Transitional cell carcinoma of the prostate

Clinicopathologic study of 50 cases

Abstract

BACKGROUND

This study was performed to identify clinical and histologic features most significant in predicting outcome in patients with transitional cell carcinoma (TCC) of the prostate without invasive bladder carcinoma.

METHODS

The histologic and clinical material from 50 patients with prostatic TCC without invasive bladder carcinoma were studied. The tumors were divided into the following locoregional categories: 1) TCC in situ (CIS) of the prostatic urethra; 2) CIS of the prostatic ducts and acini; 3) TCC with stromal invasion; 4) TCC with extraprostatic extension and/or seminal vesicle involvement; and 5) lymph node metastases. The Kaplan-Meier method was used to generate survival distributions for the locoregional categories, and comparison of survival curves was accomplished with the log rank test.

RESULTS

The 5-year disease specific survival rate was 52%. The 5-year disease specific survival rates for the locoregional categories were as follows: CIS of the prostatic urethra and prostatic ducts and acini (n = 19), 100%; TCC with stromal invasion (n = 21), 45%; TCC with extraprostatic extension and seminal vesicle involvement (n = 3), 0%; and lymph node metastases (n = 7), 30%. There was a significant difference in disease specific survival when patients with CIS were compared with patients with stromal invasion, extraprostatic extension and seminal vesicle involvement, and lymph node metastases (P = 0.0001).

CONCLUSIONS

This study demonstrates that patients with prostatic TCC involving prostatic stroma, extraprostatic tissues, seminal vesicles, and lymph nodes have a significantly poorer 5-year disease specific survival than patients with CIS. Cancer 1998;82:703-7. © 1998 American Cancer Society.

There is no accepted staging system for transitional cell carcinoma of the prostate. Currently, prostatic transitional cell carcinoma is staged according to the Jewett-Strong and TNM systems for the urinary bladder, and classified as Stage D1 and T4 disease, respectively. The application of these staging systems implies that prostatic transitional cell carcinoma is a high stage neoplasm with a poor prognosis.

Few studies have focused specifically on the identification of histologic prognostic indicators, and in turn, the biologic behavior of prostatic transitional cell carcinoma in the absence of invasive bladder carcinoma. Studies have suggested that prostatic stromal invasion is a poor prognostic histologic finding.1-5 However, a recent study by Pagano et al reported that patients with transitional cell carcinoma in situ (CIS) involving the prostatic ducts and acini had just as poor survival as those with prostatic stromal invasion.6 The prognostic significance of other histopathologic features also remains unclear.

In this study, we analyzed the clinicopathologic characteristics of 50 patients with prostatic transitional cell carcinoma, assessing a wide range of histopathologic and clinical features. All patients had been treated by radical cystoprostatectomy, and none had invasive bladder carcinoma. The study was performed to identify those factors most significant in predicting outcome in patients with prostatic transitional cell carcinoma.

MATERIALS AND METHODS

The clinical information and histologic material from 50 patients with a diagnosis of transitional cell carcinoma of the prostate without invasive bladder carcinoma who were treated by radical cystoprostatectomy at the Mayo Clinic between 1965 and 1995 were reviewed. This included patients with CIS of the bladder; patients with coexistent or a previous history of bladder carcinoma invasive into the lamina propria or beyond were excluded from this study. We documented the initial date of diagnosis, a detailed history of previous and/or recurrent transitional cell CIS of the upper urinary tract, urinary bladder, and nonprostatic urethra, and previous treatment for CIS.

