Bladder cancer is the seventh mostly common diagnosed tumor in men worldwide and 297,300 new cases predicted to this year [1]. According to the National Cancer Institute, 74,000 new cases are estimated for 2012 in the United States, in addition to 15,000 deaths ascribed to the disease within the same period [2]. In Brazil, bladder cancer is the tenth most frequent neoplasia in men with 6,210 predicted new cases for 2012 [3]. The most frequent histological type is urothelial carcinoma, found in 90% of the cases. Among these, 20-25% is found in its muscle-invasive form. These cases require radical cystectomy, regardless of improved knowledge about anatomy, surgical techniques, and mainly about post-operative care, the specific survival period remains on a 50-60% basis in five years [4].

Estimating treatment efficacy is fundamental to advise patients of prognoses and indicate adjuvant therapies. Therefore, several prognosis-predictive factors have been recently investigated, they can be grouped into the following: clinical, molecular, and pathological this one is considered to be the most important [5]. The insufficiency of each pathological factor to be used isolatedly in prognosis prediction played an important role to stimulate the grouping, which gave origin to the nomograms [5-9]. The difficulty to obtain data standardization from diverse institutions and the regional differences found in neoplasia are reflected in many nomograms as well as the subsequent questioning over its actual usefulness [10,11]. Not many studies are available on urothelial cancer after radical cystectomy, so we noticed the urge to evaluate them and, afterwards, elaborate reliable methods to stratify recurrence as well as death risks.

We surveyed the retrospective analysis of medical files of all patients who underwent radical cystectomy and pelvic lymphadenectomy in the 2006-2009 period at the Clinical Hospital Medical School. We include all bladder urothelial carcinoma cases. The following exclusion factors were considered for this study: metastatic disease, palliative surgeries, non-urothelial carcinoma, patients under neoadjuvant therapy, and patients with insufficient medical records. Patients were evaluated during the postoperative period within 3-month intervals through laboratory and image tests. The upcoming of secondary lesions over one month after the surgical procedure was considered as recurrence or relapse, as they were not evident when the treatment was indicated.

The following data were collected: clinical stage (pT), lymphonodal disease (pN), histological grade, lymphovascular invasion (LVI), perineural invasion (PNI), carcinoma in situ (CIS), recurrence-free survival (RFS), cancer-specific survival (CSS), and overall survival (OS). The SPSS software version 17.0 (Chicago, IL) was used in the analysis. Descriptive statistics (average, standard deviation, 95% confidence interval) was employed to characterize patients in the groups, in addition to two other methods for the study on survival: logistic regression to determine the “Odds Ratio” between groups, and survival analysis with calculation of the average time, and log-rank test significance. The significance level adopted was 5%.

Following methodological criteria, we found 128 patients: 20 females (15.6%) and 108 males (84.4%) aged 41-84 with an average 67 years of age. The urinary derivation in use was 85 ileal conduit (66%) and 39 orthotopic neobladder (31%) cases, and four cases involving other reconstruction types, such as ureterosigmoidostomy (3%). Pelvic lymphadenectomy was noticed at an average removal rate of 15.14 (± 8.9) lymph nodes per surgery and an average follow-up of 17.3 months (15.24 to 19.44).

Table 1 shows a higher pT2 stage tumor in 53.1% of the cystectomy specimens of our case study. Its presence proved to be responsible for an enhanced recurrence of the disease, death derived from cancer, and general death (respectively 2.3X, p=0.032; 4.44X, p=0.001 and 4.2X, p=0.002) (Table 2). The Kaplan-Meier analysis shows a significant survival improvement of 30 months for the organ-confined disease (70%/50%, 85%/55% and 80%/50%, respectively for RFS, CSS and OS) (Figure 1).

Table 1: Demographics data features of 128 radical cystectomies.

Table 2: RFS, CSS and OS univariate analysis according to pathological feature.

Figure 1: RFS (A), CSS (B) and OS (C) according to grouped stage.

