The Current Role of Circulating Biomarkers in Non-Muscle Invasive Bladder Cancer

Michael Rink; Heidi Schwarzenbach; Malte W. Vetterlein; Sabine Riethdorf; Armin Soave; on behalf of the EAU Young Academic Urologists—Urothelial Cancer Working party


Transl Androl Urol. 2019;8(1):61-75. 

In This Article



Origin of cfDNA in the urine. cfDNA clearance from the blood is warranted by liver and kidney, and its half-life is variable ranging from 15 minutes to several hours.[26] cfDNA has to pass through the renal filtration system to be ultimately released into the urine. This kidney barrier has been shown to be permeable for DNA molecules, but only complexes smaller than 6.4 nm in diameter and with a molecular weight ≤70 kDa corresponding to DNA of about 100 bp in size can pass through it and enter the nephron. Thus, cfDNA fragments of 50–100 bp in size and those which are only partially protected by histones can reach the urine, but possibly the non-globular shape or deformability of cfDNA may allow the passage of longer fragments through the barrier. However, it should also considered that the presence of apoptotic and necrotic urinary tract cells is another important source for cfDNA in the urine.[27] In this regard, Su et al. reported the presence of low-molecular weight cfDNA in size of 150–250 bp as well as high-molecular weight cfDNA longer than 1 kb in urine. These findings suggest that the low-molecular weight cfDNA stems from the blood circulation, and the high-molecular cfDNA originates mostly from cells shed into the urinary tract.[28]

The history and introduction of ctDNA analyses in UCB. As previously reported,[24] in UCB, cfDNA was initially analyzed in urine.[29] Although urine, particularly from UCB patients, is well eligible for cfDNA analyses, the fragmentation of cfDNA may be higher in urine than in plasma or serum, and therefore, disturb the analyses. Extensive research on ctDNA in plasma and serum of UCB started at the beginning of this century. At this time, the studies by von Knobloch et al.[30] and Utting et al.[31] showed that microsatellite instability (MSI) assessed by fluorescence PCR cannot only be detected in cfDNA isolated from urine but also from serum and plasma of UCB patients. Simultaneously, Dahse et al. evaluated TP53 alterations as a potential marker for a non-invasive diagnosis of recurrences or residuals in superficial UCB patients, but they only re-detected TP53 mutations from the primary tumor in 25% of plasma/serum samples using direct genomic sequencing.[32] Apparently, the former sequencing method was not enough sensitive. In the same year, Domínguez et al. reported that p14ARF promoter hypermethylation or MSI in plasma was associated with recurrence in UCB patients.[33] In particular, further small studies revealed hypermethylation of APC, GSTP1, TIG1, DAPK, p16 and cadherin promoters in serum cfDNA and its association with clinico-pathologic features.[34–38]

ctDNA in NMIBC.Table 1 presents an overview on ctDNA data in NMIBC. Recent advances in DNA profiling techniques have improved detection of tumor-associated genomic aberrations in peripheral blood. To date, most studies have applied polymerase chain reaction (PCR)-based methods or next-generation sequencing (NGS) approaches. An important study on ctDNA in NMIBC was carried out by Birkenkamp-Demtröder et al. in 2016. Using droplet digital PCR (ddPCR), this research group detected somatic variants of ctDNA, including deletions, insertions, inversions as well as intra- and inter chromosomal translocations, in both plasma and urine of NMIBC patients, and demonstrated that low levels of ctDNA in NMIBC are no barrier for their clinical utility.[10] Thereupon, Puntoni et al. measured the serum levels of VEGF by a quantitative sandwich enzyme immunoassay and found that they are a significant predictor of overall survival and may help to identify such high-risk NMIBC patients who may benefit from more aggressive therapy.[44]

Frequent activating (hotspot) mutations have been identified in the fibroblast growth factor receptor (FGFR) and phosphatidylinositol 3-kinase (PIK3). Their dysregulations are potentially accountable for the initiation and progression of NMIBC, since their signaling pathways regulate cell proliferation, differentiation, migration, angiogenesis and tumorigenesis.[45,46] In a recent study, Christensen et al. carried out ddPCR analyses and screened ctDNA for these hotspot mutations in urine and plasma from NMIBC and MIBC patients undergoing radical cystectomy. They demonstrated that high levels of ctDNA in serial urine supernatants from the NMIBC cohort were associated with later disease progression. In the plasma samples, high levels of ctDNA were associated with recurrence in patients undergoing radical cystectomy. Increased levels of FGFR3 and PIK3CA mutated ctDNA in urine and plasma are indicative for later progression and metastasis in bladder cancer, and a positive correlation of ctDNA levels between urine and plasma was observed. However, the associations of urine and plasma ctDNA with the patient risk factors were not thoroughly congruent. The authors emphasize the observation, that urine supernatant ctDNA may also originate from renal clearance of ctDNA in the circulation, and consequently, its presence may not be bladder cancer specific.[39] Kim et al. examined the relevance of urine cfDNA levels of topoisomerase 2-alpha, a DNA gyrase isoform that plays an important role in the cell cycle, as a noninvasive diagnostic marker for bladder cancer. Receiver operating characteristics (ROC) curve analysis revealed that the area under the ROC curve (AUC) was 0.701 with a sensitivity of 63%, specificity of 70%, positive predictive value of 48% and negative predictive value of 82% for detecting NMIBC.[40]

