EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 available at www.sciencedirect.com journal homepage: euoncology.europeanurology .com Review - Bladder Cancer Diagnostic Accuracy of Novel Urinary Biomarker Tests in Non–muscle-invasive Bladder Cancer: A Systematic Review and Network Meta-analysis Ekaterina Laukhtina a,b,1, Sung Ryul Shim c,1, Keiichiro Mori a,d, David D‘Andrea a, Francesco Soria e, Pawel Rajwa a,f, Hadi Mostafaei a,g, Eva Compérat h, Alessia Cimadamore i, Marco Moschini j, Jeremy Yuen-Chun Teoh k, Dmitry Enikeev b, Evanguelos Xylinas l, Yair Lotan m, Juan Palou n, Paolo Gontero e, Marko Babjuk a,o, J. Alfred Witjes p, Ashish M. Kamat q, Morgan Roupret r, Shahrokh F. Shariat a,b,o,s,t,u,v,*, Benjamin Pradere a, European Association of Urology–Young Academic Urologists (EAU-YAU): Urothelial Carcinoma Working Group a Department of Urology, Comprehensive Cancer Center, Medical University of Vienna, Vienna, Austria; b Institute for Urology and Reproductive Health, Sechenov University, Moscow, Russia; c Department of Preventive Medicine, Korea University College of Medicine, Seoul, Korea; d Department of Urology, The Jikei University School of Medicine, Tokyo, Japan; e Division of Urology, Department of Surgical Sciences, San Giovanni Battista Hospital, University of Studies of Torino, Turin, Italy; f Department of Urology, Medical University of Silesia, Zabrze, Poland; g Research Center for Evidence Based Medicine, Tabriz University of Medical Sciences, Tabriz, Iran; h Department of Pathology, Comprehensive Cancer Center, Medical University of Vienna, Vienna, Austria; i Section of Pathological Anatomy, Marche Polytechnic University, School of Medicine, United Hospitals, Ancona, Italy; j Department of Urology and Division of Experimental Oncology, Urological Research Institute, Vita-Salute San Raffaele, Milan, Italy; k S.H. Ho Urology Centre, Department of Surgery, The Chinese University of Hong Kong, Hong Kong, China; l Department of Urology, Bichat-Claude Bernard Hospital, Assistance-Publique-Hôpitaux de Paris, Université de Paris, Paris, France; m Department of Urology, University of Texas Southwestern, Dallas, TX, USA; n Fundació Puigvert, Department of Urology, Autonomous University of Barcelona, Barcelona, Spain; o Department of Urology, Hospital Motol, Second Faculty of Medicine, Charles University, Prague, Czech Republic; p Department of Urology, Radboud University, Nijmegen Heyendaal, The Netherlands; q Department of Urology, M.D. Anderson Cancer Center, University of Texas, Houston, TX, USA; r Urology, GRC n°5, Predictive Onco-Urology, Ap-Hp, Hôpital Pitié-Salpêtrière, Sorbonne Université, Paris, France; s Karl Landsteiner Institute of Urology and Andrology, Vienna, Austria; t Department of Urology, Weill Cornell Medical College, New York, NY, USA; u Department of Urology, University of Texas Southwestern, Dallas, TX, USA; v Hourani Center for Applied Scientific Research, Al-Ahliyya Amman University, Amman, Jordan Article info Abstract Article history: Context: During the past decade, several urinary biomarker tests (UBTs) for bladder can- Received 16 August 2021 cer have been developed and made commercially available. However, none of these is Received in revised form recommended by international guidelines so far. 23 September 2021 Objective: To assess the diagnostic estimates of novel commercially available UBTs for Accepted 5 October 2021 diagnosis and surveillance of non–muscle-invasive bladder cancer (NMIBC) using diag- nostic test accuracy (DTA) and network meta-analysis (NMA). Associate Editor: Evidence acquisition: PubMed, Web of Science, and Scopus were searched up to April Gianluca Giannarini 2021 to identify studies addressing the diagnostic values of UBTs: Xpert bladder cancer, Adxbladder, Bladder EpiCheck, Uromonitor and Cxbladder Monitor, and Triage and Keywords: Detect. The primary endpoint was to assess the pooled diagnostic values for disease Urine cytology 1 These authors contributed equally. * Corresponding author at: Department of Urology, Comprehensive Cancer Center, Vienna General Hospital, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria. Tel. +4314040026150; Fax: +4314040023320. E-mail address: shahrokh.shariat@meduniwien.ac.at (S.F. Shariat). https://doi.org/10.1016/j.euo.2021.10.003 2588-9311/Ó 2021 The Author(s). Published by Elsevier B.V. on behalf of European Association of Urology. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 928 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 Bladder cancer recurrence in NMIBC patients using a DTA meta-analysis and to compare them with Urinary biomarkers cytology using an NMA. The secondary endpoints were the diagnostic values for high- Accuracy grade (HG) recurrence as well as for the initial detection of bladder cancer. Evidence synthesis: Twenty-one studies, comprising 7330 patients, were included in the quantitative synthesis. In most of the studies, there was an unclear risk of bias. For NMIBC surveillance, novel UBTs demonstrated promising pooled diagnostic values with sensitivities up to 93%, specificities up to 84%, positive predictive values up to 67%, and negative predictive value up to 99%. Pooled estimates for the diagnosis of HG recurrence were similar to those for the diagnosis of any-grade recurrence. The analysis of the num- ber of cystoscopies potentially avoided during the follow-up of 1000 patients showed that UBTs might be efficient in reducing the number of avoidable interventions with up to 740 cystoscopies. The NMA revealed that diagnostic values (except specificity) of the novel UBTs were significantly higher than those of cytology for the detection of NMIBC recurrence. There were too little data on UBTs in the primary diagnosis setting to allow a statistical analysis. Conclusions: Our analyses support high diagnostic accuracy of the studied novel UBTs, supporting their utility in the NMIBC surveillance setting. All of these might potentially help prevent unnecessary cystoscopies safely. There are not enough data to reliably assess their use in the primary diagnostic setting. These results have to be confirmed in a larger cohort as well as in head-to-head comparative studies. Nevertheless, our study might help policymakers and stakeholders evaluate the clinical and social impact of the implementation of these tests into daily practice. Patient summary: Novel urinary biomarker tests outperform cytology with the potential of improving routine clinical practice by preventing unnecessary cystoscopic examina- tions during the surveillance of non–muscle-invasive bladder cancer. Ó 2021 The Author(s). Published by Elsevier B.V. on behalf of European Association of Urology. This is an open access article under the CC BY license (http://creativecommons. org/licenses/by/4.0/). 