Original Study

Significance of Positive Urine Cytology on Progression and Cancer-Specific Mortality of NoneMuscle-Invasive Bladder Cancer Fumitaka Koga, Shuichiro Kobayashi, Yasuhisa Fujii, Junichiro Ishioka, Minato Yokoyama, Yasukazu Nakanishi, Yoh Matsuoka, Noboru Numao, Kazutaka Saito, Hitoshi Masuda, Kazunori Kihara Abstract This prospective study, consisting of 326 patients with nonemuscle-invasive bladder cancer (NMIBC) who underwent 597 tumor resections, demonstrated that a voided urine cytology result that is positive for atypical cells is a predictor of disease progression, independent of histological grade. Time-dependent covariate Cox models appear more sensitive in detecting risk factors for NMIBC progression than conventional time-fixed models. Voided urine cytology may be useful for NMIBC risk stratification. Background: Positive results from voided urine cytology (VUC) indicate the fragility of the intercellular adhesion of bladder cancer cells, a critical biological process for invasion and metastasis, along with the presence of atypical cells. Few studies have focused on the prognostic role of VUC in nonemuscle-invasive bladder cancer (NMIBC). Methods: Between 2000 and 2010, 326 patients diagnosed pathologically with Ta or T1 bladder urothelial carcinoma underwent 597 transurethral resections of bladder tumor (TURBTs). Clinicopathological data were prospectively collected at each TURBT. Reports of cells of class IIIb or greater were considered positive VUC results. Muscle-invasive or metastatic recurrences were considered progression. Risk factors for progression and cancer-specific mortality (CSM) were determined using time-fixed and time-dependent Cox models. Variables at the study entry and at each TURBT were used for time-fixed and time-dependent models, respectively. Results: The 5-year cumulative progression and CSM rates were, respectively, 7% and 5% (median follow-up, 46 months). The 5-year cumulative progression and CSM rates for patients with positive VUC were 20% and 15%, respectively, compared with 2% (P < .0001) and 2% (P ¼ .0002), respectively, for patients with negative VUC results. A positive VUC result was a significant and independent risk factor for progression and CSM in the time-fixed and time-dependent models. In time-dependent models, 7 predictors for progression or CSM were identified (positive VUC results, T1 disease, lack of intravesical instillation, higher prior recurrence rate, higher histological grade, male gender, and advanced age), whereas 3 predictors were identified in time-fixed models (positive VUC, T1 disease, and higher prior recurrence rate). VUC results consistently outperformed histological grade as a prognostic predictor. Conclusion: Positive VUC results predict the progression and CSM of NMIBC, independent of and outperforming histological grade. Clinical Genitourinary Cancer, Vol. -, No. -, --- ª 2013 Elsevier Inc. All rights reserved. Keywords: Cancer-specific survival, Progression, Urinary bladder, Urothelial carcinoma, Voided urine cytology

Introduction Bladder cancer is one of the most common genitourinary malignancies and, for patient management, is classified into muscle-invasive Department of Urology, Tokyo Medical and Dental University Graduate School, Tokyo Submitted: Jun 6, 2013; Revised: Jul 25, 2013; Accepted: Jul 31, 2013 Address for correspondence: Fumitaka Koga, MD, PhD, Tokyo Metropolitan Cancer and Infectious diseases Center, Komagome Hospital, 3-18-22 Honkomagome, Bunkyo-ku, 113-8677, Tokyo, Japan E-mail contact: [email protected]

1558-7673/$ - see frontmatter ª 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clgc.2013.07.007

bladder cancer and nonemuscle-invasive bladder cancer (NMIBC). NMIBC accounts for approximately 70% of new urothelial bladder cancer cases and generally has a favorable prognosis.1 However, 40% to 80% of NMIBC patients develop intravesical tumor recurrences, and approximately 15% of recurrent tumors progress to muscleinvasive or metastatic disease.1,2 Once the disease progresses, the prognoses of patients are generally unfavorable.2 Prediction of stage progression is therefore important for the management of NMIBC patients. Voided urine cytology (VUC) is a simple, noninvasive, and relatively inexpensive test for detecting urothelial carcinoma (UC).

Clinical Genitourinary Cancer Month 2013

-1

Prognostic Significance of Cytology in NMIBC In combination with cystoscopy, VUC is currently the standard method for detecting UCs and monitoring patients for recurrent UCs.3 Positive VUC results indicate not only the presence of atypical cells but also highlight the fragility of UC cells’ intercellular adhesion. Given that the loss of intercellular adhesion is one of several critical biological processes by which cancer cells acquire invasive and metastatic potential,4 positive VUC results may be useful in predicting the progression of NMIBC to muscle-invasive or metastatic disease. A number of studies reported risk factors for stage progression of NMIBC, including histological grade, tumor stage, number of tumors, presence of concurrent carcinoma in situ (CIS), prior recurrence rate, tumor size, response to intravesical instillation of bacillus Calmette-Guérin (BCG), advanced age, and tumor location.5-9 Some papers have reported the prognostic role of VUC testing in NMIBC, although few studies have focused on the topic. Zieger et al. reported that a positive VUC result is associated with stage progression of stage Ta NMIBC.10 Okajima et al. demonstrated that a positive VUC result has a significant impact on cancerspecific mortality (CSM) in T1 bladder cancer patients.11 Both of these were retrospective studies, and the latter was conducted using a nationwide Japanese patient database.11 All of the published studies focusing on risk factors for progression and CSM of NMIBC were, to our knowledge, conducted using time-fixed analysis, with clinicopathological parameters at the study entry used for analysis.5-11 Given that NMIBC patients usually develop disease progression after multiple recurrences,1,2 and that many clinicopathological parameters are different at each NMIBC recurrence in the same patient, risk evaluation incorporating timedependent covariates at each NMIBC recurrence may be more appropriate for predicting future progression and CSM than risk evaluation based on a time-fixed analysis.12 In the present study, we prospectively investigated the potential role of positive VUC results in predicting stage progression and CSM of NMIBC in a single institutional patient cohort. Risk factors for these events were analyzed using both time-fixed and timedependent Cox proportional hazards models.

