Prognostic value of inflammation or granuloma after intravesival BCG in non-muscle-invasive bladder cancer

Authors


Abstract

Objective

  • To evaluate the prognostic value of inflammation or granuloma after intravesical bacille Calmette-Guérin (BCG) treatment in non-muscle-invasive bladder cancer (NMIBC).

Materials and Methods

  • Patients with NMIBC treated with intravesical BCG over a 5-year period were identified.
  • The correlations between histopathological results and disease recurrence and progression were assessed, with survival analysis performed using the Kaplan–Meier method. Other relevant variables were also evaluated using univariate and multivariate analysis.
  • A log-rank test was performed to compare time-to-event between groups.

Results

  • A total of 215 patients were treated with BCG for NMIBC and the median follow-up was 32 months.
  • Granuloma was identified in 60 patients and inflammation in 125 patients. In 18 patients there was no evidence of either (normal histology group). A total of 12 patients did not have biopsies and were subsequently excluded.
  • The mean recurrence-free survival rate was significantly higher in the granuloma and inflammation groups (65 months [95% CI: 58–72] and 56 months [95% CI: 49–63], respectively) than in the normal histology group (20 months [95% CI: 6–34]; log-rank P < 0.001).
  • On the multivariate analysis, the absence of inflammation/granuloma was significantly associated with recurrence (log-rank P < 0.001).
  • The progression-free survival rate was higher in the granuloma and inflammation groups (75 months [95% CI: 71–79] and 82 months [95% CI: 78–86], respectively) compared with the normal histology group (33 months [95% CI: 17–48]; log-rank P < 0.001).
  • On multivariate analysis, the absence of inflammation/granuloma was significantly associated with recurrence (log-rank P < 0.001).

Conclusion

  • Inflammation or granuloma in histology samples after intravesical BCG treatment for NMIBC are positive markers of response and their absence increases the risk of recurrence and progression.

Introduction

Bladder cancer is one of the commonest cancers, with a worldwide incidence of 10.1 per 100 000 males and 2.5 per 100 000 females [1]. It is estimated that 70–80% of bladder cancers are non-muscle-invasive [2]. Bladder cancer has a high recurrence rate of ∼65% and a progression rate of ∼30%, which presents a big challenge to health organizations [3, 4].

A live strain of Mycobacterium bovis, BCG, is a well-known immunotherapy agent and is the most common intravesical therapy for treating high grade non-muscle-invasive bladder cancer (NMIBC). It was first used as a vaccine for tuberculosis in humans in 1921 [5] and Morales et al. [6] used it intravesically for the first time in 1976 to treat bladder cancer. Intravesical BCG has been shown to reduce the recurrence rate and the risk of progression to carcinoma invading bladder muscle [7-9]. Although the exact mechanism by which intravesical BCG affects bladder cancer is not fully understood, it is believed to induce an immune response which, in turn, has an anti-tumour effect [10-12]. In follow-up cystoscopy, inflammation and granuloma are commonly found in histological samples after intravesical BCG, but the significance of this in terms of recurrence and progression prevention is not clear.

In the present study, we evaluated whether the presence of inflammation or granuloma in histological samples after intravesical BCG can act as a marker for response.

Patients and Methods

Between 2005 and 2011, patients who underwent intravesical BCG treatment for NMIBC were identified and retrospectively evaluated. They had all undergone initial transurethral resection of bladder tumour and were diagnosed with TCC with grade 3 and/or T1 tumour and/or carcinoma in situ (CIS).

Patients with primary bladder TCC were treated with intravesical BCG instillations; the assigned regime was a 6-week induction course and then further maintenance courses of three instillations every 3–4 months for a year, as tolerated.

If muscle was not included in the initial resection, then early relook cystosocopy was performed. All patients were followed-up with rigid cystoscopies and bladder biopsies 6–8 weeks after each BCG cycle. Biopsies were usually targeted if recurrence was found, but random biopsies were taken even if the bladder looked normal. After intravesical BCG, we assessed the correlation between the presence of granuloma and/or inflammation on histopathological review and disease recurrence and progression, which we defined as tumours of a higher grade or stage than the primary tumour.