1. Introduction ture to allow implementation in clinical practice. Moreover, these have not yet been approved by either the FDA or the Cystoscopy is the standard examination for the initial detec- EMA. In this context, the use of a diagnostic test accuracy tion and follow-up of non–muscle-invasive bladder cancer (DTA) analysis associated with a meta-analysis and a net- (NMIBC) [1]. Urine cytology is the only widely used urinary work meta-analysis (NMA) is a new interesting approach to biomarker; however, despite high sensitivity in high-grade give an overall assessment of the test accuracy and to com- (HG) tumors and carcinoma in situ (CIS), its sensitivity pare these tests giving better evidence in this field [7]. remains poor in low-grade (LG) tumors and related to the A specific pooled DTA analysis would expand upon the pathologist analysis [2]. In the past decades, several diagnos- current evidence and deliver useful information to stake- tic urinary biomarkers have been developed with the aim to holders and policymakers of health care agencies to facili- detect recurrences while avoiding unnecessary cystoscopies. tate the decision-making process on the value of the tests Repeated cystoscopies are uncomfortable for the patient and in the different NMIBC disease settings followed by their are one of the reasons for the high cost associated with blad- implementation into clinical practice [8]. der cancer, especially NMIBC [3]. Some of these urinary To fill this gap in knowledge, we conducted a systematic biomarkers, such as nuclear matrix protein 22 (NMP22), review, DTA, and NMA to assess the estimates of the novel bladder tumor antigen (BTA), UroVysion (fluorescence commercially available UBTs for diagnosis and surveillance in situ hybridization), and ImmunoCyt/uCyt+ [4], have been of NMIBC. approved by the Food and Drug Administration (FDA) and European Medicines Agency (EMA); these are commercially 2. Evidence acquisition available as urinary biomarker tests (UBTs). However, none of these UBTs are commonly used in daily practice; these 2.1. Protocol are rarely reimbursed by payors, and have not been imple- mented in guidelines due to poor specificity and a lack of This systematic review and meta-analysis was conducted demonstrable clinical benefits, with some exceptions such according to the Preferred Reporting Items for Systematic as atypical cytology and cystoscopic findings [1,5]. Reviews and Meta-analyses (PRISMA) of DTA studies [9]. Recently, several novel UBTs, such as Xpert bladder cancer The study protocol was registered on the International (Cepheid; Sunnyvale, California, USA), adxbladder (Arquer Prospective Register of Systematic Reviews (PROSPERO; Diagnostics; Sunderland, UK), Bladder EpiCheck (Nucleix; registration ID CRD42021248896). Rehovot, Israel), Uromonitor (U-Monitor; Porto, Portugal), Cxbladder Monitor, and Triage and Detect (Pacific Edge; 2.2. Literature search Dunedin, New Zealand), have become commercially avail- able for surveillance and detection of bladder cancer [6]. PubMed, Web of Science, and Scopus were searched to However, data on their diagnostic accuracy are still imma- identify reports published up to April 2021 and addressing EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 929 the diagnostic value of select UBTs in NMIBC (Xpert index test was defined as the value of NMIBC detection with bladder cancer, Adxbladder, Bladder EpiCheck, Uromonitor, the novel UBTs. NMIBC detection with cystoscopy and Cxbladder Monitor, and Triage and Detect). The keywords histopathology was used as a reference. Discrepancies were used in our search strategy are reported in the Supplemen- resolved through discussion and consensus. tary material. Initial screening was performed indepen- dently by three investigators based on the titles and 2.6. Statistical analyses abstracts of the article to identify ineligible reports. Reasons for exclusions were noted. Potentially relevant reports were 2.6.1. Meta-analysis of diagnostic test accuracy subjected to a full-text review, and the relevance of the Pooled Se, Sp, PPV and NPV, and diagnostic odds ratio (DOR) reports was confirmed after the data extraction process. were calculated for each UBT. The pooled DOR is a single Any discrepancies during the primary and secondary litera- indicator of test performance that pooled measure of the ture screening were resolved by referring to the senior performance of a diagnostic test. It is defined as the ratio author. The primary endpoint was to assess pooled diagnos- of the odds of the TP relative to the odds of the FP. We also tic values of the novel UBTs for recurrence during follow-up performed subgroup analyses for the detection of HG recur- of NMIBC using a DTA meta-analysis and to compare these rence of NMIBC. We used the bivariate random-effect model UBTs with cytology using an NMA. The secondary endpoints for analysis and pooling of the diagnostic performance mea- were to assess their diagnostic values in the detection of HG sures across studies, as well as comparisons between differ- recurrence as well as for the primary diagnosis of bladder ent index tests. The bivariate model estimates pairs of logit- cancer. transformed Se and Sp from studies, incorporating the cor- relation that might exist between specificity from studies 2.3. Inclusion and exclusion criteria [7,11]. Forest plots with 95% confidence interval (CI) were calculated and depicted. We developed a hierarchical sum- The population, intervention, control, and outcomes (PICO) mary receiver operating curve and calculated the area in this study were decided by the coauthors as follows: under the curve (AUC) to examine the diagnostic accuracy patients who were included in detection (primary detec- of each UBT. Heterogeneity among the outcomes of tion) or were undergoing follow-up for recurrent disease included studies in this meta-analysis was evaluated using and with detected urothelial carcinoma recurrence at the the Cochrane’s Q test and the I2 statistic. Significant hetero- cystoscopy or histopathology compared with patients with- geneity was indicated by p < 0.05 in the Cochrane’s Q tests out NMIBC. We analyzed diagnostic differences for the and a ratio of >50% in I2 statistics. All statistical analyses value of NMIBC detection in initial diagnosis or diagnosis were performed using R version 4.0.3 (2020; R Foundation of recurrence during follow-up. Studies were eligible if for Statistical Computing, Vienna, Austria). The statistical these reported data on the following: true positive (TP), true significance level was set at p < 0.05. negative (TN), false positive (FP), false negative (FN), sensi- tivity (Se), specificity (Sp), accuracy, positive predictive 2.6.2. Network meta-analysis value (PPV), or negative predictive value (NPV). In case of The NMA was performed for the diagnostic values of the not reporting NPV, PPV, TP, TN, FP, or FN, these were calcu- five UBTs that were used for the detection of NMIBC recur- lated from known variables (Se and Sp). rence and compared with cytology. For the assessment of We excluded every study that did not evaluate the diag- the diagnostic values, arm-based analyses were performed nostic accuracy of UBTs compared with reference methods to estimate the odds ratio (OR) of the recurrence detection (cystoscopy and/or histopathology). We also excluded stud- and 95% credible interval (CrI) from Se, Sp, PPV, and NPV ies reporting data on laboratory developing tests (eg, in the included manuscripts. MCM5) as well as reviews, letters to editors, editorials, As described previously, first, a prior distribution (prior study protocols, case reports, brief correspondence, and probability) was selected [12–14]. Second, the likelihood articles not published in English. References of all papers was calculated from the present data, and a Bayesian hierar- included were scanned for additional studies of interest. chical model was created in NMA. Third, prior distribution and likelihood were entered as input to the Markov chain 2.4. Data extraction Monte Carlo (MCMC) simulation, and a distribution that best Three investigators independently extracted the following converges the posterior distribution was set. The probability information from included articles: author’s