Reading Tumor Activity Through Urine RNA: Bladder Cancer Detection Reaches 95% Sensitivity

For patients with bladder cancer, tumor removal is often just the beginning of a long period of monitoring. Cancer may recur, and treatment may not work, making repeated examinations part of life. A Stanford University team is now trying to obtain more clues from RNA fragments in urine: beyond identifying whether cancer is present, the researchers are exploring whether tumor biological activity can predict treatment response. The study was published in *Nature Medicine* on October 2.

The technique, called uRARE-seq, analyzes cell-free RNA in urine—RNA fragments released by cells that are not contained within intact cells. The researchers adapted the sequencing method to account for the scarcity and fragmentation of urine RNA, capturing gene signals that are uncommon in healthy urine but more abundant when tumors are present, then building a model to identify bladder cancer. RNA reflects how genes are functioning, so it may also provide information about tumor status.

The study analyzed 683 urine samples, rather than 683 patients with bladder cancer. Among the 251 samples used to develop the model, detection of localized bladder cancer reached 95% sensitivity and 90% specificity. Performance was similar when the fixed model was applied to a separate validation cohort comprising 142 patients and 140 non-cancer controls. These two percentages describe the ability to identify cancer and rule out non-cancer cases, respectively; they do not mean that a positive result carries a 95% probability of being cancer.

In comparisons of paired samples, uRARE-seq detected cancer more readily than urine cytology and also outperformed urine DNA analysis that does not require prior tumor information. The study also found that “field effect” mutations carried by some non-tumor bladder cells did not noticeably interfere with RNA test performance. Such mutations may complicate DNA interpretation; the RNA method instead looks for disease signals in overall gene expression, providing another route to identification.

Clues more directly relevant to treatment decisions emerged in patients receiving intravesical bacillus Calmette–Guérin (BCG) immunotherapy. Patients who later achieved a complete molecular response already had stronger T-cell and immune signals in their urine before treatment, while nonresponders showed signals more associated with cell proliferation. Using pretreatment samples from 114 patients, the study developed biomarkers to predict responses to BCG and chemotherapy. This supports the possibility of pretreatment stratification, but does not yet prove that choosing treatment based on the test can improve patient outcomes.

Urine RNA can also track minimal residual disease after treatment. In its announcement about the same study, Resero Bio stated that patients who achieved a complete molecular response after surgery or BCG treatment had a lower risk of recurrence than those with detectable tumor RNA. The study also explored the identification of tumor grade and muscle invasion status. These results suggest that urine testing could connect diagnosis, treatment, and monitoring, but the disappearance of molecular signals still does not guarantee that cancer will not recur.

The next step toward routine care is larger, multicenter prospective validation. The team said it is planning such trials to assess clinical utility; current evidence remains insufficient to establish practices that replace cystoscopy or adjust treatment based on the test. Resero Bio said its platform is currently for research use and that it holds an exclusive commercial license for RARE-seq. The study also disclosed that several authors hold equity in the company, two coauthors are company employees, and there are related patent interests. Further validation will be an important basis for assessing the technology’s practical value.