Granuloma is defined as a focus of chronic inflammation consisting of a microscopic aggregation of macrophages that are transformed into epithelium-like cells, surrounded by a collar of mononuclear leukocytes, principally lymphocytes, and occasionally plasma cells [13].

Other relevant variables (age, sex, grade, presence of CIS, stage, lymphovascular invasion, multifocal disease, treatment compliance, early relook cystoscopy) were also evaluated using univariate and multivariate analysis. The survival analysis was performed using the Kaplan–Meier method. A log-rank test was performed to compare time-to-event between groups, and a P value of 0.05 was considered to indicate statistical significance.

Results

A total of 221 patients received intravesical BCG treatment and the median (range) follow-up was 32 (5–89) months. Six patients were excluded as they were found to have associated upper tract TCC.

Granuloma was identified in 60 patients and inflammation in 125 patients. There was no evidence of either in 18 patients (normal histology group), and biopsies were not taken in 12 patients who were deemed unfit for general anaesthetic or had significant other illnesses and were subsequently excluded. The original histological findings of the different groups are shown in Table 1.

Table 1. Original histology among different groups
Tumour grade/stageGranuloma group, n (%)Inflammation group, n (%)Normal histology group, n (%)
  1. Grade 2 histology was either associated with T1 or CIS.
Grade 215 (25)24 (19)4 (22)
Grade 331 (52)71 (57)9 (50)
Ta23 (38)45 (36)4 (22)
T123 (38)50 (40)9 (50)
CIS26 (43)49 (39)8 (44)
Total6012518

Recurrence was found in 17% of the granuloma group (10/60 patients), 38% of the inflammation group (47/125 patients) and 71% in the normal histology group (10/14 patients). The mean recurrence-free survival rate was significantly higher in the granuloma and the inflammation groups (65 months [95% CI: 58–72] and 56 months [95% CI: 49–63], respectively) than in the normal histology group (20.2 months [95% CI 6–34]; P < 0.001). When comparing the granuloma group with the inflammation group, recurrence-free survival was significantly longer in the granuloma group (P = 0.008; Fig. 1, Table 2).

Figure 1.

Kaplan–Meier cumulative recurrence-free survival analysis.

Table 2. Confidence intervals and log-rank tests for mean recurrence-free survival
 Means for recurrence-free survivalPairwise comparisons
Histology groupEstimate, monthsse95% CINo inflammationInflammationGranuloma
Lower bondUpper bondChi-squared testPChi-squared testPChi-squared testP
Normal histology206.9634  18.42<0.00131.84<0.001
Inflammation563.7496318.42<0.001  6.940.008
Granuloma653.8587231.84<0.0016.940.008  
Overall573.05163      

The median (range) age at diagnosis was 76 (35–91) years. Age at diagnosis had no association with recurrence-free survival (95% CI: 0.972–1.019; P = 0.677). Male sex was associated with better recurrence-free survival; recurrence was found in 54/171 men (31.5%) and in 16/32 women (50% [chi-squared test = 5.186; P = 0.023]).

At presentation, 31 patients had multifocal disease. There was no association between multifocal disease at diagnosis and recurrence-free survival (Table 3, Fig. 2)

Table 3. Disease recurrence in patients with single and multifocal tumours
 No recurrence, n (%)Recurrence, n (%)Total, n (%)
  1. The assumption of proportional hazard is not met and hence log-rank was not significant.
Single tumour119 (65.7)59 (34.3)172 (100)
Multifocal tumour19 (61.3)12 (38.7)31 (100)
Figure 2.

Kaplan–Meier cumulative recurrence-free survival analysis for single and multifocal tumours at presentation.

