Cytologic and cystoscopic predictors of recurrence and progression in patients with low-grade urothelial carcinoma




Patients with low-grade urothelial carcinoma (LGUC) are at risk of recurrence and must undergo lifelong surveillance. To date, cytology and cystoscopy are the gold standard for the detection of de novo and recurrent LGUC. The objective of the current study was is to further characterize the role of cytology and cystoscopy in determining the risk of recurrence and progression in these patients.


The authors retrospectively identified patients with LGUC who had urine cytology within 2 months of biopsy, and data were abstracted from their electronic charts. Electronic medical records were reviewed for cystoscopic findings and histologic and cytologic follow-up data over a 5-year period. Statistical analysis was performed with chi-square tests.


In total, 76 patients were identified who had histologic follow-up material available, and 49% of those patients demonstrated progression or recurrence of urothelial carcinoma. The initial presence of multiple lesions on cystoscopy was associated with any recurrence or progression (67.7% vs 31%; P = .002), tumor size >2 cm was associated with initial positive or suspicious urine cytology (23.8% vs 3.7%; P = .076), and positive or suspicious initial cytology was associated with high-grade recurrence (58.3% vs 19.4%; P = .009).


Cystoscopic findings, such as the presence of multiple lesions, together with concurrent positive or suspicious urine cytology, were associated with recurrence or progression of LGUC. These findings may help to identify high-risk patients. Cancer (Cancer Cytopathol) 2012;121:398–402. © 2012 American Cancer Society.


An estimated 73,510 new cases of bladder cancer were diagnosed, and 14,880 deaths were attributed to bladder cancer in 2012 according the American Cancer Society.[1] Although the majority of patients with newly diagnosed bladder cancer, approximately 70% to 80%, present with nonmuscle-invasive disease, an estimated 50% to 70% of patients are expected develop a recurrence, and 10% to 30% will progress to muscle-invasive disease.[2, 3] It is believed that the characteristically high recurrence rate among patients with bladder cancer arises from a “field change” in the normal urothelium, which puts the entire urothelial mucosa at risk for tumor formation.[4] Therefore these patients must undergo lifelong surveillance for tumor recurrence.

The current gold standard for the diagnosis and treatment of superficial, Ta and T1 bladder tumors is cystoscopy with transurethral resection (TUR), whereas urine cytology remains the most specific adjunctive tool for the surveillance of bladder cancer.[4] In addition, urinary biomarkers, such as bladder tumor antigen, ImmunoCyt, nuclear matrix protein-22 (Diagnocure Inc., Quebec City, Quebec, Canada), and fluorescent in situ hybridization (FISH), have demonstrated significant promise for the routine surveillance of patients with a history of bladder cancer. However, with the exception of FISH by UroVysion (Abbott Molecular, Des Plaines, Ill), which is gaining a wide acceptance as a urine adjuvant test, the role of such biomarkers has not yet been well defined.[4, 5]

Studies have demonstrated that cystoscopic findings at the time of TUR, including the size of the largest tumor and the number of tumors, are important prognostic factors in recurrence and progression of superficial (Ta and T1) bladder cancer.[6] In addition, it has been reported that positive urine cytology after TUR predicts recurrence.[11] Although those studies included superficial tumors of all histologic grades, others have observed that, when taking into account low-grade lesions alone, tumor size was not associated with a risk of grade progression or recurrence.[15] In the current study, our objective was to further characterize the role of cytology and cystoscopic findings in disease recurrence and progression among patients with superficial low-grade urothelial carcinoma (LGUC) of the urinary bladder.


All patients who were diagnosed with LGUC on bladder biopsy or transurethral resection of bladder tumor (TURBT) specimens from a 5-year period (2006-2010) were identified using the CoPath (Cerner, Kansas City, Mo) anatomic pathology database. These diagnoses were made according to the 2004 World Health Organization (WHO)/International Society of Urological Pathology (ISUP) system for grading urothelial cancer.[16] A subspecialty pathology sign-out system is in place at our institution; therefore, the majority of cases that were included were diagnosed by a pathologist with urologic pathology expertise. Patients were excluded if they had a concurrent histologic diagnosis of high-grade UC (HGUC) or UC in situ (UCIS) or if there was no urine cytology performed either at or within the 2 months before the diagnosis of LGUC. If more than 1 specimen was accessioned from the same patient within the study period, then the first case from that patient was included, and the remaining subsequent cases from the same patient were included only within the follow-up data for that patient.

