Key points

Introduction

Urothelial carcinoma is a multifocal disease of the urothelium that can develop anywhere in the urinary tract, including the prostatic urethra. The lining of the prostatic urethra and the prostatic ducts have the same transitional cell urothelium as the bladder; therefore, it is not surprising that a patient who develops non-muscle invasive bladder cancer (NMIBC) can develop urothelial carcinoma in the prostatic urethra or ducts, either as an extension from the primary process or as a de novo entity.

Knowledge of urothelial cancer of the prostate is pertinent to the management on NMIBC for 3 reasons: the presence and depth of prostate urothelial cancer will affect prognosis ; its presence in the prostatic urethra may suggest understaging of the primary tumor in the bladder ; and NMIBC managed conservatively can recur in the prostatic urethra.

This article offers an overview of urothelial cancer in the setting of NMIBC, and its incidence, diagnosis, and management.

Introduction

Urothelial carcinoma is a multifocal disease of the urothelium that can develop anywhere in the urinary tract, including the prostatic urethra. The lining of the prostatic urethra and the prostatic ducts have the same transitional cell urothelium as the bladder; therefore, it is not surprising that a patient who develops non-muscle invasive bladder cancer (NMIBC) can develop urothelial carcinoma in the prostatic urethra or ducts, either as an extension from the primary process or as a de novo entity.

Knowledge of urothelial cancer of the prostate is pertinent to the management on NMIBC for 3 reasons: the presence and depth of prostate urothelial cancer will affect prognosis ; its presence in the prostatic urethra may suggest understaging of the primary tumor in the bladder ; and NMIBC managed conservatively can recur in the prostatic urethra.

This article offers an overview of urothelial cancer in the setting of NMIBC, and its incidence, diagnosis, and management.

Definitions, classifications, and staging

The prostate contains glands that consist of ducts and terminal acini that are in continuity with the transitional urothelium of the prostatic urethra and bladder. The ducts and acini are surrounded by the stroma, consisting of connective tissues and smooth muscle cells. The urothelium and the stroma are separated only by a thin basement membrane. The differences in anatomic and functional structure of the prostate, as well as the absence of muscularis propria, alter the natural history, staging, and management of urothelial cancer that occurs in this area. Involvement of prostatic urethra, ducts, acini, and stroma have all been described in patients with non-muscle invasive bladder transitional cell carcinoma. These entities each have distinctly different clinical significance and prognosis.

Conceptually urothelial cancer of the prostatic urethra can be confusing, so it is important to establish classifications and definitions before further discussion on diagnosis and management. A significant contributor to this problem is the different anatomic structure of the prostate and the fact that prostatic urothelial cancer can arise primarily in the prostate or as a direct extension of a primary bladder tumor.

As a result, prostatic urothelial cancer is included in two separate staging systems: first, a staging system for primary urothelial cancer arising in the prostate, and second, a staging system for primary urothelial cancer of the bladder that extends into the prostate. Table 1 illustrates the similarities and differences between the staging systems for these 2 anatomic sites.

It should be noted that the American Joint Cancer Committee staging system was only changed in 2010 to reflect these differences. Before the 2010 revision, any involvement of the prostate in a patient with bladder cancer was classified as stage T4a bladder cancer, regardless of depth of invasion. This classification was justly criticized, as convincing survival data suggest that primary prostatic urothelial cancer, as opposed to bladder cancer directly invading the prostate, carries a different prognosis, as does involvement of the prostatic urethra with urothelial cancer of different depths of invasion. Examining the first question, Pagano and colleagues compared survival in patients with contiguous spread of bladder cancer to the prostate with those with a noncontiguous urothelial cancer arising within the prostate. Only 7% of the patients with a contiguous tumor were alive at 5 years compared with 46% of patients with a noncontiguous tumor. Examining the second question of prostatic stromal involvement in comparison with lesser degrees of tumor invasion, Esrig and colleagues investigated 143 men who underwent cystectomy and were found to have prostatic involvement; 5-year overall survival rates with and without stromal invasion were 36% and 71%, respectively. In addition, men without stromal invasion had a prognosis no worse than men without any prostatic involvement at all. Similarly, Solsona and colleagues found patients with prostatic stromal involvement to have a hazard ratio of 2.89 for cancer-specific mortality in comparison with those without stromal involvement. Herr and Donat investigated patients with NMIBC who had been treated with intravesical bacillus Calmette-Guérin (BCG) and had late recurrences in the prostate. Prognosis was highly dependent on depth of invasion, whereby survival rate was 82% in patients with prostatic urethral or ductal involvement, compared with 48% with those with stromal involvement. Examining the question of lesser degrees of prostatic involvement, Barocas and colleagues investigated 3-year overall survival among 162 cystoprostatectomy patients with prostatic involvement. Three-year survival was 59% if the prostatic urethral involvement was carcinoma in situ (Tis pu), 52% if there was ductal involvement, and only 17% if there was stromal involvement.