name, publica- of stable distribution and the area under the posterior distri- tion year, number of the patients, tumor stage and grade, bution function could be determined through the MCMC presence of CIS, previous intravesical therapy, cutoff value simulation. Finally, statistical reasoning for the treatment of biomarker, recurrence rates, as well as Se, Sp, and the effect was performed with the determined posterior distri- numbers of TP, FP, FN, and TN for the main outcomes (the bution. For the MCMC simulation, we selected the random- value of NMIBC detection in initial diagnosis or diagnosis effect model that had four chains, 5000 burn-ins, 50 000 of recurrence during follow-up). All discrepancies regarding iterations, and an interval of 5 to sufficiently remove the data extraction were resolved by consensus with the effect of initial values, increase the iterations and extraction coauthors. interval, and minimize the MCMC error and the deviance information criterion variation with almost no variations 2.5. Risk of bias assessment and stability of various plots [12–14]. For the consistency test, we performed node-splitting assessments to determine The risk of bias of included studies was evaluated according the association between the direct and indirect evidence. to the risk of bias using the revised Quality Assessment of To assist in the interpretation of diagnostic performance, Diagnostic Accuracy Studies tool (QUADAS-2) [10]. The the surface under the cumulative ranking curve (SUCRA) 930 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 was used to calculate the probability of each UBT, being the assessment, urologists and/or pathologists were blinded to most effective diagnostic method based on a Bayesian UBT results. approach using probability values, and the larger was the The prevalence of disease for each UBT is reported in SUCRA value, the better was the rank of the intervention Table 1. The pooled recurrence rate was calculated at 18%, [15,16]. Publication bias was assessed with funnel plots. with significant heterogeneity across studies with rates ranging from 6% [36] up to 70% [32]. The pooled diagnostic values calculated for 1000 patients 3. Evidence synthesis as well as net benefits, the number of cystoscopies avoided (TN + FN), and the risk of missing recurrences by avoiding 3.1. Study selection and characteristics cystoscopy (TN) for each UBT are presented in Figure 1 and Supplementary Table 1. The calculations were based The literature search identified 2231 unique references. on the pooled recurrence rate and on 5-yr recurrence rates Among them, 190 records were removed due to duplication according to risk classification [1,38]. UBTs might be effi- and 1763 articles were excluded due to unrelated outcomes cient in reducing the number of avoidable interventions, during the screening process (Supplementary Fig. 1). Of the with up to 740 cystoscopies avoided for 1000 patients. For 278 full-text articles assessed for eligibility, 252 were the detection of HG recurrence, the number of the excluded based on the selection criteria. cystoscopies avoided was similar among UBTs, with up to Twenty-seven studies were included in the qualitative 790 cystoscopies being potentially avoided. For the detec- synthesis (Table 1). Twenty-one studies, comprising 7330 tion of recurrence at 5 yr in low-, intermediate-, and high- patients, were included in the quantitative synthesis. All risk groups, the number of the cystoscopies avoided the studies included in the quantitative synthesis reported reduced to 510, 491, and 474, respectively. The number of on the diagnostic value of UBTs for the detection of recur- cancers missed (FN) are also presented in Supplementary rence during surveillance: ten studies on Xpert bladder can- Table 1. cer [17–26], five studies on Bladder EpiCheck [26–30], three on Adxbladder [31–33], two on Uromonitor [34,35], and 3.2. Meta-analysis of diagnostic test accuracy for NMIBC recurrence two on Cxbladder Monitor [36,37]. The summary of the risk of bias and applicability con- 3.2.1. Xpert bladder cancer cerns is presented in Supplementary Figure 2. Overall qual- Ten studies provided data on the diagnostic values of the ity of the included studies was deemed satisfactory. In most Xpert bladder cancer UBT for the diagnosis of recurrence of the studies, there was an unclear risk of bias as to refer- during NMIBC follow-up [17–26]. The pooled Se, Sp, PPV, ence standards because included studies did not specify and NPV were 0.72 (95% CI 0.63–0.80), 0.76 (95% CI 0.72– whether, during cystoscopy and/or histopathological 0.81), 0.43 (95% CI 0.32–0.54), and 0.92 (95% CI 0.90– Fig. 1 – Net benefits and interventions avoided for the urinary biomarker tests for the detection of recurrence during the follow-up of 1000 patients with NMIBC based on the pooled recurrence rate. FP = false positive; FN = false negative; NMIBC = non–muscle-invasive bladder cancer; TN = true negative; TP = true positive. Table 1 – Characteristics of included studies reporting the diagnostic estimates of urinary biomarkers for diagnostics of NMIBC during the diagnosis of recurrence within follow-up Author (publication Study Number of T stage, n (%) Grade, n (%) CIS, n (%) Previous intravesical Cutoff of biomarker Sensitivity Specificity Recurrence rate, n (%) Follow-up, year) design patients therapy, n (%) (%) (%) median (range) Xpert bladder cancer Cancel-Tassin (2021) P 500 Ta: 360 (72%) LG: 287 (57.4%) 47 (9.4%) No BCG 6 wk before 0.5 All: 72.7 73.7 All: 44 (8%) NR [17] T1: 88 (17.6%) HG: 194 (38.8 %) HG: 92.3 HG: 13 (29.5%) LG: 64.5 LG: 31 (70.5%) Cowan (2021) [18] P 429 Ta: 324 (75.5%) LG: 178 (41.5%) 37 (8.6%) BCG: 149 (34.7%) NR All: 60.3 76.5 All: 58 (13.5%) In case of positive test: 12 mo (±90 d) T1/T2: 46 (10.7%) HG: 230 (53.6%) chemo: 63 (14.7%) HG: 87.0 HG: 23 LG: 35 D’Elia (2021) [21] P 416 (1015 NR LG: 126 (75%) NR NR 0.5 All: 52.38 78.39 168 samples (16.5%) 3–48 mo samples) HG: 42 (25%) LG: 42.86 HG: 80.95 Elsawy (2021) [19] P 181 Ta: 25 (13.8%) LG: 33 (18.2%) Exclusion BCG or chemo NR 73.7 79.6 19 (10.4%) In cystoscopy- EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 (epirubicin) negative patients:9 (5– 19) mo T1: 156 (86.2) HG: 148 (81.8%) Elsawy (2021) [20] P 254 Ta: 8 (3.1%) LG: 51 (20.1%) Exclusion BCG: 21 (8.3%) NR 85.9 72.3 84 (35%) 12 (3–25) mo T1: 53 (20.9%) HG: 10 (3.9%) Chemo (epirubicin): 40 (15.7%) Hurle (2020) [22] P 106 Only Ta, T1a LG: 106 (100%) Exclusion NR 0.4 29.85 94.12 NR 18.8 (0–56.5) mo HG: exclusion 0.5 30 90.24 Pichler (2018) [23] P 140 Ta: 110 (78.6%) LG: 97 (69.3%) 8 (5.7%) BCG: 33 (23.6%) NR All: 84 All: 91 43 (30.7%) NR T1: 22 (15.7%) HG: 43 (30.7%) MMC: 26 (18.6%) Previous Previous intravesical intravesical therapy: 100 therapy: 92 No previous No previous intravesical intravesical therapy: 79 therapy: 90 Smrkolj (2020) [24] P 54 NR NR NR No BCG or chemo 3 mo 0.5 66.7 95.2 NR NR before 0.4923 75 95.2 Van Valenberg (2018) P 239 NR NR NR BCG: 83 (35%) NR All: 74 All: 80 All: 43 (18%) NR [25] Chemo: 69 (29%) HG: 83.3 HG: 75.8 HG: 24 LG: 19 Trenti (2020) [26] P 487 Ta: 341 (70%) LG: 336 (69%) 59 (12%) BCG: 122 (25.1%) 0.5 All: 66.30 76.47 All: 92 (21.3%) NR T1: 87 (18%) HG: 151 (31%) MMC: 37 (7.6%) HG: 78.95 HG: 38 (8.8%) LG: 57.41 LG: 54 (12.5%) Bladder EpiCheck D’Andrea (2019) [27] P 357 Ta: 219 (61.3%) LG: 182 (51%) 36 (10%) BCG: 70 (20.2%) 60 All: 67.3 All: 88 All: 49 (13.7%) NR T1: 97 (27.2%) HG: 170 (47.6%) MMC: 111(31.1%) HG: 88.9 HG: 84.1 HG: 18 (5.1%) Both: 59 (16.5%) LG: 40 LG: 81.6 LG: 