Key points

Screening, diagnosis, and evaluation in bladder cancer screening

Bladder cancer (BCa) is a heterogeneous disease with a variable natural history. Most patients (70%) present with superficial tumors (stages Ta, T1, or carcinoma in situ). However, 3 out of 10 patients present with muscle-invasive disease (T2–4) with a high risk of death from distant metastases. Moreover, roughly between 50% and 70% of superficial tumors do recur, and approximately 10% to 20% of them progress to muscle-invasive disease. However, BCa has a relatively low ratio of mortality versus incidence of new cases. In consequence, there is the danger of overdiagnosis and overtreatment.

Hence, the goal and challenge of screening would be to detect the disease at an earlier stage, consequently improving morbidity and survival, but, more importantly, to be able to identify the tumors that are more likely to become muscle-invasive cancers. Such early detection of tumors could allow for earlier curative intervention and could potentially preclude the need for unnecessary surgical treatment or chemotherapy and lower the costs associated with treatment.

To date, an estimated 75,000 new cases of urinary BCa will be diagnosed in the United States (56,390 men and 18,300 women) in 2014. In the same year, approximately 16,000 new deaths (11,170 men and 4410 women) are expected. BCa is the fourth most common cancer and is 3 times more common in men than in women in the United States. It has also been previously reported that the age-adjusted incidence of BCa seems to be increasing over time: from 21.0 to 25.5 per 100,000 person-years between 1973 and 2009 (+0.2% per year, P = .001).

Two landmark studies have evaluated the effect of screening for BCa. Messing and colleagues performed an important assessment based on 1575 cases (≥50-year-old men screened at home using hematuria dipsticks) and 509 controls (nonscreened). Those who showed positive results underwent cystoscopy (n = 283), and 21 (7.4%) of them were diagnosed with BCa. The primary results of that study indicated that earlier detection of BCa could result in a lower proportion of invasive cancers among high-grade and/or muscle-invasive diseases in screened versus nonscreened men (10% vs 60%, P = .002) and a significant reduction in mortality caused by the disease (0% vs 20%, P = .02). Britton and colleagues examined 2356 men aged 60 years and older for dipstick hematuria. The test was positive in 20% of men, and BCa was ultimately diagnosed in 17 individuals (5.3%). Of those, 9 patients had high-risk non–muscle-invasive BCa. At a 7-year follow-up, 5 out of 9 patients progressed to muscle-invasive disease and 3 out of 9 died of the disease.

Partly owning to an overall low incidence of the disease (25.5 per 100,000 person-years in 2009), screening for BCa is currently not recommended as a standard of care during routine clinical practice. The challenge being that a clearly defined high-risk population needs to be identified, so as to avoid the usual harms in screening, such as unnecessary diagnostic-related treatments (ie, cystoscopy and biopsy, transurethral resection of bladder tumor [TURBT], intravesical chemotherapy), and overdiagnosis. Wu and colleagues generated a model based on large case-control data (n = 678 cases and n = 678 controls) for the prediction of BCa risk using established risk factors, such as smoking and well-known occupational exposure (eg, diesel, aromatic amines, dry cleaning fluids, radioactive materials, arsenic). At internal validation, the model demonstrated an area under the curve of 80%. However, the model is impeded by the lack of external validation.

Similarly, Vickers and colleagues using data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial attempted to create a risk score in order to identify those at higher risk of developing BCa. The study comprised 49,873 persons for the training set and 99,746 individuals for the external validation set. The investigators showed that the trade-off between the number of patients screened and invasive/high-grade tumors avoided was more optimal when restricting screening to a high-risk population instead of the whole population (57 vs 38 per 100,000), hence supporting the strategy to screen a high-risk population only.

Although large-scale screening programs may seem beneficial and critical, the US Preventive Services Task Force has deemed existing evidence to be insufficient in assessing the balance of benefits and harms of screening. Experts have suggested that improvements have to be made with regard to the identification of what should be defined as high-risk populations before a randomized-controlled trial could be designed to validate the hypothesis of an improvement in screening. It is noteworthy that in recent years, little progress in improving survival and reducing mortality for patients with muscle-invasive BCa has been seen. Indeed, despite well-known risk factors (ie, smoking, occupational exposure), as well as the developments in BCa management over time, it did not lead to better cancer control outcomes. In consequence, it may be important to adequately address the challenges of screening in BCa in the near future.

