North American study and meta-analysis evaluating performance of Bladder EpiCheck®, a FDA cleared test, in non-muscle invasive bladder cancer recurrence Authors: Neil Fleshner, Herbert Barton Grossman, Ryan Berglund, Jason Hafron, Brant Inman, Lawrence Karsh, Kelvin Moses, Daniel Saltzstein, Anup Shah, Jonathan Wright, Johannes Alfred Witjes, Yair Lotan Published: 2025 ABSTRACT AbstractBackgroundBladder EpiCheck (BE) is a novel methylation-based PCR urine test for the detection of non-muscle invasive bladder cancer (NMIBC) recurrences.ObjectiveWe present the results of a North American study evaluating BE and meta-analysis of literature.MethodsA prospective, blinded, multicenter study was conducted in North America. Voided urine was collected from NMIBC patients prior to cystoscopic surveillance. BE testing was performed centrally. For the meta-analysis, a PUBMED search was performed to identify all published peer-reviewed clinical studies of BE for NMIBC surveillance.ResultsIn this study, 674 patients were enrolled of which 449 were included. Overall sensitivity was 67% (95%CI 58%-74%), specificity was 84% (80%-88%), PPV was 65% (57%-73%) and NPV was 85% (81%-89%). For high-grade (HG) recurrence, sensitivity was 77% (65%-85%) and NPV was 95% (92%-97%).In patients with negative cystoscopy and cytology at the first study visit, risk of subsequent recurrence in 12 months was 5.3 (2.7–10.3) times higher in patients with positive BE vs. negative BE (p < 0.0001). In patients with negative cystoscopy and equivocal cytology, BE was positive in 75–89% of those with later HG recurrence, with PPV of 42% (15%-72%)-63% (38%-84%).The meta-analysis included 7 studies and 1564 patients. Overall sensitivity was 82% (66–92%), HG sensitivity was 91% (82–95%), specificity was 85% (80–88%), PPV was 60% (55–64%) and HG NPV was 98% (97–99%).ConclusionsThe consistently strong performance of BE indicate that a positive test could improve timely disease recurrence detection and a negative test could rule-out HG disease. Furthermore, the low rate of false positive results, potentially minimizes unnecessary downstream procedures and patient anxiety.Keywords: non-muscle invasive bladder neoplasms [C04.588.945.947.960.500], non-muscle invasive bladder neoplasms [C12.900.820.945.500], urine [A12.207.927], neoplasm recurrence, local [C04.697.655], real-Time polymerase chain reaction [E05.393.620.500.706], biomarkers, tumor [D23.101.140] Introduction Bladder cancer is the 6th most common cancer in the United States and 4th most common cancer in men.1,2 Recurrence rates of non-muscle invasive bladder cancer (NMIBC) depend on stage and grade and other tumor factors (multifocality, prior recurrence, carcinoma in situ), and can be as high as 30–70%, requiring close surveillance of patients with cystoscopy and urine cytology. 2 Cystoscopy is associated with minor risks such as urinary tract infection, bleeding, and pain and can be costly. 3 Cystoscopy is imperfect, and carcinoma in situ (CIS) and small tumors can be missed. 4 Furthermore, it can be difficult to distinguish between CIS and benign inflammation – which can result from intravesical therapy, prior instrumentation, and infection. 5 Cytology is an adjunct biomarker useful for identifying recurrences in patients undergoing cystoscopy with a history of high-grade (HG) tumors. Cytology has a poor sensitivity for low-grade (LG) tumors and 50–80% sensitivity for HG tumors. 6 There is also interobserver variability between pathologists assessing cytology specimens, and indeterminate results like “atypical” and “suspicious” can leave clinicians and patients confused. Many studies have evaluated urine-based tumor markers to improve the surveillance and detection of bladder cancer. 7 While urine-markers have not yet demonstrated the ability to replace cystoscopy, they have been recommended in the AUA/SUO guidelines for various clinical scenarios, such as adjudication of equivocal cytology, and to assist with evaluating response to BCG therapy. 2 The main hurdles to urine biomarker adoption into the standard-of-care are accuracy, reproducibility, and clinical utility 8 and lack of randomized studies demonstrating effective use of markers. Different clinical utility scenarios could be proposed according to each test's characteristics, for example, a rule-in test would require high specificity and a rule-out test for HG tumors would require high HG sensitivity. Bladder EpiCheck (BE) is an objective novel methylation-based PCR urine test for detection of bladder cancer recurrence that demonstrated overall sensitivity of 68%, HG sensitivity of 89% and specificity of 88%. 