Clinical, demographic and histopathological prognostic factors for urothelial carcinoma of the bladder.

INTRODUCTION: Our aim is to evaluate the influence of clinical and histopathological parameters, including age, gender, tumor stage, grade, tumor differentiation, necrosis, lymphovascular/perineural invasion (LVI/PNI) and concomitant carcinoma in situ (CIS), on outcomes of patients with urothelial carcinoma of the bladder (UCB).
MATERIAL AND METHODS: A total of 84 patients who underwent radical cystectomy (RC) (n = 11) and radical cystoprostatectomy (n = 73) for muscle-invasive bladder cancer at our hospital between 2007-2013, were included in the study.
RESULTS: The mean age of patients at diagnosis was 66.1, of whom 75 were males and 9 were females. Of the 84 patients, 38 were ≤65 years and 46 were >65 years. Mean tumor diameter was 3.66 cm. There were 38 cases which showed divergent differentiations. Concomitant CIS was observed in 30 tumors, 41 cases showed tumor necrosis, 44 PNI and 61 LVI. The rate of overall survival (OS) in patients aged ≤65 years was statistically significantly higher than in those aged >65 years. A negative statistical relationship was found between OS with lymph node metastasis (LNM) and tumor differentiation. On the other hand, necrosis did not remain significant on multivariate analysis. No statistically significant relationship was found between smoking, tumor stage, PNI, LVI and concomitant CIS and OS.
CONCLUSIONS: In this study, advanced age, LNM, tumor differentiation were found to be independent prognostic risk factors associated with OS after RC. These additional factors, which may explain the different clinical course in patients with similar tumor stage and lymph node status, should be taken into consideration in treatment planning.

INTRODUCTION

Urothelial carcinoma of the bladder (UCB) is the second most common genitourinary malignancy and is associated with a heterogeneous clinical outcome [1-4]. Radical cystectomy (RC) with bilateral pelvic lymph node dissection (PLND) is currently the gold standard treatment for muscle-invasive UCB [1, 5–9]. Unfortunately, 40% of patients with organ-confined disease at the time of cystectomy subsequently suffer recurrence. Several studies have evaluated the risk factors for recurrence and survival after cystectomy. Advanced pathologic stage, nodal involvement, grade and urinary obstruction have been reported as prognostic factors for survival and recurrence. However, some bladder cancer cases of similar stage and grade have demonstrated variable clinical outcomes after RC, so many attempts have been made to determine new and reliable prognostic factors [10-14].

The aim of the present study is to evaluate the influence of clinical and detailed histopathological parameters including age, gender, tumor stage, grade, tumor differentiation, necrosis, lymphovascular invasion (LVI), perineural invasion (PNI) and concomitant carcinoma in situ (CIS), on outcomes of patients with UCB treated with RC.

MATERIAL AND METHODS

A total of 84 patients who underwent RC (n = 11) and radical cystoprostatectomy (n = 73) for muscle-invasive bladder cancer (MIBC) at our institute between 2007-2013, were included in the study. Of the total, 79 underwent standard PLND. Bladder cancer was diagnosed histopathologically by transurethral resection in all patients before cystectomy. RC and standard PLND were performed using the standard technique. Surgical specimens were re-examined by 2 genitourinary pathologists applying a standardized reporting protocol. Tumor staging and grading were standardized according to the American Joint Committee on Cancer and World Health Organization. Tumor differentiation, depth of tumor invasion, necrosis, LVI, PNI and concomitant CIS were assessed histopathologically.

Statistically analyses of prognostic effects of age (65 years), gender, smoking status, pathologic tumor stage, lymph node metastasis (LNM), tumor differentiation, LVI, PNI and necrosis on overall survival (OS) were performed. Univariate OS after RS were estimated using the Kaplan-Meier method and log-rank statistics. Multivariate Cox regression models addresses OS after RS. The chi-square test was used to determine correlations among the variables. Statistical significance was set at p <0.05. Statistical analyses were performed with SPSS v.15.

RESULTS

The mean age at diagnosis was 66.1 (min. 42, max. 84) and there were 75 (89.3%) males and 9 (10.7%) females. Of the 84 patients, 38 (45.2%) were under 65 years, whereas 46 (54.8%) were over 65 years. Mean tumor diameter was 3.66 cm (min: 0.70 cm, max: 8 cm). The average overall follow-up time was 17.6 months (SD ±15.1). At the time of analysis, 33 (39.3%) patients were alive with disease, whereas 51 (60.7%) were dead. In 75 patients with available habitual data, 64 (85.3%) were recorded as heavy smokers.

