The aim of the present prospective study was to assess the diagnostic benefit of UroVysion (Vysis-Abbott Laboratories, Downers Grove, IL) in the follow-up of patients with a history of high-grade non–muscle-invasive urothelial carcinoma of the bladder (NMIBC). An unselected cohort of 25 patients with a history of high-grade NMIBC was prospectively followed up by office-based cystoscopy, cytology, and UroVysion in 210 events.
The sensitivity and specificity for standard combined cystoscopy and cytology were 78% and 83%, respectively. UroVysion yielded a considerably higher detection rate with 94% and 93%, respectively. In 89% of the follow-up events of patients with a history of previous carcinoma in situ (CIS) and negative cystoscopy but a positive UroVysion finding, CIS recurrence was noticed within 5 months.
UroVysion is a worthwhile approach in patients with previous CIS, a high risk for the development of CIS, or previous unequivocal cytology suggestive of CIS, especially during or shortly after instillation therapy.
According to the 2008 guidelines of the European Association of Urology for non–muscle-invasive urothelial carcinoma of the bladder (NMIBC), the risk of recurrence and progression as determined by clinical and histopathologic parameters should be reflected in follow-up strategies and intervals of cystoscopy and cytology.1,2 While this approach is an improvement of rather rigid schemes regarded as the “gold standard” for 30 years, it does not pose a true advance. No real progress in follow-up approaches has been established.3
Combined cystoscopy and cytology have been regarded as the gold standard for diagnosis and follow-up of urothelial carcinoma, despite their well-known limitations. The most important oncologic disadvantage of urethrocystoscopy is its low sensitivity in patients undergoing bacille Calmette-Guérin therapy owing to high-grade lesions, especially carcinoma in situ (CIS). Cytology is very useful in high-grade tumors and CIS, but inflammation following instillation therapy impairs results because it is difficult to distinguish malignant from reactive nonneoplastic cells.4,5 CIS is a significant risk factor for progression, especially in patients with pT1G3 urothelial carcinoma.1 It has been proposed that muscle-invasive urothelial bladder cancer arises from flat dysplasia via subsequent development of CIS and invasive tumors.6 Early detection of CIS has been reported to be beneficial for patients; survival seems to be prolonged by early cystectomy.7 Although studies have not been able to show a direct linear relationship between the delay of cystectomy and prognosis, a significantly worse outcome is observed if the delay exceeds 3 months following transurethral resection of CIS.8
Thus, a real need for a noninvasive urinary marker with high sensitivity and specificity in the detection of CIS is warranted. Despite numerous and recent developments, no single marker has been established for broad clinical application to date.9 One of the most promising tests is the commercially available UroVysion (Vysis-Abbott Laboratories, Downers Grove, IL). This multicolor fluorescence in situ hybridization (FISH) assay is the first molecular urine test using DNA probes on chromosomal abnormalities. It detects aneuploidy of chromosomes 3, 7, and 17 and loss of the p16 gene at the 9p21 locus. The combined sensitivity of 4 probes for paraffin-embedded urothelial carcinoma was initially described as 95%,10 which was confirmed in the development study of UroVysion by Sokolova et al11 in 2000. The sensitivity values of the probes for aneuploidy of chromosomes 3, 7, and 17 were reported as 73.7%, 76.2%, and 61.9%, respectively. The sensitivity of the assay was increased by adding the probe for 9p21, although its overall sensitivity for homozygous deletion was only 29%. It was added because 9p21 is a common feature of noninvasive papillary urothelial carcinoma of the bladder. In contrast, hyperdiploidy is a common feature of CIS and invasive urothelial carcinoma, but not noninvasive papillary urothelial carcinoma of the bladder.6 Nevertheless, while UroVysion has a low sensitivity for low-grade tumors owing to their lack of chromosomal abnormalities,12 it detects high-grade carcinomas and CIS with reliable sensitivity.13,14
UroVysion has been discussed to be valuable in the follow-up of urothelial carcinoma of the bladder by tailoring intervals and intravesical treatment strategies according to risk stratification and aiding in decision making.3 However, a standardized collection of urine or bladder washing, sophisticated laboratory equipment, and experience with FISH are necessary, preventing broad routine application. The aim of the present study was to evaluate the use of UroVysion in the follow-up of patients with a history of high-grade NMIBC and CIS in urologic outpatient care.