9 This manuscript presents the results of a prospective multicenter cohort study conducted to evaluate the performance of BE in patients with NMIBC undergoing surveillance in North America. Additionally, we summarize herein the body of evidence from the literature that led to BE's inclusion as an option in the EAU NMIBC guidelines. 10 Materials and methods A prospective, blinded, multicenter, cohort study was conducted at 11 centers in USA and Canada (NCT02700464). The primary endpoint of the study was to determine the sensitivity and specificity of the BE test as compared to the gold standard (GS) defined by cystoscopy, cytology and, if indicated, pathology. Secondary endpoints were non-inferiority of BE sensitivity and specificity vs. UroVysion's (matched and all samples) in the baseline visit (T0). Men and women aged 22 or older undergoing cystoscopic surveillance for NMIBC were enrolled within 12 months of their last TURBT. Patients planned for cystectomy or chemoradiation or those unable to give a urine sample of 45 mL, were excluded. Patients were managed according to the standard of care with cystoscopy, cytology and pathology performed locally by each site, and were followed up for up to 3 consecutive surveillance visits (Baseline – T0, 1st Follow-up visit – T1, 2nd follow-up visit – T2). Collection of urine prior to cystoscopy at each visit was used for cytology (T0, T1 and T2), BE testing (T0, T1 and T2) and UroVysion testing (T0). Physicians and patients were blinded to the results of BE, and the personnel in the lab performing BE testing was blinded to patient data, including disease status. Study enrollment in each site started after local or central ethics committee approval was granted and all patients provided written informed consent prior to any study procedures. Bladder EpiCheck urine test At each visit, voided urine was collected prior to cystoscopy and sent for cytology and BE evaluation. Barbotage urine samples were excluded to align with the FDA intended use of voided urine. BE detects methylation patterns in urine cell DNA associated with urothelial carcinoma. The test requires at least 10 mL of urine. Urine is kept in refrigeration (2–8°C) for up to 5 days of collection, and then centrifuged to generate a cell pellet. DNA is extracted from the pellet using the BE kit (Nucleix Ltd., Israel) and digested using methylation sensitive enzymes provided in the BE kit. The digested DNA runs with proprietary primers and probes for 15 methylation loci, supplied in the BE kit, on the ABI® 7500 Fast Dx PCR. The minimal amount of DNA required is 5 ng per sample. The PCR output is analyzed by the BE proprietary software that calculates an EpiScore based on the detected methylation levels and reports a qualitative positive/negative result (Figure S1). BE Kit includes several controls: two run controls (No Template Control [NTC] and Undigested Control [UC]) for DNA contamination, function of PCR mixes components, and instrument detectors; and three internal sample controls (Digestion Control [DC], Digestion Control Reference [DCR], and Internal Reference [IR]) for quality of the sample (e.g., sufficient amount of DNA) and successful digestion. In case of an issue, the software will give an invalid result. For this study, two central CLIA-certified laboratories in the U.S. were used for BE testing. Diagnostic gold standards The GS was considered positive if any of the following was present: (a) a positive cystoscopy confirmed by a biopsy showing urothelial carcinoma, (b) a urine cytology positive for high-grade urothelial carcinoma (HGUC), or (c) a urine cytology suspicious for high-grade urothelial carcinoma (SHGUC). All other results were considered GS negative. Statistical analysis The analyses were performed on all eligible patients who were enrolled in the study and had at least one visit with all data available for GS determination and BE test result. Sensitivity, specificity, positive and negative predictive values are presented along with exact two-sided 95% confidence intervals (CI). Analyses