Real-World Evaluation of Uromonitor® for Bladder Cancer Detection and Surveillance.

BACKGROUND

Surveillance of non-muscle-invasive bladder cancer (NMIBC) relies on cystoscopy and urine cytology, both of which have well-recognised limitations. Molecular urine assays have been developed to reduce the burden of invasive surveillance, yet their real-world clinical utility remains uncertain. Uromonitor® is a quantitative PCR-based assay targeting hotspot variants in the TERT promoter, FGFR3, and KRAS, which are frequently altered in urothelial carcinoma. We evaluated the performance of Uromonitor® in routine clinical practice and assessed its technical reproducibility.

METHODS

Uromonitor® diagnostic test accuracy was retrospectively calculated from samples from patients undergoing investigation for suspected bladder cancer (n = 64) or surveillance (n = 30) following a prior diagnosis at a tertiary referral centre between 2021 and 2023. Uromonitor® results were compared with histology where available (n = 49, 52%), or with contemporaneous cystoscopy and urine cytology findings (n = 45, 48%). This pragmatic dual reference standard reflects routine clinical practice but may introduce some heterogeneity in diagnostic accuracy verification. A prospective in-house verification cohort was used to assess inter-laboratory reproducibility. Discordant cases underwent orthogonal next-generation sequencing (NGS) analysis. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy were calculated for the Uromonitor® against the standard of care.

RESULTS

Ninety-four patients were included in the clinical performance analysis. Overall sensitivity, specificity, PPV, NPV and overall accuracy for Uromonitor® were 38%, 88%, 63%, 72% and 70%, respectively. Sensitivity was higher in the diagnostic setting (47%; 95% CI 27.3-68.3%) than during surveillance (23%; 95% CI 8.2-50.2%). Several false-negative cases in the verification cohort harboured variants either detectable by NGS at variant allele frequencies below or slightly above the assay's limit of detection or variants not covered by the assay hotspot design. Inter-laboratory reproducibility was excellent, with 100% concordance observed in the verification cohort.

CONCLUSIONS

In a real-world clinical setting, Uromonitor® demonstrated high specificity but limited sensitivity for detection of bladder cancer, particularly during surveillance. A negative result does not reliably exclude recurrence. Assay sensitivity thresholds and restricted variant coverage appear to be key contributors to false-negative results. These findings highlight the need for cautious clinical integration of Uromonitor®. It is unclear whether this approach has sufficient sensitivity in surveillance to safely reduce cystoscopy frequency. This underscores the need for further refinement of urine-based molecular assays, including a need for enhanced sensitivity and broader mutational coverage before routine clinical adoption.