Screening, diagnosis, and evaluation in bladder cancer screening

Bladder cancer (BCa) is a heterogeneous disease with a variable natural history. Most patients (70%) present with superficial tumors (stages Ta, T1, or carcinoma in situ). However, 3 out of 10 patients present with muscle-invasive disease (T2–4) with a high risk of death from distant metastases. Moreover, roughly between 50% and 70% of superficial tumors do recur, and approximately 10% to 20% of them progress to muscle-invasive disease. However, BCa has a relatively low ratio of mortality versus incidence of new cases. In consequence, there is the danger of overdiagnosis and overtreatment.

Hence, the goal and challenge of screening would be to detect the disease at an earlier stage, consequently improving morbidity and survival, but, more importantly, to be able to identify the tumors that are more likely to become muscle-invasive cancers. Such early detection of tumors could allow for earlier curative intervention and could potentially preclude the need for unnecessary surgical treatment or chemotherapy and lower the costs associated with treatment.

To date, an estimated 75,000 new cases of urinary BCa will be diagnosed in the United States (56,390 men and 18,300 women) in 2014. In the same year, approximately 16,000 new deaths (11,170 men and 4410 women) are expected. BCa is the fourth most common cancer and is 3 times more common in men than in women in the United States. It has also been previously reported that the age-adjusted incidence of BCa seems to be increasing over time: from 21.0 to 25.5 per 100,000 person-years between 1973 and 2009 (+0.2% per year, P = .001).

Two landmark studies have evaluated the effect of screening for BCa. Messing and colleagues performed an important assessment based on 1575 cases (≥50-year-old men screened at home using hematuria dipsticks) and 509 controls (nonscreened). Those who showed positive results underwent cystoscopy (n = 283), and 21 (7.4%) of them were diagnosed with BCa. The primary results of that study indicated that earlier detection of BCa could result in a lower proportion of invasive cancers among high-grade and/or muscle-invasive diseases in screened versus nonscreened men (10% vs 60%, P = .002) and a significant reduction in mortality caused by the disease (0% vs 20%, P = .02). Britton and colleagues examined 2356 men aged 60 years and older for dipstick hematuria. The test was positive in 20% of men, and BCa was ultimately diagnosed in 17 individuals (5.3%). Of those, 9 patients had high-risk non–muscle-invasive BCa. At a 7-year follow-up, 5 out of 9 patients progressed to muscle-invasive disease and 3 out of 9 died of the disease.

Partly owning to an overall low incidence of the disease (25.5 per 100,000 person-years in 2009), screening for BCa is currently not recommended as a standard of care during routine clinical practice. The challenge being that a clearly defined high-risk population needs to be identified, so as to avoid the usual harms in screening, such as unnecessary diagnostic-related treatments (ie, cystoscopy and biopsy, transurethral resection of bladder tumor [TURBT], intravesical chemotherapy), and overdiagnosis. Wu and colleagues generated a model based on large case-control data (n = 678 cases and n = 678 controls) for the prediction of BCa risk using established risk factors, such as smoking and well-known occupational exposure (eg, diesel, aromatic amines, dry cleaning fluids, radioactive materials, arsenic). At internal validation, the model demonstrated an area under the curve of 80%. However, the model is impeded by the lack of external validation.

Similarly, Vickers and colleagues using data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial attempted to create a risk score in order to identify those at higher risk of developing BCa. The study comprised 49,873 persons for the training set and 99,746 individuals for the external validation set. The investigators showed that the trade-off between the number of patients screened and invasive/high-grade tumors avoided was more optimal when restricting screening to a high-risk population instead of the whole population (57 vs 38 per 100,000), hence supporting the strategy to screen a high-risk population only.