were performed using the SAS® version 9.4 (SAS Institute, Cary NC, USA) software. Nominal p-values and non-adjusted CI are presented. For the evaluation of the primary endpoint, one visit was included in the analysis for each patient, comparing the BE result of the urine collected on that visit to the GS of that visit. GS status and availability of a valid BE result were used to define which of the visits was selected for the analysis: the first GS positive visit with a valid BE result was used, and if all visits were GS negative, the last GS negative visit with a valid BE result was used. A patient with a GS positive visit completed the study at that visit. Statistical assumptions, including sample size calculations are detailed in the supplement. To determine if BE could be used to anticipate future bladder cancer recurrences in patients with negative GS evaluations (i.e., negative cystoscopy and negative urine cytology), we evaluated patients whose first visit cystoscopy and cytology were negative, and had at least one more visit where cystoscopy, cytology, and BE were performed. Outcomes were compared between patients with a positive BE vs. negative BE on the first visit. Cox models were constructed to estimate the probability of a future positive GS with the BE test result as the sole covariate. Cumulative overall sensitivity and specificity estimates calculated performance of all valid visits together (T0, T1 and T2). In this calculation a subject may appear more than one time. The calculation was done using generalized estimating equations (GEE) with Subject ID used as a repeated factor. This model accounts for potential dependency between multiple measurements within a subject. The confidence intervals are calculated using Clopper-Pearson score method. The meta-analysis followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A PUBMED search was performed on March 28, 2024 to identify all published peer-reviewed clinical studies of BE in NMIBC surveillance. The term “Bladder EpiCheck” was searched with no time limitations. All studies reporting original BE performance in NMIBC surveillance were included in the analysis: prospective or retrospective studies reporting true positive (TP), true negative (TN), false positive (FP), false negative (FN), and/or sensitivity and specificity. In case of not reporting TP, TN, FP, or FN, these were calculated from known variables (sensitivity and specificity) or by correspondence with the authors. Reviews, meta-analyses, case reports, modelling papers, comments, and papers not reporting the required information to calculate TN/TP/FN/FP were excluded. The most up-to-date publication was used for centers that reported on the same patients more than once. Forest plots show the results of the individual studies in the meta-analysis with 95% Two-Sided Clopper-Pearson Confidence Intervals. Meta Analysis Estimates Based on Generalized Linear Mixed Model with Paper Random Effect were calculated for the pooled performance measures. Chi-squared test for heterogeneity was performed for these measures using the R package mada (Meta-Analysis of Diagnostic Accuracy). Results In total, 674 patients were prospectively enrolled from August 31 2016 until March 28 2019 of which 449 were included in this analysis, 138 patients were excluded due to barbotage urine collection method on the visit selected for analysis (Figure 1). Patient demographics and disease history are shown in Tables 1 and S1. Median age was 71 (range 37–93) and 81% were male, ethnicity, race and sex distributions were similar to those reported for the Medicare population in the USA. 47% of the patients had history of recurrence, 33% and 64% had history of T1/Tis and HG in the last recurrence, respectively, and 62% were treated with instillations since last TURBT. One-hundred thirty-one (29%) patients completed the study at T0, 75 (17%) at T1 and 243 (54%) at T2. The median time in the study was 182 (range 0–903) days. CONSORT table of the North American clinical study. Table 1.Patient demographics in the North American clinical study.DemographicsIncluded in analysisMedicare US bladder cancer patients 2021 (prevalence) 11 N449730,044 Age, y Median (range)71 (37–93)NRSex(N/%)(N/%) Male365 (81%)551,797 (76%) Female84 (19%)178,247 (24%)Ethnicity, N574,406 Not Hispanic or Latino431 (96%)NR Hispanic or