The Institutional Electronic Medical Record for each patient was reviewed for pertinent patient information, including cystoscopic findings at the time of histologic sampling (eg tumor size, presence or absence of multiple tumors), patient demographics, and subsequent histologic and cytologic follow-up data from the time of biopsy. Urine cytology diagnoses were made based on commonly accepted morphologic criteria. Cases were diagnosed as positive when cells demonstrated a high nuclear-to-cytoplasmic ratio, marked nuclear hyperchromasia, coarse chromatin, and an irregular nuclear outline.[17] Criteria, including rare, small, highly atypical cells or degenerated atypical cells with intact nuclear outlines, were used to diagnosis urine cytology specimens as suspicious for UC, and cases that were considered atypical included those with rare, mildly atypical cells and that did not meet the other criteria listed above.[17]

Statistical analysis of the data was performed subsequently using chi-square tests and Fisher exact tests with the Interactive Statistical Pages (available at:; [accessed August 2012]). Fisher exact tests were used when there were cells that contained <5 items.


In total, 87 patients with a histologic diagnosis of LGUC were identified. The patient population was comprised of 29 women and 58 men, and the mean patient age was 70 years (age range, 27-89 years). In 57 patients, bladder biopsy specimens were obtained; and, in the remaining 30 patients, TURBT specimens were procured. The great majority of specimens were confined to urothelium without lamina propria invasion (83 specimens; 95.4%). In 1 specimen (1.1%), there was lamina propria invasion; 1 specimen (1.1%) was considered suspicious for invasion; and, in 2 specimens (2.2%), invasion could not to be assessed because of marked cautery artifact. There were no specimens in which muscularis propria invasion was identified.

The corresponding cytology specimens, which were obtained at the time of biopsy or within the 2 months before biopsy, consisted of 86 (98.9%) bladder barbotages and 1 (1.1%) voided urine. The cytologic diagnosis was negative for malignancy in 60 specimens (70%), atypical in 12 specimens (13.8%), suspicious for UC in 4 specimens (4.6%), and positive for UC in 9 specimens (10.3%). Two specimens (2.3%) were unsatisfactory for diagnosis because of low cellularity (see Table 1). FISH with the UroVysion Bladder Cancer Kit (Abbott Molecular) was performed on only 5 urine specimens and was positive in 3 specimens, negative in 1 specimen, and not unable to be assessed in 1 specimen because of insufficient sample size.

Table 1. Summary of Cytology Data
VariableNo. of Specimens (%)
  1. Abbreviations: UC, urothelial carcinoma.
Total no. of cytology specimens87
Barbotage86 (98.9)
Voided1 (1.1)
Cytology diagnosis 
Negative for UC60 (70)
Atypical urothelial cells12 (13.8)
Suspicious for UC4 (4.6)
Positive for UC9(10.3)
Unsatisfactory for diagnosis2 (2.3)
Average time between cytology and biopsy, d13

According to cystoscopy reports, multiple (>1) bladder lesions were noted in 36 patients (41.3%), whereas a single lesion was observed in 48 patients (55.1%). No discrete lesions were identified in 3 patients (3.4%). Of the 51 tumors for which a measurement of the largest lesion was provided, 21 lesions (41.1%) measured ≥2 cm, and 27 lesions (52.9%) measured <2 cm (see Table 2). After the initial biopsy, subsequent cytologic sampling was performed in 79 of 87 patients, and subsequent histologic sampling was performed in 76 of 87 patients. Seven patients had neither cytologic nor histologic follow-up sampling and, thus, were excluded from our analysis of disease recurrence/progression. Of the 76 patients who had histologic follow-up, 30 (39.5%) had recurrence of LGUC, and 19 (25%) demonstrated progression to HGUC and/or UCIS on subsequent histologic sampling. Of these 76 patients, 12 (15.8%) demonstrated both LGUC and HGUC on subsequent biopsies (see Table 3). Subsequent cytology sampling was performed in 79 of 87 patients. Of these 79 patients, 38 (48.1%) had negative urine cytology results, whereas 41 (51.9%) had 1 or more subsequent urine cytology result that was abnormal and either atypical, positive, or suspicious (see Table 3). During the follow-up period, only 1 patient went on to undergo cystectomy, and 1 patient underwent radical nephrectomy after HGUC was identified. There were no patient deaths because of disease.