Although large-scale screening programs may seem beneficial and critical, the US Preventive Services Task Force has deemed existing evidence to be insufficient in assessing the balance of benefits and harms of screening. Experts have suggested that improvements have to be made with regard to the identification of what should be defined as high-risk populations before a randomized-controlled trial could be designed to validate the hypothesis of an improvement in screening. It is noteworthy that in recent years, little progress in improving survival and reducing mortality for patients with muscle-invasive BCa has been seen. Indeed, despite well-known risk factors (ie, smoking, occupational exposure), as well as the developments in BCa management over time, it did not lead to better cancer control outcomes. In consequence, it may be important to adequately address the challenges of screening in BCa in the near future.

Future considerations for screening

Many unknowns remain before a large-scale program of BCa screening may be put forward. For one, few studies have examined the natural history of screened early stage untreated BCa. The appreciation of how such tumors behave may assist clinicians in understanding the effects of screening and overdiagnosis/overtreatment. Other randomized studies with large sample sizes that assess urine tests and the incidence of BCa, and staging, with follow-up information on clinical and cancer control outcomes are needed. A uniform definition of what constitutes a high-risk population is necessary in order to reduce the potential harms associated with overdiagnosis/overtreatment. Moreover, it may be important to invest in less invasive and costly diagnostic tools but equally sensitive and specific. For example, traditional cystoscopy may not be adequate if population-based screening is envisaged, as the procedure is invasive and not cost-efficient. In order to reduce the number cystoscopies required to diagnose one patient with BCa, the combination of dipstick testing and molecular markers have been suggested. It may also be worthwhile to test the effect of less toxic treatments in patients diagnosed in the earlier stages to prevent progression versus more invasive and toxic treatments reserved for patients diagnosed in the more advanced stages. This testing may be useful in evaluating the trade-offs and costs associated with earlier treatment modalities.

Presentation and investigation

The most common presenting symptom of BCa is gross hematuria. In patients with carcinoma in situ, additional unexplained urinary frequency, urgency, or irritative voiding symptoms may indicate the possibility of BCa. The prevalence of BCa ranges between 13% and 35% in patients presenting with gross hematuria and between 5% and 10% in patients presenting with microscopic hematuria. The proportion of patients with asymptomatic hematuria ranges between 0.2% and 21.0%, when a direct relationship with age has been shown.

The current practice includes urine dipstick or microscopic urinalysis for hematuria, urine cytology, and other tests for urine biomarkers, although their value has not yet been established. Patients with positive results are referred for further evaluation, which generally include cystoscopy, followed by biopsy if necessary, and radiological investigation. Imaging is important in the evaluation for patients presenting with hematuria, primarily to visualize the upper urinary tract, as cystoscopy inspects the lower tracts. Options for imaging include ultrasonography, intravenous urogram (IVU), computed tomography (CT) urography, MRI, or a combination of these.

Bladder tumor markers

Urine cytology, a simple and noninvasive procedure, in combination with cystoscopy is superior to cystoscopy alone in identifying high-grade lesions (ie, carcinoma in situ). It has a good sensitivity for detecting high-grade BCa but a poor sensitivity for detecting low-grade tumors (between 7%–17%). In recent years, several novel developments of urine markers for the detection of BCa have been investigated ( Table 1 ). For example, of 79 patients with BCa, the nuclear matrix protein 22 (NMP-22) assay was positive in 44 patients (sensitivity 56% and specificity 86%). Similarly, fluorescence in situ hybridization (FISH, Vysis Inc., Downers Grove, IL) (ie, UroVysion assay designed to detect aneuploidy for chromosomes 3, 7, 17 and loss of the 9p21 locus via FISH) demonstrated a promising range of sensitivity (70%–86%) and specificity (66%–93%) in a pooled analysis of 4 articles as part of a meta-analysis.

However, urine cytology remains superior in specificity compared with most other markers (up to 98%, see Table 1 ). Moreover, no biomarker has demonstrated superior clinical utility over cytology and cystoscopy combined. Hence, experts agree that it is highly unlikely that a new marker could achieve the accuracy of what cytology and cystoscopy can offer. Currently, the use of urine cytology is recommended by national guidelines despite the paucity of randomized trials that actually examined its use in clinical practice. For patients with noninvasive disease, urine cytology also represents an important surveillance tool.