Latino15 (3%)23,898 (4%) Not Reported3 (1%)–Race White411 (92%)525,695 (92%)^ Black19 (4%)27,211 (5%)# Asian11 (2%)12,192 (2%)^ American Indian or Alaska Native3 (0.7%)826 (0.1%)^ Not Reported5 (1%)–Disease HistoryN (%)History of Recurrence Primary239 (53%) Recurrent210 (47%)Stage in last recurrence Ta301 (67%) T188 (20%) Tis60 (13%) CIS present95 (21%)Grade in last recurrence Low Grade/PUNLMP163 (36%) High Grade286 (64%)Instillation Treatment prior to enrollment <90 days from last instillation to first visit201 (45%) >90 days from last instillation to first visit74 (16%) Unknown time from last instillation to first visit3 (1%) No Instillation Treatment prior to enrollment171 (38%)^Non-Hispanic; #Including Hispanic. Patient demographics in the North American clinical study. ^Non-Hispanic; #Including Hispanic. One-hundred thirty-eight (138, 31%) patients experienced a recurrence, 64 (14%) HG, 43 (10%) LG and 31 (69%) cytology positive. Overall sensitivity was 67% (CI 95%: 58%-74%), specificity was 84% (80%-88%), PPV was 65% (57%-73%) and NPV was 85% (81%-89%). In the subpopulation with HG recurrence, sensitivity was 77% (65%-85%) and NPV was 95% (92%-97%) (Table 2). Performance was statistically different by grade but was not impacted by sex, smoking history, occupational risk or time from the last intravesical instillation treatment. In a post-hoc analysis assessing the cumulative performance of all valid visits’ samples (n = 898), the prevalence was 15%, overall sensitivity was 67% (58%-74%), HG sensitivity was 77% (65%-85%), specificity was 81% (77%-84%), PPV was 42% (35%-48%), overall NPV was 93% (90%-95%) and HG NPV was 98% (96%-99%). Table 2.Bladder EpiCheck performance in the North American clinical study.Overall performancen/N% (95% CI)Sensitivity92/13867% (58%-74%)Specificity262/31184% (80%-88%)PPV92/14165% (57%-73%)NPV262/30885% (81%-89%)Performance by graden/NSensitivityNPVHigh Grade49/6477% (65%-85%)95% (92%-97%)Low Grade20/4347% (33%-61%)93% (89%-95%)Positive by cytology23/3174% (57%-86%)97% (95%-99%)PPV, positive predictive value; NPV, negative predictive value. Bladder EpiCheck performance in the North American clinical study. PPV, positive predictive value; NPV, negative predictive value. There was evidence that BE could anticipate future bladder cancer recurrences in patients with negative cystoscopy and cytology since the risk of any recurrence and HG recurrence in 12 months was 5.3 (HR = 5.3 [2.7-10.3], P < 0.0001) and 5.1 (HR = 5.1 [1.5–17.5], p = 0.0106) times higher, respectively, in BE positive patients (n = 56) compared to BE negative patients (n = 228)(Figures 2(a) and (b)). The PPV for such patients was 38% (n/N = 21/56, 25%-52%) and the median time from a positive BE test result to GS conversion to positive was 182 days. Anticipatory positive. Solid line represents the cumulative incidence of a recurrence in patients with a positive Bladder EpiCheck result and negative gold standard (cystoscopy and cytology) result at T0. Dotted Line represents the cumulative incidence of a recurrence in patients with a negative Bladder EpiCheck result and negative gold standard (cystoscopy and cytology) result at T0. (a) Time from T0 with Negative Cystoscopy and Cytology to Any Recurrence at Later Visits. (b). Time from T0 with Negative Cystoscopy and Cytology to High-Grade Recurrence at Later Visits. There were 53, 40, and 21 equivocal visits (visit with a negative cystoscopy and atypical urothelial cells (AUC) result in cytology) at T0, T1 and T2, respectively, all were without biopsy and were considered GS negative at that visit (Table 3). BE was positive in 19 (36%), 12 (30%) and 6 (29%) of equivocal visits, respectively. In patients with at least one visit after the equivocal visit, 16 (30%) of T0 and 6 (15%) of T1 became GS positive in a later visit. BE was already positive at the equivocal visit in 12 (75% [48%-93%]) and 5 (83% [36%-99.6%]), respectively, of the patients who eventually converted to GS positive. Nine and 4 of the T0 and T1 visits, respectively, became HG in a later visit of which 8 (89% [52%-99.7%]) and 3 (75% [19%-99.4%]) were already positive by BE at the equivocal visit. BE PPV was 63% (38%-84%) and 42% (15%-72%) for later recurrence, respectively. Table 3.Equivocal visits – patients with a negative cystoscopy and equivocal cytology in the North American clinical