex-Seq for use in frequently encountered clinical scenarios, as well as in normal, cancer free, individuals.

In Specific Aim 1, we will determine the biological occurrence of bladder cancer-associated mutations in normal individuals and establish age-defined thresholds across several different frequently mutated genes.

In Specific Aim 2, we will determine if the presence or absence of NMIBC tumor mutations in uDNA is predictive of recurrence at the conclusion of intravesical BCG therapy, potentially reducing the need for repeated and unpleasant cystoscopies.