Likewise, the presence of pathological lymph nodes, found in 36% of the specimens, resulted in a greater recurrence rate, death derived from cancer, and overall death (respectively, 3.2X, p=0.003; 2.4X, p=0.0371 and 2.3X, p=0.038). The lympho-nodal metastasis determines worsened survival for 50%, 60% and 55%, respectively for RFS, CSS and OS within 30 months (Figure 2).

Figure 2: RFS (A), CSS (B) and OS (C) according to pathological lymph node.

The LVI diagnosis found in 46% of the surgical pieces under analysis showed a significant association with cancer-related death (p=0.041). It was found not relevant in our study in terms of disease recurrence (p=0.08). The Kaplan-Meier analysis showed that the presence of lymphovascular invasion determines a significant reduction in RFS, CSS, and OS (respectively 50%, 60% and 60%) (Figure 3).

Figure 3: RFS (A), CSS (B) and OS (C) according to lymphovascular invasion.

Perineural invasion found in 29% of the specimens was associated with death derived from cancer and general death (respectively p=0.026 and p=0.012). However, no significant difference was found when recurrence is under analysis. In the presence of PNI, a significant reduction in RFS, CSS and OS was noticed (40%, p<0.05) (Figure 4).

Figure 4: RFS (A), CSS (B) and OS (C) according to PNI presence/absence.

Recurrence cases were not diagnosed in the low histological grade group (8.6% of cases). A significant difference was noticed in comparison with the high grade group (p=0.016). No association was observed with the high histological grade and specific/general mortality (respectively p=0.259 and p=0.481). Likewise, the presence of CIS did not show any association with death, recurrence or survival (p>0.05).

The overall analysis concerning the most significant factors for the recurrence-free survival showed 70% of survival in the absence of >pT2 and pN+ stages, 55% in the presence of only one factor, and 40% in the presence of both factors (p=0.005) (Figure 5).

Figure 5: RFS according to the presence or absence of >pT2 and pN+ stages.

Together, the >pT2, pN+, LVI+ and PNI+ factors allowed the stratification of 3 risk groups regarding death caused by cancer. A 90% CSS level was observed in the absence of factors; 65% in the presence of one or two factors, and 50% in the presence of 3 or more of the factors described (p<0.05) (Figure 6).

Figure 6: CSS according to the presence or absence of >pT2, pN+, LVI+ and PNI+ stages.

The OS analysis presented an 85% survival rate in the absence of stage factors >pT2, N+ and PNI+; 65% in the presence of only one factor, and 45% in the presence of two or more factors (p<0.05) (Figure 7).

Figure 7: OS according presence or absence of >pT2, pN+, LVI+ and PNI+.

The factors were submitted to the proportional Cox multivariate analysis through which we noticed that >pT2 and pN+ were the independent variables in the prognosis in relation to recurrence and death, respectively.

The bladder urothelial cancer is an extremely prevalent disease, the incidence of which can appear in two ways: muscle-invasive and nonmuscle invasive. Local invasion, lymphatic involvement, and distant metastases are typical of muscle-invasive tumors for which the radical cystectomy with pelvic lymphadenectomy is considered to be the most effective treatment. It is known that the characteristics derived from the radical cystectomy are fundamental for prognosis determination. CIS prevalence was observed in the surgical piece in 28.1%. In the literature, it was found in 25 to 53.6% [6,12]. A study by Nuhn et al. analyzed 3,973 patients who underwent radical cystectomy. The study was not able to demonstrate any association of the CIS diagnosis (43.8%) with recurrence or death due to cancer [13]. In our study, we observed CIS in 28.1% of the surgical pieces. The RFS, CSS and OS were not changed due to its presence. Despite its relevance for urothelial cancer prognosis without muscle invasion, the histological grade is still controversial in the prognosis of tumors with invasion. In the literature, it is possible to find publishing material defending a high histological grade (84.6% of cystectomies) as a prognostic factor, although most authors have not confirmed it [5-7,14]. In our case study, we identified a high grade association with recurrence, although it was not relevant for the other analyses.