The pathologic tumor stages were 4 (4.8%), 8 (9.5%), 20 (23.8%), 37 (44%) and 15 (17.9%) for Ta, T1, T2, T3 and T4 respectively. Of the patients with Ta and T1 tumors who had an extensive mass, which could not be totally excised by TUR, or intensive gross hematuria, underwent RC. Of the total 84 cases, 79 underwent standard PLND and LNM was detected in 25 patients (29.8%).

Of the 84 cases, 82 were high grade on histopathological examination. Both of the low grade tumors were stage Ta and exhibited no tumor necrosis, CIS, LVI and PNI. One of the patients was alive and the other one died of a non-tumoral cause.

The histologic type was pure urothelial carcinoma (UC) in 46 (54.8%) cases. Of the 38 (45.2%) cases which showed divergent differentiations or components, 26 (68.4%) had squamous differentiation, 7 (18.4%) sarcomatoid, 1 (2.6%) glandular differentiation, 1 (2.6%) clear cell, 1 (2.6%) neuroendocrine, 1 (2.6%) micropapillary and 1 (2.6%) squamous plus sarcomatoid components.

Concomitant CIS was observed in 30 (30.7%) tumors. 41 (48.8%) cases showed tumor necrosis, 44 (52.4%) PNI and 61 (72.3%) LVI. Demographic, clinical and pathological characteristics are summarized in Table 1.

Table 1

Univariate analysis of demographic, clinical and pathological characteristics for overall survival

Clinicopathologic factors	Category	n (%)	p values
Age	>65	46 (54.8%)	<0.001
	≤65	38 (45.2%)
Gender	Male	75 (89.3%)	0.23
	Female	9 (10.7%)
Tobacco consumption	Absence	11 (14.7%)	0.81
	Presence	64 (85.3%)
Pathologic stage	Ta	4 (4.8%)	0.15
	T1	8 (9.5%)
	T2	20 (23.8%)
	T3	37 (44%)
	T4	15 (17.9%)
Lymph node status	N0	54 (64.3%)	0.001
	N1	25 (29.8%)
	Nx	5 (6%)
Histopathologic Differentiation	Absence	46 (54.8%)	0.011
	Presence	38 (45.2%)
LVI	Absence	23 (27.7%)	0.37
	Presence	61 (72.3%)
PNI	Absence	40 (47.6%)	0.06
	Presence	44 (52.4%)
UCIS	Absence	54 (64.3%)	0.24
	Presence	30 (35.7%)
Tumor necrosis	Absence	43 (51.2%)	0.025
	Presence	41 (48.8%)

The relationship of tumor necrosis with pathologic tumor stage and LNM was evaluated. Accordingly, tumor necrosis was found in 25% (1/4) of Ta tumors, 25% (2/8) of T1 tumors, 50% (10/20) of T2 tumors, 48.6% (18/37) of T3 tumors and 66.7% (10/15) of T4 tumors. No statistically significant relationship was found between tumor necrosis and pathologic tumor stage (p = 0.32). Tumor necrosis was found in 60% (15/35) of N1 cases, 46.3% (25/54) of N0 cases and 20% (1/5) of Nx cases. No statistically significant relationship was found between tumor necrosis and LNM (p = 0.21).

The evaluation of OS data revealed that 55.6% (5/9) of female patients, and 37.3% (28/75) of male cases were alive. There was no statistically significant relationship between OS and gender (p = 0.23).

In this study, 57.9% (22/38) of the patients aged ≤65 years and 23.9% (11/46) of patients aged >65 were alive. The rate of OS in patients aged ≤65 years was statistically significantly higher than those aged >65 years (p <0.001) (Figure 1).

Figure 1

Kaplan-Meier curves of overall survival stratified according to age (≤65, <65).

On the other hand, 36.4% (4/11) of non-smokers and 45.3% (29/64) of smokers were alive. No statistically significant relationship was noted between smoking and OS (p = 0.81).

With regard to pathological tumor stage, 75% (3/4) of Ta patients, 75% (6/8) of T1 patients, 40% (8/20) of T2 patients, 37.8% (14/37) of T3 patients, and 13.3% (2/15) of T4 patients were alive. The cause of death in the Ta and T1 patients was not related to the primary tumor. No statistically significant relationship was found between OS and tumor stage (p = 0.15).

In this study, 51.9% (28/54) of the patients with no LNM, 16% (4/25) of the patients with metastasis and 20% (1/5) of the patients who did not undergo lymph node dissection were alive. A negative statistical relationship was found between LNM and OS in patients undergoing lymph node dissection (p = 0.001) (Figure 2).

Figure 2

Kaplan-Meier curves of overall survival stratified according to lymph node status.