Table 2. Summary of Cystoscopy Findings
VariableNo. of Patients (%)
Total no. of cystoscopy reports available87
Multiple lesions identified36 (41.3)
Single lesion identified48 (55.2)
No discrete lesion observed3 (3.4)
No. of patients with lesion measurement provided51
≥2 cm21 (41.2)
<2 cm27 (52.9)
Table 3. Summary of Histologic and Cytologic Follow-Up Data
ResultsNo. of Patients (%)
  1. Abbreviations: AU, atypical urothelial cells; HGUC, high-grade urothelial carcinoma; LGUC, low-grade urothelial carcinoma.
Follow-up biopsy results, N = 76 
Recurrent LGUC30 (39.5)
Progression to HGUC19 (25)
No recurrence27 (35.5)
Follow-up urine cytology results, N = 79 
Negative38 (48.1)
Positive, suspicious, or AU41 (51.9)

When cystoscopic lesions measured ≥2 cm rather than <2 cm, initial urine cytology tended to be positive or suspicious for UC (5 of 21 lesions [23.8%] vs 1 of 26 lesions [3.8%], respectively; P = .076). However, patients who had lesions ≥2 cm rather than <2 cm were not more likely to demonstrate recurrence/progression of either LGUC or HGUC (4 of 16 patients [25%] vs 13 of 25 patients [52%], respectively; P = .113). In addition, patients who had multiple lesions on cystoscopy tended to have abnormal (atypical or higher) urine cytology on follow-up compared with patients who had single lesions on initial cystoscopy (20 of 32 patients [62.5%] vs 18 of 44 patients [40.1%], respectively; P = .063). Likewise, patients who had multiple lesions on cystoscopy were significantly more likely to have a recurrence/progression of either LGUC or HGUC than those who had single lesions (21 of 31 patients [67.7%] vs 13 of 42 patients [31%], respectively; P = .002) (see Table 4).

Table 4. Cytoscopic Predictors of Recurrence or Progression
VariableRecurrence or Grade Progression: No. of Patients/Total No. (%)P
Tumor size, cm  
≥24/16 (25) 
No. of lesions  
Single21/31 (67.7) 
Multiple13/42 (31).002

When considering the initial urine cytology diagnosis, patients who had an initial cytologic diagnosis of positive or suspicious for UC had a tendency toward histologic recurrence of either LGUC or HGUC during follow-up compared with those who had an initial negative or atypical cytology diagnosis (9 of 12 patients [75%] vs 28 of 62 patients [45.2%], respectively; P = .112). Specifically, those who had an initial cytologic diagnosis of positive or suspicious for UC were significantly more likely to demonstrate HGUC on subsequent histologic sampling compared with those who had an initially negative or atypical cytologic diagnosis (7 of 12 patients [58.3%] vs 12 of 62 patients [19.4%], respectively; P = .009) (see Table 5).

Table 5. Cytologic Predictors of Recurrence or Progression
Initial Urine CytologyNo. of Patients/Total No. (%)P
  1. Abbreviations: HGUC, high-grade urothelial carcinoma; UC, urothelial carcinoma.
Any recurrence of UC  
Positive/suspicious9/12 (75) 
Negative/atypical28/62 (45).112
Progression to HGUC  
Positive/suspicious7/12 (58.3) 
Negative/atypical12/62 (19.4).009


Bladder cancer, which is the fourth most common cancer among men and the eighth leading cause of cancer-related death,[1] is well known for its propensity to recur, giving rise to the concept of a “field effect” in which the entire urothelium is at increased risk for carcinogenesis. Determining which patients are at the greatest risk of recurrence, however, continues to be a challenge.