20 (5.6%) Pierconti (2021) [28] R 375 T1: 269 (71.7%) HG: 269 (71.7%) 106 BCG: 305 (81.3%) 60 HG: 73 HG: 70.1 111 (29.6%) In case of (28.3%) negative cytology: 12 mo MMC: 70 (18.7%) Trenti (2019) [29] P 243 Ta: 165 (68%) LG: 154 (63.4%) 49 BCG: 83 (34.2%) 60 All: 62.3 All: 86.3 All: 69 (32.1%) NR (20.2%) T1: 29 (12%) HG: 89 (36.6%) MMC: 11 (4.5%) HG: 83.3 HG: 79.5 HG: 30 (14.0%) LG: 46.1 LG: 74.4 LG: 39 (18.1%) Trenti (2020) [26] P 487 Ta: 341 (70%) LG: 336 (69%) 59 (12%) BCG: 122 (25.1%) 60 All: 64.13 82.06 All: 92 (21.3%) NR T1: 87 (18%) HG: 151 (31%) MMC: 37 (7.6%) HG: 78.95 HG: 38 (8.8%) LG: 53.70 LG: 54 (12.5%) Wasserstrom (2016) [30] R 222 NR NR NR NR 60 All: 90 All: 83 All: 40 (18%) NR HG: 95 HG: 19 (8.6%) LG: 84 LG: 19 (8.6%) (continued on next page) 931 932 Table 1 (continued) Author (publication Study Number of T stage, n (%) Grade, n (%) CIS, n (%) Previous intravesical Cutoff of biomarker Sensitivity Specificity Recurrence rate, n (%) Follow-up, year) design patients therapy, n (%) (%) (%) median (range) ADXBLADDER Biaûçûçlek (2021) [32] P 119 pTaLG: 80 (67%) NR NR 0.985 All: 73.5 33.3 83 (69.7%) In recurrence patients: 46 (18– 117) mo; in nonrecurrence patients: 41 (22–72) mo pTaHG: 14 (12%) HG: 81.8 pT1LG: 2 (2%) LG: 72.2 pT1HG: 23 (19%) CIS: 50 Ta: 72.6 T1-T2: 100 Gontero (2021) [33] P 503 TaLG: 203 (40%) CIS all: 63 (13%) BCG: 288 NR All: 51.9 66.4 54 (10%) NR (57%) TaHG: 143 (28%) CIS alone: 7 (1%) Chemo: HG: 58.8 105 (21%) EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 T1: 136 (27%) LG: 44.1 CIS all: 60 CIS alone: 100 Ta: 45.2 T1: 75 LG pTa: 44.1 Solitary: 44.8 Multiple: 52.4 Roupret (2020) [31] P 1431 TaLG: 738 (51.6%) 145 (10.1%) BCG: 534 NR All: 44.9 71.1 127 (8.9%) NR (37.3%) TaHG: 376 (26.3%) Chemo: HG: 73 424 (29.6%) T1: 267 (18.7%) LG: 30.2 pTa: 38.3 pT1: 75 pT2: 100 CIS all: 71.4 pTaLG: 30.2 Non-pTaLG: 75.6 Solitary: 45.9 Multiple: 42.1 Uromonitor Batista (2019) [34] P 185 (122 Ta: 32 (62.7%) LG: 25 (51%) 5 (9.8%) NR NR Uromonitor: 93.2 34/122 (28%) NR follow-up) T1: 12 (23.5%) HG: 24 (49%) All: 73.5 T2: 1 (2%) HG: 75 LG: 62.5 CIS: 100 Ta: 53.8 T1: 71.4 Uromonitor- 83.3 V2: 100 Sieverink (2020) [35] P 77 Ta: 44 (57%) LG: 32 (42%) 20 (26%) BCG: 16 (21%) NR Uromonitor- 85.4 29 (37.7%) 6 mo V2: 93.1 T1: 11 (14%) HG: 45 (58%) MMC: 34 (44%) PUNLMP: 2 (3%) Synergo: 21 (27%) Cxbladder Monitor Koya (2020) [36] R 309 Ta LG: 257 (83%) NR NR 100 79.7 13 (6.2%) 35 mo Tis or HG: 52 (17%) Lotan (2017) [37] P 803 Ta: 675 (80%) LG: 508 (50%) 101 NR NR 90.7 38.8 151 (14.9%) NR (10%) T1: 161 (20%) HG: 453 (45%) BCG = bacillus Calmette-Guérin; CIS = carcinoma in situ; HG = high grade; LG = low grade; MMC = mitomycin C; NMIBC = non–muscle-invasive bladder cancer; NPV = negative predictive value; NR = not reported; P = prospective; PPV = positive predictive value; PUNLMP = papillary urothelial neoplasm of low malignant potential; R = retrospective.Additionally, PPV and NPV can be found in Supplementary Table 1. EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 933 Fig. 2 – Forest plot for pooled sensitivity of five different urinary biomarker tests (Xpert bladder cancer, ADXBLADDER, Bladder EpiCheck, Uromonitor, and Cxbladder Monitor) for diagnostics of NMIBC recurrence during follow-up. CI = confidence interval; df = degree of freedom; IV = inverse variance; NMIBC = non–muscle-invasive bladder cancer. 0.94), respectively (Fig. 2–5). The Cochrane’s Q tests and Q tests and I2 tests revealed significant heterogeneity I2 tests revealed significant heterogeneity among studies among studies only for PPV. The pooled DOR was 19.80 for each variable. (95% CI 9.18–42.72; Supplementary Fig. 9). Xpert bladder The pooled DOR was 9.09 (95% CI 5.67–14.57; Supple- cancer for the detection of HG recurrence of NMIBC reached mentary Fig. 3). The test reached an AUC of 80.7% for the an AUC of 79.5% during follow-up. detection of recurrence during follow-up (Supplementary Fig. 4). 3.2.2. Bladder EpiCheck In the subgroup analysis of patients with HG recurrence, Five studies provided data on the diagnostic values of the the forest plots revealed that the pooled Se, Sp, PPV, and Bladder EpiCheck UBT for the diagnosis of recurrence during NPV were 0.88 (95% CI 0.79–0.96), 0.75 (95% CI 0.73– NMIBC follow-up [26–30]. The forest plots revealed that the 0.78), 0.18 (95% CI 0.08–0.28), and 0.99 (95% CI 0.98– pooled Se, Sp, PPV, and NPV were 0.74 (95% CI 0.57–0.85), 1.00), respectively (Supplementary Fig 5–8). The Cochrane’s 0.84 (95% CI 0.80–0.88), 0.48 (95% CI 0.42–0.54), and 0.94 934 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 Fig. 3 – Forest plot for pooled specificity of five different urinary biomarker tests (Xpert bladder cancer, ADXBLADDER, Bladder EpiCheck, Uromonitor, and Cxbladder Monitor) for diagnostics of NMIBC recurrence during follow-up. CI = confidence interval; df = degree of freedom; IV = inverse variance; NMIBC = non–muscle-invasive bladder cancer. (95% CI 0.90–0.97), respectively (Fig. 2–5). The Cochrane’s Q In the subgroup analysis of patients with HG recurrence, tests and I2 tests revealed significant heterogeneity among the forest plots revealed that the pooled Se, Sp, PPV, and studies only for Se and NPV. The pooled DOR was 15.57 NPV were 0.80 (95% CI 0.70–0.90), 0.78 (95% CI 0.69– (95% CI 6.71–36.13). Bladder EpiCheck for diagnostics of 0.86), 0.38 (95% CI 0.21–0.55), and 0.94 (95% CI 0.89– NMIBC recurrence reached an AUC of 86.7% during follow- 1.00), respectively (Supplementary Fig. 5–8). The up. Cochrane’s Q tests and I2 tests revealed significant hetero- EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 935 Fig. 4 – Forest plot for pooled positive predictive value of five different urinary biomarker tests (Xpert bladder cancer, ADXBLADDER, Bladder EpiCheck, Uromonitor, and Cxbladder Monitor) for diagnostics of NMIBC recurrence during follow-up. CI = confidence interval; df = degree of freedom; IV = inverse variance; NMIBC = non–muscle-invasive bladder cancer. geneity among studies in terms of Sp, PPV, and NPV. The 3.2.3. ADXBLADDER pooled DOR was 14.74 (95% CI 4.80–45.27). Bladder Three studies provided data on the diagnostic values of the EpiCheck for diagnostics of HG recurrence reached an AUC ADXBLADDER test for the diagnosis of recurrence during of 86.6% during follow-up. NMIBC follow-up [31–33]. The forest plots revealed that 936 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 Fig. 5 – Forest plot for pooled negative predictive value of five different urinary biomarker tests (Xpert bladder cancer, ADXBLADDER, Bladder EpiCheck, Uromonitor, and Cxbladder Monitor) for diagnostics of NMIBC recurrence during follow-up. CI = confidence interval; df = degree of freedom; IV = inverse variance; NMIBC = non–muscle-invasive bladder cancer. the pooled Se, Sp, PPV, and NPV were 0.57 (95% CI 0.39–0.73), was 1.95 (95% CI 1.46–2.61). ADXBLADDER for diagnostics 0.62 (95% CI 0.51–0.72), 0.29 (95% CI 0.08–0.67), and 0.82 of NMIBC recurrence reached an AUC of 59.7%. (95% CI 0.53–0.95), respectively (Fig. 2–5). The Cochrane’s Q In the subgroup analysis of patients with HG recurrence, tests and I2 tests revealed significant heterogeneity among the pooled Se, Sp, PPV, and NPV were 0.71 (95% CI 0.63– studies in terms of Se, Sp, PPV, and NPV. The pooled DOR 0.79), 0.76 (95% CI 0.65–0.88), 0.37 (95% CI 0.26–0.99), EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 937 and 0.93 (95% CI 0.87–0.99), respectively (Supplementary ure 6. The results of the current NMA revealed that the Se Fig. 5–8). There was significant