Overview of Urine Tumor Markers
More than 30 urinary biomarkers have been reported for use in bladder cancer diagnosis, but only a few are commercially available; the remainder are still being tested. [1, 2] Commercially available tests include the following:
Fluorescence in situ hybridization (FISH)
Nuclear matrix protein (NMP-22)
Newer, voided urine assays (ie, bladder tumor antigen [BTA stat, BTA TRAK], NMP-22, fibrin/fibrinogen degradation products [FDP]) are being used for the detection and surveillance of urothelial carcinoma. These tests have high false-positive and false-negative rates. In the future, other newer assays based on telomerase and microsatellite analysis may prove to be a better detection method than urinary cytology. 
Chromosomal alterations have been associated with urothelial carcinoma. One encouraging test is a multitarget interphase FISH assay called UroVysion that consists of probes to the centromeres of chromosomes 3, 7, 17, and 9p21. Aneuploidy of chromosomes 3, 7, and 17 and deletion of chromosome 9 have been associated with high sensitivity and specificity to detect bladder cancer. Often, this is an anticipatory positive result with a positive finding preceding visual evidence of bladder tumor.
The use of additional urine markers such as UroVysion (FISH), BTA, and NMP-22 in the initial diagnosis of bladder cancer is controversial. [4, 5, 6] All of these assays may yield false-positive and false-negative results.
Genetic Aberrations in Bladder Cancer
The study of genetic aberrations commonly associated with urothelial carcinoma provides a more objective assessment for diagnosing and detecting recurrent disease. Homozygous loss of band 9p21, the site for the tumor suppressor gene P16, is a known early genetic event in the development of papillary carcinoma and urothelial carcinoma in situ (CIS). 
Increased chromosomal instability and aneuploidy have been implicated in tumor progression. A study by Sokolova et al of 9 genetic markers for detecting urothelial carcinoma showed that polysomy of chromosomes 3, 7, and 17 and deletion of 9p21 were the most sensitive and specific markers, detecting 95% of recurrent urothelial carcinomas.  Halling et al established that a threshold of 5 or more cells with polysomy was 84% sensitive and 92% specific for detecting recurrent urothelial cancer. 
Fluorescence In Situ Hybridization
A commercial FISH assay (UroVysion Bladder Cancer Kit, Abbott Molecular; DesPlaines, Ill), which includes probes for the centromeres for chromosomes 3, 7, and 17 and has a locus-specific probe for 9p21, was developed to screen for recurrent urothelial carcinoma and was approved by the US Food and Drug Administration (FDA) in 2005. The intended use is as an aid for initial diagnosis of bladder cancer in patients with hematuria and subsequent monitoring for tumor recurrence in patients previously diagnosed with bladder cancer.
Initial comparisons of urine cytology with FISH for detecting bladder cancer recurrence showed that FISH yielded a greater sensitivity.  FISH is 42-83% sensitive for detecting pTa and pT1 lesions and 92-100% sensitive for pT2-4 invasive lesions in patients with known bladder cancer, while urine cytology yields sensitivities of 24-50% for pTa and pT1 lesions and 78-85% for pT2-4 invasive lesions. 
For suspected new cases of urothelial carcinoma, cytology yields a reported diagnostic sensitivity of 48%. 
Laudadio et al found that FISH is considerably more sensitive and only slightly less specific than cytology in diagnosing urothelial carcinoma. FISH analysis yielded a high sensitivity for detecting new cases of urothelial carcinoma, as well as recurrences. FISH detected 95% of cases with high-grade carcinoma, while cytology detected 41% of such cases. FISH yielded an overall specificity of 65%, compared to 93% with cytology. These researchers recommended FISH as a useful initial diagnostic tool in patients suspected of both new and recurrent bladder cancer.
Nuclear Matrix Protein–22
Nuclear matrix, first described in 1974, is the nonchromatin structure that supports nuclear shape and organizes DNA. It also takes part in DNA replication and transcription, as well as RNA processing. [12, 13, 14]
NMP-22 is involved in the proper distribution of chromatin to daughter cells during cell division and is found in the nuclear matrix of all cell types. NMP-22 is thought to be released from the nuclei of tumor cells after they die and can be detected in the urine. Research has found that persons with bladder cancer may have urinary NMP-22 levels up to 25 times that in healthy persons. 