It has been demonstrated that tumor multicentricity, or the presence of multiple tumors, and tumor size are key risk factors for recurrence and progression in patients with nonmuscle-invasive bladder cancer.[4] Researchers from the European Organization for Research and Treatment of Cancer have derived a scoring system to determine the risk of recurrence and progression in patients with Ta or T1 bladder cancer. In this scoring system, patients with multiple tumors are given more points, which correspond to an increased risk of recurrence or progression.[18] Not surprisingly, in our study population, which specifically included patients with low-grade, nonmuscle-invasive tumors, we similarly observed that tumor multicentricity was an important predictor of disease progression or recurrence. Patients who had multiple lesions on cystoscopy were significantly more likely to have a recurrence/progression of either LGUC or HGUC than the patients who had single lesions on cystoscopy. These patients, as would be expected, also tended to have abnormal urine cytology during subsequent follow-up surveillance.

The European Organization for Research and Treatment of Cancer scoring system also incorporates tumor size ≥3 cm in greatest dimension as an important risk factor in determining recurrence or progression in patients with superficial bladder cancers. On the other hand, others have observed no significant correlation between tumor size and the likelihood of recurrence or grade progression in patients with LGUC.[15] In our current study, similar to the findings reported by Miyamoto et al, we were unable to demonstrate a clear correlation between tumor size and disease recurrence or progression. However, it is important to note that, for many of the patients in our study, an exact measurement of tumor size was not provided within the cystoscopic report, making it difficult to draw conclusions about tumor size and the risk of progression or recurrence in our patient population. We did observe that, when the reported tumor size was ≥2 cm, cytology at the time of LGUC diagnosis tended to be positive. One possible explanation is that the larger the tumor, the more likely the tumor is to harbor small areas of high-grade cytology, which can be identified in urine cytology specimens. Likewise, the diagnoses of LGUC were not subjected to a consensus review, making the initial under grading of tumors a potential contributing factor to our findings.

Until now, the role of urine cytology has been mainly in the detection and surveillance of bladder cancer, however, unfortunately, its use in the detection of low-grade tumors is limited by poor sensitivity. Although the reported specificity of urine cytology approaches 94% to 100% for the detection of LGUC, the reported sensitivity is only approximately 35% to 65%.[19, 20] Conversely, the role of urine cytology as a potential predictor of recurrence or progression in patients with low-grade, nonmuscle-invasive tumors has yet to be fully explored; hence, the most current prognostic scoring systems for patients with Ta or T1 bladder cancer focus primarily on gross and histologic tumor characteristics and do not take urine cytology into account.[21] A few small studies have demonstrated that positive urine cytology in patients with superficial (Ta or T1) tumors is associated with disease recurrence, suggesting that urine cytology perhaps does have a role to play in prognostication.[11, 13, 14] However, none demonstrated an association of urine cytology with disease progression. Further, those studies included patients who had high-grade tumors, a known independent risk factor for recurrence and progression.

In our study of patients with LGUC, we observed that patients who had urine cytology diagnosed as either positive or suspicious for UC at the time of diagnosis of a low-grade bladder tumor tended to have a greater probability of recurrence during follow-up compared with those who had an initial negative or atypical cytology diagnosis. Moreover, these patients were significantly more likely to have demonstrated progression to HGUC on subsequent histologic sampling. Although further study with larger sample sizes and longer term follow-up is needed, these findings suggest an important role for urine cytology not only in detection and surveillance of patient with bladder cancer but also in risk assessment and prognosis.

In summary, we assessed the cytoscopic findings, cytologic findings, and follow-up data from 87 patients who were diagnosed with nonmuscle-invasive LGUC and observed that the presence of multiple tumors on cystoscopy at the time of diagnosis was a significant risk factor for disease recurrence or progression and for the development of abnormal urine cytology on subsequent surveillance. We also have documented that urine cytology tends to be positive when tumors measure ≥2 cm and that a positive urine cytology at the time of diagnosis of LGUC predicts disease recurrence and progression. Our current findings emphasize the importance of proper documentation of cystoscopy findings, including size and number of tumors, and also the recognition of urine cytology not only as a tool for diagnosis and surveillance but also as a prognostic indicator in patients with nonmuscle-invasive LGUC.


No specific funding was disclosed.


The authors made no disclosures.