DNA methylation is a common early event in carcinogenesis and thus may represent a potential risk factor. Although epigenetic alterations are not unique for one cancer type, there are tumor suppressor genes that are frequently methylated and down-regulated in bladder cancer.[47] In 2005, Friedrich et al. underlined the usefulness of gene methylation as a prognostic marker in NMIBC patients. They investigated the methylation status of a large panel of 20 genes in microdissected tumor samples using methylation sensitive real-time PCR. They could identify six highly methylated genes (SOCS-1, STAT-1, BCL-2, DAPK, TIMP-3, E-Cadherin), that were associated with tumor recurrence. In addition, methylation of TIMP-3 predicted prolonged disease-free survival.[48] In 2012, Reinert et al. demonstrated that methylation levels of cfDNA can also be measured in urine of NMIBC patients. They found that methylation levels of EOMES, HOXA9, POU4F2, TWIST1, VIM, and ZNF154 in urine specimens may be promising diagnostic biomarkers for disease surveillance. Performing MethyLight PCR, they detected significant hypermethylation of all six markers in NMIBC, achieving sensitivity in the range of 82–89% and specificity in the range 94–100%. For the use in disease surveillance, the evaluation of cfDNA hypermethylation in urine revealed sensitivities of 88–94% and specificities of 43–67%.[41] Roperch et al. showed that the analysis of the mutation status of FGFR3 cfDNA using allele specific PCR combined with cfDNA methylation status of HS3ST2, SLIT2 and SEPTIN9 in urine improved the risk stratification of NMIBC patients, and may be a useful strategy in diagnosis and surveillance. Using a logistic regression analysis, they found a sensitivity of over 90% and an AUC of over 0.80 for diagnosis and follow-up.[42] Finally, Shindo et al. investigated the clinical utility of cfDNA methylation in urine for detection of intravesical recurrence of NMIBC patients who had undergone transurethral resection. Using bisulfite pyrosequencing, they analyzed the methylation status of four microRNA genes (miR-137, miR-124-2, miR-124-3, and miR-9-3), and found that elevated levels of cfDNA methylation in urine are strongly associated with later radical cystectomy, and may also be a useful tool for detecting and predicting recurrence.[43]


CTC detection techniques. As previously reported,[24] CTC are rare cells in the peripheral circulation. In consequence, elaborate techniques are necessary for the enrichment, detection and analysis of CTC. Currently, more than 50 assays are available, of which enrichment and detection methods base on physical properties, e.g., cell size, plasticity density or dielectrophoretic mobility or on antigen expression.[49,50] Several elaborate reviews addressed specific advantages, challenges and future opportunities of different CTC assays.[50–52] Here, we summarize specifications and differences between the currently most relevant CTC platforms.

