Introduction:

Over 2 million cystoscopies are performed annually in the United States and Europe for detection and surveillance of bladder cancer. Patients diagnosed with intermediate and high-risk non-muscle invasive bladder cancer are at high risk for disease progression and recurrence. Decisions regarding cystoscopy surveillance schedules, need for adjunct imaging technologies, and use of intravesical therapies are based upon presumed risk of cancer recurrence and progression. Thus, accurate risk stratification is critical in determining management strategies for patients with bladder cancer. Current risk-stratification models are constructed from clinical and histopathologic data, and thus rely on the availability of a thorough cystoscopy, histopathologic diagnosis, and time to establish a patient’s individual rate of recurrence.  Urinary biomarkers have shown promise for bladder cancer detection however their clinical utility remains unknown. We aimed to develop a urinary biomarker model capable of detecting patients with increased risk bladder cancer.

Methods:

With IRB approval, voided urine specimens were collected at Veterans Affairs Palo Alto Health Care System from subjects undergoing bladder cancer screening or surveillance cystoscopy between 2016 and 2019. All patients underwent white-light flexible cystoscopy. Subjects with bladder tumors identified on cystoscopy underwent transurethral resection of bladder tumor. Samples were categorized based upon tissue histopathologic diagnosis. Tumors were sub-categorized by stage and grade, and operative reports reviewed for tumor size and focality. AUA risk stratification defined as: low risk, solitary low-grade Ta < 3cm not recurrent within one year; high risk, high-grade T1, recurrent high-grade Ta, high-grade Ta > 3cm or multifocal, presence of CIS, BCG failure with high-grade disease, variant histology, LVI, or high-grade prostatic urethral involvement; intermediate risk, all others; was determined for all patients. Urine specimens (n=257) from 181 subjects were evaluated for expression of a 3-mRNA panel (ROBO1, WNT5A, CDC42BPB) for bladder cancer detection previously identified at VAPAHCS and the GeneXpert® Bladder Cancer Assay 5 mRNA panel (ABL1, CRH, IGF2, ANXA10, UPK1B). All mRNA expression was determined using the GeneXpert Dx automated multiplex RT-PCR platform using 2mL of urine per panel. Stepwise logistic regression analysis of the cycle threshold values of ABL1, CRH, IGF2, ANXA10, UPK1B, ROBO1, WNT5A, and CDC42BPB was done to create a diagnostic model to detect intermediate and high-risk bladder cancer. Ten-fold cross-validation was used to generate a receiver operating curve and a positivity threshold selected. Sensitivity and specificity for detection of intermediate and high-risk bladder cancer was determined.

Results:

Urine specimens were collected from 76 patients undergoing screening evaluation for bladder cancer, 99 patients with a history of bladder cancer undergoing surveillance cystoscopy, and 6 patients found to have cancer on screening cystoscopy and then underwent subsequent surveillance cystoscopies. There was a total of 65 diagnoses of bladder cancer (27 low grade, 38 high grade). Twelve patients had low risk disease, 22 had intermediate risk disease, and 31 had high risk disease by AUA criteria. Stepwise logistic regression was used to create a diagnostic model for the detection of increased risk bladder cancer. ROBO1, CRH, and IGF2 were confirmed to correlate with the presence of intermediate and high-risk disease. Setting a P_(IRBC) = .0903 as the cutoff for a positive test, the three-marker panel had a sensitivity of 93% (95% CI, 85%-98%) and specificity of 80% (95% CI, 74%-85%) for intermediate and high-risk disease. Among screening patients, 95% of intermediate and high-risk bladder cancers were detected, with all high-risk cancers (31/31) detected and 19 of 22 intermediate risk cancers detected.

Conclusion:

A 3-marker urinary mRNA panel evaluating the expression of ROBO1, CRH, and IGF2 allows for automated, early identification of intermediate and high-risk bladder cancer.  Biomarker based risk-stratification may allow clinicians to better triage cystoscopy scheduling for screening, guide frequency of surveillance cystoscopy, and identify patients who may benefit from adjunct imaging technology.

Funding: N/a