study.Number of subjects with an equivocal visit*Bladder EpiCheck positiveLater Gold Standard positiveBladder EpiCheck anticipatory positiveLater high-gradeBladder EpiCheck high-grade anticipatory positiveBladder EpiCheck PPV for later recurrences (CI95%)T0531916129863% (38%-84%)T14012654342% (15%-72%)T2216N/AN/AN/AN/AN/A* Visit was defined as equivocal if cystoscopy was negative and cytology result was atypical urothelial cells (AUC).PPV, positive predictive value; N/A, not applicable. Equivocal visits – patients with a negative cystoscopy and equivocal cytology in the North American clinical study. * Visit was defined as equivocal if cystoscopy was negative and cytology result was atypical urothelial cells (AUC). PPV, positive predictive value; N/A, not applicable. To compare BE and cytology performance, patients with both valid BE and a cytology result on the selected visit were compared to pathologically confirmed cancers. Overall sensitivity was 62% (59%-75%) for BE and 53% (44%-61%) for cytology. HG sensitivity in this sub-analysis was 73% (59%-85%) and 43% (29%-58%), respectively. BE specificity was 84% (80%-88%), however, cytology specificity could not be calculated, because all cytology HGUC and SHGUC were considered as GS positive in our analysis. BE demonstrated non-inferiority to UroVysion (data not shown). In the meta-analysis, of the 25 BE peer-reviewed articles appearing in the search, 7 unique clinical studies on NMIBC surveillance were identified,9,12–17 including a sponsored prospective multicenter European clinical study and 6 independent studies (Figure 3). In a pooled analysis 1564 patients with NMIBC undergoing surveillance were included, 357 had an event (241 HG and 116 LG). BE demonstrated an overall sensitivity of 82% (66–92%), HG sensitivity of 91% (82–95%) and specificity of 85% (80–88%) (Table 4 and Figures 4(a), (b), and (c)). PPV was 60% (55–64%) and HG NPV was 98% (97–99%). Chi-squared test for equality of overall sensitivity and specificity rejected the null hypothesis of equality (p < 0.001 and p = 0.0326, respectively) and the Chi-squared test for equality of HG sensitivity did not reject the null hypothesis of equality (p = 0.197). Three of the studies focused on high-risk patients or patients with history of HG,13,14,17 driving the overall sensitivity up. When these studies are excluded from the pooled analysis, overall sensitivity is lower at 73% (57%-85%), and overall specificity is similar at 85% (79%-89%) (Figures 4(d) and (e)). Results were similar between the North American clinical study and the pooled analysis, with overlapping CI95% for overall sensitivity 67% (58%-74%) vs. 73% (57%-85% [only mixed studies]), specificity 84% (80%-88%) vs. 85% (80–88% [all studies])/85% (79%-89% [only mixed studies]) and HG sensitivity 77% (65%-85%) vs. 91% (82–95%), respectively. Literature review for Bladder EpiCheck evidence. Table 4.Meta-analysis performance of Bladder EpiCheck in NMIBC surveillance.Published clinical studiesStudy designcompany sponsoredStudy populationIncluded in the pooled analysis?number of evaluable patientsRecurrence (N)disease prevalenceHG recurrence (HG N)HG disease prevalenceno cancerBE true positive (n)BE HG true positive (HG n)BE true negativesensitivityHG sensitivityspecificityNPVHG NPVPPVWitjes et al. Eur Urol Oncol. 2018 9 Pr, Mc, Sv yes NMIBC surveillance yes 353 44 12% 18 5% 309 30 16 272 68% 89% 88% 95% 99% 45% D'Andrea et al. BJU Int. 2019 18 Pr, Mc, SvyesNMIBC surveillanceno – same cohort as Witjes et al.3574914%185%308331627167%89%88%94%99%47%Trenti et al. Cancer Cytopathol. 2020 12 Pr, Sc, Sv no NMIBC surveillance yes 432 92 21% 38 9% 340 59 30 279 64% 79% 82% 89% 98% 49% Trenti et al. Cancer Cytopathol. 2019 19 Pr, Sc, SvnoNMIBC surveillanceno – same patients as Trenti et al. 20202156932%3014%146432512662%83%86%83%97%68%Pierconti et al. Urol Oncol. 2021 13 Re, Sc, 1y FU no HG NMIBC surveillance with instillations yes 375 111 30% 111 30% 264 104 104 208 94% 94% 79% 97% 97% 65% Pierconti et al. J Clin Pathol. 2021 20 Pr, Sc, SvnoNMIBC surveillanceno – same patients as Pierconti et al. 202137412734%12734%24712112120995%95%85%97%98%76%Pierconti et al. Urol Oncol. 2022 21 Pr, Sc, SvnoHG NMIBC surveillance with instillationsno – same patients as Pierconti et al. 202120515174%15174%541421424394%94%80%83%86%93%Cochetti et al. Urol Oncol. 