heterogeneity among stud- (OR 68.2, 95% CrI 7.51–1.26e + 03), Sp (OR 0.41, 95% CrI ies in terms of Sp, PPV, and NPV. The pooled DOR was 0.04–4.34), PPV (OR 1.31, 95% CrI 0.15–10.1), and NPV (OR 6.19 (95% CI 1.94–19.84). ADXBLADDER for diagnostics of 14.0, 95% CrI 3.28–116) of the Uromonitor test were signif- HG recurrence reached an AUC of 74.5%. icantly higher than those of any other tests for the detection of NMIBC recurrence (Fig. 7). The results indicated consis- 3.2.4. Uromonitor tency between the direct and indirect evidence in all out- Two studies provided data on the diagnostic values of the comes. As such, the consistency model was applied to the Uromonitor test for the diagnosis of recurrence during current study (all p > 0.05). NMIBC follow-up [34,35]. The forest plots revealed that An exploratory analysis of the SUCRA values of the diag- the pooled Se, Sp, PPV, and NPV were 0.93 (95% CI 0.79– nostic performance of different UBTs indicated that the 0.98), 0.79 (95% CI 0.62–0.90), 0.67 (95% CI 0.36–0.89), Uromonitor test might be ranked first in terms of Se, PPV, and 0.96 (95% CI 0.86–0.99), respectively (Fig. 2–5). The and NPV (0.90, 0.73, and 0.98, respectively), while cytology Cochrane’s Q tests and I2 tests did not reveal significant ranked first in terms of Sp (0.94; Supplementary Table 2). heterogeneity among studies in terms of Se, Sp, PPV, or The funnel plot identified two studies over the pseudo– NPV. The pooled DOR was 63.82 (95% CI 15.05–270.52). 95% CI for Se, four studies for Sp, four studies for PPV, and Uromonitor reached an AUC of 92.4% for the detection of one study for NPV (Supplementary Fig. 10). recurrence. A subgroup analysis for HG recurrence was not feasible due to the lack of data. 3.4. Performance in primary diagnosis Seven studies reported on the diagnostic values of the novel 3.2.5. Cxbladder Monitor UBTs for the detection of NMIBC during initial diagnosis Two studies provided data on the diagnostic values of [34,39–44]. Table 2 summarizes the characteristics of Cxbladder Monitor for the diagnosis of recurrence [36,37]. included studies and their main outcomes. The pooled Se, Sp, PPV, and NPV were 0.91 (95% CI 0.85– In the studies on initial diagnosis setting, the Xpert blad- 0.95), 0.61 (95% CI 0.21–0.90), 0.16 (95% CI 0.09–0.28), der cancer test has been shown to have Se varying between and 0.98 (95% CI 0.82–0.99), respectively (Fig. 2–5). There 73% and 79% regardless of the presence of hematuria was significant heterogeneity among studies in terms of [39,40]. In terms of Sp, the Xpert bladder cancer test had Sp, PPV, and NPV. The pooled DOR was 12.00 (95% CI the highest estimates of 90% in patients with hematuria 1.63–88.64). Cxbladder Monitor reached an AUC of 91.7% [40]; in the overall population, it ranged from 77% [40] to for the detection of recurrence. A subgroup analysis for 84% [39]. The ADXBLADDER test demonstrated Se up to HG recurrence was also not feasible. 87% in patients diagnosed with HG NMIBC, compared with 48–55% in patients diagnosed with LG tumors [41,42]. Sim- 3.3. Network meta-analysis ilarly, Cxbladder (mix of Triage and Detect) reached the Thirteen studies were available for NMA [17–19,21,23–26, highest Se value of 97% in patients with HG tumors [43,44]. 30,33–35,37]. The networks of eligible comparisons are In general, there was a lack of data on the use of the graphically represented in network plots on the diagnostic novel UBTs in the initial diagnosis setting. Additionally, values of UBTs for the diagnosis of NMIBC recurrence in Fig- some of the studies poorly reported the clinicopathological characteristics of patients diagnosed with NMIBC. Hence, despite the promising results of some tests, large-scale studies are definitely warranted to confirm and validate these tests in this indication. 3.4.1. Combinations Four studies reported the diagnostic value of different com- binations of UBTs for the detection of recurrence during NMIBC follow-up (Supplementary Table 3) [21,23,26,29]. Nevertheless, data on benefits of combinations are limited, and no formal analysis can be performed. More studies are still needed in the future and should implement their cost. 3.5. Discussion We conducted the first DTA and NMA of the diagnostic esti- mates of the most promising novel UBT. Our results demon- strate that many markers have good performance characteristics for the detection of recurrence during surveillance of bladder cancer. Although the Uromonitor test had the highest diagnostic values among all the tests in our NMA, our results should be Fig. 6 – Evidence network plot of diagnostic value of cytology and five different urinary biomarker tests (Xpert bladder cancer, ADXBLADDER, interpreted with caution due to the limited number of Bladder EpiCheck, Uromonitor, and Cxbladder Monitor) for the detection of included studies; confirmation by future well-designed tri- NMIBC recurrence. NMIBC = non–muscle-invasive bladder cancer. als evaluating multiple UBTs focusing on the recurrence of 938 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 Fig. 7 – Forest plots representing five different urinary biomarker tests (Xpert bladder cancer, ADXBLADDER, Bladder EpiCheck, Uromonitor, and Cxbladder Monitor) for the detection of NMIBC recurrence compared with cytology. CrI = credible interval; NMIBC = non–muscle-invasive bladder cancer. NMIBC is necessary. Compared with our results among the vious meta-analyses did not assess the possibility of avoid- novel UBTs, one of the most studied FDA-approved ing unnecessary cystoscopies [1]. biomarkers, NMP22, was shown to have lower diagnostic In general, NMIBC has been shown to have considerable estimates (pooled Se of 71% and Sp of 80%) [45]. Similarly, symptomatic health-related quality of life and economic pooled Se and Sp of other FDA-approved tests such as BTA burden [3,47]. The reduction of cystoscopy frequency as (58% and 79%, respectively) and ImmunoCyt/uCyt+ (75% interventional procedures is not only to potentially improve and 76%, respectively) were also lower [46]. Moreover, pre- the quality of life, but also to improve the value-based care EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 939 Table 2 – Characteristics of included studies reporting the diagnostic estimates of urinary biomarkers for diagnostics of NMIBC during initial diagnosis Author Study Number of Diagnosed Stage, Grade, CIS, Sensitivity (%) Specificity (%) (publication year) design patients BCa, n (%) n (%) n (%) n (%) Xpert bladder cancer Van Valenberg (2020) [39] P 828 59 (7%) NR NR NR All: 78 84 Microhematuria: 73 Macrohematuria: 79 Wallace (2018) [40] P 370 49 (10.8%) Hematuria: Hematuria: Hematuria: All: 73 All: 77 Ta: 14 (64%) LG: 8 (36%), CIS alone: 1 Hematuria: 73 Hematuria: 90 (5%) T1: 2 (9%) HG: 14 (64%) CIS all: 1 (5%) Greater T1: 3 (14%) ADXBLADDER Anastasi (2020) [41] P 91 40 MIBC: 3/37 (8.1%) LG: 21/37 (57%) 7 (18.9%) All: 60 88.2 NMIBC: 34/37 HG: 16/37 (43%) HG: 87.5 (91.9%) LG: 47.6 Dudderidge (2019) [42] P 856 74 NMIBC: 57 LG: 29, 8 All: 73 68.4 MIBC: 16 HG: 42 HG NMIBC: 86 pTa: 40 pT1 and >: 34 LG NMIBC: 55 pT1 and >: 97.0 pT2 and >: 100 CIS: 88 pTa: 53 Uromonitor Batista (2019) [34] P 63 28 (44.4%) NR NR NR Uromonitor: 50 100 Uromonitor-V2: 80 93.3 Cxbladder (mix of Triage and Detect) O’Sullivan (2012) [43] P 485 66 (13.6%) Ta: 37 LG: 32 (48.5%) 2 (3%) All: 81.8 All: 85.1 HG: 80.5 T1: 16 HG: 29 (43.9%) HG: 96.6 LG: 79 T2: 9 Mixed: 4 (6%) LG: 68.8 T3/greater: 2 Davidson (2019) [44] P 571 44 Ta: 21 LG: 12 1 All: 95.5 All: 34.3 Macrohematuria: 