There was no significant association between tumour grade at diagnosis (P = 0.390) or between CIS (P = 0.835) and recurrence-free survival, while T stage was found to be significantly associated with recurrence-free survival (P = 0.001). Only six patients had lymphovascular invasion, of whom four patients experienced recurrence. Lymphovascular invasion was significantly associated with recurrence (P = 0.012).

An early relook was carried out in 22 patients and recurrence was found in seven of these (31.8%). There was no association between early relook and recurrence-free survival (P = 0.397).

Ten patients did not complete the BCG induction, but after those patients were excluded, there was no difference among groups with regard to treatment compliance. In all, 71% of patients completed the year's course of BCG treatment. Not completing BCG induction (seven patients in the inflammation group and three in the normal histology group) was significantly associated with recurrence (P = 0.008); seven of these patients had recurrence. There was a positive association between the completion of BCG induction and the presence of granuloma (P = 0.015).

On multivariate analysis, the absence of inflammation/granuloma, female sex and T1 stage remained significantly associated with recurrence (Table 4).

Table 4. Univariate and multivariate analysis of significant variables for recurrence-free survival
VariableUnivariate analysisMultivariate analysis
P95% CIP
Histology (absence of inflammation/granuloma)<0.0010.057–0.297<0.001
Age0.677
Female sex0.0231.064–3.3480.030
Early relook0.397
Tumour grade0.390
Tumour stage0.0011.241–3.2500.005
CIS0.835
Lymphovascular invasion0.0120.738–6.6370.157
Completing BCG induction0.0080.263–1.3170.197

Progression was found in 3/60 patients (5%) in the granuloma group, 10/125 patients (8%) in the inflammation group and 11/18 patients in the normal histology group. Progression-free survival was longer in the granuloma and inflammation groups (75 months [95% CI: 71–79] and 82 months [95% CI: 78–86], respectively) than in the normal histology group (33 months [95% CI: 17–48]; P < 0.001). There was no significant difference between the granuloma and the inflammation groups (P = 0.542 [Fig. 3, Table 5]).

Figure 3.

Kaplan–Meier cumulative progression-free survival analysis.

Table 5. Confidence intervals and log-rank tests for mean progression-free survival
 Means for progression-free survivalPairwise Comparisons
Histology group Estimate, monthsse95% CINo inflammationInflammationGranuloma
Lower bondUpper bondChi-squared testPChi-squared testPChi-squared testP
Normal histology337.91748  50.981<0.00136.265<0.001
Inflammation822.1788650.981<0.001  0.3720.542
Granuloma752.2717936.265<0.0010.3720.542  
Overall792.07583      

Age, grade, stage, CIS and multifocal disease were not associated with progression (P = 0.622, 0.969, 0.526, 0.879 and 0.652, respectively). Sex, lymphovascular invasion and completion of induction BCG treatment were associated with disease progression on univariate analysis (P = 0.051, 0.073 and 0.034, respectively). None of the patients for whom an early relook was carried out experienced disease progression. On multivariate analysis, only the presence of inflammation or granuloma were significant variables (P < 0.001 [Table 6]).

Table 6. Univariate and multivariate analysis of significant variables on recurrence free survival
VaribaleUnivariateMultivariate
Sig.95% CISig.
Histology (Absence of inflammation/granuloma)<0.0010.015–0.195<0.001
Age0.622
Sex0.0510.756–4.3130.219
Early relook0.0000.979
G Grade0.969
T Stage0.526 
CIS0.879
Lymphovascular invasion0.0730.374–7.5800.497
Multifocal disease0.652
Completing BCG induction0.0340.142–1.6740.253

A total of 14 patients went on to have radical cystectomy, of whom one died from their cancer.

During the period of follow-up, a total of 42 patients died, of whom eight died from their cancer. The overall survival rate was significantly different between the normal histology and the inflammation groups (P = 0.042) and between the normal histology and the granuloma groups (P = 0.032 [Fig. 4]).

Figure 4.

Kaplan–Meier cumulative overall survival analysis.