2022 14 Pr, Sc, Sv no High Risk NMIBC PDD surveillance yes 40 18 45% 8 20% 22 18 8 20 100% 100% 91% 100% 100% 90% Peña et al. J Clin Med. 2022 15 Pr, Sc, Sv no NMIBC surveillance yes 57 13 23% 12 21% 44 12 12 36 92% 100% 82% 97% 100% 60% Ragonese et al. Clin Genitourin Cancer. 2022 16 Pr, Sc, Sv no NMIBC surveillance yes 209 71 34% 53 25% 138 57 48 117 80% 91% 85% 89% 97% 73% Pierconti et al. Cancer Cytopathol. 2023 22 Re, Sc, 1y FUnoHigh Risk NMIBC surveillanceno – same patients as Pierconti et al. 202129012744%12744%163127127115100%100%71%100%100%73%Caño Velasco et al. Actas Urol Esp. 2023 17 Pr, Sc, Sv no High Risk NMIBC surveillance yes 98 8 8% 1 1% 90 7 1 82 88% 100% 91% 99% 100% 47% Pooled analysis (only unique studies appearing in bold) 1564 357 23% 241 17% 1207 287 15% 1014 82% 91% 85% 94% 98% 60% BE, Bladder EpiCheck; FU, follow-up; HG, high-grade; Mc, multicenter; NMIBC, non-muscle invasive bladder cancer; NPV, negative predictive value; PDD, Photodynamic diagnosis; PPV, positive predictive value; Pr, prospective; Re, retrospective; Sc, single center; Sv, single visit. Meta-analysis performance of Bladder EpiCheck in NMIBC surveillance. BE, Bladder EpiCheck; FU, follow-up; HG, high-grade; Mc, multicenter; NMIBC, non-muscle invasive bladder cancer; NPV, negative predictive value; PDD, Photodynamic diagnosis; PPV, positive predictive value; Pr, prospective; Re, retrospective; Sc, single center; Sv, single visit. Meta-analysis forest plots. (a) Overall Sensitivity (All). (b) High-Grade Sensitivity. (c) Overall Specificity (All). (d) Overall Sensitivity (Only studies with mixed population). (e) Overall Specificity (Only studies with mixed population). Table 5 details sensitivity by stage in the North American clinical study, the European multicenter study and pooled results from all peer-reviewed publications reporting results by stages. Table 5.Bladder EpiCheck sensitivity by stage.StageNorth American Clinical studyEuropean StudyLiterature (pooled, including European study)n/N% (95% CI)n/N% (95% CI)Number of studies reporting% (n/N)Ta38/6559% (46%-70%)14/2752% (32–71)29,1467% (26/39)T17/1164% (35%-85%)7/7100% (59–100)39,14,17100% (9/9)T21/333% (6%-79%)NRNR1 14 100% (1/1)CIS^10/1377% (50%-92%)3/3100% (29–100)39,14,1690% (28/31)^Studies did not differentiate between concomitant or stand-alone CIS.NR, not reported. Bladder EpiCheck sensitivity by stage. ^Studies did not differentiate between concomitant or stand-alone CIS. NR, not reported. Discussion This North American clinical study demonstrated BE overall sensitivity of 67%, HG sensitivity of 77% and specificity of 84%. In the cumulative analysis, specificity was slightly lower (81%) compared to the longitudinal analysis, but overall and HG sensitivity were identical, due to study design (patients with a GS positive completed the study, so the number of positive events was the same in the two analyses). The main difference between the longitudinal and cumulative results are PPV and NPV, since these measures are impacted by disease prevalence which was 15% in the cumulative analysis vs. 31% in the longitudinal analysis. PPV, NPV and HG NPV were 65%, 85%, 95% for the longitudinal analysis and 42%, 93% and 98% for the cumulative analysis, respectively. The cumulative predictive values probably better represent the per-visit predictive values of the test, as they do not inflate disease prevalence beyond a single visit. The findings in this study are consistent with the meta-analysis which included the European multicenter clinical study. 9 In the pooled analysis of 1564 patients with NMIBC under surveillance, the pooled HG sensitivity and HG NPV were 91% and 98%, respectively. Equally important was the high specificity of 84% in this clinical study which was corroborated in the meta-analysis with pooled specificity of 85%. One main criticism of urine markers is the low specificity resulting in low overall PPV. With a high PPV, BE provides a benefit by raising the specter of missed malignancy suggesting the need for further evaluation with biopsies, enhanced cystoscopy or meticulous surveillance monitoring. One could also consider upper tract imaging if there has not been any recent evaluation. High specificity is very important in clinical decision-making to avoid dealing with false positive results, since every 5% drop in specificity could translate to 30% more false positive cases. There were only 3 MIBC recurrences in the study, of which 2 were missed by the test. It is not possible to assess the performance of the test in such a small sample size and a larger study in such patients would be required. The AUA/SUO guidelines currently identify a role for urine markers to help adjudicate patients with equivocal cytology. 