Incidence

When discussing the incidence of urothelial cancer involving the prostate it should be realized that isolated primary prostatic urothelial cancer is rare; instead this tumor is almost always associated with the presence of, or history of, urothelial cancer of the bladder. There is a moderate amount of data available on the incidence of urothelial cancer arising within the prostate in patients with bladder cancer, with reported overall incidence rates of 6% to 48%, with a 7% to 17% incidence of prostatic stromal involvement. Most of these data come from the cystoprostatectomy literature, which is heavily focused on muscle invasive bladder cancer and only includes NMIBC that has failed more conservative therapy. Therefore, much of the information is not directly applicable to the patient with newly diagnosed NMIBC.

The incidence figures for NMIBC-associated prostatic involvement also vary depending on the method by which the prostatic urethra is assessed (visual inspection, cold-cup biopsy, resectoscope biopsy, or pathologic analysis of the cystoprostatectomy specimen) and the timing and indications for the sample collection (diagnosis, investigation of isolated positive cytology, following intravesical therapy, or at the time of cystectomy).

In regard of the different sampling methodologies, the incidence of prostatic involvement reported following cold-cup biopsy is 24% to 27%. In these series Tis constitutes 69% to 100% of prostate pathology, and only 1 case of stromal invasion has been found. However, with cold-cup biopsy there is clearly a high frequency of inadequate sampling of the ducts and stroma. Most studies use a resectoscope to sample the prostate. Studies using a resectoscope loop biopsy have found a 6% to 36% rate of prostatic involvement with urothelial cancer. Many of these studies were designed to monitor outcomes of intravesical therapy whereby stromal and sometimes ductal involvement were exclusion criteria; therefore, establishing the true incidence of ductal involvement can be difficult. In a series of 49 patients undergoing transurethral resection (TUR) for a bladder tumor and with biopsy of their prostatic urethra, Schellhammer and colleagues reported that 18 (37%) had prostatic involvement. Of these patients, 13 had identifiable ducts in the specimen and 7 of 13 (56%) harbored ductal urothelial cancer. Cystectomy data, although heavily focused on muscle invasive bladder cancer, have been reported in a few series that allow for subanalysis of NMIBC. In a cystectomy series of 192 patients, Nixon and colleagues found 42 men with Tis, or T1, bladder disease, 8 (19%) of whom had prostatic involvement. Similarly, in series by Liedberg and colleagues and Mazzucchelli and colleagues, 5 of 19 men (26%) and 26 of 70 men (37%), respectively, with NMIBC were found to have prostatic involvement. Although these studies reported the overall rates of prostatic stromal involvement, this was not specified for the subset with NMIBC.

To sum up, the incidence rates of prostatic urethral involvement at initial diagnosis range from 6% to 37% and are dependent on sampling method and patient risk. The frequency of recurrence of urothelial cancer in the prostate following therapy for NMIBC ranges from 8% to 48%, and is explored in more detail in the next section.

Risk factors for prostatic urethral involvement

If diagnosed, prostatic involvement may drastically affect both prognosis and management of NMIBC. However, in most cases identification and diagnosis requires transurethral biopsy or resection of the prostatic urethra. As tissue biopsy of the prostate is not without potential morbidity, a moderate amount of work has gone into trying to identify which patients might benefit from sampling of the prostate. Numerous factors have been evaluated.

Much information can be gained from the cystectomy literature. Nixon and colleagues looked at 192 prostatectomy samples, of which 30 (15%) had prostatic urethral involvement. Patients with tumor multifocality and bladder carcinoma in situ (CIS) had rates of prostatic involvement of 34% and 31%, respectively. In a study by Mazzucchelli and colleagues, 94 out of 248 (38%) cystoprostatectomy specimens had prostatic urethral involvement. Trigonal or bladder neck tumor location, pT and pN stage, and previous history of recurrence were significant risk factors. Based on their data, a formula was constructed that uses number of tumor foci in bladder, history of prior recurrence, and location of bladder tumor, which was able to predict with 61% sensitivity and 78% specificity whether there was a concomitant prostatic involvement. Other series have found numerous other factors associated with prostatic urethral involvement, including multifocal CIS, trigonal CIS, periurethral tissue invasion, ureter invasion, vascular invasion, history of bladder tumor recurrence, and history of prior bladder intillations.