Archival hematoxylin and eosin stained slides were examined in all cases, and all neoplasms were determined to be transitional cell carcinomas. Five locoregional categories of prostatic transitional cell carcinoma were defined as follows: 1) CIS of the prostatic urethra and periurethral glands; 2) CIS of the prostatic ducts and acini; 3) transitional cell carcinoma with urethral submucosal and prostatic stromal invasion; 4) involvement of seminal vesicles and extraprostatic extension; and 5) lymph node metastasis. Each patient was assigned to one of these categories. Categories 1 and 2 were compared with Categories 3-5 for statistical analysis (noninvasion vs. invasion/metastasis). Other factors assessed were grade of invasive tumors (according to the World Health Organization grading system), preoperative and postoperative systemic chemotherapy and radiotherapy, previous history of urinary bladder/upper urinary tract CIS, previous treatment for CIS, concurrent CIS (at the time of cystoprostatectomy), concurrent prostatic adenocarcinoma, serum prostate specific antigen concentration, and subsequent urethrectomy for recurrent neoplasm.

The outcome variables were disease specific and overall survival, both of which were determined from the time of diagnosis of prostatic transitional cell carcinoma to the time of death. The Kaplan-Meier method was used to generate survival distributions,6 and the locoregional categories were compared using the log rank test. Univariate analysis was used to evaluate the significance of the parameters other than the locoregional categories in overall and disease specific survival.

RESULTS

Patients ranged in age from 52-87 years (mean age, 66 years), and were followed for a mean of 69 months (range, 3-197 months). Twenty-one patients (42%) had a previous history of transitional cell CIS of the bladder, and 10 patients (20%) had received intravesical mitomycin, thiotepa, or bacille Calmette-Guérin (BCG) prior to undergoing radical cystoprostatectomy. Thirty-two patients (64%) had CIS of the bladder identified in the radical cystoprostatectomy specimen. Twelve patients (24%) had no prior history of urinary bladder CIS, and no CIS detected in the urinary bladder resection specimen. The majority of patients (86%) had flat lesions, and 14% of the tumors had a papillary component. Five invasive tumors (10%) were Grade 2 and 26 invasive tumors (52%) were Grade 3. CIS in all cases was high grade. Concurrent prostatic adenocarcinoma was identified in 4 patients (8%). The serum prostate specific antigen concentration was measured in 3 patients with values of 4.2, 1.8, and 1.5 ng/mL; none had prostatic adenocarcinoma. Subsequent urethrectomy for recurrent CIS was performed in 10 patients (20%). Four patients (8%) with prostatic stromal invasion received preoperative radiotherapy; 3 of these patients died of carcinoma. No patients received preoperative chemotherapy. After surgery, 3 patients (6%) with metastases received systemic chemotherapy, and 2 of these patients, both of whom had distant metastases, died of carcinoma whereas the third patient with regional lymph node involvement was alive and free of disease at last follow-up. Three patients (6%) developed transitional cell carcinoma of the upper urinary tract after undergoing cystoprostatectomy (Table 1). Nineteen patients (38%) died of transitional cell carcinoma, and 15 (30%) died of other causes. The remaining 16 patients (32%) were alive and free of disease at last follow-up.

Table 1. Pathologic and Clinical Features of 50 Patients with Transitional Cell Carcinoma of the Prostate

Feature	No. of patients (%)
CIS: carcinoma in situ.
Previous history of bladder CIS	21 (42)
Previous treatment for bladder CIS	10 (20)
Concurrent bladder CIS	32 (64)
No previous history of or concurrent bladder CIS	12 (24)
Concurrent prostatic adenocarcinoma	4 (8)
Subsequent urethrectomy for CIS	10 (20)
Subsequent transitional cell
carcinoma of upper urinary tract	3 (6)
Preoperative radiation	4 (8)
Postoperative chemotherapy	3 (6)

The diagnosis of transitional cell carcinoma was confirmed in all cases (Figs. 1 (64K) and 2 (59K)). The number of patients for each of the locoregional categories was as follows: CIS involving the prostatic urethra and periurethral glands, 4 patients (8%); CIS involving the prostatic ducts and acini, 15 patients (30%); urethral submucosal and prostatic stromal invasion, 21 patients (42%); extraprostatic extension and/or seminal vesicle involvement, 3 patients (6%); and regional lymph node metastases, 7 patients (14%) (Table 2). No patients had distant metastases at the time of cystoprostatectomy. The 5-year overall survival rate was 40%. The 5-year overall survival rates for the locoregional categories were as follows: CIS of the prostatic urethra, periurethral glands, and prostatic ducts and acini, 62%; urethral submucosal and prostatic stromal invasion, 35%; extraprostatic extension/seminal vesicle involvement, 0%; and lymph node metastases, 30%. The 5-year disease specific survival rate was 52%.