2 In this prospective North American clinical study, BE identified at the equivocal visit 75–89% of later detected HG disease. The evaluation of equivocal cytology could be aided by BE because a negative test indicates the probability of CIS is low. Furthermore, the high PPV for future recurrences in this population (42–63%), reinforces its utility in this setting. Multiple reports have shown the potential for “anticipatory findings” with urine markers, apparently false positive results that actually indicate an early or undetected tumor. In this study, patients with negative cystoscopy and cytology who had a positive BE result at their first study visit had a ∼5 times higher risk of any/HG recurrence, compared to those with a negative BE result. This suggests that BE's specificity may be higher than the reported 84%, since 38% of the apparent false positive results of the test are actually false negative results of both cystoscopy and cytology. With such significant risk of recurrence, a positive BE test in the face of a negative cystoscopy could warrant further evaluation with more sensitive tools, such as enhanced cystoscopy, to detect recurrence 6 months earlier, or help guide intensification of surveillance for patients who are on long intervals between cystoscopies. An earlier detection could potentially identify BCG non-responders, as recommended in the AUA/SUO guidelines for older urine markers. For example, detecting CIS prior to development of T1, or detecting T1 prior to the development of muscle invasive disease, allowing for earlier intervention with potentially improved oncological outcomes. The study's robust eligibility, keeping inclusion broad with minimal exclusion criteria, resulted in a study population representative of the USA bladder cancer population including age, sex, race, and ethnicity and had good balance between patients with and without a history of recurrence, history of HG and LG as well as patients with and without recent instillations. One of the concerns with urine marker studies is the lack of independent validation. The finding of similar performance in the North American study results and the meta-analysis supports the robustness of BE and the generalizability of results across diverse populations. The fact that BE testing was performed in local labs with a commercially distributed FDA cleared kit and still had similar performance to a central lab supports the ability to use it on a widespread basis. BE's specificity was very similar across publications and among the published papers. Such robust results can be explained by the very limited susceptibility of the test to interference by common urinary conditions and urine constituents, such as infection, inflammation, hematuria, and recent instillation treatments demonstrated in analytical interference studies, 23 the North American clinical study, and the European multicenter clinical study. 9 The main limitation to this analysis was the exclusion of 138 patients due to collection of barbotage urine instead of voided urine. Importantly, omitting these patients from the analysis did not bias the population demographics and disease history (Table S1). Only 1.6% of patients were excluded from the analysis due to insufficient DNA. An additional 2% were excluded from the analysis due to missing BE result for other reasons (e.g., failure of internal control). The results were robust and consistent between the North American clinical study and the prior studies in the literature. The results of randomized controlled-trials, such as the German UroFollow study 24 and the ongoing American “Replace Cysto” study, 25 are needed to guide use of BE in clinical practice. Conclusions The consistently strong performance of BE indicate that the test could have a potential utility both as a rule-in as well as a HG rule-out test. A strong anticipatory positive signal could translate into earlier detection of recurrence, potentially leading to