No patient in the granuloma group died from their cancer, while three patients in the inflammation group vs five patients in the normal histology group died from their cancer, which was found to be significant (P < 0.001 [Fig. 5]).

Figure 5.

Kaplan–Meier cumulative cancer-specific survival analysis.

Discussion

The present study has the limitations inherent in any retrospective series of patients. Furthermore, 12 patients were not biopsied after intravesical BCG treatment, because of either a lack of fitness to undergo a general anaesthetic procedure or the presence of serious medical conditions influencing their short-term survival.

Patients' pre-treatment BCG immunization status was not available and the Mantoux test was not performed in any patient, as this is not our usual practice and it would not have changed the treatment decision. It has been proposed that four BCG induction instillations would be enough for immunized patients while six would be required for non-immunized patients [14]. In our series, a failure to complete the BCG induction was associated with a lower presence of granuloma and also a higher risk of recurrence. This result could possibly have been influenced by the patients' previous immunization status.

Histology at biopsy, sex and T stage were all significantly associated with recurrence; however, CIS and tumour grade, multifocal disease and lymphovascular invasion, which have been reported to be significant in the literature [15-18], were not found to be significant on multivariate analysis. The presence of inflammation and/or granuloma was found to have a significant impact on recurrence. Previous studies did not include this as a variable and this might explain the findings from our series.

There was a slight difference in initial histology among the groups: the normal histology group had more T1 histology, which was found to be a significant variable in terms of recurrence. This difference may have influenced this finding. On multivariate analysis, the presence or absence of inflammation or granuloma was also found to be significant.

There is no agreement in the literature with regard to the prognostic value of inflammation or granuloma in histology samples after intravesical BCG treatment. Kelley et al. [19] reported on 62 patients and found recurrence in only 24% of patients with granuloma over 12 months follow-up. They concluded that there was an association between granuloma and response. Herr et al. [20], in their series of 86 patients, and Schellhammer et al. [21], in their series of 28 patients, reached similar conclusions. In a series of 105 patients, Bassi et al. [22] concluded that the presence of granuloma could not be considered a prognostic factor. Torrence et al. [23], in their series of 104 patients, initially reported a better prognosis in the presence of granuloma, but the statistical significance was lost on longer follow-up. Pieras-Ayala et al. [24], in their series of 99 patients, reported no significant association between response and the presence of granuloma or inflammation.

In the present series, there was a significant difference in recurrence-free survival between histology samples in which inflammation or granuloma was present and those in which it was absent. The difference was more evident when granuloma was present. In the absence of inflammation or granuloma, the median recurrence-free survival was 20.2 months, while in the presence of inflammation or granuloma recurrence-free survival was almost three times as long.

Similar findings were seen in the progression analysis with regard to inflammation and granuloma; although there was no difference between inflammation and granuloma on progression, their presence was associated with significantly better outcomes. The presence or absence of inflammation or granuloma was the only significant factor among the variables on the multivariate analysis.

The mechanism by which BCG exhibits its anti-tumour activity is not fully understood. It is believed that an immune response develops in the bladder wall which ultimately clears the cancer or prolongs recurrence/progression-free survival [10-12]. This hypothesis is supported by the findings of the present series.

In the absence of inflammation or granuloma, the risk of recurrence and progression is significantly higher. This necessitates a closer surveillance policy or perhaps a low threshold for alternative therapy. The practice of random biopsies after BCG therapy could help in establishing whether inflammation or granuloma is present, which could act as a prognostic marker. In our series, 16 patients had targeted biopsies and the rest had random biopsies (mean 3.6, range 1–6 biopsies). From this retrospective series, it was not possible to recommend an ideal number of biopsies.

In conclusion, the presence of inflammation or granuloma in histology samples after intravesical BCG for NMIBC are positive markers of response and their absence increases the risk of recurrence and progression.

Conflict of Interest

None declared.

Abbreviations
NMIBC

non-muscle-invasive bladder cancer

CIS

carcinoma in situ

Ancillary