Figure 1. Transitional cell carcinoma in situ within the prostatic ducts. The ducts are expanded by the neoplastic transitional cells.

Figure 2. Invasive transitional cell carcinoma of the prostate.

Table 2. The Number of Patients in the Locoregional Categories of Prostatic Transitional Cell Carcinoma and Survival Data

		5-year survival (%)
Locoregional category	No. of patients (%)	Overall	Disease specific
CIS: carcinoma in situ. a  P = 0.0001 when compared with all other patients.
Prostatic urethral CIS and	4 (8)
prostatic duct and acini CIS	15 (30)	62	100a
Prostatic stromal invasion	21 (42)	35	45
Extraprostatic extension	3 (6)	0	0
Lymph node metastasis	7 (14)	30	30
All cases	50 (100)	40	52

The 5-year disease specific survival rates for the locoregional categories were as follows: CIS of the prostatic urethra, periurethral glands, and prostatic ducts and acini, 100%; urethral submucosal and prostatic stromal invasion, 45%; extraprostatic extension/seminal vesicle involvement, 0%; and lymph node metastases, 30% (Fig. 3 (5K)).

Figure 3. Comparison of disease specific survival among the locoregional categories. A statistically significant difference was observed between carcinoma in situ and invasive transitional cell carcinoma (P = 0.0001).

There was a difference in disease specific survival when patients with CIS of the prostatic urethra, periurethral glands, and prostatic ducts and acini (noninvasive locoregional categories) were compared with patients with urethral submucosal and prostatic stromal invasion, extraprostatic extension/seminal vesicle involvement, and lymph node metastases (invasive/metastatic locoregional categories) (P = 0.0001). There was no difference in disease specific survival among the patients in the noninvasive locoregional categories or among the patients in the invasive/metastatic locoregional categories.

No differences in disease specific survival were identified for the other clinical and histopathologic features examined including age, previous history of urinary bladder CIS, previous history of intravesical chemotherapy or BCG therapy, concurrent urinary bladder CIS, papillary component to the tumor, concurrent prostatic adenocarcinoma, preoperative and postoperative systemic chemotherapy and radiation therapy, subsequent urethrectomy for recurrent CIS, or subsequent upper urinary tract transitional cell carcinoma. The invasive/metastatic locoregional categories were evaluated separately to identify any histopathologic features that could identify a subset of patients with better prognosis; no such features were identified.

DISCUSSION

In this study, we found a 52% 5-year disease specific survival rate for patients with transitional cell carcinoma of the prostate, and identified that the extent of locoregional spread was the strongest predictor of patient survival outcome. This was evidenced by a 5-year disease specific survival rate of 100% for patients with CIS compared with a 45% 5-year disease specific survival rate for patients with urethral and prostatic stromal invasion.

The anatomic barrier to tumor spread in prostatic transitional cell carcinoma is the basement membrane of the prostatic urethra, periurethral glands, and prostatic ducts and acini. The importance of stromal invasion has been identified previously. In one study, 2 of 18 patients (11%) with CIS of the prostatic ducts developed metastases, compared with 5 of 5 patients (100%) with prostatic stromal invasion.2 Schellhammer et al reported 15 patients with prostatic transitional cell carcinoma who also had low stage bladder carcinoma. Five of 10 patients with CIS of the prostatic ducts and acini died of cancer within 5 years, compared with 4 of 5 patients with prostatic stromal invasion.3 Schellhammer et al concluded that patients with noninvasive transitional cell carcinoma of the prostate associated with low stage bladder carcinoma have a survival predicted on the basis of radical cystectomy for low stage bladder carcinoma alone. However, for patients with invasive prostatic transitional cell carcinoma associated with a low stage bladder carcinoma, the 5 year survival rate was determined by the stage of the prostate carcinoma and was low. Esrig et al also found that patients with low stage bladder tumors (T1, Tis, Ta, and T0) and prostatic transitional cell carcinoma with stromal invasion had a worse prognosis than patients with low stage bladder tumors and prostatic transitional cell carcinoma without prostatic stromal invasion.5 In their study, the probability for nonrecurrence in the group of patients with low stage bladder tumors and prostatic stromal invasion was approximately 60% compared with 90% for patients without stromal invasion. In addition, in patients treated by radical cystoprostatectomy, they could find no 5-year survival difference between patients with prostatic urethral CIS and patients with CIS involving the prostatic ducts.