Materials and Methods
Patients and Clinical Pathway
The prospective study comprises 210 follow-up events from a consecutive unselected cohort of 25 patients with a history of high-grade NMIBC. Follow-up consisted of standard cystoscopy, cytology, and UroVysion in a single urologic outpatient office every 3 months for the first 2 years and every 6 months for the next 2 years in accordance with the current European Association of Urology guidelines.2 The study was approved by institutional ethical review boards. Informed consent was given by all patients.
The indication for transurethral resection was a suspect finding in cystoscopy. No biopsy was performed during the cystoscopy. In case of a positive UroVysion assay or cytology and negative cystoscopy, urography and ultrasound of the upper urinary tract were performed, and, in case of negative findings, a subsequent visit was scheduled for 3 months thereafter. If the subsequent cystoscopy revealed a “suspicious” lesion and the following transurethral resection showed recurrence confirmed by histologic examination, prior UroVysion and/or cytology was regarded as true-positive presuming the existence of undetected CIS or precursor lesion at that time. In the current literature, these events are termed as anticipatory positive, aiming to show that UroVysion seems to detect almost half of recurrent bladder cancers before they are visible by cystoscopy.14 If the repeated tests were negative, the initial test was assumed to be a false-positive. The present study did not try to detect high-grade NMIBC earlier in the course, nor did it include a second group for comparison of different follow-up regimens and intervals. Adjuvant therapy with mitomycin C and bacille Calmette-Guérin was initiated according to current guidelines.
Cystoscopy and Transurethral Resection of Bladder Tumors
All cystoscopies were performed by 1 highly experienced urologist (E.B.). Transurethral resection was performed with photodynamic diagnosis (PDD) by experienced surgeons in 1 institution. The resected specimens were processed according to standard pathologic procedures and diagnosed by 2 experienced uropathologists (A.H. and S.S.) based on the current American Joint Committee on Cancer TNM classification.15 Tumor grading was assessed according to the 2004 WHO/International Society of Urologic Pathology consensus classification.16
Cytology and UroVysion
Samples from voided urine and bladder washings were fixed in an end concentration of 50% ethanol and transported from the outpatient practice to the pathology laboratory within 48 hours. From each specimen, cytocentrifuged specimens were performed and stained with Papanicolaou. Cytology and UroVysion analysis was performed by 2 experienced uropathologists (A.H. and S.S.) and UroVysion also by 1 experienced molecular biologist (W.D.), all blinded to other results. Cytology was reported according to a standardized 4-tier classification with negative, atypical, suspicious, and positive categories. Cases with suspicious findings were put in the positive category, while atypical findings were categorized as negative.
UroVysion was performed on the same slides as used for cytology according to the producers’ instructions. Each specimen was scanned for morphologically abnormal cells, and a minimum of 25 such cells were analyzed. If fewer than 25 urothelial cells of interpretable quality were found, the specimen was judged noninterpretable. According to Sokolova et al,11 the test was considered positive if at least 5 cells of the 25 abnormal cells showed gains of 2 or more chromosomes 3, 7, or 17 or if more than 12 of the 25 cells showed no 9p21 signals. The presence of rare cells with a tetraploid pattern and fewer than 4 cells with other aberrations was regarded as FISH-negative, according to Zellweger et al.17
Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for high-grade recurrences were calculated. The Fisher exact test and χ2 test were used to evaluate the association between categorical variables. To assess the independent diagnostic value of cystoscopy, cytology, and UroVysion, stepwise multivariable logistic regression based on the Akaike information criterion (AIC) for tumor recurrence was used.18 The model minimizing the AIC has the highest potential to maximize external validation and is hence termed the optimal model. Analyses were performed with the open-source software R-project (version 2.7.2; http://www.rproject.org). All statistical tests were performed at the 2-sided α = .05 level of statistical significance.
The mean follow-up was 27 months (range, 1–48 months). We recorded 210 follow-up events (median, 8; interquartile range, 4–10), including cystoscopy, cytology, and UroVysion. One patient had recurrence on the first follow-up after inclusion. We could not evaluate 25 events (11.9%) by UroVysion owing to insufficient amounts of cells, and these events were, therefore, excluded from the analysis.
Cytology was positive in 41 (22.2%) of 185 events, UroVysion in 46 events (24.9%), and cystoscopy in 42 events (22.7%), leading to 42 transurethral resections. In 25 (60%) of 42 transurethral resections, recurrence was confirmed; the remaining 17 events revealed benign lesions. In 16 (64%) of 25 histologically proven tumors, concomitant CIS or CIS only was seen (CIS only, 4; Ta, 4; T1, 6; and T2, 2).