32.8 T1: 9 HG: 28 Macrohematuria: Microhematuria: 95.1 42.6 T2: 10 Microhematuria: 100 Tx: 3 BCa = bladder cancer; CIS = carcinoma in situ; HG = high grade; LG = low grade; MIBC = muscle-invasive bladder cancer; NMIBC = non–muscle-invasive bladder cancer; NR = not reported; P = prospective. of NMIBC. Therefore, this is the first study to show that the number of studies. Further studies might help gain insight novel UBTs might be efficient in reducing the number of into that field (NCT04100733). Although we were not able avoidable cystoscopies, with up to 740 cystoscopies avoided to perform an analysis among LG tumors, it is one of the for 1000 patients. Nevertheless, the quality-of-life analyses indications proposed by some authors during the follow- as well as the cost effectiveness of each UBT still need to be up [31,33,49–51]. Future prospective studies are necessary assessed specifically in each health care system. Such anal- to prove the concept that the currently available markers yses for EpiCheck have shown potential cost effectiveness of could play an important role as an alternate to cystoscopy this approach [48]. Moreover, as the included studies still in patients with LG NMIBC (eg, UroFollow trial) [52]. suffer from limited follow-up, we also reported the predic- Our results demonstrate that the novel UBTs have supe- tive DTA of each UBT according to the 5-yr recurrence rate rior Se and NPV for the detection of recurrences to cytology. within each established NMIBC risk group. This information A combination of these different tests and/or combination might be useful for interpreting the benefit of these tests with cytology might be of interest in select scenarios where according to the desired time length, early and long term, the patient benefits from both high Se (ie, UBT) and high Sp and to define their utility according to the follow-up. (ie, cytology). All these UBTs need to be tested as reflex test- Our subgroup analysis on the detection of HG recurrence ing in patients with equivocal findings such as unclear cys- of NMIBC demonstrated similar diagnostic estimates to toscopic finding or atypical cytology [53]. Nevertheless, those in the overall population for three available UBTs further studies are needed to integrate all criteria needed (Xpert bladder cancer, Bladder EpiCheck, and ADXBLAD- to usher a UBT into widespread use, including cost, ease of DER). However, these results might be limited due to the use, and patient acceptance [54]. 940 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 Although the novel UBTs showed promising results in their potential value in preventing unnecessary cysto- the NMIBC setting, there is a lack of data for their use in scopies. However, there are not enough data to support the initial diagnosis setting [55]. The diagnostic estimates their use in the initial diagnosis setting. Our study might and specificity of Xpert bladder in patients with hematuria help policymakers and stakeholders, such as payors and were promising but need validation [39,40]. To date, the patient advocacy groups, assess the implementation of guidelines do not recommend any UBT routinely in the ini- these tests into daily practice. tial diagnosis [56], but ongoing randomized trials are awaited to determine the benefit of biomarkers in this large Author contributions: Benjamin Pradere had full access to all the data in patient group (NCT03988309). Therefore, cystoscopy the study and takes responsibility for the integrity of the data and the remains as the gold standard in the investigation of visible accuracy of the data analysis. hematuria. One could foresee a point of care test used by general practitioners and other nonurological specialists to select patients for fast-track referral to the urologist [57]. Study concept and design: Pradere. Our study has several potential limitations. The main Acquisition of data: Laukhtina, Mori, Pradere. limitation was the heterogeneity of patient populations in Analysis and interpretation of data: Laukhtina, Mostafaei, Pradere. terms of inclusion criteria and clinicopathological features. Drafting of the manuscript: Laukhtina, Mori, D‘Andrea, Soria, Teoh, Pradere. Moreover, the studies demonstrated high heterogeneity Critical revision of the manuscript for important intellectual content: Com- regarding the prevalence of recurrence rate, which might pérat, Cimadamore, Moschini, Enikeev, Xylinas, Lotan, Palou, Gontero, be an important confounder. Different conditions and previ- Babjuk, Witjes, Kamat, Roupret, Shariat. ous intravesical therapy may also influence the results. No Statistical analysis: Shim, Rajwa. study was precise enough to assess the impact of previous Obtaining funding: Shim, Rajwa. instillation on the diagnostic performance. As there was Administrative, technical, or material support: None. measurable significant heterogeneity across the studies, Supervision: Shariat, Pradere. we used the random-effect model to account for the hetero- Other: None. geneity among studies; nevertheless, our conclusions should still be interpreted with caution. The difference in Financial disclosures: Benjamin Pradere certifies that all conflicts of patient populations and reference standards in the included interest, including specific financial interests and relationships and affili- studies, as well as absence of data on blinding to patholo- ations relevant to the subject matter or materials discussed in the manu- gists or urologists, is likely to be the main reason for the script (eg, employment/affiliation, grants or funding, consultancies, notable risk of bias. Additionally, most of the studies did honoraria, stock ownership or options, expert testimony, royalties, or not report the cutoff of the UBT used. We were able to per- patents filed, received, or pending), are the following: Ekaterina Laukhtina form subgroup analyses for HG recurrence detection only, is supported by the EUSP Scholarship of the European Association of Urol- and not for LG NMIBC due to the lack of data in the litera- ogy (EAU). Keiichiro Mori is supported by The Uehara Memorial ture. Nevertheless, as most UBTs had data on all recurrences Foundation. and HG recurrences, it is unlikely that the analyses in LG recurrence would have yielded effect results. The included studies assessed the diagnostic performance of UBTs; how- Funding/Support and role of the sponsor: None. ever, data on its oncological impact are still lacking. CIS was underevaluated in many studies, which impacts the propor- Appendix A. Supplementary data tion of FP and FN findings, especially in patients under surveillance for CIS. The protocol of used cystoscopy follow-up was not reported in the majority of included Supplementary data to this article can be found online at studies. The cost effectiveness of individual UBTs that https://doi.org/10.1016/j.euo.2021.10.003. belong to national agencies was out of scope in the present meta-analysis. Still, well-designed large-scale trials com- References paring all UBTs head to head are required to confirm the [1] Witjes JA, Bruins HM, Cathomas R, Compérat E, Cowan NC, Efstathiou findings of the present study and to propose detailed JA, et al. EAU Guidelines on Muscle-invasive and Metastatic Bladder follow-up protocols using UBTs. Although UBTs have the Cancer. EAU Guidel. Edn. Present. EAU Annu. Congr. Milan 2021, potential to be used widely in the near future, cystoscopy 2021. cannot completely be abandoned and will continue to have [2] Karakiewicz PI, Benayoun S, Zippe C, Lüdecke