In contrast to these studies, Pagano et al reported there was no difference in survival when patients with CIS involving the prostatic ducts and acini were compared with patients with prostatic stromal invasion.6 They based this conclusion on 12 patients with prostatic transitional cell carcinoma who also had superficial bladder cancer (pTis, pT1, and pT0) treated by radical cystoprostatectomy. They reported 3-year survival estimates of 50% and 40% for ductal and acinar CIS and prostatic stromal invasion, respectively. The 3-year survival for patients with prostatic urethral CIS was 100%. It is unclear why the findings of Pagano et al differ from the findings in the current study as well as the findings of other researchers.

We did not identify a difference in survival between patients with urethral submucosal and prostatic stromal invasion, extraprostatic extension/seminal vesicle involvement, and lymph node metastases, but it is important to note that there were small numbers of patients in the latter two locoregional categories. Additional numbers of patients in each of these categories will be necessary to determine whether a difference in survival exists.

Our study demonstrates that prostatic transitional cell carcinoma frequently is associated with urinary bladder CIS. Thirty-eight of the patients in our study (76%) had a previous history of bladder CIS or were found to have bladder CIS at the time of cystoprostatectomy. All 25 patients with prostatic transitional cell carcinoma in the study by Wishnow and Ro had multifocal CIS, and Johnson et al reported that 14 of 20 patients (70%) with prostatic transitional cell carcinoma also had CIS of the bladder.2, 4 Conversely, patients with bladder carcinoma treated by radical cystoprostatectomy often are found to have prostatic transitional cell carcinoma. The frequency of prostatic involvement ranges from 5-43%.3, 8-10 This variation likely is due to sampling rather than true variation in patients with prostatic transitional cell carcinoma associated with bladder carcinoma.1

Ten of the patients in our series (20%) were found to have recurrent CIS of the residual urethra after surgery. This required completion urethrectomy, and although subsequent urethrectomy did not affect survival, the procedure is associated with morbidity. Wishnow and Ro found that four of ten patients in their report required subsequent urethrectomy.3 Hardeman et al found that of 30 patients with prostatic involvement by transitional cell carcinoma, 37% developed urethral recurrence compared with 4% of patients without prostatic involvement.11 They recommended prophylactic urethrectomy only in patients with transitional cell carcinoma involving the prostatic urethra. Our study supports this recommendation.

We identified concurrent prostatic adenocarcinoma in 4 of the patients in the current study (8%). Johnson et al found that 3 of 9 patients (33%) with prostatic transitional cell carcinoma also had adenocarcinoma,4 whereas Algaba et al found that 1 of 5 patients (20%) had adenocarcinoma.12 The difference in the prevalence of prostatic adenocarcinoma in these studies most likely is due to the sampling. In our study, only selected portions of the prostate were examined, and the prostate was not embedded completely. More extensive sampling of the prostate most likely would have increased the number of patients found to have adenocarcinoma.

This study demonstrates that the extent of locoregional involvement of prostatic transitional cell carcinoma is a strong predictor of patient survival, and the anatomic barrier to disease progression and a significantly worse prognosis is the basement membrane of the prostatic urethra, periurethral glands, and prostatic ducts and acini. We found that prostatic transitional cell carcinoma frequently was associated with urinary bladder CIS, and urethral recurrence of CIS was a common occurrence.