The NMP-22 BladderChek test (Alere; Waltham, Mass) is an in vitro immunoassay intended for the qualitative detection of NMP-22 in urine. Unlike cytologic analysis, the NMP-22 test does not depend on intact cells and does not require expert analysis or laboratory time. The test provides an absolute positive or negative test result, much in the same manner as a pregnancy test.
It is a painless and noninvasive assay that provides results within 30 minutes (thus allowing performance during an office visit), and its cost is less than half that of cytology. It is the only in-office test approved by the FDA in 2000 for the diagnosis of bladder cancer.
Grossman et al reported that when combined with cystoscopy, the NMP-22 test improves the detection of recurrence in patients with a history of bladder cancer.  Initial cystoscopy alone detected 91% of the cancers. The combination of the NMP-22 test with cystoscopy increased overall sensitivity to 99%. The NMP-22 test was significantly more sensitive than cytologic analysis of voided urine.
Of concern with the NMP-22 assay is its variability of performance in detecting bladder cancer. A multicenter study by Shariat et al assessed the variability in the diagnostic performance of NMP-22 for detecting recurrence and progression in patients with previous Ta, T1, and/or CIS and found that the manufacturer cutoff of 10 U/mL detected 57% of cases with a 19% false-positive rate.  For each NMP-22 cutoff assessed, NMP-22 had a higher sensitivity for detecting grade III and stage T2 or greater bladder cancer than for detecting any cancer.
No optimal cutoffs for detecting any or aggressive bladder cancer could be derived based on NMP-22 values. The authors concluded that there is a substantial degree of heterogeneity in the diagnostic performance of NMP-22 applied to populations from different institutions. There was no clearly defined NMP-22 cutoff, but there was a continuum of risk for recurrence and progression.
Bladder Tumor Antigen (BTA) STAT and TRAK
The BTA STAT & TRAK tests (Polymedco; Cortlandt Manor, NY) use monoclonal antibodies to detect complement factor H-related protein and complement factor H in voided urine specimens. These factors are found in bladder cancer cell lines and inhibit the complement cascade to prevent cell lysis.BTA STAT is a point of care qualitative assay with an average sensitivity and specificity of 68.7% (53-89%) and 73.7% (54-93%), respectively.BTA TRAK is a quantitative enzyme-linked immunosorbent assay with similar sensitivity and specificity of 62% (17-78%) and 73.6% (51-95%), respectively.  Studies have shown wide ranges of sensitivity and specificity with BTA testing. Additionally, the specificity of both of these tests can be significantly decreased, as false positives have been noted to occur in the setting of hematuria, urolithiasis, inflammation, recent instrumentation, other genitourinary malignancies, and intravesical BCG therapy. 
The ImmunoCyt/uCyt+ test (Scimedex; Denville, NJ) was originally developed in 1997 by Fradet and Lockhard as an immunohistochemical test with 3 fluorescent monoclonal antibodies directed at urothelial cell antigens found on exfoliated cells. Two antibodies, LDQ10 and M344, are directed against mucins, specifically glycoproteins found on epithelial cell surfaces in malignancy, and labeled with fluorescein. [19, 20, 21] The other antibody is labeled Texas red and directed against a high molecular weight glycosylated form of carcinoembryonic antigen 19A211. M344 and CEA 19A211 have been found to be expressed in 71% and 90% of Ta-T1 tumors, respectively. 
However, the test is limited in that it requires processing in laboratories with properly trained personnel. Additionally, it requires a minimum of 500 negative cells on the slide in order for the sample to be deemed negative.  As is common with other protein-based assays, false positives are common in the setting of urinary tract infection, urolithiasis, and benign prostatic hyperplasia.