G, Boman H, Sanchez- an important role during the follow-up of NMIBC. Moreover, Carbayo M, et al. Institutional variability in the accuracy of urinary cytology for predicting recurrence of transitional cell carcinoma of decision curve analyses will be needed to help clinicians the bladder. BJU Int 2006. https://doi.org/10.1111/j.1464-410X. assess the net benefit of UBTs in this setting. In the mean- 2006.06036.x. time, our results could help health care agencies and stake- [3] Lee LJ, Kwon CS, Forsythe A, Mamolo CM, Masters ET, Jacobs IA. holders decide whether these new UBTs might be of value Humanistic and Economic Burden of Non-Muscle Invasive Bladder Cancer : Results of Two Systematic Literature Reviews. Clin to patients in their respective care area. Outcomes Res 2020:693–709. [4] Soria F, Droller MJ, Lotan Y, Gontero P, D’Andrea D, Gust KM, et al. An up-to-date catalog of available urinary biomarkers for the 4. Conclusions surveillance of non-muscle invasive bladder cancer. World J Urol 2018;36:1981–95. https://doi.org/10.1007/s00345-018-2380-x. [5] Chang SS, Boorjian SA, Chou R, Clark PE, Daneshmand S, Konety BR, Our study supports the promising role of the novel com- et al. Diagnosis and Treatment of Non-Muscle Invasive Bladder mercially available UBTs for the diagnosis of recurrence Cancer: AUA/SUO Guideline. J Urol 2016;196:1021–9. https://doi. during the follow-up of NMIBC. Their performance supports org/10.1016/j.juro.2016.06.049. EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 941 [6] Wolfs JRE, Hermans TJN, Koldewijn EL, van de Kerkhof D. Novel urinary [24] Smrkolj T, Cegovnik Primozic U, Fabjan T, Sterpin S, Osredkar J. The biomarkers ADXBLADDER and bladder EpiCheck for diagnostics of performance of the Xpert Bladder Cancer Monitor Test and voided bladder cancer: A review. Urol Oncol Semin Orig Investig 2021;39: urinary cytology in the follow-up of urinary bladder tumors. Radiol 161–70. https://doi.org/10.1016/j.urolonc.2020.11.014. Oncol 2021. https://doi.org/10.2478/raon-2020-0072. [7] Shim SR, Kim SJ, Lee J. Diagnostic test accuracy: application and [25] Van VFJP, Hiar AM, Wallace E, Bridge JA, Mayne DJ, Beqaj S, et al. practice using R software. Epidemiol Health 2019. https://doi.org/ Prospective Validation of an mRNA-based Urine Test for 10.4178/epih.e2019007. Surveillance of Patients with Bladder Cancer. Eur Urol [8] Leeflang MMG, Deeks JJ, Gatsonis C, Bossuyt PMM. Systematic reviews 2018;75:853–60. https://doi.org/10.1016/j.eururo.2018.11.055. of diagnostic test accuracy. Ann Intern Med 2008;149:889–97. [26] Trenti E, Pycha S, Mian C, Schwienbacher C, Hanspeter E, Kafka M, https://doi.org/10.7326/0003-4819-149-12-200812160-00008. et al. Comparison of 2 New Real-Time Polymerase Chain Reaction – [9] McInnes MDF, Moher D, Thombs BD, McGrath TA, Bossuyt PM, Based Urinary Markers in the Follow-Up of Patients With Non – Clifford T, et al. Preferred Reporting Items for a Systematic Review Muscle-Invasive Bladder Cancer. Cancer Cytopathol 2020:341–7. and Meta-analysis of Diagnostic Test Accuracy Studies: The https://doi.org/10.1002/cncy.22246. PRISMA-DTA Statement. JAMA 2018;319:388–96. https://doi.org/ [27] D’Andrea D, Soria F, Zehetmayer S, Gust KM, Korn S, Witjes JA, et al. 10.1001/jama.2017.19163. Diagnostic accuracy, clinical utility and influence on decision- [10] Whiting PF, Rutjes AWS, Westwood ME, Mallett S, Deeks JJ, Reitsma making of a methylation urine biomarker test in the surveillance of JB, et al. Quadas-2: A revised tool for the quality assessment of non-muscle-invasive bladder cancer. BJU Int 2019;123:959–67. diagnostic accuracy studies. Ann Intern Med 2011. https://doi.org/ https://doi.org/10.1111/bju.14673. 10.7326/0003-4819-155-8-201110180-00009. [28] Pierconti F, Martini M, Fiorentino V, Cenci T, Capodimonti S, [11] Reitsma JB, Glas AS, Rutjes AWS, Scholten RJPM, Bossuyt PM, Straccia P, et al. The combination cytology/epichek test in non Zwinderman AH. Bivariate analysis of sensitivity and specificity muscle invasive bladder carcinoma follow-up: Effective tool or produces informative summary measures in diagnostic reviews. J useless expence? Urol Oncol Semin Orig Investig 2021;39:131. Clin Epidemiol 2005;58:982–90. https://doi.org/10.1016/j. e17–21. https://doi.org/10.1016/j.urolonc.2020.06.018. jclinepi.2005.02.022. [29] Trenti E, D’Elia C, Mian C, Schwienbacher C, Hanspeter E, Pycha A, [12] Shim S, Kim S, Lee J, Rücker G. Network meta-analysis: application et al. Diagnostic predictive value of the Bladder EpiCheck test in the and practice using the R software. Epidemiol Health 2019;41: follow-up of patients with non–muscle-invasive bladder cancer. e2019006. https://doi.org/10.4178/epih.e2019006. Cancer Cytopathol 2019;127:465–9. https://doi.org/10.1002/ [13] Spiegelhalter DJ, Best NG, Carlin BP, Van Der Linde A. Bayesian cncy.22152. measures of model complexity and fit. J R Stat Soc Ser B Stat [30] Wasserstrom A, Frumkin D, Dotan Z, Bukin E, Gadish T, Hanuka S, Methodol 2002;64:583–616. https://doi.org/10.1111/1467-9868. et al. Mp13-15 Molecular Urine Cytology – Bladder Epicheck Is a 00353. Novel Molecular Diagnostic Tool for Monitoring of Bladder Cancer [14] Lu G, Ades AE. Combination of direct and indirect evidence in mixed Patients. J Urol 2016;195:. https://doi.org/10.1016/j.juro.2016.02. treatment comparisons. Stat Med 2004;23:3105–24. https://doi. 2496e140. org/10.1002/sim.1875. [31] Roupret M, Gontero P, McCracken SRC, Dudderidge T, Stockley J, [15] Salanti G, Ades AE, Ioannidis JPA. Graphical methods and numerical Kennedy A, et al. Diagnostic Accuracy of MCM5 for the Detection of summaries for presenting results from multiple-treatment meta- Recurrence in Nonmuscle Invasive Bladder Cancer Followup: A analysis: An overview and tutorial. J Clin Epidemiol 2011. https:// Blinded, Prospective Cohort, Multicenter European Study. J Urol doi.org/10.1016/j.jclinepi.2010.03.016. 2020;204:685–90. https://doi.org/10.1097/JU.0000000000001084. [16] Rücker G, Schwarzer G. Ranking treatments in frequentist network [32] Biaûçûçlek Ł, Czerwińska K, Fus Ł, Krajewski W, Sadowska A, meta-analysis works without resampling methods. BMC Med Res Radziszewski P, et al. MCM5 urine expression (ADXBLADDER) is a Methodol 2015. https://doi.org/10.1186/s12874-015-0060-8. reliable biomarker of high-risk non- muscle-invasive bladder [17] Cancel-Tassin G, Roupret M, Pinar U, Gaffory C, Vanie F, Ondet V, cancer recurrence: A prospective matched case-control study. et al. Assessment of Xpert Bladder Cancer Monitor test performance Cancer Biomarkers 2021;30:139–43. https://doi.org/10.3233/CBM- for the detection of recurrence during non-muscle invasive bladder 200316. cancer follow-up. World J Urol 2021. https://doi.org/10.1007/ [33] Gontero P, Montanari E, Roupret M, Longo F, Stockley J, Kennedy A, s00345-021-03629-1. et al. Comparison of the performances of the ADXBLADDER test and [18] Cowan B, Klein E, Jansz K, Westenfelder K, Bradford T, Peterson C, urinary cytology in the follow-up of non-muscle-invasive bladder et al. Longitudinal Follow-up and Performance Validation of a cancer: a blinded prospective multicentric study. BJU Int mRNA-based Urine Test (XpertÒ Bladder Cancer Monitor) for 2021;127:198–204. https://doi.org/10.1111/bju.15194. surveillance in Non-Muscle Invasive Bladder Cancer Patients. BJU [34] Batista R, Vinagre J, Prazeres H, Sampaio C, Peralta P, Conceição P, Int 2021. https://doi.org/10.1111/bju.15418. et al. Validation of a Novel, Sensitive, and Specific Urine-Based Test [19] Elsawy AA, Awadalla A, Elsayed