Cystoscopy showed an overall sensitivity of 68% and specificity of 89%, PPV of 60%, and NPV of 92%. Cytology yielded a sensitivity of 68%, specificity of 89%, PPV of 61%, and NPV of 92%.
UroVysion showed a sensitivity of 95%, specificity of 93%, PPV of 76%, and NPV of 99%. Sensitivity, specificity, PPV, and NPV for the standard follow-up scheme consisting of combined cystoscopy and cytology were 78%, 83%, 54%, and 94%, respectively Table 1. In the 12 patients with a history of concomitant CIS, UroVysion showed sensitivity, specificity, PPV, and NPV results of 96%, 94%, 88%, and 98%, respectively. In 8 (89%) of 9 follow-up events with negative cystoscopy and a positive UroVysion finding, CIS recurrence developed within 5 months.
Patients were grouped in 2 categories according to their disease: (1) high-grade papillary carcinoma without CIS and (2) high-grade non–muscle-invasive urothelial carcinoma–CIS. In case of subsequent histologically proven CIS in group 2, associated events, including anticipatory positive UroVysion tests, were excluded from the calculation of sensitivity and specificity to have groups as homogeneous as possible. Sensitivity, specificity, PPV, and NPV for UroVysion were comparable in both groups (P = .078) Table 2.
Univariate logistic regression demonstrated cystoscopy, cytology, UroVysion, and combined cystoscopy/cytology to be statistically significant for prediction of recurrence (P < .05). UroVysion showed the highest predictive value (AIC, 76 vs 143 vs 141 vs 139 in comparison with cystoscopy, cytology, and cystoscopy/cytology respectively) Table 3.
The multivariable logistic analysis demonstrated that UroVysion as the third variable significantly adds to the information that cystoscopy and cytology provide about subsequent histologically proven cancer. The best multivariable logistic model for predicting cancer recurrence included UroVysion and cytology as predictors. The odds for cancer recurrence increased more than 130-fold in UroVysion-positive cases (odds ratio, 131.4; 95% confidence interval, 27.0–638.5; P < .01). The odds for cancer recurrence increased almost 5-fold in positive cytology (odds ratio, 4.7; 95% confidence interval, 1.3–17.2; P = .02) Table 4.
Initial reports on UroVysion trials reported this novel technique to outperform cytology sensitivity in all stages and grades of bladder cancer.19,20 Subsequent data, however, demonstrated that UroVysion has only a slightly better sensitivity and lower specificity compared with cytology. Further workgroups even postulated no improvement over cytology or quantitative cytology in the diagnosis of recurrent NMIBC.5,13,14,20–23 However, certain critical aspects need to be taken into consideration in interpreting these studies Table 5.5,11,13,14,17,19–22,24–36
First, the more recent studies are based on retrospective data with large cohorts but only very short median follow-up, eg, 6.3 months in the trial by Ferra et al24 with accordingly low prevalence of recurrence. This creates a problem in that the PPV depends on the prevalence of recurrence in the study population. The mean follow-up in the present study was 26 months, resulting in a higher PPV than in studies with shorter follow-up periods. Thus, the present study on UroVysion reflects the real situation of NMIBC quite well and features the longest follow-up, including repeated UroVysion, of a study population with a history of high-grade NMIBC reported to date.
Second, most studies did not take into account the anticipatory positive events. The end point in these studies has been subsequent transurethral resection and histopathology of the specimen. Transurethral resection was performed only for suspected cystoscopic findings, however, negating the possible presence of cystoscopically invisible lesions. Furthermore, resections were not performed with PDD and, thus, did not reflect the currently most advanced detection rates of CIS and respective precursor lesions. It has been shown that UroVysion detects recurrent urothelial carcinoma before visualization by cystoscopy.25 In general, white-light cystoscopy has been viewed as insufficient in the diagnosis of CIS and PDD has become more established recently in addition to white-light cystoscopy, as it detects significantly more CIS.2 Taking anticipatory positive events into consideration, UroVysion showed improved sensitivity, specificity, PPV, and NPV compared with combined cystoscopy and cytology. In 90% of our patients with a history of CIS and positive UroVysion assay, despite negative conventional cystoscopy, confirmed CIS developed within 5 months as demonstrated by PDD-aided resection. Yoder et al25 reported the natural history of anticipatory positive findings in 250 patients. In 63% of patients with positive UroVysion, recurrence developed within 29 months. As the mean follow-up was 27 months in the present study, the prevalence of recurrences was comparably high.