CellSearch®. Still, there is only one standardized platform for CTC detection, i.e., the semi-automated CellSearch® system, which has been cleared by the Food and Drug Administration (FDA) for the analysis of blood from patients with metastatic breast, prostate and colorectal cancer.[50,53–55] With CellSearch®, CTC are enriched from venous blood by positive selection using epithelial cell adhesion molecule (EpCAM) antibodies coated to ferric nanoparticles. Indeed, it was demonstrated that 96% of bladder tumors in cystectomy specimens express EpCAM.[56] To identify CTC, the enriched cells will be immunostained with anti-keratin antibodies. 4,6-diamino-2-phenylindole (DAPI) is used for staining of the nuclei and of leukocytes are excluded by CD45 positivity. Subsequently, an automated fluorescence microscope scans selected cells and the presented images are further evaluated by experienced tumor-biologists.[50] To be classified as CTC, selected cells must meet specific morphological criteria, including a minimum diameter of 4μm, a round or oval shape and a visible nucleus within the cytoplasm.[57] Numerous studies showed high sensitivity, specificity and reproducibility of CTC detection by CellSearch®.[50,53,58,59] The presence of CTC seems to be associated with metastatic disease on FDG-PET-CT imaging studies.[56] In addition, further phenotypical and molecular characterization of CTC has been established,[50] including analysis of antigens, which might be relevant for targeted therapy, e.g., human epidermal growth factor receptor (HER2),[12] programmed death ligand-1 (PD-L1),[60] as well as analyses of transcriptomes[61–63] and genomic aberrations using fluorescence in situ hybridization (FISH) and/or PCR-based techniques.[64–67] However, the standardized CellSearch® system implicates critical limitations: it is possible that this assay fails to detect cells that completely lost EpCAM and/or keratin expression, which occurs during epithelial-mesenchymal transition (EMT). In UCB and other solid malignancies, EMT is essential for the metastatic process.[20,25,68] Thus, CellSearch® potentially misses parts of the most dedifferentiated and—from an oncological perspective—most interesting cells.[57] Thus, assays, which allow CTC capturing independent of EpCAM expression may be advantageous.[69] In addition, it remains a point of continuing discussions, whether all selected cells fulfilling the criteria for CTC are able to initiate metastasis.[57,70]

CELLection™ Epithelial Enrich Dynabeads®. Corresponding to CellSearch®, CELLection™ Epithelial Enrich Dynabeads® is another EpCAM-dependent assay for positive enrichment and detection of CTC. Dynabeads® are uniform, super-paramagnetic polymer beads with a diameter of 4.5 μm coated with monoclonal EpCAM antibodies. In contrast to CellSearch®, captured cells are subsequently lysed and analyzed for CD45 and keratin expression using reverse transcription-polymerase chain reaction (RT-PCR).[21] CTC are defined as keratin-positive cells without CD45 expression.[71]

AdnaTest®. Corresponding to CellSearch® and CELLection™, the AdnaTest® is an immunological essay, which positively captures and selects CTC from peripheral blood.[21] In contrast to CellSearch® and CELLection™, it does not rely solely on EpCAM-dependent CTC capture, but on various epithelial markers, e.g., EpCAM, epidermal growth factor receptor (EGFR) or HER2 by an antibody-mix (anti-EpCAM, anti-HER2, anti-EGFR) linked to magnetic particles.[72] Following immuno-magnetic enrichment, RT-PCR and multiplex PCR analyze cancer-specific transcripts, e.g., EMT-related and tumor stem cell-related markers like PI3Kα, TWIST1 AKT2 and ALDH1.[72] For a positive CTC status, at least one of the cancer-specific transcripts must exceed a pre-defined threshold.[21] The AdnaTest® has been evaluated and commercialized for colon, prostate, ovarian, breast and bladder cancer.[72–75]

AccuCyte®-CyteFinder®. The AccuCyte®-CyteFinder® is an EpCAM-independent density-based cell separation platform using two complementary technologies. The AccuCyte® system uses a unique separation tube and collector device to separate the buffy coat from red blood cells and plasma. The whole buffy coat is completely harvested without cell lysis or wash steps, which may cause loss of a relevant number of CTC.[76] The CyteFinder® is an automated scanning digital fluorescent microscope and image analysis system, which allows imaging of cells after staining with specific antibodies, e.g., anti-EpCAM, anti-EGFR, anti-CD45, anti-keratin. After definite classification as CTC, the integrated CytePicker™ device is used to retrieve CTC, and these cells can be further characterized by genomic analysis.[76] The AccuCyte®-CyteFinder® has thus far been used for detection and characterization of CTC in prostate and bladder cancer.[69,77]

CTC in NMIBC.Table 2 presents selected studies on CTC in NMIBC treated with TURBT. During TURBT bladder cancer cells can be released to the irrigation fluid and urine and potentially can be washed out into the blood stream when vessels are lanced during resection. TURBT itself might cause measurable seeding of CTC into the vascular system in patients with muscle invasive and NMIBC.[78,83] The number of detected CTC was higher in blood withdrawn from a venous catheter placed in the inferior cava vein compared to blood withdrawn from peripheral veins.[78] However, the oncologic impact of intra-operatively released CTC remains currently undefined.