In the transurethral literature it has been reported that macroscopic tumor in the urethra correctly identifies prostatic urethral involvement with a sensitivity of 83% and a specificity of 95%. However, Mungan and colleagues reported a series with an overall prostatic urethral involvement rate of 6.2% whereby macroscopic involvement was seen in 3.5% of cases. In the other 2.7% the urethra appeared normal. The same study found that on univariate analysis, stage, grade, and multifocality of the bladder tumors correlated with prostatic involvement, not macroscopic appearance of the prostatic urethra. On multivariate analysis, only bladder tumor multifocality significantly correlated with prostatic involvement. In this study the risk of developing prostatic involvement in the presence of multiple bladder tumors was 16 times higher than the risk conferred by stage or grade.

Cystoscopic inspection, therefore, has a high specificity in correctly identifying a macroscopic tumor but a low sensitivity in correctly identifying the presence of tumor when the urethra appears normal, especially in a high-risk patient. The risk factors that correlate best with the presence of prostatic urethral carcinoma include the presence of multifocal bladder tumors, CIS, and tumors located on the trigone or bladder neck.

The European Association of Urology (EAU) 2011 guidelines on NMIBC recommend biopsy of the prostatic urethra for cases of bladder-neck tumor, when bladder CIS is present or suspected, when there is positive cytology without evidence of tumor in the bladder, or when abnormalities of the prostatic urethra are visible. The National Comprehensive Care Network (NCCN) recommends considering a TUR biopsy of the prostatic urethra if the bladder tumor is sessile, or there is suspicion of high-grade bladder tumor or bladder CIS.

Finally, whereas bladder tumor parameters can predict the presence of a prostatic urethral cancer, at least one study has investigated the converse, that is, the presence of prostatic urethral involvement as a risk factor for understaging of the primary bladder tumor. Huguet and colleagues evaluated 62 patients with Tis, T1, or Ta bladder cancer that had failed TUR and BCG and eventually went onto cystoprostatectomy. In their series, initial prostatic involvement was the only significant risk factor of upstaging of the bladder tumor to muscle or stromal invasive disease, with a hazard ratio of 12.2 (95% confidence interval 2.2–65.5).

Diagnostic method

There are numerous reports of the different methods used to diagnose prostatic urothelial cancer. Most describe cold-cut biopsy or TUR, although fine-needle aspiration and transrectal ultrasound-guided biopsy have been used.

As mentioned in the previous section, the finding of a cystoscopically visible papillary lesion in the prostatic urethra on initial inspection has a high sensitivity and specificity for the presence of urothelial cancer. However, one should be cautious in patients with a prior history of intravesical BCG treatment. A study by Orihuela and colleagues reported that 48% of biopsies of suspicious lesions were negative. The patient with a normal-appearing urethra, in the absence of risk factors (multiple bladder tumors, CIS, tumor at the bladder neck, and so forth), has a relatively low absolute risk of the presence of tumor (∼6%); however, approximately 43% of prostatic urethral tumors are not visually apparent.

Several studies have examined the sensitivity and specificity as well as the most appropriate technique with which to perform a prostatic urethral biopsy with the resectoscope loop. Donat and colleagues performed a prospective TUR loop biopsy on 246 male patients who subsequently underwent cystoprostatectomy, and found the sensitivity and specificity for Tis to be 53% and 77%, for T1 57% and 89%, and for stromal invasion 56% and 93%. Liedberg and colleagues reported a similar 66% sensitivity and 89% specificity for any prostatic urothelial cancer in their 175-patient cystoprostatectomy series.

Various investigators have studied how extensively and where the prostatic urethra ought to be sampled. Hillyard and colleagues initially described cold-cup biopsies and TUR biopsies of the lateral lobes. Taylor and colleagues, in a long-term follow-up of the same series, described additionally taking TUR resections proximal to the verumontanum. Liedberg and colleagues described transurethral loop biopsies from the bladder neck to the verumontanum, obtained at the 4- and 8-o’clock positions. This approach is supported by Sakamoto and colleagues who, based on pathologic evaluation of cystoprostatectomy specimens, found that firstly, 93% of urothelial cancer involving the prostatic urethra is superficial enough to be identified with a superficial resection and secondly, that 84% are located at the 5- and 7-o’clock positions. In their series, only 1 patient had deep stromal involvement without a superficial tumor being identified. Given that most of the tumors are detected superficially in the prostate, a complete TUR of the prostate is not indicated. Parenthetically, should a TUR of the prostate be indicated for lower urinary tract symptoms at the time of a bladder tumor resection, a recent meta-analysis suggests that this is quiet safe, with a low risk of tumor implantation in the prostate resection bed.

Accordingly, the technique most used to sample the prostatic urethra is a TUR at the 5- and 7-o’clock positions just proximal to the verumontanum, as this allows the surgeon to obtain a specimen of adequate depth and size to allow for pathologic assessment of mucosa, ducts, and stroma. This procedure is in agreement with the 2011 EAU guideline, which suggests that the biopsy should be taken “from abnormal areas and from the precollicular area (between 5 and 7 o’clock position) using a resection loop.”