The use of Immunocyt/uCyt+ improves sensitivity at a minimum of 15% over cytology alone and ranges from 53.8-94.1%, with greater improvements in low-grade tumors. This improves further with the combination of cytology and Immunocyt/uCyt+. Specificity is slightly lower than that of cytology, at 61-80.7%. The negative predictive value is improved with the use of Immunocyt/uCyt+, at 81-96.2%, whereas the positive predictive value is worse (26-63.2%).  In a recent analysis, Comploj et al. noted that the sensitivity of Immunocyt/uCyt+ increased from 63% for a pTa tumor to 80% for a pT1 tumor and in combination with cytology it increased from 65% to 88%. Conversely, poorer sensitivity was noted in T2 tumors (68%), likely because the 19A211, M344, and CEA 19A211 antigens are not found in muscle-invasive tumors. 
Recently, nomograms using immunocytology (Immunocyt/uCyt+) have been created for the detection of bladder cancer. According to the authors, the addition of immunocytology improved the diagnostic accuracy of the nomogram and outperformed cytology with a significantly higher sensitivity and negative predictive value.
Fibroblast growth factor receptor 3 (FGFR3) belongs to a family of tyrosine kinase receptors and is encoded by the FGFR3 gene. Specific point mutations in various domains result in constitutive activation of the receptor and have been found in approximately 50% of urothelial carcinomas.  The frequency of these mutations is high in low-grade pTa tumors and low in pT1-4 tumors. The presence of the FGFR3 mutation is a selective marker for favorable disease, with a low recurrence rate and improvement in disease-specific survival. [23, 24, 25]
Predictive Biosciences (Lexington, Mass) developed a urine-based FGFR3 assay using ultradeep amplicon sequencing, allowing a high clinical sensitivity of 55.8% and specificity of 100%, similar to that of tissue.  They have implemented the use of this assay in a multianalyte diagnostic assay, CertNDx, used for the evaluation of hematuria and monitoring bladder cancer recurrence. For the diagnosis of urothelial carcinoma, they analyze urine for the presence of mutant FGFR3, quantified matrix metalloproteinase 2 (MMP-2), and hypermethylation of TWIST1 and NID2. This allows for the presence of 2 biomarker cutoff values. The presence of FGFR3 provides a high positive predictive value of 95.2%, while the lack of all 4 biomarkers provides a negative predictive value of 98.2%.  Furthermore, unlike other assays, this is not affected by the degree of hematuria or presence of other urinary tract diseases. 
CxBladder (PacificEdge, Dunedin, New Zealand) is a recently released urine-based assay consisting of 5 mRNA markers, CDC2, HOXA13, MDK, IGFBP5, and CXCR5.  The addition of CXCR5, or a mediator of neutrophil migration to sites of inflammation, allows the reduction of false positives secondary to the presence of acute or chronic inflammation. The sensitivity of this assay is superior to that of NMP-22 and cytology, at 83%, with a specificity of 85%. Interestingly, the specificity for high-grade tumors was 97% while the specificity for low-grade tumors was 69%. This offers a potential adjunct to cystoscopy for the diagnosis of urothelial carcinoma.
Role of Urine Markers
Several reviews have been performed to assess the myriad urine markers proposed for bladder cancer surveillance. They note that none of the markers has been proven sensitive and specific enough to replace cystoscopy.  The combination of multiple urinary biomarkers with cytology has been studied, with encouraging results for both the diagnosis and surveillance of bladder cancer. [31, 32] While commercially available urinary markers are promising, the clinical evidence is insufficient to warrant the substitution of the cystoscopic follow-up scheme with any of the currently available urine marker tests. [33, 34] If FISH and NMP-22 are considered to have some utility when used to complement or replace cytology, a dilemma arises when their results conflict with each other. Of particular interest is how to treat a patient with positive cytology and/or FISH findings when cystoscopy findings are negative.
Other Potential Biomarkers
The field of urothelial carcinoma tumor markers is an area of significant interest as a means to potentially improve cancer detection, as well as the cost and anxiety associated with surveillance. Several potentially clinically relevant biomarkers are under investigation, including bladder cancer–specific nuclear matrix protein (BCLA-1 and BCLA-4), Lewis X antigen, Aurora kinase A, carcinoembryonic antigen-related cell adhesion molecule (CEACAM1), epigenetic markers, microsatellite analysis miRNA markers,  DD23, quanticyte nuclear karyometry, cytokeratins, hyaluronic acid, and survivin.  However, further studies are required to determine their application in the diagnosis and monitoring of urothelial carcinoma.