Prior to TURBT, CTC are detectable in a relevant number of patients with NMIBC. In the majority of studies, the presence of CTC was not associated with clinico-pathologic characteristics in NMIBC. Using the CELLection™ assay, 44% of 54 patients with pT1 high-grade UCB had presence of CTC. In total, 92% of CTC showed expression of survivin, which was measured by RT-PCR. In patients with presence of survivin-expressing CTC, the intravesical tumor tissue expressed survivin in 82% of patients. A positive CTC status was an independent risk factor for reduced recurrence-free survival in multivariable analysis [odds ratio (OR): 16.7; 95% confidence interval (CI): 3.6–77.5].[71] A long-term follow-up evaluation (median follow-up: 9 years) of this cohort corroborated that a positive CTC status was associated with decreased recurrence-free survival (CTC+ vs. CTC−: 23 vs. 89 months; P value <0.001).[79]

Other studies using CellSearch® showed that CTC were present in 18–20% of patients with NMIBC with a mean CTC number of one to 1.5 per 7.5 mL blood. After a median follow-up of up to 24 months, a positive CTC status was associated with inferior oncologic outcomes, i.e., reduced recurrence-free and progression-free survival.[80,81] The presence of CTC was associated with increased tumor stage, since CTC were found in 8 patients (32%) with pT1 and in no patients with pTa tumors.[80] In addition, the presence of CTC was associated with presence of carcinoma in situ (CIS), since CTC were found in 5 patients (62.5%) with CIS vs. 3 patients (8.3%) without CIS.[80] Due to the small sample size, multivariable analysis could not be performed. However, in Kaplan-Meier analysis recurrence-free survival was reduced in patients with presence of CTC compared to CTC-negative patients (6.5 vs. 21.7 months; P value <0.001).[80] In the currently largest prospective study including 102 patients with high-risk pT1 bladder cancer treated with TURBT plus intravesical BCG immuno-therapy, a positive CTC status was associated with several clinico-pathologic characteristics, i.e., female gender, a tumor size exceeding 3 cm, presence of CIS, tumor multi-focality and presence of lympho-vascular invasion (LVI).[81] In multivariable analysis adjusting for established outcome prognosticators, the presence of CTC was an independent predictor for reduced recurrence-free survival [hazard ratio (HR): 2.92; 95% CI: 1.38–6.18] and the strongest predictor for progression-free survival (HR: 7.17; 95% CI: 1.89–27.21). In addition, a positive CTC status had a positive and negative predictive value of 75% and 93% for disease progression, respectively.[81]

A recent study comparing detection rates of CTC between two assays found higher detection rates with CELLection™ (44.4%) vs.CellSearch® (19.8%) in 155 high-risk NMIBC patients treated with TURBT. Peripheral blood of 101 patients (65.2%) was analyzed with CellSearch®, and of 54 patients (34.8%) with CELLection™. There was no difference in age, gender, presence of CIS, tumor multi-focality and tumor size between the CellSearch® group and the CELLection™ group. Both, CTC detected with CELLection™ vs. CellSearch®, had a negative impact on disease recurrence and disease progression.[82] However, the authors concluded that—although comparing reliability and efficacy between these two approaches is difficult—CellSearch® seems to be more reliable and more efficient to correlate with recurrence-free and progression-free survival.[82] Especially in patients with high-risk NMIBC, CTC may therefore be helpful for identifying those patients, who need more aggressive treatment, e.g., systemic chemotherapy and/or early RC.

In a bladder cancer cohort consisting of 83 patients and 29 controls, the AdnaTest® detected CTC in 6.7% of patients with NMIBC, 15% of patients with MIBC and 18.7% of patients with metastatic disease.[72] Transcripts for the epithelial marker HER2, EMT-related marker PI3Kα and tumor stem cell-related marker ALDH1 were present in 6.7%, 3.3% and 10% of NMIBC patients, respectively.[72] The authors concluded that detection of stem cell-related as well as EMT-related transcripts in patients with missing epithelial transcripts may indicate presence of a subgroup of CTC that could be missed by epithelial marker-dependent methods.[72]

To circumvent EpCAM-dependent selection of CTC, a novel method, i.e., selection-free AccuCyte®-CyteFinder® system has been described.[76] This method allows identifying EpCAM-negative as well as EpCAM-positive CTC using high throughput imaging without need for an initial selection step for CTC capture.[69] This platform detected CTC in 29 bladder cancer patients with non-muscle invasive, muscle invasive and metastatic disease.[69] When applying the definition of CTC as any keratin-positive and white blood cell marker-negative cell, 25% of NMIBC patients had presence of CTC. Interestingly, all NMIBC patients with presence of CTC had pT1 disease. In contrast, when applying the more rigorous definition of CTC with the additional requirement of EpCAM-positivity, no CTC were detected in any NMIBC patient.[69] Future studies are warranted to evaluate the prognostic utility of this assay, especially to further characterize the impact of CTC without EpCAM expression on survival.