Perspectives on screening

The current recommendations do not support routine BCa screening because of insufficient evidence and lack of understanding of the effects of screening in the case of overdiagnosis and overtreatment. However, the results of existing studies suggest that BCa screening may be important in high-risk populations. The challenge remains that experts need to be able to diagnose patients at an early stage of the disease and, at the same time, identify the tumors that are likely to progress to muscle-invasive disease. Currently, the combination of urine cytology and cystoscopy remains the gold standard for diagnosing patients with BCa. Less invasive urine biomarkers have been investigated over time, but their performance remains subpar with respect to specificity compared with cytology alone. It is unlikely that a new marker will be used to replace the conventional urine cytology and cystoscopy.

Endoscopic examination of the lower urinary tract

The cornerstone of the diagnosis and subsequent management of BCa is the cystoscopic examination of the lower urinary tract. Specifically, white light cystoscopy (WLC) remains the gold standard, despite its limitations. Recently, new optical diagnostic methods have been designed to improve the accuracy of WLC: fluorescence cystoscopy, narrow-band imaging (NBI), and optical coherence tomography (OCT).

White light cystoscopy

WLC allows examination of the urethra and the bladder. It can be performed with either rigid or flexible endoscopes, depending on the clinical scenario. For screening and diagnostic office-based purposes, flexible cystoscopy is typically used. Although it is a more comfortable instrument than the rigid cystoscope, it has the disadvantages of having a small irrigation port and lacks a separate working sheath. It is also more costly and inclined to failure.

The sensitivity and specificity of WLC range from 62% to 84% and 43% to 98%, respectively. Its sensitivity is lower for small papillary bladder tumors and carcinoma in situ. Moreover, the accuracy of WLC has been shown to be operator dependent. Given the known drawbacks of an office-based flexible cystoscopy under local anesthesia, it is recommended to perform a thorough reassessment using the rigid cystoscope under general anesthesia when patients are brought to the operating room to resect a newly diagnosed bladder tumor.

Moreover, evidence suggests that the operator’s ability to estimate the stage or grade based on the visual appearance on WLC is limited. It is also difficult to differentiate inflammatory lesions because of the previous instillation therapy from carcinoma in situ (CIS), as both will present as erythema. Although WLC remains the current standard for diagnosis and follow-up, further improvements of the technique are needed.

Fluorescence cystoscopy

Fluorescence cystoscopy or photodynamic diagnosis (PDD) can improve the visualization of BCa compared with conventional WLC and, consequently, reduced the rates of residual tumor at the first cystoscopy. PPD relies on the intravesical administration of endogenous or exogenous photosensitizing agents that cause abnormal or rapidly proliferating cells to fluoresce under a specific wavelength of light. The solution containing the photosensitizing agents is instilled in the bladder thru a transurethral catheter before surgery. By illuminating the mucosa with blue-violet light, the neoplastic cells appear red or pink against a blue background. However, false or artifact fluorescence can occur with inflammation, recent TUR, or tangential illumination of the bladder mucosa. False positives are frequent in patients who recently have had intravesical therapy, especially bacillus Calmette-Guérin.

Multiple studies have demonstrated that PDD, in addition to WLC, improves the detection rate of BCa. The tumor detection rate for WLC alone is 73% to 96% versus 90% to 96% for WLC plus PDD. The difference is even more striking for carcinoma in situ: 23% to 68% versus 91% to 97%, respectively. However, the false-positive detection rate of PDD is higher (9%–63% vs 7%–47%, respectively). A recent meta-analysis of 18 studies showed that the rate of residual tumor was significantly decreased after PDD (odds ratio [OR]: 0.28; 95% confidence interval [CI], 0.15–0.52; P <.0001), whereas the recurrence-free survival was higher at 12 and 24 months in the PDD groups relative to the WLC-only groups (log-rank = 0.00002). Several researchers have also investigated the value of PDD in the evaluation of patients with positive urine cytology in the absence of disease after standard investigations. For example, Ray and colleagues showed that BCa was detected by PDD in 32% of patients with confirmed positive urinary cytology and negative WLC. Finally, the same group also examined the role in the treatment of patients with multifocal tumors, who are at a 1.5-fold increased risk of recurrence. They showed in a series of 18 patients that the sensitivity of PDD for the detection of tumor was 97.8% compared with 69.6% for WLC.