A study by Bassi et al. point to the pT and OS association, but Canter et al., in their research involving 356 patients, documented deterioration in RFS, CSS and OS in >pT2 stage [14,15]. In the literature, we found prevalence of the >pT2 stage in 41.3% to 56.2% of the cases, and 53.1% in our case study [6,16]. In 2009, Manoharan et al. demonstrated that the confined disease presented an enhanced survival with OS and RFS, respectively at 74% vs 52% and 91% vs 56% [17]. In our study, we found that the ≤ pT2 stages presented better evolution than the >pT2 stages (70%/50%, 85%/55% and 80%/50%, respectively for RFS, CSS and OS). Both univariate and multivariate analyses revealed the association of >pT2 stage with survival.

In the previous studies, the presence of lympho-nodal metastasis is associated with 56.7% and 49.5% of RFS and OS within five years [6,16]. In our study, we found 50%, 60% and 55% of RFS, CSS and OS, respectively. The univariate and multivariate analysis revealed that pathological lymph nodes make up an independent factor for cancer prognosis either in recurrence or death.

The lymphovascular invasion is found in the literature in 30 to 50% out of the radical cystectomy specimens [15,18,19]. In 2006, a review made by Algaba suggested the association of LVI with disease recurrence [20]. In our case study, the presence of LVI determined a significant reduction in RFS from 60% to 50% within 30 months. The worst prognosis in the evaluation of its effect regarding CSS and OS was found when LVI was detected (75% to 60% and 70% to 60%, respectively). The univariate analysis demonstrated its association with cancer-specific mortality. We did not succeed in confirming its independence either in recurrence or death. However, Canter et al. found a significant reduction in RFS, CSS and OS [15]. In their multicenter study, Bolenz et al. concluded that LVI determined a greater number of recurrence and death [21].

Up to the present moment few publications are focused on PNI, which is found in 5.9% to 47.7% of the specimens [14,22]. Its real role in the prognosis is still controversial. While evaluating prognostic factors in post-radical cystectomy, Bassi et al. found that the presence of PNI determined a worse OS. However, such findings did not resist the multivariate analysis [14]. Leissner et al. demonstrated a greater recurrence in the group with PNI presenting 30% of disease-free survival when compared to 55% of the group without PNI [22]. In our case study, we found 28.9% of prevalence in surgical pieces, with a relation to both specific and general mortality. The evaluation through Kaplan-Meier showed a significant reduction in RFS, CSS and OS. Meanwhile, such findings did not confirm the independence analysis.

The evaluation of the isolate prognostic factors has been abandoned due to the nomograms, since the association of factors has proven to be more accurate [6]. The use of still unacclaimed parameters as prognostic factors leads to the questioning on the actual nomogram reliability and applicability [10,11]. Manoharan et al. [17] demonstrated the importance of the pT and pN stages. They have been gathered in three groups: Organ-confined disease, extravesical disease, and lymphonodal disease. Such stratification showed to be practical, but it does not take LVI into account for cancer prognosis, something that has also been found in their research [17]. While evaluating recurrence in the presence of the >pT2 and pN+ stages, we observed disease-free survival at the rates of 70%, 55% and 40%, respectively in the absence of factors, presence of one factor, and presence of both factors. The inclusion of LVI+ and PNI+ to the >pT2 and pN+ stages allowed us to categorize three different groups regarding CSS. Survival rates of 90%, 65% and 50% within 30 months were found in the absence of factors, presence of one or two factors, and presence of three or more factors, respectively. Regarding OS, three risk groups were obtained through 85%, 65% and 45% of general survival within 30 months, respectively in the absence of factors, presence of one factor, and presence of two or more factors out of: >pT2, pN+ and PNI+.

Although LVI and PNI did not appear to be an independent factor in our study, they were revealed as important prognostic factors as far as the survival time is concerned. We are aware of the shortcomings presented by a retrospective study so a prospective study involving a larger case study might attest to this supposition.

Our study confirmed the importance of the pT and pN stages as a prognostic factor. We agree with recent publications that include LVI as a prognostic factor, and we suggest that further PNI studies are carried out, as this subject has not been fully developed so far.