A, Abdullateef M, Abol-Enein H. for Recurrence Surveillance of Patients With Non-Muscle-Invasive Prospective Validation of Clinical Usefulness of a Novel mRNA- Bladder Cancer in a Comprehensive Multicenter Study. Front Genet based Urine Test (XpertÒ Bladder Cancer Monitor) for surveillance 2019;10:1–15. https://doi.org/10.3389/fgene.2019.01237. in Non Muscle Invasive Bladder Cancer. Urol Oncol 2021;39:77. [35] Sieverink CA, Batista RPM, Prazeres HJM, Vinagre J, Sampaio C, e9–77.e16. https://doi.org/10.1016/j.urolonc.2020.07.013. Leão RR, et al. Clinical validation of a urine test (uromonitor-V2Ò ) [20] Elsawy AA, Awadalla A, Abdullateef M, Ahmed AE, Abol-enein H. for the surveillance of non-muscle-invasive bladder cancer Can repeat biopsy be skipped after initial complete resection of T1 patients. Diagnostics 2020;10. https://doi.org/10.3390/ bladder cancer? The role of a novel urinary mRNA biomarker. Urol diagnostics10100745. Oncol Semin Orig Investig 2021. https://doi.org/10.1016/j.urolonc. [36] Koya M, Osborne S, Chemaslé C, Porten S, Schuckman A, Kennedy- 2021.02.009. Smith A. An evaluation of the real world use and clinical utility of [21] D’Elia C, Folchini DM, Mian C, Hanspeter E, Schwienbacher C, the Cxbladder Monitor assay in the follow-up of patients previously Spedicato GA, et al. Diagnostic value of Xpert Ò Bladder Cancer treated for bladder cancer. BMC Urol 2020;20:1–9. https://doi.org/ Monitor in the follow-up of patients affected by non-muscle 10.1186/s12894-020-0583-0. invasive bladder cancer : an update. Ther Adv Urol 2021:1–7. [37] Lotan Y, O’Sullivan P, Raman JD, Shariat SF, Kavalieris L, Frampton C, https://doi.org/10.1177/https. et al. Clinical comparison of noninvasive urine tests for ruling out [22] Hurle R, Casale P, Saita A, Colombo P, Maria G, Lughezzani G, et al. recurrent urothelial carcinoma. Urol Oncol Semin Orig Investig Clinical performance of Xpert Bladder Cancer (BC) Monitor, a mRNA 2017;35:531.e15–22. https://doi.org/10.1016/j.urolonc.2017.03.008. - based urine test, in active surveillance (AS) patients with [38] Ritch C, Velasquez M, Kwon D, Becerra M, Soodana-Prakash N, recurrent non - muscle - invasive bladder cancer (NMIBC): results Atluri V, et al. Use and Validation of the AUA/SUO Risk Grouping for from the Bladder Cancer Italian Active Surveillance. World J Urol Nonmuscle Invasive Bladder Cancer in a Contemporary Cohort. J 2020;38:2215–20. https://doi.org/10.1007/s00345-019-03002-3. Urol 2020;203:05–511. [23] Pichler R, Fritz J, Tulchiner G, Klinglmair G, Soleiman A, Horninger [39] Van Valenberg FJP, Hiar AM, Wallace E, Bridge JA, Mayne DJ, Beqaj S, W, et al. Increased accuracy of a novel mRNA-based urine test for et al. Validation of an mRNA-based Urine Test for the Detection of bladder cancer surveillance. BJU Int 2018;121:29–37. https://doi. Bladder Cancer in Patients with Haematuria. Eur Urol Oncol org/10.1111/bju.14019. 2020;4:93–101. https://doi.org/10.1016/j.euo.2020.09.001. 942 EUROPEAN UROLOGY ONCOLOGY 4 (2021) 927–942 [40] Wallace E, Higuchi R, Satya M, Mccann L, Sin MLY, Bridge JA, et al. Non-Muscle-invasive Bladder Cancer - A National Study of Patient, Development of a 90-Minute Integrated Noninvasive Urinary Assay Provider, and Facility Factors. Urology 2019:112–9. https://doi.org/ for Bladder Cancer Detection. J Urol 2018;199:655–62. https://doi. 10.1016/j.urology.2019.04.036.Overuse. org/10.1016/j.juro.2017.09.141. [50] Brummelhuis ISG, Witjes JA, Messing EM. The conundrum of recurrent [41] Anastasi E, Maggi M, Tartaglione S, Angeloni A, Gennarini G, low-grade tumours: To treat or to observe? Curr Opin Urol Leoncini PP, et al. Predictive value of MCM5 (ADXBLADDER) 2018;28:557–62. https://doi.org/10.1097/MOU.0000000000000550. analysis in urine of men evaluated for the initial diagnosis of [51] Contieri R, Paciotti M, Lughezzani G, Buffi N, Frego N, Diana P, et al. bladder cancer: A comparative prospective study. Diagn Cytopathol Long-term Follow-up and Factors Associated with Active 2020;48:1034–40. https://doi.org/10.1002/dc.24530. Surveillance Failure for Patients with Non-muscle-invasive [42] Dudderidge T, Stockley J, Nabi G, Mom J, Umez-Eronini N, Hrouda D, Bladder Cancer: The Bladder Cancer Italian Active Surveillance et al. A Novel, non-invasive Test Enabling Bladder Cancer Detection in (BIAS) Experience. Eur Urol Oncol 2021. Urine Sediment of Patients Presenting with Haematuria—A [52] Benderska-Söder N, Hovanec J, Pesch B, Goebell PJ, Roghmann F, Prospective Multicentre Performance Evaluation of ADXBLADDER. Noldus J, et al. Toward noninvasive follow-up of low-risk bladder Eur Urol Oncol 2019;3:42–6. https://doi.org/10.1016/j.euo.2019.06. cancer – Rationale and concept of the UroFollow trial*. Urol Oncol 006. Semin Orig Investig 2020;38:886–95. https://doi.org/10.1016/j. [43] O’Sullivan P, Sharples K, Dalphin M, Davidson P, Gilling P, Cambridge urolonc.2020.01.006. L, et al. A multigene urine test for the detection and stratification of [53] Konety B, Shore N, Kader AK, Porten S, Daneshmand S, Lough T, bladder cancer in patients presenting with hematuria. J Urol et al. Evaluation of Cxbladder and Adjudication of Atypical Cytology 2012;188:741–7. https://doi.org/10.1016/j.juro.2012.05.003. and Equivocal Cystoscopy(Figure presented.). Eur Urol [44] Davidson PJ, McGeoch G, Shand B. Inclusion of a molecular marker 2019;76:238–43. https://doi.org/10.1016/j.eururo.2019.04.035. of bladder cancer in a clinical pathway for investigation of [54] Bensalah K, Montorsi F, Shariat SF. Challenges of Cancer Biomarker haematuria may reduce the need for cystoscopy. N Z Med J Profiling. Eur Urol 2007. https://doi.org/10.1016/j.eururo.2007.09. 2019;132:55–64. 036. [45] Fu L, Wang R, Yin L, Zhang P, Tian Y, Niu M. Nuclear matrix protein [55] Woldu SL, Souter L, Boorjian SA, Barocas DA, Lotan Y. Urinary-based 22 (NMP22) test in the diagnosis of bladder cancer: A meta- tumor markers enhance microhematuria risk stratification analysis. Int J Clin Exp Med 2016;9:7965–75. according to baseline bladder cancer prevalence. Urol Oncol Semin [46] hou R, Gore JL, Buckley D, Fu R, Gustafson K, Griffin JC, et al. A Orig Investig 2021. https://doi.org/10.1016/j.urolonc.2021.03.022. Systematic Review and Meta-analysis Annals of Internal Medicine [56] Bellmunt J, Hussain M, Gschwend JE, Albers P, Oudard S, Castellano Urinary Biomarkers for Diagnosis of Bladder Cancer 2015. D, et al. Adjuvant atezolizumab versus observation in muscle- https://doi.org/10.7326/M15-0997. invasive urothelial carcinoma (IMvigor010): a multicentre, open- [47] Leal J, Luengo-Fernandez R, Sullivan R, Witjes JA. Economic Burden label, randomised, phase 3 trial. Lancet Oncol 2021;22:525–37. of Bladder Cancer Across the European Union. Eur Urol https://doi.org/10.1016/S1470-2045(21)00004-8. 2016;69:438–47. https://doi.org/10.1016/j.eururo.2015.10.024. [57] Khadhouri S, Gallagher KM, MacKenzie KR, Shah TT, Gao C, Moore S, [48] Lotan Y, Gakis G, Manfredi M, Morote J, Mostafid H, Porpiglia F, et al. et al. The IDENTIFY Study: The Investigation and Detection of Alternating Cystoscopy with Bladder EpiCheck Ò in the Surveillance Urological Neoplasia in Patients Referred with Suspected Urinary of Low-Grade Intermediate-Risk NMIBC: A Cost Comparison Model. Tract Cancer; A multicentre observational study. BJU Int 2021. Bl Cancer 2021:1–9. https://doi.org/10.3233/blc-211528. https://doi.org/10.1111/bju.15483. [49] Han D, Lynch K, Chang J, Sirovich B, Robertson D, Swanton A, et al. Overuse of Cystoscopic Surveillance Among Patients With Low-risk