The evaluation of the relationship between tumor differentiation and OS revealed that 26.9% (7/26) of patients with squamous differentiation and 14.3% (1/7) of patients with sarcomatoid differentiation were alive. One patient with micropapillary differentiation and another with clear cell differentiation were alive, whereas those with tumors exhibiting squamous and sarcomatoid differentiation, glandular and neuroendocrine differentiation were not alive. A negative statistically significant relationship was found between tumor differentiation and OS (p = 0.01) (Figure 3).

Figure 3

Kaplan-Meier curves of overall survival stratified according to the histological differentiation of tumors.

In addition, 26.8% (11/41) of patients with tumor necrosis and 51.2% (22/43) of those without tumor necrosis were alive. A negative statistical relationship was found between tumor necrosis and OS (p = 0.025) (Figure 4).

Figure 4

Kaplan-Meier curves of overall survival stratified according to tumor necrosis.

No statistically significant relationship was found between PNI, LVI and concomitant CIS and OS (p = 0.06, p = 0.37, p = 0.24, respectively).

On univariate analysis, variables such as age (>65), LNM, tumor necrosis, differentiation status were all demonstrated to be significant prognostic factors affecting the OS. On the other hand, necrosis did not remain significant on multivariate analysis (Table 2).

Table 2

Multivariate analysis of parameters predicting overall survival

Variable	Levels	Hazard Ratio	95% Cl Lower Bound	95% Cl Upper Bound	p value
Age (>65)	≤65 years	2.969	1.550	5.684	0.001
	>65 years
Lymph node metastasis	N0	2.204	1.223	3.970	0.009
	N1
Differentation	Negative	2.116	1.173	3.818	0.013
	Positive
Tumor necrosis	Negative	1.601	0.878	2.917	1.124
	Positive

DISCUSSION

UCB is the second most common tumor of the genitourinary tract [1-4]. Globally, UCB is the seventh most common cancer in males and seventeenth in females [1, 5]. Recently, the pathologic tumor stage and the presence of LNM have been reported to be the most important prognostic factors [1, 2, 12]. However, reports of different clinical outcomes, in patients with similar stages of disease following RC, have prompted the investigation of other factors that may affect prognosis.

A number of studies have reported that the prognosis of UC in females is much worse than that in males [5, 15–18]. A large European epidemiological study of 1.2 million patients reported that the 5-year cancer-specific mortality was 30% lower in females, which, however, was not the case in bladder carcinomas. The study also demonstrated that UC followed a more aggressive clinical course in females than that in males [5]. Horstmann et al. reported that, in a MICB series of 455 patients, 129 of whom were females, the 10-year survival was lower in females compared to that in males [19]. Aggressive tumor biology in females is considered to be responsible for shorter survival [5]. In our series, where most patients were males (89.3%), there were only 9 female patients and there was no statistically significant difference between OS and gender (p = 0.23). The absence of a statistically significant relationship between gender and OS can be attributed to the small number of female patients in this study.

In a study by Mitra et al., where 259 tumors with tumor differentiation were compared with pure UCB, the OS was lower in patients with differentiation and aged >65 years [13]. In our series, the rate of OS was higher in patients aged ≤65 years (57.9%) compared to those aged >65 years (23.9%) and the difference was statistically significant (p <0.001).

Previous studies have demonstrated that tobacco consumption and the number of cigarettes smoked per day are associated with advanced tumor stage and grade in newly-diagnosed UCB [14, 20]. A study of 1506 patients with UCB by Rink et al., reported the association between smoking and cancer-specific mortality, which, however, lost its significance on multivariate analysis [14]. In the mentioned study, cumulative cigarette exposure was associated with aggressive tumor biology in patients treated with RC. Lifetime cumulative exposure to cigarette carcinogens leads to cumulative molecular changes in UCB, thus affecting the biologic and clinical behavior of the tumor [21]. In our study, of the 75 patients with available data on tobacco consumption, 64 were current smokers. Even though no statistically significant relationship was noted between tobacco consumption and OS, we believe that our series is not adequately powered to make such a comparison, due to the imbalance between the number of patients (p = 0.81).

Tumor stage and LNM have been reported to be the most important prognostic factors in UC [1, 8, 22]. In our series, even though the rates of OS decreased significantly as the tumor stage increased, no statistically significant relationship was found between tumor stage and OS, probably due to the small number of patients in our series (p = 0.15).

It has been reported in the literature that 25% of the patients undergoing cystectomy had LNM [23]. In our series, the rate of LNM was 29.8%, which is consistent with the literature. There was also a negative statistical relationship between LNM and OS (p = 0.001).