Patients and Methods

2

-

Between January 2000 and June 2010, 340 Japanese patients underwent transurethral resections of bladder tumors (TURBTs) a total of 611 times for therapeutic purposes and were pathologically diagnosed with Ta or T1 (N0M0) UC of the bladder, based on the 2002 tumor, node, metastasis (TNM) classifications,13 at the University Hospital of Tokyo Medical and Dental University. Patients with primary CIS were excluded from this study. VUC testing was performed before each TURBT without any interventions in the urinary tract. Reports of cells of class IIIb or greater were considered positive VUC results; class IIIb was defined as cytologic findings suspicious of, but not conclusive for, malignancy with advanced dysplasia.14 Clinicopathological data, including age, gender, tumor number, tumor diameter, prior recurrence rate, T stage, concurrent CIS, histological grade, and VUC, were collected prospectively at each TURBT. Tumors were histologically graded according to the 1973 World Health Organization grading system.15 Of the 340 patients, 14 patients who underwent cystectomy after the initial TURBT were excluded. Thus, 326 patients (60 women and

Clinical Genitourinary Cancer Month 2013

266 men) and 597 therapeutic TURBTs were included in this prospective study. The institutional review board approved this study, and informed consent was obtained from each participant. The patients routinely received single-dose intravesical chemotherapy with doxorubicin (Adriamycin) or mitomycin-C immediately after TURBT. A second TURBT was conducted in patients whose tumors were pathologically diagnosed as T1 and whose TURBT specimens at the initial TURBT were histologically shown to not contain the muscle layer. In the present series, T1 disease was diagnosed after 113 TURBTs in 84 patients and 59 second TURBTs were performed, with no diagnosis of muscle-invasive bladder cancer (MIBC). The patients were followed up with VUC and cystoscopy every 3 months for the first 2 years after TURBT, every 6 months for the next 3 years, and yearly thereafter. Of the 326 patients, 246 (82%) had the following intravesical instillations during their follow-up examinations: doxorubicin (n ¼ 110); mitomycin-C (n ¼ 95); BCG (n ¼ 21); doxorubicin and mitomycin-C (n ¼ 9); and BCG and doxorubicin or mitomycin-C (n ¼ 11). The primary and secondary endpoints were time to progression and time to CSM, respectively. Progression was defined as the presence of muscle-invasive disease or metastatic recurrence. The time to progression and time to CSM were calculated from the date of the initial TURBT for the time-fixed analysis and from the date of each TURBT for the time-dependent covariate analysis.12 Variables associated with the time to progression and time to CSM were analyzed using a Cox proportional hazard model. For the multivariate analysis, variables were selected using the backward stepwise method. Clinicopathological data collected at the study entry were used as variables for the time-fixed analysis. For the timedependent covariate analysis, the following variables were handled as time-dependent covariates: number of tumors, tumor diameter, prior recurrence rate, T stage, presence or absence of concurrent CIS, grade, and VUC results. Continuous variables were expressed as median values and ranges. Categorical data and continuous variables were compared using the two-tailed Fisher exact test or c2 test, and the Wilcoxon rank sum test, respectively. Kaplan-Meier curves were used to estimate eventtime distributions, and differences between curves were analyzed using the log-rank test. Statistical analyses were performed with JMP 8 (SAS Institute Inc, Cary, NC) and R.2.11.0 (The R Foundation for Statistical Computing, Vienna, Austria). A P value < .05 was considered to indicate statistical significance.

Results Patient and tumor characteristics of the 326 enrolled patients are shown in Table 1. In all cases, tumors were histologically determined as UC. The positive VUC rate for this cohort was 28% (90/326) at study entry. Positive VUC was significantly associated with male gender, multiple disease, tumor size  3 cm, primary disease, T1 disease, presence of concurrent CIS, and higher histological grade (Table 1). The median follow-up period for the whole cohort was 46 months. Of the 326 patients, 29 were lost to follow-up (median follow-up period, 40 months; range, 7 to 92 months) and 15 died of causes other than bladder cancer (median time to death, 40 months; range, 8 to 109 months). During the follow-up, 19 patients progressed to muscle-invasive or metastatic disease, and 11 patients

Fumitaka Koga et al Table 1 Patient and Tumor Characteristics of the 326 Enrolled Patients According to Cytology Results

Figure 1 Time-to-Progression Curves (A) and Time-to-CSM Curves (B) Based on VUC Results at Study Entry

VUC Negative, n (%)

Age

.08

70

136 (42)

45 (33)

91 (67)