Among the non-invasive urine tests, urine tests with molecular biomarkers are promising tools for the diagnosis of BC [16]. However, these tests are still not well established in daily clinical routine and the standard diagnostic workup of BC. Further evaluation of the usefulness of these biomarkers in complex clinical situations is needed before they are recommended for widespread clinical use. BC patients often exhibit haematuria, which is also the most common manifestation of other urological diseases [11, 12]. Hence, it is important to evaluate the usefulness of molecular biomarkers of BC in the context of haematuria interference.
The BTA stat test is a molecular urine test for BC that is currently subject to the most attention. The BTA stat test specifically recognizes bladder tumour-associated antigen (hCFHrp) in urine through monoclonal antibodies. This antigenic protein is believed to be isolated from the urine of patients with BC but cannot be detected in the urine of most healthy individuals [9]. Previous studies have compared the sensitivity of the BTA stat test and cytology detection in patients with BC diagnosed by cystoscopy, and the sensitivity of the BTA stat test was found to be superior to that of urine cytology and bladder irrigation cytology [17-19]. However, data regarding its effectiveness in real-world clinical situations are limited. Specifically, its specificity under haematuria interference situations needs further observations. In this study, we investigated the effect of haematuria on the sensitivity and specificity of the BTA stat test in a clinical cohort.
Previous studies have shown that the sensitivity of the BTA stat test ranges between 57% and 83% [20]. In our study, the sensitivity of the BTA stat test in detecting patients without haematuria BC was 80.77%, which is consistent with previous studies. When we analysed only patients with haematuria, the overall sensitivity of the BTA stat test was 97.22%. This finding illustrates the effect of haematuria on its sensitivity. Because of false-positives, haematuria patients may exhibit higher sensitivity. Our research shows that haematuria had a significant effect on the specificity of the BTA stat test in each group. In non-BC with haematuria patients, the specificity of the BTA stat test was significantly lower than that in patients without haematuria who had the same conditions (the data show that except for the UTI and renal cancer groups, the reasons may be that urinary tract infections also significantly affect the specificity of the BTA stat test and that the sample size of the renal carcinoma group was too small). Furthermore, compared with other studies, our study showed a lower specificity. In a multicentre US study, the specificity of the BTA stat test in patients with BPH, urolithiasis and UTI was 88.5%, 50% and 76%, respectively [21]. In this study, the specificity of the BTA stat test decreased to only 31.89% and 15.52% in patients with urolithiasis and UTI, respectively. In patients with prostatic cancer and renal cancer, the specificity of the BTA stat test was only 44.87% and 48.39%, respectively, which may be because we did not exclude patients with haematuria from our total enrolment in these data. In addition, our study found that the specificity of the BTA stat test was correlated with the severity of haematuria. The specificity of the BTA stat test in patients with haematuria detected under a microscope was significantly higher than that in patients with gross haematuria.
Our results suggest that haematuria leads to high positive rates. Considering that hCFHrp is a serum factor, the BTA stat test can possibly concomitantly detect serum proteins that cause haematuria. A study supports our conclusion; the Makito Miyake group built an experimental haematuria model. These authors added whole blood to BTA-negative urine samples and found that spiking BTA-negative urine samples with as little as 1 µl whole blood/10 ml urine was enough to produce a positive BTA stat test result [22]. These studies confirm that the BTA stat test is very unsatisfactory for the diagnosis of BC. On the one hand, most patients with BC present with haematuria, and gross or microscopic haematuria is often the first symptom in BC patients [11, 12]. On the other hand, haematuria is the most common symptom in urology patients. In many benign or tumorous urologic diseases, haematuria is often the only manifestation [23]. Even 9% to 18% of normal individuals have some degree of haematuria [12]. This fact indicates that the BTA stat test may be prone to detect particularly high false-positives in haematuria patients.
We analysed patients with haematuria symptoms in a clinical setting and found that the specificity of the BTA stat test was only 15.82% in all groups of patients with haematuria. These data indicate that when only patients with haematuria symptoms who were suspected of having BC were considered, the specificity of diagnosing BC with the BTA stat test was only 15%. In addition, the proportion of BC patients with haematuria is very small, especially patients with microscopic haematuria. Previous studies have shown that only approximately 1% of patients with microscopic haematuria actually have BC [10,24,25]. In this case, if the BTA stat test has such a high false-positive rate for haematuria, this test lacks further directivity, and its clinical utility and benefit are low. The cost of screening includes the cost of labelling each patient and the cost of evaluating patients with false-positive results. Adding any test to the clinical assessment increases the costs; thus, the cost-effectiveness needs to be balanced by the benefits.
Optimal screening strategies require not only identifying a method with high sensitivity in a population with significant disease prevalence but also a method with reasonable specificity [26]. The specificity of the marker plays a key role because patients with false-positive results require additional “unnecessary” tests. Our results suggest that the BTA stat test cannot achieve ideal effective specificity and that using this test to screen and monitor BC may lead to “unnecessary” costs. Therefore, the BTA stat test is not recommended for screening or diagnosing BC in patients with haematuria.
In addition, in the case of follow up of urinary tumour markers, a negative predictive value is also very important if the goal is to avoid a more invasive cystoscopy based on a urinary tumour marker report. In our study, the negative predictive value of the BTA stat test was 97.09%, which is similar to the results reported in previous studies. Our study also shows that the negative predictive value of the BTA in patients with haematuria was significantly higher than that in patients without haematuria, and the difference was statistically significant (P<0.05). Nevertheless, notably, because the negative predictive value is affected by the prevalence, even if the diagnostic performance is very poor, this value can be as high as 90% or greater when the prevalence is relatively low. Currently, it is not our objective to use the negative predictive value to measure the diagnostic value of diagnostic experiments. After identifying high-risk groups, the negative predictive value can play its greatest role, but this requires large-scale prospective studies to determine the actual prevalence in high-risk groups. If the prevalence is still low in high-risk groups, the cost and actual benefits of screening need to be considered. Further studies can compare and analyse the specific costs involved in the clinical practice of qualitative urinary tract testing. However, there are no available data related to this problem.
Various biomarkers have been developed for the evaluation of BC, including protein-based and gene-based biomarkers. Similar to BTA, the use of the protein marker NMP22 is also hampered by haematuria [27-28]. Notably, these results do not indicate that the high false-positive rate of protein-based makers in patients with haematuria is directly caused by red blood cells in the urine. Other factors that promote haematuria may also promote tumour cell shedding, thereby increasing the positive detection rate of these makers. Hence, protein-based biomarkers need to be tested under specific circumstances. The gene-based detection of BC, such as the UroVysio test and Cxbladder test, performs better under haematuria interference [29-30]. However, these testing methods are far more complex, and their high cost is the greatest obstacle to their widespread use.
There are some limitations in our study. The numbers of patients in the T3, T4, and undifferentiated cancer groups are low. The sensitivity of 100% in these patient groups may not be statistically significant. The sample size in the renal cancer group is small, which may also cause some bias. We provide and retain these data to allow readers to evaluate the sample size and clinical data.