UCB is a tumor manifesting clinical and morphological differences and has a distinct capacity for histological differentiation [13]. Squamous and glandular differentiation are the most common variants of UCB [13]. The rate of the variants of UC, which have been reported to be poor prognostic factors, varies over a wide range (from 7% to 81%) [24, 25]. This situation has been considered to be caused by differences in disease stage, sampling techniques and pathologic evaluation [12]. Furthermore, there is no standard technique for quantitative measurement of the extent of tumor differentiation [12]. In a study of 1984 patients with UCB by Xylinas et al., 488 patients had UC variants, with squamous differentiation being the most common (227 patients) and glandular differentiation, the second most common (75 patients) [12].

In patients with differentiation, the disease followed a more aggressive course biologically [12]. These patients were at significantly higher risk for disease recurrence and cancer-specific mortality compared to those with pure UCB patients. In addition, non-squamous variants were found to be associated with worse prognosis compared to pure UCB and UCB with squamous differentiation. In our series, squamous differentiation was the most common (26 patients, 68.4%), followed by sarcomatoid differentiation (7 patients, 18.4%). A statistically significant difference was found in OS between pure UCB and UCB with variant histology (p = 0.01). There was no statistically significant difference in OS between the subtypes of squamous and non-squamous differentiation (p = 0.34).

Previous studies have reported that the rate of LVI ranges from 35% to 55% in MIBC [8, 22, 26, 27]. It has been reported that routine HE staining may be enough to assess vascular invasion [26]. In our study, LVI was assessed on HE sections and the rate of LVI was found to be 72% (61 cases). The high rate of LVI is probably due to the difficulty in distinguishing LVI from tissue retraction artifacts. Thus, we believe that, the assessment of LVI carried out by HE staining, should be supported by the use of other immunohistochemical markers.

The prognostic role of LVI in UCB remains controversial. Some studies have reported that LVI is a poor prognostic factor, whereas other studies have suggested that LVI is not a prognostic predictor of LNM, survival or recurrence [8, 22, 26]. Some of the studies, which identified LVI as a poor prognostic factor, reported that LVI did not remain significant on multivariate analysis [8, 22, 27]. No statistically significant relationship was noted between LVI and OS in our study (p = 0.37).

Controversial results have been reported in studies investigating the effects of PNI on prognosis. Some studies have reported that PNI is associated with LNM and distant metastases, whereas other studies have demonstrated that tumor stage, LNM, urethral obstruction, LVI and PNI are significant prognostic factors on univariate analysis; however, only tumor stage and LNM are independent prognostic factors on multivariate analysis [10]. A retrospective review of 125 patients by Hong et al., reported that LVI and PNI were prognostic factors on univariate analysis, whereas only the tumor stage and vascular invasion were statistically significant on multivariate analysis [10]. In this study, there was a small difference in OS between patients with and without PNI, and no statistically significant relationship was noted (p = 0.06).

A large series of 1425 patients with UC of the upper urinary tract by Zigeuner et al., reported that extensive tumor necrosis (>10% of the tumor) was associated with aggressive biology. However, this association was not observed in the presence of focal necrosis [11]. In a series of 98 patients with UCB by Ord et al., tumor necrosis was associated with advanced tumor stage and nodal metastasis, and tumor necrosis was found to be the independent prognostic factor on univariate and multivariate analyses [28]. In the present study, 73.2% of the patients with tumor necrosis were not alive at the time of follow-up. There was a negative statistical relationship between tumor necrosis and OS (p = 0.025) and also the tumor necrosis in N1 patients was higher than that in N0 patients, which however did not reach the level of statistical significance (p = 0.21). Similarly, advanced tumor stage was associated with an increased rate of tumor necrosis, but no statistically significant relationship was found due to the small number of patients in our study (p = 0.32).

Numerous cytogenetic, molecular, genetic and immunohistologic studies revealed similar molecular changes in CIS and invasive UC [4]. The presence of isolated or concomitant CIS carries a higher risk of the disease progressing to MIBC [29]. There are numerous studies reporting that the presence of CIS and concomitant non-invasive UC following RC is associated with a poor clinical course [4]. On the other hand, a study by Nuhn et al. of 3973 patients treated with RC, reported no association between concomitant CIS and clinical outcome and the prognostic value of concomitant CIS in UCB could not be confirmed [4]. Similarly in our study no statistically significant difference was noted in OS between the patients with and without concomitant CIS (p = 0.24).

CONCLUSIONS

Advanced age (>65), LNM, tumor differentiation and tumor necrosis were found to be independent prognostic risk factors associated with OS after RC. Tumor necrosis did not remain significant on multivariate analysis. The presence of concomitant CIS had no effect on prognosis. These additional factors, which may explain the different clinical course in patients with similar tumor stage and lymph node status, should be taken into consideration in treatment planning.