better outcomes. The test could help adjudicate equivocal cytology and identify BCG-non responders, as currently recommended in the AUA/SUO guidelines for older urine markers. A negative test could be used to rule-out HG disease, and identify patients who could delay a cystoscopy. Furthermore, the low rate of false positive results, potentially minimizes unnecessary downstream procedures and patient anxiety. Randomized studies will help to expand our understanding of use cases and provide additional evidence for test use. Statements and declarations Author contributions: Author conception performance of work interpretation of data writing the article Neil Fleshner X X X H. Barton Grossman X X X Ryan Berglund X Jason Hafron X X Brant Inman X X X Lawrence Karsh X Kelvin Moses X Daniel Saltzstein X Anup Shah X Jonathan Wright X J Alfred Witjes X Yair Lotan X X Author conception performance of work interpretation of data writing the article Neil Fleshner X X X H. Barton Grossman X X X Ryan Berglund X Jason Hafron X X Brant Inman X X X Lawrence Karsh X Kelvin Moses X Daniel Saltzstein X Anup Shah X Jonathan Wright X J Alfred Witjes X Yair Lotan X X Author conception performance of work interpretation of data writing the article Neil Fleshner X X X H. Barton Grossman X X X Ryan Berglund X Jason Hafron X X Brant Inman X X X Lawrence Karsh X Kelvin Moses X Daniel Saltzstein X Anup Shah X Jonathan Wright X J Alfred Witjes X Yair Lotan X X Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The North American study and the meta-analysis were funded by Nucleix. The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Neil Fleshner: leadership or fiduciary role: Verity Pharmaceuticals, Point Biopharma. Consultant or Advisory Role: Janssen, Astellas; H. Barton Grossman: Consultant or Advisory Role: Nucleix; Ryan Berglund: No conflicts; Jason Hafron: Consultant or Advisory Role: Astellas Pharma, Dendreon Pharmaceuticals, Eli Lilly and Company, Immunis.AI, Janssen Biotech, Myriad Genetics, Myovant Sciences, Pfizer, Urogen Pharma, Promaxo, Lynx DX. Payments or honoraria: Astellas Pharma, Amgen Inc; Bayer; Blue Earth Diagnostics; Dendreon Pharmaceuticals, Janssen Biotech, Lantheus, Merck, Myriad Genetics, Myovant Sciences, Pfizer, Procept-Biorobotic, Progenics Pharmaceuticals, Tolmar Pharmaceuticals Inc; Urogen Pharma; Brant Inman: Consultant or Advisory Role: Combat Medical; Lawrence Karsh: Consultant or Advisory Role: Abbvie, Astellas. Astra-Zeneca, Aurora Oncology, Bayer, Bristol Meyers Squibb, Dendreon, Ferring, Genentech/Roche, Janssen, Myovant. Pfizer,Sanofi, UroGPO,and Vaxiion. Payments or honoraria: Astellas, Astra Zeneca, Bayer, Ferring, Janssen, Merck, Myovant, and Pfizer. Stock options: Swan Valley Medical; Kelvin Moses: No conflicts; Daniel Saltzstein: Consultant or Advisory Role: Myovant Sciences, Janssen, Beyer, AstraZeneca, Lantheus; Anup Shah: Patents issued or pending and stocks: Sonomotion; Jonathan Wright: Consultant or Advisory Role: Immunity Bio; Pacific Edge. Clinical trial support: Merck, SeaGen, Pacific Edge, Nucleix, Janssen. Royalties: UpToDate; J Alfred Witjes: Editorial Board member of this journal, but was not involved in the peer-review process nor had access to any information regarding its peer-review.; Consultant or Advisory Role: Nucleix, Astellas, medac, Janssen, BMS, Combat, JeMedis, thermosome, CAPS, photocure.; Yair Lotan: Editorial Board member of this journal, but was not involved in the peer-review process nor had access to any information regarding its peer-review. Consultant or Advisory Role: Nanorobotics, C2I genomics, Photocure, Astra-Zeneca, Merck, Fergene, Abbvie, Nucleix, Ambu, Seattle Genetics, Hitachi, Ferring Research, verity pharmaceutics, virtuoso surgical, Stimit, Urogen, Vessi medical, CAPs medical, Xcures, BMS, Nonagen, Aura Biosciences, Inc., Convergent Genomics, Pacific Edge, Pfizer, Phinomics Inc, CG oncology, Uroviu, On target lab. Stock options: Vessi Medical, CAPs medical, C2I genomics, Nanorobot, Phinomics Inc, Uroviu. Data availability: North American study - the data supporting the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Meta-analysis - the data supporting the findings of this study are available within the article and/or its supplementary material. Supplemental material: Supplemental material for this article is available online. References