Moonen et al14 reported no improvement over cytology or quantitative cytology in the diagnosis of recurrent non–muscle-invasive bladder tumors by UroVysion. In contradiction to this report, we found an improvement of combined cystoscopy/cytology by UroVysion in the follow-up of patients with a history of high-grade NMIBC. Notably, positive UroVysion was significantly related to subsequent CIS recurrence. Decreased follow-up intervals and possibly a photodynamic evaluation of patients with negative cystoscopy but positive UroVysion should be considered.
UroVysion has lower sensitivity for low-grade tumors, possibly owing to the absence of shed cells and of chromosomal changes.25 In contrast, UroVysion detects high-grade carcinomas and CIS with reliable sensitivity and specificity values of up to 70% and 95%, respectively, as reported by Gudjónsson et al.13 The authors concluded that UroVysion may be a valuable supplement. Our data suggest UroVysion to be an effective tool in the follow-up of high-grade NMIBC, especially in patients with solitary and concomitant CIS. This finding reflects other studies on UroVysion. The current literature provides no evidence, however, as to whether UroVysion is valuable in the follow-up of patients by possibly detecting recurrences significantly earlier and, thus, allowing more timely intervention.
To our best knowledge, the present study is the first prospective evaluation of UroVysion in the follow-up of patients with a history of high-grade NMIBC. It demonstrates significant improvement of follow-up strategies in a high-risk cohort and the feasibility of interpreting UroVysion positivity in the absence of positive cystoscopy and cytology findings as precursor lesions, initiating closer follow-up. The current study also is the first report of the application of UroVysion in an office-based outpatient setting. UroVysion specimens were transported more than 100 km. While unsuccessful UroVysion tests in the present study were mostly due to a lack of evaluable cells and not insufficient hybridization, which is reflected by other current reports on UroVysion,24 the failure rate was about 2 times higher than previously published.13 While the shipment from an outpatient practice to the diagnosing institute seems to have an impact on the quality of the specimen, it does not seem to affect the diagnostic value of UroVysion, as the method proved sufficiently reliable in our hands with an overall test success rate of 88.1%.
The greatest limitation of the present study, however, is the small population, which might result in a biased selection of patients with specific genetic alterations measured by UroVysion. Furthermore, studies dealing with CIS are always biased, because it can be very focal and, despite PDD, not detectable, especially in patients undergoing intra-vesical therapy.
A crucial aspect in introducing a novel diagnostic technique into the clinical routine is cost. Currently, reimbursement for UroVysion in Germany is about €170, including costs of the assay, labor, and infrastructure. Cystoscopy is calculated at about €70 and cytology at about €25 (GOÄ, 2002, FNA 2122-4). Huge cost disparities between Europe and the United States hamper interpretation of these figures. Chen et al37 reported that the average allowable charge for UroVysion under Medicare is $900 and the standard care, consisting of cystoscopy plus cytology, is about $400. The conclusion was that cost per positive recurrence event detected in the surveillance care for NMIBC is much higher by using UroVysion testing than that of standard care. In the present study, the costs of UroVysion vs cystoscopy/cytology are similarly linked. With that in mind, the conclusion should be transferable to our study. But UroVysion may ultimately lead to a decreased frequency of follow-up cystoscopies in many patients. In the present study, the follow-up regimen has not been adjusted according to the UroVysion results. Further studies are warranted to clarify these issues.
While the value of UroVysion seems to be limited in NMIBC in general, it may be a worthwhile approach in patients with previous CIS, a high risk for the development of CIS, or previous unequivocal cytology suspicious for CIS, especially during or shortly after instillation therapy. Thus, tighter follow-up schemes or photodynamic assessment should be considered in positive UroVysion and concomitant negative cystoscopy and cytology results in the surveillance of patients with a history of high-grade NMIBC. The benefits of UroVysion must be balanced against its costs.
Upon completion of this activity you will be able to:
list the common chromosomal changes observed in urothelial carcinoma.
recognize the limitations of urethrocystoscopy and cytology in identification of flat in situ high-grade urothelial neoplasia.
describe the methods used in the UroVysion FISH test.
define the clinical indication, potential application, and benefits of using the UroVysion FISH test in urothelial carcinoma.
The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit ™ per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.
The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.
Questions appear on p 661. Exam is located at www.ascp.org/ajcpcme.